Advances in Child Development and Behavior Volume 16

ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR

VOLUME 16

Contributors to This Volume Elizabeth Bates Marjorie Beeghly-Smith

Marc H. Bornstein Inge Bretherton Martha J. Farah Tiffany M. Field Stephen M. Kosslyn Michael E. Lamb Sandra McNew David S. Palermo Mitchell Robinson Robert S. Siegler Tedra A. Walden

ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR

edited by Hayne W. Reese

Lewis P. Lipsitt

Department of Psychology West Virginia University Morganrown, West Virginia

Department of Psychology Brown University Providence, Rhode Island

VOLUME 16

@

1982

ACADEMIC PRESS A Subsidiary of Harcourt Brace Jovanovich, Publishers

New York London Paris. San Diego San Francisco *SBo Paulo *Sydney .Tokyo .Toronto

BY ACADEMIC PRESS,INC. ALL RIGHTS RESERVED. NO PART O F THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, 1NCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

COPYRIGLIT @ 1982,

ACADEMIC PRESS, INC.

I 1 1 Fifth Avenue, New York, New York 10003

UuifedKiiiPdotn Editiori ~irblisliedbv A C ADEM~CPRESS, INC. (LONISON) 24/28 Oval Road, London N W l

LIBRARY OF

7DX

CONGRESS CATALOG CARD

NUMBER:63-23237

ISBN 0-12-009716-8 PRINTED IN THE UNITED STATES OF AMERICA

82 83 84 85

9 8 7 6 5 4 3 2 1

LTD.

Contents List of Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

Boyd R. McCandless (1915-1975) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

Preface .....................................................................

XIII

Erratum ....................................................................

xv

...

The History of the Boyd R. McCandless Young Scientist Awards: The First Recipients DAVID S. PALERMO Text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I

Social Bases of Language Development: A Reassessment

I. 11. 111. IV. V. VI.

VII. VIII. IX. X.

ELIZABETH BATES, INGE BRETHERTON, MARJORIE BEEGHLY-SMITH, AND SANDRA McNEW Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Brief History of the Marriage: Love and Epistemology .............. Cognitive Inputs to Language ....................... .............. Social Inputs to Language ............................................... Verbal Interaction: “Motherese” . . . . . . . . . . ............... Conceptual and Methodological Confounds in Social-Causal Theories of Language Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Direction of Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Genetic Confounds . . . . . . . . . . . . . . . . . . . . . .............. Threshold Effects . . . . . . . . . . . . . . . . . . . . . . .............. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . .............. References ..............................................

8 8 12 15

35 48 54

59 61 64 68

Perceptual Anisotropies in Infancy: Ontogenetic Origins and Implications of Inequalities in Spatial Vision I. 11. 111. IV.

MARC H. BORNSTEIN Introduction .......................................................... Two Classes of Perceptual Anisotropy ..................................... Two Classes of Perceptual Anisotropy in Infancy ............................ Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V

77 79 85 108 115

vi

Contents

Concept Development MARTHA J. FARAH AND STEPHEN M. KOSSLYN I. Introduction . . . . . . . . . . . . ........................ 11. Information-Processing Theones . . . . . . . . . . . . . . . . . . . . . . 111. The Contents of Concept Representations ........................ IV. The Format of Concept Representations.. ......................... V. The Organization of Concept Representations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Conclusions ...................... References ..............................................

125

130 160 164

Production and Perception of Facial Expressions in Infancy. and Early Childhood TIFFANY M. FIELD AND TEDRA A. WALDEN 1. Adult Facial Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. The Infancy Literature ................................ ............... Ill. The Child Literature. . ............................ ............... Expressions by the Authors . . . . . . . . . . . . . . . . IV. Studies of Infant and C V. An Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References ..............................................

17 I 174 179

208

Individual Differences in Infant Sociability: Their Origins and Implications for Cognitive Development MICHAEL E. LAMB I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Measuring Sociability ....... IV. V. VI.

Explaining the Relationship between Sociability and Cognitive Performance . . . . . . Origins of Individual Differences in Sociability ............... Conclusion ..........................

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

213 215 226 230 236 237

The Development of Numerical Understandings ROBERT S. SIEGLER AND MITCHELL ROBINSON ................... ........................ 242 11. An Initial Study of Number Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 Ill. Preschoolers’ Knowledge of Counting ........................ 250 IV. Preschoolers’ Knowledge of Numerical ........................ 267 V. Preschoolers’ Knowledge of Addition . . 287 VI. Conclusions: The Development of Nu . . . . . . . . . . . . . . . 299 References . . . . . . . . . ........................ 308

I. Introduction . . . . .

Contents

vii

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

313

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

323

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

327

This Page Intentionally Left Blank

List of Contributors Numbers in parentheses indicate the pages on which the authors' contributions begin.

ELIZABETH BATES Department of Psychology, University of Colorado, Boulder, Colorado 80309 ( 7 ) MARJORIE BEEGHLY-SMITH Department ofPsychology, University of Colorado, Boulder, Colorado 80309 (7) MARC H. BORNSTEIN Department ofPsychology, New York University, New York, New York 10003 (77) INGE BRETHERTON' Department of Psychology, University of Colorado, Boulder, Colorado 80309 (7) MARTHA J . FARAH Department of Psychology and Social Relations, Harvard University, Cambridge, Massachusetts 02138 (125) TIFFANY M. FIELD Mailman Center f o r Child Development, University of Miami Medical School, Miami, Florida 33101 (169) STEPHEN M. KOSSLYN2 Department of Psychology and Social Relations, Harvard University, Cambridge, Massachusetts 02138 (125) MICHAEL E. LAMB Department of Psychology, University of Utah, Salt Lake City, Utah 841 12 (213) SANDRA McNEW Department of Psychology, University of Colorado, Boulder, Colorado 80309 ( 7 ) DAVID S . PALERMO Department of Psychology, The Pennsylvania State University, University Park, Pennsylvania 16802 ( I ) MITCHELL ROBINSON Department of Psychology, Carnegie-Mellon University, Pittsburgh, Pennsylvania 15213 (241) ROBERT S . SIEGLER Department of Psychology, Carnegie-Mellon University, Pittsburgh, Pennsylvania 15213 (241) TEDRA A . WALDEN3 Mailman Center f o r Child Development, University of Miami Medical School, Miami, Florida 33101 (169) 'Present address: Department of Human Development and Family Studies, Colorado State University. Fort Collins, Colorado 80523. *Present address: Program in Linguistics and Cognitive Science, Brandeis University, Waltham, Massachusetts 02554. 'Present address: Department of Psychology and Human Development, Peabody College, Vanderbilt University, Nashville, Tennessee 37203. ix

BOYDR. MCCANDLESS

Boyd R. McCandless (1 915-1 975)

This volume is dedicated to the memory of Boyd R. McCandless, who was a beloved teacher and important influence on the lives of both present editors of the Advances in Child Development and Behavior. This publication was founded by one of the present editors (LPL) and another McCandless student, Charles C. Spiker. Boyd himself was an author in Volume 10, published in 1975, the year of his death. It seemed fitting to us that a volume be dedicated to Boyd McCandless’ memory, for he toiled tirelessly to build this field of child development and behavior, and he influenced numerous teachers, researchers, and students. Early in 1979, as we thought about the content that we would choose to appear in such a dedicatory volume, it occurred to us that nothing could be more fitting than to invite contributions from the recipients of the Boyd R. McCandless Young Scientist Award created by the Division on Developmental Psychology of the American Psychological Association shortly after Boyd’s untimely death. The history and intent of this Award are described in David S. Palermo’s contribution. The first awards were made at the 1978 meeting of the American Psychological Association, and we decided to invite contributions from these recipients and the second year’s recipients, who were to be announced at the 1979 meeting of the American Psychological Association. We did invite those young scholars, most of whom did not know Boyd but whose scientific lives in child development could not help but have been affected by his nurturance of the field. This sort of intergenerational continuity in science and among scientists always fascinated Boyd. He was rather a keen observer of the history of child development and delighted in reciting the intellectual pedigrees of colleagues. In keeping with the tradition of the Advances, the contributions to this memorial volume are not organized around any theme, other than child development and behavior in general. Indeed, the diversity of topics in this volume is greater than that of some earlier volumes. We feel that this diversity makes a highly appropriate memorial to Boyd McCandless, whose own scholarly work covered a wide range of topics. He was a codeveloper, with Alfredo Castaneda and David S. Palermo, of the children’s form of the Manifest Anxiety Scale. McCandless, Castaneda, and Palermo used the CMAS, as it came to be called, in five studies published in 1956. This landmark series of studies was among the earliest research demonstrating the utility of learning theory for an understanding of child development and behavior, and helped establish experimental child psychology as a science in its own right. Among the other landmarks in the development of xi

xii

Boyd R . MrCandless

that science were Boyd’s publications with Charles C. Spiker on the operational definition of intelligence, in 1954, and on experimental child psychology research, in 1956. Boyd also had a long-continuing interest in delinquency and personality disorders, as reflected, for example, by a 1944 report on predelinquent children coauthored with Sidney W. Bijou, a 1962 paper on delinquency coauthored with John w.McDavid, and a 1972 paper on intervention. He also published extensively on intelligence, measurement, and education; but by far his major research interest and largest volume of publications were in the area of normal personality and social development. Our purpose here, however, is not to review Boyd’s work but rather to memorialize it and his personal contribution to the science of child development and behavior and to the practitioners of that science. To the memory of those contributions we dedicate this volume. Hayne W. Reese Lewis P. Lipsitt

Preface

The present volume in Advances in Child Development and Behavior is dedicated to the memory of Boyd R. McCandless and contains articles by recipients of the Boyd R. McCandless Young Scientist Award (see the dedication preceding this Preface and the History of the Boyd R. McCandless Young Scientist Awards). As in previous volumes, no attempt was made to organize the articles around a particular theme or topic, thus continuing the tradition of the Advances of providing a place for publication of critical reviews and scholarly speculation. 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 in being knowledgeable in areas that are peripheral to their primary focus of interest. Moreover, journal space is often simply too limited to permit publication of more speculative kinds of analyses that might spark expanded interest in a problem area or stimulate new modes of attack on a problem. This publication is intended to ease the burden by providing scholarly technical and speculative articles in which recent advances in the field are summarized and integrated, complexities are exposed, and fresh viewpoints are offered. The articles should be useful not only to the expert in the area but also to the general reader. Manuscripts are solicited from investigators conducting programmatic work on problems of current and significant interest. Contributions often deal 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 are accepted for review if submitted first in outline form. Whether invited or submitted, all articles receive careful editorial scrutiny, often with outside review. Invited contributions are automatically accepted for publication in principle, but usually require revision before final acceptance. Submitted contributions 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 benefited as well from the ...

Xlll

xiv

Preface

facilities of the Center for Advanced Study in the Behavioral Sciences at Stanford, where one of us (LPL) was located during the early work on the volume. We also wish to thank Dr. Robert K . Moore for his editorial assistance. With the publication of the present volume, Lewis P. Lipsitt will have stepped down as coeditor and Hayne W. Reese will be the sole editor. Correspondence should therefore be addressed to the latter. Hayne W. Reese Lewis P. Lipsitt

Erratum Advances in Child Development and Behavior Volume 15 In the article by William Fowler entitled “Cognitive Differentiation and Developmental Learning,” the text on pages 180 and 181 should appear on pages 178 and 179. The text on pages 178 and 179 should appear on pages 180 and 181.

xv

This Page Intentionally Left Blank

THE HISTORY O F THE BOYD R. McCANDLESS YOUNG SCIENTIST AWARDS: THE FIRST RECIPIENTS

David S.Palerrno DEPARTMENT OF PSYCHOLOGY THE PENNSYLVANIA STATE UNlVtRSlTY UNIVERSITY PARK, PtNNSYLVANIA

In 1974, the Executive Committee of Division 7 of the American Psychological Association discussed the possibility of providing some kind of recognition for outstanding young developmental psychologists. The Committee thereupon voted to establish a set of awards for that purpose. In the following year, Willard Hartup, who was president of the Division, asked me to present a proposal for implementing the awards to the Executive Committee at the 1976 APA meeting. I asked several members of Division 7 on the Pennsylvania State University campus to meet with me to consider the best way to proceed, the criteria for making such awards, and the general procedures for implementing the wishes of the Executive Committee. That group, which included Joseph Britton, Frank DiVesta, Dale Harris, and John Withall, developed a plan which I, in turn, presented to the Executive Committee at its 1976 meeting. The Committee offered advice concerning the proposed implementation procedures, and with the Committee’s comments in mind, I revised the proposal and presented it at the 1977 meeting, where it was approved by the Executive Committee and was subsequently set into motion, when the membership voted approval at the 1977 Business Meeting of Division 7. On December 5 , 1975, the year that the Pennsylvania State group was meeting to formulate the plans for these awards, Boyd R . McCandless died. At a memorial service honoring Boyd on December 8 , Walter Hodges said, Boyd McCandless gave lifc, hope, love, and knowledge to many. Such gifts will never perish. He is in each of our hearts, our intellects, our beings. Those of us he touched are legion. People all over the world have been nourished by his warmth, his advice, his reflection, his collaboration, his steadfast friendship. Intellectually and spiritually it felt so good to be with this man. His nurturance was always available.. . . His scholarly contributions also reflect his great spirit. He rose above his own human vulnerabilities to do research, to write, to speak, to teach about areas of great concern in our own and other societies. Honored by his profession many times. . . he never lost touch with his basic affection and concern for students, colleagues, friends, and family.

I ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR. VOL. 16

Copynght 0 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-W971&8

2

Duvid S. Palermo

Boyd Rowden McCandless was born in St. John, Kansas, on August 18, 1915. He obtained his bachelor’s degree (in English) at Fort Hays (Kansas) State College in 1936. Thereafter he found his way to the historic Child Welfare Research Station at the University of Iowa, where he earned his master’s degree in 1938 and his Ph.D. in 1941. Iowa became the central locus of his career, both because of the imprint on him of his Iowa training and the faculty members there with whom he was so close, and because he returned there in 1950 to become Director of the Station for a decade. Boyd’s death was a shock to a large number of developmental psychologists, who felt a strong sense of loss both personally and professionally. Boyd had been an important influence on the field of developmental psychology. His influence began when he was a graduate student in the Iowa Child Welfare Research Station where, among other activities, he participated as a research assistant in Kurt Lewin’s pioneering leadership studies. After completing his graduate studies in Iowa, he went to Wayne County Training School in Michigan, where he expanded his clinical and research experience for 2 years. After a 4-year stint in the U.S. Maritime Service during World War 11, he took a position at San Francisco State University, where he rose from Assistant to Associate Professor in the 2 years he was there (1946-1948). During that period, he encouraged two young men, Charles Spiker and Alfred0 Castaneda, to go on to graduate work in child psychology. In 1948 he moved to Ohio State University as Associate Professor. He remained there until 195 1, when he was appointed Director and Professor of the Iowa Child Welfare Research Station. While in Iowa, with the help of his former students Castaneda and Spiker, he developed a nationally recognized program in experimental child psychology. In 1960, he left Iowa to spend 2 years in Pakistan, where he was Professor and Chairman of the Psychology Department’s Institute for Education and Research at the University of Punjab. He returned to this country in 1962 as Professor of Education and Psychology and Director of the University School Clinic Complex at Indiana University. He also served as Chairman of the Special Education Department from 1963 to 1965. In 1966, he moved to Emory University, where he remained as Professor of Psychology and Education and Director of the Educational Psychology program until his death. Boyd’s influence on developmental psychology is clearly evident in his research contributions to the literature, his several excellent textbooks, his distinguished editorship of the first ten volumes of Developmental Psychology (1970-1974), and his presidency, in 1955, of Division 7 of the American Psychological Association. Boyd was also a Diplomate in Clinical Psychology. As is evident from Boyd’s record, he was a leader, as both teacher and researcher, in experimental, clinical, and educational aspects of developmental psychology.

Boyd R . MrCandless Young Scientist Awards

3

The loss of Boyd McCandless was, however, a very personal one to the many students and colleagues whose lives he influenced at the universities where he had taught. In addition, he had affected many others through his editorial labors and the individual contacts he made at meetings and conferences, where he expanded his friendships so naturally and easily. While Boyd was an obvious leader in the field at large, his influence on the younger members of the discipline was particularly notable. Boyd engendered in his students an excitement, a vigor, and a genuine effervescence which led to an unusually hard-working, interested, and productive group who spread into the various specialized areas of the field. Boyd was eclectic, and the diversity of his students’ interests reflects that characteristic. It was Boyd’s sincere interest in each student as a person about whom he cared, regardless of his or her interests, that brought out the best in those students. The qualities that enabled Boyd to encourage and stimulate students and others beginning their careers in the developmental field made it natural for us to accept Lewis Lipsitt’s suggestion to me that the Division 7 awards be named in Boyd’s honor. At its meetings in 1976, the Executive Committee accepted with enthusiasm the suggestion that the newly established award be called the Boyd R. McCandless Young Scientist Award. As approved in 1977, the Boyd R. McCandless Young Scientist Award was to take the form of an individually printed citation to be presented at the annual business meeting of the Division, following the presentation of the G. Stanley Hall Award for outstanding contributions to developmental psychology by a senior member of the discipline. It was stipulated that the criteria for the McCandless Awards should include evidence of a distinguished theoretical contribution, programmatic research effort, or dissemination of scientific developmental information to those outside as well as within the profession. The award is made on the basis of evidence of a continued effort rather than a single outstanding piece of work, and the contributions of the potential candidates in the year preceding the award are the primary focus of the selection committee’s attention, although the literature for at least 3 preceding years is also examined. Candidates are eligible for the Award only during the 7-year period following the receipt of their doctoral degree. A committee, appointed each year by the president of the division, is charged with the task of nominating one to five persons for the Award to the Executive Committee, which makes the final decision. Once the procedure for making the awards had been established, Frances D. Horowitz, who had assumed the presidency of Division 7 at the 1977 meeting, asked me to chair the nominating committee for the first 2 years of the awards. In addition, she asked Rachel K. Clifton, Martin L. Hoffman, Lewis P. Lipsitt, and Richard D. Odom to serve on the first selection committee. As a group, we solicited the names of young persons who fulfilled the criteria that had been established. In that year, 27 nominated persons were found to be eligible. The

4

Duvid S. Palerrno

curricula vitae of those persons were collected and examined by the five commit-

tee members independently. The members each selected the ten best candidates and rank-ordered the ten. The quality of the candidates and the difficulty of the task of making distinctions among them is reflected in the fact that 19 of the 27 persons nominated were on the combined list of the committee. There was, however, very good agreement among the committee members about the top four candidates. The committee therefore recommended to the Executive Committee of the Division that Marc H . Bornstein, Stephen M. Kosslyn, Michael A. Lamb, and Michael P. Maratsos be awarded the first Boyd R. McCandless Young Scientist Awards. At the 1978 American Psychological Association meeting in Toronto, I had the opportunity of making the presentations to those four outstanding persons. It was the culmination of an exhilarating but difficult taskexhilarating because once we turned our attention to the accomplishments of the young developmental psychologists, it became clear that we had a healthy, vigorous discipline with a large number of very talented and productive young researchers who are pushing to new frontiers our knowledge about developmental processes. I personally was particularly pleased by this, perhaps obvious, fact because it reflects so well the spirit of Boyd McCandless and perpetuates his tradition of encouraging the younger members of the profession. The decisions in the second year of the awards were no less difficult and no less exciting. The selection committee consisted of Rachel Clifton, Michael Lamb, Richard Odom, and Charlotte Patterson, and me, remaining as chairperson. Nominations were solicited from the committee members and the general membership. Out of the 30 nominees who were eligible, three emerged as consensual stars among the outstanding group of emerging scholars who came to our attention. They were Elizabeth Bates, Tiffany Field, and Robert S. Siegler. Again, I had the opportunity of presenting the awards to these individuals at the I979 American Psychological Association meeting in New York. Let me sketch briefly the backgrounds of the seven developmental psychologists who have achieved this distinction. They are clearly promising developmental psychologists in the tradition that Boyd McCandless so naturally encouraged among those who preceded them. Marc H. Bornstein was born on November 23, 1947, in Boston, Massachusetts. He received his B.A. degree from Columbia College in 1969 and his M.S. and Ph.D. degrees from Yale University in 1973 and 1974, respectively. His first position was at Princeton University, after a postdoctoral fellowship year spent at Yale and at the Max-Planck Institut in Munich. He is currently in the Department of Psychology at New York University. He has received a number of awards for his teaching and research, which has focused on visual perception and, in particular, the color vision capabilities of infants. In addition, however, he has been concerned with visual and time sensitivity in different social-cultural settings.

Boyd R . M~~Cunillrss Young Scientist A,rards

5

Stephen M. Kosslyn was born on November 30, 1948, in Santa Monica, California. He received his B.A. degree from the University of California at Los Angeles in 1970 and his Ph.D. degree from Stanford University in 1974. His first teaching position was at Johns Hopkins University. He joined the faculty at Harvard University in 1977, where he has remained. He has received grant and fellowship support from the National Science Foundation nearly continuously since he was an undergraduate student. His research efforts have focused on mental representations, with special concern for mental imagery and how children use mental imagery in their cognitive processing. Michael E. Lamb was born on October 10, 1953, in Lusaka, Northern Rhodesia (Zambia). He received his B.A. degree from the University of Natal, Durban, South Africa in 1972. In 1973 he came to the United States and enrolled at Johns Hopkins University, where he received an M.A. degree in 1974. The following year, 1975, he received a M.Phi1. degree from Yale University, and in 1976 he was awarded his Ph.D. degree, also from Yale University. He joined the faculty at the University of Wisconsin in 1976 as Assistant Professor and in 1978 moved to the University of Michigan at the same rank. In 1981 he accepted a position as Professor of Psychology, Psychiatry, and Pediatrics at the University of Utah. In 1976, 2 years prior to receiving the McCandless Award, he was presented the American Psychological Association Young Scientist Award. His research has been directed toward understanding the socialization processes of young children, with special interest in the father’s role in the context of socialization. Michael P. Maratsos was born on June 26, 1945, in San Francisco, California. He did his undergraduate work at Stanford University, where he received his B.A. degree in 1967. He moved to Harvard University from Stanford and was awarded his A.M. in 1968 and his Ph.D. degree in 1972. Since that time, he has been on the faculty in the Institute of Child Development at the University of Minnesota. His research efforts have been directed primarily toward an understanding of the acquisition of language by young children, which has led him to a serious study of linguistics as well as developmental psychology. He was unable to contribute to this volume because of other commitments. Among the 1979 awardees, the first, in alphabetical order, was Elizabeth Ann Bates. She was born in Wichita, Kansas, on July 26, 1947. She received her B.A. degree from Saint Louis University in 1968. She began her graduate work at the University of Connecticut and then moved to the University of Chicago, where she received her M.A. in 1971 and her Ph.D. in 1974 from the Committee on Human Development. Since 1974, she has been a faculty member at the University of Colorado. In addition, she has spent several research summers at the Institute of Psychology, CNR in Rome, Italy. Her research has been focused on the interrelations among cognitive development, symbolic capacity, and the acquisition and pragmatic use of language.

6

David S. Palermo

Tiffany Field was born on January 16, 1942, in La Crosse, Wisconsin. She was an undergraduate student at the University of Cincinnati, where she obtained a B.S. degree in 1963. She obtained an O.T.R. degree at Tufts University in 1965 and an M.A. at the same university in 1973. Between those two degrees she held several positions: she was a faculty member in the Group Psychotherapy and Group Work Training Institute of the Psychiatric Institute Foundation, Washington, D.C., Director of the Rehabilitation Center at the Psychiatric Institutes of America in Washington, and lecturer and consultant to the Educateurs at the Marseille School for Retarded and Emotionally Disturbed Children in France. In 1973 she moved to the University of Massachusetts, where she was awarded a Ph.D. in 1976. After 2 years as a faculty member at the University of Massachusetts and Research Director in the Department of Pediatrics at the Bay State Medical Center, she joined the faculty at the University of Miami Mailman Center for Child Development. Her research in recent years has focused on the psychophysiological and psychological characteristics of high-risk infants. Robert S. Siegler was born on May 12, 1949, in Chicago, Illinois. He attended the University of Illinois, where he received his B.A. degree in 1970. He enrolled at the State University of New York at Stony Brook in 1970 and was awarded a Ph.D. in 1974. He moved from there to become a member of the faculty at Carnegie-Mellon University. He spent one year at the Institute of Child Development at Minnesota, and recently he joined the faculty at the University of Chicago. His research has centered on the cognitive development of children, with particular emphasis on the development of scientific reasoning and related aspects of formal operational thinking and knowledge. AS evidenced in this volume, these seven persons are making significant contributions to developmental psychology. I look forward to the future development of these young scholars as they are joined by others who receive the honor bestowed upon them by Division 7 of the American Psychological Association through the Boyd R. McCandless Young Scientist Awards. Clearly, developmental psychology is in good hands, and our knowledge of child development will grow rapidly if these persons are representative of our future.

SOCIAL BASES OF LANGUAGE DEVELOPMENT: A REASSESSMENT

Elizabeth Bates, Inge Bretherton, I Marjorie Beeghly-Smith, and Sandra McNew DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF COLORADO BOULDER, COLORADO

I. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. A BRIEF HISTORY OF THE MARRIAGE: LOVE AND EPISTEMOLOGY

.

111. COGNITIVE INPUTS TO LANGUAGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1V. SOCIAL INPUTS TO LANGUAGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8 8 12 15

A. ATTACHMENT AND LANGUAGE .................................. B. PREVERBAL INTERACTION AND LANGUAGE . , . , . . . . . . . . . . . . . . . . . .

17 18

V. VERBAL INTERACTION: "MOTHERESE" . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

35

V1. CONCEPTUAL AND METHODOLOGICAL CONFOUNDS IN SOCIAL-CAUSAL THEORIES OF LANGUAGE DEVELOPMENT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. INTERNAL VERSUS EXTERNAL CAUSES . . . . . . . . . . . . . . . . . . . . . . . . . . . B. STRUCTURE VERSUS MOTIVATION . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48 49 50

VII. DIRECTION OF EFFECTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

54

VIII. GENETIC CONFOUNDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59

IX. THRESHOLD EFFECTS , . . . . . . , , , . . , , . . . , , , . , . . . . . . . . . . . . . . . . . . .

61

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

64

REFERENCES . . . . . . . . .

68

'Present address: Department of Human Development and Family Studies, Colorado State University, Fort Collins. Colorado 80523. 7 ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR, VOL 16

Copynght 0 1982 by Academic Press. Inc All rights of repduction in any form reserved. ISBN 0-12-0097168

8

Elizabeth Bates et a1

I. Introduction Around 1960, a “marriage” was formed between child language research and linguistic theory. By 1970, that marriage was in a state of crisis, and researchers in the two fields began to move in separate directions in their quest for descriptive principles and explanatory mechanisms. In particular, researchers in child language began to look outside linguistic theory for the “causes” of development, seeking both cognitive and social influences on language acquisition and language structure. In the 198Os, there are now signs of a possible reconciliation, with researchers once again seeking a unified theory of language acquisition and language structure (e.g., Bates & MacWhinney, 1979, from the “functionalist” viewpoint; Pinker, 1979, 1980, from the “formalist” viewpoint). As any good marriage counselor could tell us, this effort will work only if what has been learned in the intervening years is incorporated into the “remarriage,” including the successes and failures of research on cognitive and social bases of language development. In this contribution, we shall examine one aspect of the “open marriage” of the 1970s. Of the two affairs, one with cognitive and the other with social theories, we shall concentrate on the search f o r social influences on language acquisition. The cognitive dalliance will be reviewed briefly, only to illustrate how easy that whole effort was. This will be a critical reassessment, documenting our own failures and frustrations as well as those of others. The point, however, is nor that the search for social bases for language was in vain. It is possible, in principle, that language acquisition proceeds on the basis of minimal input, buffered from the vagaries of a social world (e.g., Chomsky, 1975; Wexler & Culicover, 1980). However, in a social species such as ours it would be odd indeed if there were no relationship between heart and mind in the acquisition of what is, after all, a communicative system. Our purpose here is to point out just how difficult this kind of research is, describing some conceptual and methodological confusion that has plagued social input studies. Above all, we would like to point out some possible ways around these problems in the future, so that what we learn about social development can be incorporated into a new union between child language research and linguistics. We shall start with a brief history of the union, and the reasons for its dissolution. Then, after a still briefer review of research on cognitive inputs to language, we shall turn to the social issues.

11. A Brief History of the Marriage: Love and Epistemology Several historians have likened the 1960s union between child language and linguistics to a “paradigm shift” or “scientific revolution” in Thomas Kuhn’s

Social Bascs oJ Language Development

9

use of those terms (Kuhn, 1970). A comparison of successive reviews in Carmichael's Manual of Child Psychology (McCarthy, 1954; McNeill, 1970) makes this change extremely clear. Prior to 1954, most relevant studies assumed a passive model of language acquisition in which (1) the principal structures of language are directly available in the environment, in the form of words and word associations, and (2) the mechanisms of learning in general and language learning in particular (e.g., imitation and exercise, reinforcement and behavioral shaping) are also driven by the environment. These were, of course, the epistemological assumptions that underlay most of American psychology at the time. Chomsky 's ( 1957) publication of Synrucric Strucrures triggered a radical shift in the assumptions, methods, and basic epistemology of both linguistic theory and child language research. The new model stressed the highly abstract nature of grammar in natural languages, based on structures so far removed from actual language production (and hence the data available to the child) that no human child could ever acquire them without knowing in advance what kind of things she or he was looking for. Through careful attention to spontaneous speech in children at various stages (e.g., Brown, 1973), and through a variety of experimental probes (e.g., Berko, 1958), child language researchers soon established that Chomsky was at least partially right: children acquire rules rather than associations, and they formulate hypotheses and make creative errors that are only remotely related to the language spoken by their parents (for reviews, see also Dale, 1976; Slobin, 1979). To some extent, Chomsky was a Hegelian hero, triggering an epistemological shift that was long overdue, and one that was occurring in several other fields at the same time. For example, although Piaget had been writing in French since the 1920s, his work was not widely read in American psychology until the 1960s (Flavell, 1963; Furth, 1966). Similarly, the writings of European ethologists (e.g., Lorenz, 1965; Tinbergen, 1951) began to influence American comparative psychology around the same time, with passive-inductive general learning models giving way to research emphasizing species-specific, innate biases toward particular stimulus-response-reinforcement configurations (for reviews, see Schwartz, 1978; Shettleworth, 1972). Finally, computer technology in the 1950s had a profound influence on the studies of information processing in human adults, human children, and nonhuman species. Edward Tolman's ideas concerning goals, expectations, and mental representation (e.g., Tolman, 1932) were once considered radical and/or mystical. They soon became commonplace in the new field of experimental mentalism called cognitive psychology (e.g., Bruner, Olver, & Greenfield, 1966; G. A. Miller, Galanter, & Pribam, 1960). However, the fact that history had prepared us for Chomsky does not diminish the importance of his influence on the new field of psycholinguistics. Although parallel changes were taking place across the board in American psychology, the study of language acquisition and language processing (as Lashley, 1951, had predicted) provided the most detailed and compelling evidence for active, con-

10

Elizabeth Bates et al.

structive, rule-based approaches to the nature of knowledge. Although Chomsky’s Syntuctic Structures and Aspects of u Theory of Syntax make poor reading as love letters, they were the inspiration for the close relationship between child language and linguistics for at least a decade (Chomsky, 1957, 1965). What went wrong with this fruitful union? As so often happens, the initial romance had drawn much of its fervor from the battle against a common enemy: an empiricist approach to the acquisition process that denied the fundamental creativity of the child. When the battle was won, and the two fields were left alone to develop their relationship, some asymmetries of power and interest remained that were difficult to resolve. Psychologists had essentially abandoned their previous processing theory without developing another one. In the meantime, generative grammarians derived their theories-which were not processing models, but synchronic descriptions of a very abstract system-from their own linguistic intuitions. Neither the descriptive data of child language nor the experimental data of adult psycholinguistics were admitted as evidence against models of grammar. If psychologists could confirm the current grammar, so much the better. If, however, their data were not compatible with existing linguistic theories, then the data were deemed irrelevant (reflecting “performance ” rather than “competence”-for a discussion of this issue, see Fodor, Bever, & Garrett, 1974; Slobin, 1979). A further problem was posed by Chomsky’s very strong innateness hypothesis (e.g., Chomsky, 1975). It was a tenet of standard transformational grammar that grammars comprise abstract categories and operations that are only remotely related to other types of cognition, perception, or social functions. The grammar was defned as separate from meaning. This idea seems counterintuitive to many nonlinguists, who cannot conceive of grammar as anything other than a system for mapping meanings into sounds. Consider, however, the example of algebra. Algebra is a useful and beautiful system precisely because it does not take specific numerical content into account. We have rules for transforming and equating expressions that operate entirely on “ x ” and “ y ” without regard to what “x” and “y” stand for. Suppose that algebra were based on the “meaning” of those symbols, with rules written in the form “for every x, unless x is a prime number greater than 7, apply the following.. . . ” The system would be cumbersome, difficult to learn, and difficult to use. Mathematical formulations are evaluated on the basis of their generality, coherence, and elegance. The same principle applies to the selection of competing formulations in generative grammar. The less they are constrained by specific content or meanings, the better they are judged to be. Because grammars are so far removed from meaning, Chomsky concluded that children must have innate, abstract, and languagespecific knowledge to acquire them. This knowledge could not be derived or constructed out of nonlinguistic structures. The innateness hypothesis left developmental psychologists in something of a dilemma. Nothing in their armory of cognitive, perceptual, or social theories was deemed relevant to the language

Socit~lBases of Language Drvelupmcnl

11

acquisition problem. The most important and interesting aspects of language development involved presumably innate structures, which are best understood by linguists, who know how to examine such structures. Worst of all, the grammatical theories themselves kept changing. Of course, from one point of view this is a sign of progress. However, just as engineers would be very upset if Newtonian mechanics were deemed irrelevant (as opposed to relevant only at certain levels of analysis), so psychologists who had placed their faith in generative grammar were embarrassed by providing evidence for the psychological reality of a structure that grammarians had abandoned last week (Valian, 1979). As Newmeyer (1980) has described in his history of generative grammar, around 1968 a rift developed within generative grammar itself. Basically, the disagreement revolved around the role of semantic factors in the structural component of the grammar. One side-referred to for a few years as “generative semantics”-suggested that the “deep structure” of sentences might be described wholly in terms of abstract components of meaning (e.g., causation, location, object-instrument relations). The other side-referred to as “the standard theory, ” “the extended standard theory, ” “the revised extended standard theory,” and so on-suggested instead that basic sentence relations must be described with autonomous syntactic structures (e.g., “subject of a main clause”) that are interpreted by an independent semantic component but are not structured by that component. On the generative semantic side, more and more information that was previously considered “extragrammatical” was written directly into the grammar: presuppositions about real-world relationships, social intentions of the speaker, “deep” cognitive components underlying lexical items (e.g., “kill” = “cause to become not alive”). It is not our business here to describe the fragmentation and reorganization of linguistic theory that followed (see Newmeyer, 1980, for one version of that history). The important point for our purposes is the effect that this rift had on child language research. As Slobin put it, developmental psycholinguists had “pied-piped’’ after linguistic theory for so long that it was difficult to consider doing without it. Yet generative grammar had provided an ill fit to other aspects of developmental psychology. At this point at least a subset of generative grammarians were attempting to write grammars in the very cognitive socialperceptual terms that psychologists had had to eschew to apply linguistic theory. The result was that many child language researchers felt free to seek their own nonlinguistic explanations for linguistic phenomena, using a background of generative semantic theory, together with much more traditional approaches to cognitive and social development (e.g., Bates, 1976; Bowerman, 1976; Bruner, 1975a, 197513; Edwards, 1973; Greenfield & Smith, 1976; MacNamara, 1972). When psychologists began to look “outside” of language proper for functional explanations of language development, they were united by the rather general contrasts between “inside” and “outside. Chomsky had proposed that ”

12

Elizabeth Bares et al.

the child has innate, a priori knowledge and that such knowledge is language specific, that is, not derived from any other psychological domain. The new approach to child language did not deny the fundamental insight gained in the 1960s concerning the creative, active, selective role of the child in language acquisition. Children do apparently have some idea of what they are looking for. The debate now is centered on the nature of that a priori knowledge, and many theorists now argue that children could construct or derive successive theories of grammar out of various kinds of nonlinguistic cues that lead them in a game of “hot and cold” into the basic categories of language. The affairs began at this point. From 1970 onward, a large number of theoretical articles appeared suggesting some possible sources for these nonlinguistic clues to language structure. Some theorists emphasized cognitive factors (e.g., Bates, Camaioni, & Volterra, 1975; Bloom, 1973; Bowerman, 1973; Bruner, 1975a, 1975b; Edwards, 1973; MacNamara, 1972; Nelson, 1974; Ryan, 1974; Slobin, 1973). Most of them drew heavily on Piaget’s theory of sensorimotor and early preoperational development (e.g., Piaget, 1962), arguing that the categories and operations derived from 2 years of experience in manipulating objects and observing the results of actions gave the child a head start toward understanding the referential and relational functions of language. Other theorists suggested that the case for cognition was overstated, and that many clues to language structure may be based instead on the results of social trunsactions that prepare the child to understand communicative intentions (e.g., the difference between commands and declarations of fact), basic reference and symbolization (i.e., the naming game), the semantic relations underlying grammar (e.g., agent-action-receiver relations developed within exchange games with adult caretakers), and the subtle rules of taking turns in discourse (Bretherton & Bates, 1979; Bretherton, McNew, & Beeghly-Smith, 1981; Bruner & Sherwood, 1976; Lock, 1978, 1980; Schaffer, Collis, & Parsons, 1977; Trevarthen & Hubley, 1978). In short, many aspects of language were viewed as implicit in the shared procedures of early social exchange. This social emphasis-often quite critical of the cognitive work (e.g., Lock, 1980)constitutes the second love affair. In both cognitive and social areas, theory has outstripped data to a startling degree. However, for reasons that it is hoped will become clear in this contribution, the case for a cognitive basis to language acquisition has been easiest to establish. Several reviews of the “cognitive hypothesis” are now available (e.g., Bates & Snyder, in press; Corrigan, 1978). Since our emphasis here is on social factors, we need only a very brief review for purposes of comparison.

111. Cognitive Inputs to Language We can discern two general approaches to research on cognitive factors in language acquisition: processing hypotheses and structural hypotheses.

Processing approaches involve demonstrations that basic milestones in language development (e.g., first words and first multiword utterances) are brought about by changes in information-processing mechanisms that are not unique to language. For example, Piaget (1962) has argued that first words, and the naming function in general, are just one manifestation of a more general shift into symbolic representation processing in all areas of cognition including imitation and symbolic play, causality and problem solving, visual imagery, spatial relations, and knowledge of the permanence of objects. The strongest experimental test of this hypothesis would be to manipulate the supposed representational capacity directly and observe the resulting effects on early language. In principle, this manipulation would include ( 1) training the symbolic function in nonlinguistic areas, and/or (2) removing the hypothetical symbolic substrate for language. On ethical grounds, the latter experiments are thankfully out of the question, and the more positive manipulations required for training experiments have proved extremely difficult to design and execute (see Leonard, 1979, for a discussion of some limited results). For these reasons, research on nonlinguistic analogs to the emergence of naming and of sentences has had to rely on correlational methods: “What capacities appear to be yoked to the appearance of first words and/or first sentences in normal children? ” Also, research with language-disordered children, in turn, can provide “mirror-image correlations” (Bates, 1979) showing whether the same nonlinguistic developments are implicated when aspects of language are absent or impaired. Studies of both types are reviewed by Corrigan (1978) with particular regard to naming and object permanence, and by Bates, Benigni, Bretherton, Camaioni, and Volterra (1977, 1979), J . Miller, Chapman, and Bedrosian (1977), and Bates and Snyder (in press) for several aspects of sensorimotor and early representational intelligence. At the one-word stage, Piaget’s proposals concerning a general stage shift have nor been supported. That is, some aspects of sensorimotor functioning (in particular, object permanence in hiding games and spatial relations) show no correlational relationship to the emergence of first words. Other nonlinguistic measures, in particular, symbolic play with objects (e.g., touching a telephone receiver to the ear or a cup to the lips at 13 months), imitation of sounds or gestures, and certain aspects of problem solving through tool use, do correlate significantly with very early language. These are also the same aspects of nonlinguistic cognition that correlate with degree of deficit in language-disordered children (Snyder, 1978; for reviews, see also Johnston, 1981; Leonard, 1979). Such results suggest a revision of the orthodox Piagetian stage theory, in favor of a neo-Piagetian model (e.g., Case, 1978; Fischer, 1980) in which specific tusks are correlated at specific moments in development on the basis of a more limited set of shared structures. The idea of “local homologies” (Bates, Benigni, Bretherton, Camaioni, & Volterra, 1977) or “domain-specific relations” (Fischer, 1980) is best illustrated with one small finding that has appeared in three independent studies (Corrigan, 1978; Halpern & Aviezer, 1976; J. Miller et

14

Elizabeth Bates et al.

al., 1977): although general measures of object permanence do not relate to the appearance of first words, the specific ability to follow several invisible displacements of an object does seem to coincide with the child’s ability to comment linguistically on the disappearance and reappearance of objects (e.g., “all gone”). This particular example will be useful later in our consideration of social inputs to language acquisition. Although more evidence is available concerning relationships between linguistic and nonlinguistic development at the one-word stage, there is also some evidence for nonlinguistic analogs to the passage into multiword speech. Around 20 months, children begin to join two or more words into a single intonational contour, in other words, a primitive sentence (Branigan, 1979). Reports by Nicolich (1977), Fenson and Ramsey (1980), McCall, Eichorn, and Hogarty ( 1977), and Case (1980) have all indicated that children begin to combine rwo or more gesturuf schemes into a single, planned motor unit in play at 20 months (e.g., pouring, then drinking, then wiping the mouth in “pretend lunch”). In our own research comparing language and symbolic play, we have found significant correlations between multischeme gestures and multiword speech-even in partial correlations removing effects of IQ,vocabulary, overall amount of play, and several other potentially confounding variables (Shore, O’Connell, & Bates, 1981). This kind of evidence suggests that the transition from first words into syntax involves some kind of a more general shift in the “chunking” and planning of higher-order motor schemes, rather than the maturation of an innate and language-specific grammatical capacity. Processing studies like these are relevant only to some very general mechanisms shared by language and aspects of nonlinguistic cognition. At the level of specific grammatical structures, reasoning by analogy and correlations must ultimately break down. For example, specific principles that control the ordering of morphemes in English verb phrases (e.g., auxiliaries come before main verbs) surely have no convincing analogies outside of language proper. They, however, do bear a relationship to the meaning component of language, and hence, indirectly to cognition. At the level of grammatical structure, the cognitive approach focuses not on extralinguistic processing mechanisms, but on semantic bases for specific grammatical rules within language. For example, it has been argued that young children can acquire ordering rules for subject, verb, and object in sentences by assimilating them to prior knowledge of the actoraction-patient meanings that correlate with word order most of the time in natural conversation. Such correlations are probabilistic, not absolute. For example, in a sentence like, “John’s consideration of the problem perplexed his co-workers, the subject “consideration” is neither a “thing” nor an “actor. ” Sooner or later children must acquire a grammar that is sufficiently abstract to permit comprehension and production of sentences like this. However, a number of studies in the 1970s demonstrated that the early word order regularities of Englishspeaking children can indeed be described with semantically based rules like

”

Social Bases nf Language Development

15

“actor-action-recipient-patient” (e.g., Bowerman, 1974; Braine, 1976). Also, the acquisition of productive morphology (e.g., plurals, past and present tense) seems to be guided by the child’s knowledge of the meanings that are encoded by these inflections (e.g., Block & Kessel, 1980; Slobin, 1979). The conclusion seems to be that children use cngnitively based meanings to “crack the code” of grammar in their particular language. In this regard, Pinker (1980) concluded: “The close correspondence between syntax and semantics in early child language is probably the most robust empirical finding in developmental psycholinguistics in the past decade” (p. 39). Some investigators, including Pinker (1980) and Brown (1973), believe that children must eventually abandon such semantically based grammars for a more abstract system, and that this “syntactic shift” may involve innate grammatical capacities of the type foreseen by Chomsky. Others maintain that adults continue to map grammatical structures onto a functional or semantic “core” that can handle sentences like “John’s consideration. . .” by a process similar to metaphoric extension (Bates & MacWhinney, 1981; McNeill, 1979; Schlesinger, 1974). In either case, the cognitive approach to the acquisition of grammar has succeeded in reducing at least some of the mystery about the a priori knowledge that children bring into the acquisition process. Evidence for creative errors and rule generation in children need not always lead to conclusions about innate and language-specific structure. To summarize thus far, the cognitive approach to language acquisition has met with some success in the last decade. At the processing level, the rather general cognitive stage hypotheses of Piaget ’s orthodox theory have not been supported by the data on language. However, a neo-Piagetian approach, emphasizing specific language-cognition relationships, has received support at the level of processing mechanisms that underlie first words and first sentences. At the level of specific structures or content in grammar, evidence also suggests that children “bootstrap” their way into the language by assimilating syntactic and morphological structures to existing cognitive-semantic meanings. This does not mean that we have done away with the hypothesis that children have innate, language-specific knowledge (cf. Curtiss & Yamada, 1978; Pinker, 1979). However, many arguments about innateness have been offered by process of elimination: the source of the child’s creative errors could not be determined, and therefore the conclusion was that they must be innate. At least a few of these arguments by mystery have been eliminated. Let us turn now to the case for social inputs to language acquisition.

IV. Social Inputs to Language Piaget has been criticized for neglecting social factors in his theory of cognitive development (e.g., Trevarthen & Hubley, 1978; Turiel, 1978). In turn, language acquisition theories that rely on Piagetian theory are subject to the

16

Elizabeth Bares et al

same criticisms (e.g., Lock, 1978). As discussed in some detail by Bates et al. (1977, 1979), the criticisms of Piaget are more valid in spirit than in detail. Piaget never argued that social factors are secondary or derived, nor that the child builds a theory of the world exclusively out of interactions with nonsocial objects. Instead, the operative structures of Piaget ’s cognitive theory are intended to be abstract and relatively content free, transcending the distinction between social and nonsocial. Operations such as displacement and reversibility should be equally valid for predicting the behavior of animate or inanimate objects under transformation, and hence the child’s mastery of those operations could derive from interactions with people as well as interactions with things. However, most of Piaget ’s examples in three books on sensorimotor development concentrate on the child’s naive experiments with inanimate objects. Moreover, Piaget ignored (with a few exceptions) the types of predictions that hold only for animate beings, that is, the peculiar causal behavior of creatures that have a will of their own versus the passive behavior of blocks and bottles that can be thrown from cribs without protesting. Lock (1978), Bruner (1977), Bretherton et al. (1981), and others have stressed instead the important role of objects that answer back, social objects that foresee the child’s needs and anticipate his or her actions, in the mastery of communication and meaning. Werner and Kaplan ( I 963) made a much more explicit effort than Piaget to consider social factors as “equal partners” in tbe emergence of the symbolic capacity. They stressed the role of the “addressor-addressee” relationship as critical to the development of a concept of reference (i.e., the relationship between subject and object via a symbolic vehicle). A lengthy segment of their book is devoted to a discussion of the dissolution of symbol-referent relationships in schizophrenia, viewed as the product of affective disturbances. By recreating a stable interpersonal relationship, they contended, a therapist can, in turn, recreate the primordial dialogue that permitted language to emerge in the first place. Unfortunately, their proposals did not go much farther than this. How social relationships effect their causal role in creating and maintaining the bond between sound and meaning simply is not clear. Of the “classic” developmental theorists, Vygotsky (1962) was the most direct in elaborating a powerful causal role of social factors in the development of language and thought. Vygotsky essentially agreed with Piaget about the origins of “tool thought”: sensorimotor intelligence derives in the first year or so of life from the child’s own active exploration of things in the world. At the same time, however, he stressed the powerful social motivation of the infant-to imitate and interact with other human beings, sharing the world as she or he sees it. In the period from 2 to 5 years of age, the child acquires language as a tool for social interaction. Gradually, linguistic structures begin to accompany the child’s own private interactions with the world-talking out loud to oneself in play, using language to announce, “guide,” and anticipate actions to come and to note the

Social Bases of Lnnguage Developmenr

17

results of actions that have been completed. Through this process of accompaniment, language gradually takes on a governing role, structuring and directing tool thought. This governing role becomes more rapid and efficient, “talking aloud, ” in turn, becomes abbreviated, and language moves “inside” to become the master of thought from there on. Because the basic categories, relationships, and beliefs of the culture are embodied in language, and because language is acquired as a tool for interaction with that culture, the result is that thought becomes “socialized. Of these three classic approaches to the emergence of early language, Vygotsky’s theory is most compatible with research in the 1970s on the social bases of language. Three major lines of research in the Vygotskian mode will be considered here: (1) correlations between “attachment” and language; (2) studies of preverbal interaction between child and caretaker; and (3) studies of “motherese, the particular type of language that caretakers address to children in the passage from preverbal to verbal communication. In all three cases, investigators have documented in impressive detail the systematic and subtle nature of caretaker-child interaction. Hence, these lines of research are valid and important in their own right, regardless of their implications for language in the child. However, many researchers have gone on to stress how the complex structure of interaction could facilitate the language acquisition process. The problem is, as we shall see, that very little evidence exists that these phenomena do have a measurable effect on the child. The circumstantial evidence is there-motive and opportunity-but the “smoking gun” demonstrating a causal link from social input to language is still missing. ”

”

A . ATTACHMENT AND LANGUAGE

Psychoanalysts beginning with Freud have pointed to the importance of events during the oral period for the development of “ego.” Insofar as language development is a part of “ego,” the outcome of the oral period should have implications for success in language development in the first two years. From a similar perspective, investigators influenced by attachment theory (Ainsworth, 1973; Ainsworth, Blehar, Waters, & Wall, 1978; Bowlby, 1969, 1973) have hypothesized that cognition and language may be influenced by the quality (i.e., harmony) of the mother-child relationship. A detailed review of research on language, cognition, and quality of attachment is available in Bretherton, Bates, Benigni, Camaioni, and Volterra (1979). As reviewed by Bretherton and colleagues, investigations of language-attachment relationships have been based explicitly or implicitly on two hypotheses: 1 . Infants who can feel assured of their mother’s availability and responsiveness, especially in situations of stress, can devote themselves more fully and

18

Elizabeth Bates et a1

enthusiastically to interacting with the physical environment. In other words, because these infants can use the mother as a secure base from which to launch out into the world, they can learn more about the environment and properties of objects by teaching themselves. This hypothesis will be referred to as the uttuchment-exploration hypothesis. 2. If infant and mother can achieve an interactive style in which harmony and affective synchrony predominate, the infant has an increased opportunity to acquire cognitive and communicative skills through interaction with the mother. This hypothesis will be called the uttuchment-teaching hypothesis. Both these hypotheses would lead us to expect significant correlations between measures of quality of attachment and various measures of cognitive and language development throughout infancy and early childhood. Table I, adapted from Bretherton et ul. (1979), summarizes results across all of these studies, listing the kinds of measures used and correlations obtained. We need not repeat the details of this review here. For our purposes, the most striking conclusion that Bretherton and co-workers reached is that very little evidence exists jiw signijicunt correlutions between attuchment and language merisures. Also, curiously, certain cognitive measures (e.g., symbolic play) form significant correlations with both attachment and language even though attachment and language do not correlate with one another. In other words, the language and social measures must overlap with different aspects of the variance in certain cognitive skills. Because the studies reviewed in Table I do not include cases of severe deprivation (e.g., children of psychotic mothers), they apply only to variation within the normal range. Pentz (1975) has interpreted the largely negative results on language and attachment to mean that language development is biologically “buffered” against variability of social input within the normal range. If we take the attachment-exploration hypothesis seriously, we would have to conclude that the amount of security a child obtains in normal circumstances is enough to acquire language, and that more is not necessarily better. if, however, we adopt the attachment-teaching hypothesis, then we would have to conclude that traditional attachment measures (e.g., degrees of proximity to the mother when strangers approach, greeting behavior when mother returns from a brief separation) d o not capture those aspects of the mother’s teaching role that are relevant for language development. This is essentially the conclusion that Bretherton and co-workers reached in their review, leading us to consider other aspects of mother-infant interaction in the first year. B . PREVERBAL INTERACTION AND LANGUAGE

A quite different approach to possible relationships between language and social interaction has emerged from microanalyses of infant social behavior in

TABLE I Findings of Studies Relating Strange Situation Classifications and Ratings with Cognition and Communication ( I Age at strange situation

Relationship of strange situation classifications or ratings with *

At age (months)

Significant relationships with standardized cognitive tests. Piagetian scales. and play measures I2 DQ (Griffiths Scale) I2 12 DQ (GrifFiths Scale) 8.5, 1 1 , 15 IQ (Stanford-Binet) 30 12 DQ (Bayley) 20.5 11.5 Person permanence 10. I 1 11 Person permanence 8 . 5 . 1 1 . 14.5 14.5 11.5 Object permanence 14.5 II Object permanence 17 Means-end relations (Uzgiris-Hunt) 1 1 . I2 18

12 12

18

I2

Tool use (with mother’s help) Quality of exploration Quality of cognitively mature play (combinatonal and symbolic play) Duration of symbolic play bouts Level. breadth. and frequency of symbolic play Frequency of combinatorial play

24 20.5 I2

24 1 1 . I2

Reference

Sample size

Bell & Ainsworth (1972) Bell (1978)

23 33

Main (1973) Bell (1970) Bell (1978) Bell (1 970) Bell (1978) Bretherton, Bates. Benigni, Camaioni. & Volterra (1979) Matas, Arend. & Sroufe (1978) Main (1973) Harmon. Suwalsky. & Klein (1979)

40 33 33 33 33 25

Matas. Arend. & Sroufe (1978) Brethenon, Bates. Benigni. Camaioni , & Volterra ( 1979)

48 25

48 40 36

I1 (corzririued)

TABLE I (continued) Age at strange situation

N

0

Relationship of strange situation classifications or ratings with

Significant relationships with communicative gestures and language 12 Clarity and variety of cornmunicative signals 12 Communicative gestures

12

12 28

Observed vocabulary size and imitation of words uttered by mother Number of morphemes per utterance Percentage of utterances that are questions

At age (months)

11

18 I2

DQ (Cattell) IQ (Stanford-Binet) DQ (Bayley) 1Q (Stanford-Binet) DQ (Bayley) DQ (Bayley)

Sample size

9-12

Bell & Ainsworth (1972)

23

1 1 , 12

25

18

Brethenon, Bates, Benigni, Camaioni, & Volterra (1979) Connell (1977)

20.5 36

Main ( 1973) Pentz (1975)

40 31

Clarke-Stewart (1973) Connell ( 1977)

38 26

Bell ( 1 978)

33

No significant relationships with cognitive tests, Piagetian scales, and play measures 12 DQ (Bayley) 11.5, 18 12

Reference

14

30 24 36 24 3. 8, I 1

Matas. Arend, & Sroufe (1978) Hock (1976)

55

48 164

12 12

12 12

Object permanence, space imitation (Uzgiris-Hunt) Object permanence, combined means-end + space scale Level and duration of symbolic play Combinatorial play

No significant relationships with language measures 12 Language competence 28 MLU. imitation of maternal utterances Comprehension Percentage of utterances that are questions Number of utterances 12 Referential versus expressive style. comprehension of commands Number of different words uttered. I2 number of words per utterance 12 Comprehension, referential and nonreferential words (maternal interview and observation variables)

10. 11, 12

25

13

Bretherton, Bates, Benigni, Camaioni, & Volterra (1979) Clarke-Stewart (1973)

20.5

Main (1973)

40

10, 12

Bretherton, Bates, Benigni, Camaioni, & Volterra (1979)

25

18 28, 36

Clarke-Stewart (1973) Pentz (1975)

38 31

36 28 28, 36 18

Connell ( I 977)

44

20.5

Main ( 1 973)

40

10. 1 1 . 12

Bretherton, Bates, Benigni, Camaioni. & Volterra (1979)

25

38

"Adapted from Bretherton, Bates, Benigni, Camaioni, & Volterra (1979. Table 5.2). 'In all cases where significant relationships were found, securely attached infants (strange situation classification B) always performed at a level superior to anxiously attached infants (strange situation classifications A, C). In those studies where interactive ratings were used. proximity and contact seeking during the reunion episodes of the strange situation correlated positively with cognitive performances, whereas avoidance-resistance correlated negatively with cognitive performance.

22

Elizuberh Bures et al

the first few weeks and months of life (Brazelton, Koslowski, & Main, 1974; Kaye, 1977, 1981; Sander, 1977; Stem, Beebe, Jaffe, & Bennett, 1977). Investigators in these studies have capitalized on new technologies (e.g., videoanalysis, autonomic recording), which permit very subtle dimensions of interaction to emerge: body postures, rhythms, direction of gaze, facial expressions, etc., in reciprocal exchange between mother and child from their very first encounter. A great deal has been learned in a short time about the infant’s predisposition to social interaction (e.g., an early preference for human faces, early recognition of the human voice) and the surprising degree of organization that can be seen in mutual gaze patterns, smiling, vocalization, touching, and virtually every aspect of social exchange in the infant-caretaker relationship. Some implications of these findings for subsequent language development are summarized in volumes edited by Lock ( 1 978), Schaffer (1977), and Lewis and Rosenblum (1977). Although details vary, the logic of the argument generally goes like this: I . The complex structure of mother-infant interaction in the preverbal period must serve some function for our species. 2. A compelling analogy can be observed between preverbal and verbal interaction along several dimensions including turn taking, shared reference to objects and events, the “semantic” structure of certain games and routines such as object exchange (i.e., exercise of agent-action-recipient-patient relationships), the “pragmatic” structure of communication (e.g., “performative” or speech act functions evolving in preverbal requests, commands), and even in “syntactic” structure (nested social sequences and recursive interaction ‘‘ruI es ’’) . 3. Given the above, at least one function of preverbal interaction may be to lead the child into the structure of language. As Shatz (in press) noted in her critique of social-causal theories, this approach to language acquisition is based primarily on “existence” or “plausibility ” proofs: certain phenomena that could affect language development can be shown to exist. Hence, we hypothesize that these phenomena do have an effect on the acquisition process. Bruner and colleagues have been particularly explicit about the different roles that preverbal social interaction could play in language development (Bruner, 1975a, 1975b; Bruner & Ninio, 1978; Ratner & Bruner, 1978). The following quotations characterize some of his theoretical claims (Bruner, 1977): Language acquisition occurs in the context of an “action dialogue” in which joint undertakings are being regulated by infant and adult. The joint enterprise sets the deictic limits that govern joint reference, determines the need for a referential taxonomy, establishes the need for signal-

23 ling intent, and eventually provides a context for the development of explicit predication. Thr cvolrctictn of language irselj; notably its 14tiivrrs.crlstructures, prubrrbly reJI1,ct.sthe reyuirivnents r ~ f ’ s u cjuint / ~ crction in our spvl-1ec.ie.s.(pp. 287-288, italics added) Mother and child develop a variety of procedures for operating jointly and in support of each other. At first, these joint actions are very direct, specially geared to assistance and comfort. In time, thc two of them develop conventions and requirements about carrying out joint tasks. The struc.tura of’ these tusks mcry shupe the structure of itiiricil gruntmur by the nature of the juintly held concepts it imposes. (p. 274, italics added)

Newson ( 1978) was somewhat less explicit about the direct analogy between grammar and interaction, but was equally emphatic about the causal role of communicative exchange: From the baby’s point of view it is only by being continually involved, as a participant actor, within an almost infinite number of such sequences that he is finally brought into the community of language. In short it is only because he is treated as a communicator that he learns the essential human art of communication. (p 42)

As we shall point out in more detail later, several very distinct hypotheses are actually included in these broad theoretical proposals. All of them, however, lead us to expect significant positive correlations between quality, amount, and type of mother-infant interaction in the first year and subsequent developments in language. What is the evidence? Unfortunately, few of the investigators involved in microanalytic studies of interaction also went on to test the relationship to language in a longitudinal, correlational study. This is understandable, in that these microanalyses are in themselves exhausting and time consuming. The interaction studies are of enormous value in their own right, regardless of their implications for language. Nevertheless, some very strong predictions have been offered that must ultimately undergo an empirical test. Table I1 is a summary of all the studies that we have been able to locate that included preverbal interaction measures similar to those discussed by Bruner, Schaffer, Trevarthen, and others, and included correlations of these measures with various aspects of language and general cognitive development. Many of these studies were designed for a different purpose, for example, to investigate social class differences, high- versus low-risk infants, or very general effects of “good mothering” on “infant competence, ” without offering hypotheses about specific language structures. Hence they constitute, in some cases, only an indirect test of the social-causal model. To help the reader to evaluate these findings, we have listed the kinds of measures used in each study in as much detail as possible, with separate sections for statistically significant and nonsignificant findings. As with attachment-language findings, the results in Table I1 are extremely disappointing for social-causal theorists. The significant relationships that have

TABLE I1 Preverbal Interaction Variables Related to Later Language and Cognitive Abilities A.

Preverbal variables Reciprocal interaction

Studied by

Specific measures a

Tulkin & Covitz (1975)

x length of

Cohen & Beckwith 1979)

3 longest interactions, interaction sequences, reciprocal vocalization Social

Cohen & Beckwith 1979)

Play factor Social Play factor

Mutual

Ramey , Farran, & Campbell ( 1979)

gaze factor Mutual social factor Mutual interaction (frequency and duration)

Significant results

At age

N

Context

Related to what later variables

Pos . or neg .

At age

30 (middle class)

Home

ITPA

+

6 yr

1 mo

50

Home

+

24mo

1 mo

50

Home

Gesell DQ, Sensorimotor scale' Gesell DQ, Sensorimotor scale Receptive language " Bayley DQ Gesell DQ Bayley DQ

+ +

+

24mo 24mo 24mo

+

24mo

+ +

+

25mo 24mo 25mo

+

24mo

+

36mo

1Omo

3 mo

50

Home

8 mo

50

Home

Gesell DQ

6mo

29 (exp. group)

Lab free Play

StanfordBinet

Maternal responsivity and attentiveness

Tulkin & Covitz (1975)

Kaye (1980)

Maternal "interactiveness"

Freedle & Lewis (1977)

Tulkin & Covitz (1975)

Response to fret, reciprocal vocalization (5%) Animated facial activity to infant Initiate vocal bouts ( S ) Entertains infant, gives object

30 (middle class)

Home

ITPA, PPVT "

+,+

6yr

6mo

50

Face-toface Play

PPVT

+

34mo

3 mo

97

Home

MLU

f

2 yr

30 (middle class)

Home

PPVT

+

6 yr

97

Home

MLU

-

2 yr

30 (middle class)

Home

PPVT

-

6yr

28

Lab free Play

StanfordBinet

+

36mo

101110

10mo

VI h)

Maternal lack of responsivity

Freedle & Lewis (1977)

Allows infant to end vocal bout (%)

3 mo

Maternal restrictiveness

Tulkin & Covitz (1975)

Prohibition ratio

10mo

Physical contact

Ramey, Farran, & Campbell (1979)

Physical contact (frequency and duration)

6mo

(control group)

(conrinued)

TABLE I1 (conrinued) A.

Preverbal variables

Studied by Tulkin & Covitz (1975)

Specific measures "

X !ength

Significant results At age 10 mo

30

Total time holding infant Kay (1980)

Frequency

6 mo

50

Ramey, Farran, & Campbell ( 1979)

Frequency and duration

4 mo

28

Tulkin & Covitz (1975)

Frequency

Ramey, Farran, & Campbell (1979)

infant plays alone (frequency and duration) Time over 2 ft away, time in playpen Time over 2 ft away

h)

OI

Time no1 interacting

Tulkin & Covitz (1975)

Context Home

(lower class) 30 (middle class)

held

Maternal vocalization

N

10 mo

(control group) 30 (middle class)

Face-toface Play Lab free Play Home

Related to what later variables

Pos . or neg.

PPVT, ITPA

-

PPVT

t

Language production

-

StanfordBinet ITPA

6 mo

28 (control group)

Lab free Play

StanfordBinet

10 mo

30 (middle class) 30 (middle class)

Home

PPVT

ITPA

At age

6 yr

+

26 and 36 mo(comDined) 36mo

+

6 yr

+

36mo

B.

Preverbal variables Reciprocal interaction

Studied by Cohen & Beckwith (1979)

Specific measures

Nonsignificant results

At age

Mutual gaze factor

I mo

Mutual gaze factor Mutual social factor

3 mo

N 50

Context Home

8 mo

4 N

Burchinal & Farran (1980)

Farran & Ramey ( 1980)

Bakeman & Brown (1980)

Kaye ( 1980)

Mutual play (frequency and duration) Dyadic involvement factor Amount joint activity Amount reciprocal interaction

6 mo

51

6 mo

60

0. I. 3,

43

9 mo

6 mo

50

Lab free play Lab free Play Home feeding

Face-toface Play

Related to what later variables Gesell DQ. sensorimotor scale, receptive language Bayley DQ

24 mo

Sensorimotor scale, receptive language Bayley DQ Bayley DO Stanford-Binet

24 mo

Bayley DQ Stanford-Binet

6. 18 mo 48 mo

Bayley DQ Stanford-Binet

12, 24 mo 36 mo

Conversational ability, language production PPVT. puzzle test

26. 30 mo

25 mo

25 mo 6 . 18 mo 24 mo

34 mo

TABLE 11 (continued) B.

Preverbal variables

Studied by Rarney, Farran, & Campbell (1979)

Tulkin & Covitz (1975)

N

m

Maternal responsivity and attentiveness

Freedle & Lewis (1977)

Specific measures Mutual interaction (frequency and duration) R length of 3 longest interactions, interaction sequences reciprocal vocalization VOCd responsivity

Kaye ( 1980)

Bakeman & Brown (1980)

Cohen & Beckwith (1979)

Vocal and emotional responsivity Responsive holding factor

Nonsignificant results

At age

6 rno

N 28 (control group)

10 mo

30 (lower class)

Context

Related to what later variables

Lab free Play

Stanford-Binet

Home

ITPA, PPVT

30 (middle class)

36 mo

PPVT

6 mo

97

Home

MLU

24 mo

6 mo

50

Face-toface Play

26. 30 mo

0, 1 , 3, 9 mo

43

Home feeding

Conversational ability, language production PPVT, puzzle test Bayley DQ Stanford-Binet

1 mo

50

Home

Gesell DO, sensorimotor scale, receptive language

24 mo

34 mo 12, 24 mo 36 mo

Burchinal & Farran (1980)

Tulkin & Covitz (1975)

Social and vocal responsivity factor Contingent responsivity

3 mo

6 mo

51

Responsivity to 10 mo fretting (I), vocal I ) responsivity ( Responsivity to fretting ( I )

30

Lab free Play Home

Maternal restrictiveness

Tulkin & Covitz (1975)

Cohen & Beckwith (1979)

Entertains, gives objects to infant

Control factor

25 mo

Bayley DQ Stanford-Binet

6, 18 mo 24 mo

PPVT, ITPA

(lower class) PPVT

30 (middle class)

wvr. ITPA

Vocal responsivity (76) Maternal “interactiveness”

Bayley DQ

10 mo

3 . 8 mo

30 (lower class) 30 (middle class)

Home

50

Home

ITPA. PPVT

ITPA

Gesell DQ. sensorimotor scale, receptive language Bayley DQ

24 mo

25 mo (continued)

TABLE I1 (continued) B.

Preverbal variables

Studied by Tulkin & Covitz (1975)

C W

Maternal noti restrictiveness

Physical contact

Specific measures Prohibition ratio

Nonsignificant results

At age 10 mo

N

Context

30 (lower class) 30 (middle class)

Home

Related to what later variables ITPA, PPVT

At age 6 Y'

ITPA

Ramey, Farran. & Campbell (1979)

Absence of punishment

6mo

57

Lab free Play

Stanford-Binet

36 mo

Cohen & Beckwith (1979)

Floor freedom factor

8mo

50

Home

24 mo

Tulkin & Covitz ( I 975)

Time with no barriers

10 mo

60

Home

Gesell DQ, sensorimotor scale, receptive language Bayley DQ ITPA, PPVT

25 mo 6 Yr

Ramey, F m a n , & Campbell (1979)

Physical contact (frequency and duration) Responsive holding factor, stressful holding factor

6mo

29 (exp. group)

Lab free Play

Stanford-Binet

36 mo

1 mo

50

Home

Gesell DQ, sensorimotor scale, receptive language

24 mo

Cohen & Beckwith (1979)

2 N 1ci

8

m

% W

8 0

31

TABLE U (continued) B.

Preverbal variables

N w

Time !lor interacting

Studied by Ramey, Farran, & Campbell ( 1979)

Burchinal & Farran (1980) Bakeman & Brown (1980)

Tulkin & Covitz (1975)

Specific measures Infant plays alone (frequency and duration) Noninteraction time Independent infant activity Time over 2 ft away, time in playpen Time over 2 ft away Time in playpen

Nonsignificant results

At age

N

Context

Related to what later variables

At age

Stanford-Binet

36 mo

29 (exp. group)

Lab free.

6mo

51

0, 1.3,

43

Lab free Play Home feeding

Bayley DQ Stanford-Binet Bayley DQ Stanford-Binet

6, 18 mo 24 mo 12, 24 mo 36 mo

Home

ITPA

6 Yr

6mo

9 mo 1Omo

30 (middle class)

30 (lower class)

Play

PPVT

ITPA, PPVT

Maternal involvement

Ramey, Farran, & Campbell (1979) Burchinal & Farran ( 1980) Bakeman & Brown (1980)

Maternal teaching

2

Variability of interaction

HOME‘ maternal involvement Passive involvement Amount maternal ’ ‘effon ’ ’

6mo

57

Lab free Play

Stanford-Binet

36 mo

6mo

51

0, 1 , 3 , 9 mo

43

Lab free Play Home feeding

Bayley DO Stanford-Binet Bayley DQ Stanford-Binet

6, 18 mo 24 mo 12, 24 mo 36 mo

Gesell DQ, sensorimotor scale, receptive language Bayley DQ Bayley DQ Stanford-Binet

24 mo

Bayley DQ

12, 24 mo 36 mo

Cohen & Beckwith (1979)

Intellectual stimulation factor

8 mo

50

Home

Burchinal & Farran (1980)

Toy demonstration

6mo

51

Lab free

Bakeman & Brown (1980)

Same

Play 0, I , 3, 9 mo

43

“These measures are those reported by individual investigators. bIllinois Test of Psycholinguistic Abilities. ‘Casati-Lezine Sensorimotor Series (Kopp, Sigman, & Parmelee, 1974). “Labeling production (Beckwith & Thompson, 1976) and comprehension. ‘Peabody Picture Vocabulary Test. ’Caldwell (1970) Home Observation for Measurement of the Environment Inventory.

Home feeding

25 mo 6, 18 mo 24 mo

34

Elizobeth Butes et al.

been reported are mostly in the predicted direction: “better” relationships correlate with “better” or more precocious linguistic-cognitive development. However, nonsignificant results outweigh the positive findings. Given the wellknown bias of editors against the publication of nonresults, a conservative conclusion would be that the effects of interaction variables on language development are either very weak or very difficult to prove for some other reason. Could we argue that these studies failed to include the relevant dimensions of preverbal interaction, that is, the ones that Bruner, Schaffer, Trevarthen, and others have delineated in terms of analogies to language? Certainly the studies reported can be claimed to have face validity with the variables discussed by most social causal theories: time spent in mutual gaze, frequency and duration of interactions around a toy or other object, amount of “animated facial activity of mother to infant,” and so on. Of all these longitudinal investigations, the most relevant for social bases of language may be Kaye’s (1980) longitudinal study of 50 infants from birth through 34 months. Kaye has been one of the pioneers of microanalysis in mother-infant interaction, and his measures include many different dimensions of exchange during breast feeding, early caretaking, and play. Altogether 48 interaction variables were chosen for his longitudinal analyses, including those that had proved to be interesting in his own previous microanalytic studies and variables that had been described by other researchers. Language measures through 34 months of age included the Peabody Picture Vocabulary Test, measures of language production, and “conversational ability” (involving number of conversational turns taken by the child, linked to utterances by mother and experimenter). When effects attributable to educational status of the mothers were partialed out of 788 longitudinal correlations, only 50 relationships, 6.7%, were significant at the .05 level. Kaye’s (1980) own conclusions were as follows: Considering that many of these tasks and contingency scores were selected deliberately because of their structural similarity to one another-that is, their face validity-we have to conclude that the lack of continuity among the measures is remarkable and compelling. Beyond the social class differences within our sample (and most of our variables were free even of those differences), we have found practically no relationship between the individual performances of our mothers and infants in any one situation at one age and their performances in another situation at another age. (p. 13)

Kaye’s interpretation echoes Pentz’ (1975) explanation of similar results for the attachment-language studies: the evolutionary “fit” of mother to child has become so good that everybody does what has to be done to get communication under way. Variations within the normal range have no measurable effect on basic developments in language. We shall return later to some further methodological and conceptual problems in interpreting the disappointing results summarized in Tables I and 11. At this point, let us turn to the third major line of social-causal research: the effects of “motherese” on language acquisition.

Socicrl Buses i>f Lcinguaxe Development

35

V. Verbal Interaction: “Motherese” One of Chomsky ’s main points concerning the innateness of grammar was that linguistic input to children is much too poor to yield the rich grammatical theory that children ultimately derive: A consideration of the character of the grammar that is acquired, the degenerate quality and WdrrOW!y limited extent of the available data. the striking uniformity of the resulting grammars, and their independence of intelligence, motivation, and emotional state over a wide range of variation. leave little hope that much of the structure of the language can be learned by an organism initially uninformed as to its general character. (1965, p. 58)

However, a large body of research on the nature of adult speech to children has demonstrated that language input is much better than Chomsky believed. Comparisons of adult speech to other adults with their speech to very young language-learning children show that children receive a simplified, repetitive, and often exaggerated form of the adult code (for reviews, see Berko-Gleason & Weintraub, 1978; Snow & Ferguson, 1977). Modifications occur at every level of the system: intonation, phonology, syntax, semantics, gesture, and broader aspects of the organization of discourse. To offer just a few of the more exotic examples, Williams ( 1979) has shown that Spanish, English, and Chinese mothers tend to exaggerate precisely those phonological distinctions, in their respective languages, that prove in spectrographic analysis to be the most difficult to discriminate. Obviously, mothers do not have access to spectrographic analyses in advance, and so the process must be as “natural” as it is precise. Mandarin is a tone language, using the pitch or tone level of a word to make semantic distinctions. Mothers speaking Mandarin apparently emphasize the most difficult tone distinctions for their children. Since the semantic use of tones is a relatively rare feature in human languages, this “decision” by Chinese mothers must be based on a powerful and pervasive tendency that is applicable to m y feature of language that might pose problems for the child. Newport, Cleitman, and Gleitman (1977) have called this special type of language “motherese”-although research has shown that the language is spoken by fathers, strangers, and even 4-year-olds interacting with 2-year-olds (Shatz & Gelman, 1973) and 4-year-olds pretending to talk to 2-year-old dolls (Sachs & Devin, 1976). Because these modifications are so pervasive, some investigators have suggested that motherese evolved in our species as an unconscious, spontaneous “teaching language” (Moerk, 1975). This is essentially the same kind of plausibility argument described earlier for phenomena in preverbal interaction: motherese exists everywhere, and hence must exist for some reason; motherese ~ ~serve l asdan aid to language acquisition, and thus we hypothesize that it does have an effect. But does it? Unfortunately, the existence of the phenomenon has been documented much better than its purported function. A number of investigators

TABLE 111 Talk to Children: Summary of Effects and Noneffects Categorical adult variable General verbal stimulation

Specific adult variable Frequency/ variety of maternal speech

Total number of words spoken by

Experimental method

Age of child

Regression and correlation: infant and mother simultaneously sampled (sample of low socioeconomic status, balanced by race and sex; all firstborns)

17-17.5 mo

Cross- lagged correlations

T , = 10.512.5 mo T 2 = 1718 mo

Correlations between maternal and child measures

T, = 28 mo T , = 36

N

Child measures

Significance-direction

Studied by

Relationship of stimulation Clarke-Stewan 36 Language comto language and general (1973) petence factor, competence: positive which included Within a regression on comprehension. child competence, verbal expressed vostimulation loaded most cabulary, and highly and was particularly response to related to the language questions in a subfactor probe, as well as spontaneous measure of number of words used (part of a general competence factor) Clarke-Stewart Positive relationship; 36 Bayley mental ( 1973) apparent direction of scales effect mother to child; author extrapolates to language per se from Bayley. since language measures were not comparable from T, to T 2 . Pentz (1975) Mother at 28 mo with 31 28 and 36 mo: 36 mo comprehension: amount of speech produced; frepositive

mother

at T , ; cornlations of maternal measures at TI and T2 with child at T , (sample white and middle class)

mo

Noncontingent speech (maternal initiation) Contingency/ responsivity (proportion of child utterances to which mother responds plus proportion of maternal utterances that are responses)

W

4

Syntactic complexity

MLU

Partial correlation with child variables 6 mo later, controlling

T, = 1227 rno

Others: ns quencies of questions and imitations; response contingency; mean length of utterance (MLU) 36 mo only: rest of linguistic comprehension (syntactic contrasts) Language factor: positive

15 MLU Auxiliaries Noun phrase Inflectionsverb phrase

Clarke-Stewart ( 1973)

All ns

Pentz (1975)

All ns

Newpon, Gleitman, & Gleitman (1977)

(continued)

TABLE 111 (continued) Categorical adult variable

W

Specific adult variable

Experimental method for age and initial child level (sample = middle-class girls) Correlations between simultaneously sampled mother-child measures (sample: middle class and advanced) Correlation: simultaneously sampled mother, infant, and older children from rural Africa: Kokwet, Kenya; includes 13 mothers and 12 older children Correlation of child MLU with earlier maternal measures

Age of child

N

Child measures

Significance-direction

Studied by

Verbs-utterance Noun phrasesutterance Varies: 19-32 mo

24-42 mo

T , = 18 mo T , = 27 mo

10 Receptive score MLU Comprehensibility Typeitoken ratio Longest utterance Age 20 MLU and adjusted MLU

MLU

MLU, receptive scores, longest utterance. and age: positive. Comprehensibility and typeitoken: ns

Cross ( 1977)

Mother-infant and older child-infant: positive for both variables

Harkness (1977)

Negative

Furrow ( 1979)

Difference in maternal child MLU

Long utterances ( S ) Single-word utterances (76) Propositionshtterance

Noun phrasesutterance Noun phrase length in morphemes

Maternal at 28 mo with 36 mo comprehension: positive Others: ns Receptive scores, longest utterance, and MLU: negative Others: ns MLU for mother-child at 28 mo, for mother-child at 36 mo, and for mother at 28 mo and child at 36 mo: negative Imitation for mother-child at 28 mo: negative Others: ns Typekoken: ns Others: positive Typeltoken: negative Others: ns Receptive and comprehensibility: positive Others: ns All ns All ns

All ns. 28 mo maternal with 28 mo MLU: negative Others: ns

Pentz ( 1975)

Cross (1977)

Pentz ( 1975)

Cross (1977) Cross ( 1977) Cross (1977)

Furrow ( 1979) Newport, Gleitman, & Gleitman (1977) Cross (1977)

Pentz (1975)

(conrittued)

TABLE I11 (conrinued) Categorical adult variable

Specific adult variable Modifiers/ utterance Pronouns1 utterance

Noun/pronoun ratio Verbs/ utterance

Verb phrase length in morphemes Reverb complexity

Specific linguistic forms

Declarative

Experimental method

Age of child

N

Child measures

Significance-direction

Studied by

All ns

Furrow (1 979)

Receptive, cornprehensibility: positive Others: ns MLU: negative All ns

Cross ( 1977)

MLU: negative Mother-infant: positive Older child-infant: ns MLU and adjusted MLU All ns

Furrow (1979) Harkness (1977)

Age: positive Others: ns

Cross (1977)

All ns

Newport, Gleitman, & Gleitrnan (1977) Harkness ( 1977)

Older child-infant: positive for both Mother-infant: both ns All ns All ns

Furrow (1979) Pentz ( 1975)

Pentz ( 1975)

Cross (1977) Furrow (1979)

Questions Yes-no questions

Imperative

P

Negative imperative WH-questions

Maternal at 28 mo with 36 mo comprehension test Mother-infant and older child-infant: both ns Number of auxiliaries/ verb phrase: positive Others: ns All ns MLU: positive Number of auxiliariesiverb phrase: negative Others: ns All ns All ns Others: ns 36 mo with 36 mo maternal response contingency: negative Older child-infant: negative for adjusted MLU but ns for MLU Mother-infant: both ns All ns

Pentz ( 1 975)

Ns

Furrow (1979) Newport, Gleitman, At Gleitman ( 1977) Cross (1977)

All ns

MLU and age: negative Others: ns

Harkness ( 1977) Newport, Gleitman, & Gleitman (1977) Cross (1977) Furrow (1979) Newport. Gleitman, & Gleitman (1977) Cross (1977) Furrow ( I 979) Pentz (1975)

Harkness (1977)

Cross (1977)

(continued)

TABLE 111 (conrinued) Categorical adult variable

Specific adult variable Where is NP? What’s that?

P N

Auxiliaryfronted questions Tag questions Rising intonation questions Questions Interjections Deixis

Auxiliary deletions in yes-no questions Semantic Imitations relationship of adult utterances to child utterances

Experimental method

Age of child

N

Child measures

Significance-direction

Studied by

All ns Typdtoken and age: ns Others: negative All ns

Cross ( 1 977) Cross ( 1 977)

All ns All ns

Cross (1977) Cross (1977)

All ns All ns All ns Inflectionshoun phrase: positive Others: ns All ns All ns MLU: positive

Pentz (1975) Cross (1977) Cross ( 1977) Newport. Gleitman, & Gleitman ( 1977) Cross ( I 977) Pentz ( 1975) Furrow (1979)

All ns

Cross ( 1 977)

Older child-infant: both negative Mother-infant: both ns

Harkness (1977)

Cross ( 1977)

Paraphrases

Transformed repetitions Total repetitions (includes above types plus four others: partial exact, sequential, and nonsequential) Expansions

Experimental: expansion. modeling, and control 60.5 HR sessions/ 12 weeks Experimental: 12.5 hr sessions

Varies, 29-37 mo

12 6 measures including MLU and imitation

Varies, 30-46 mo

24 Sentence imitation

Semantic extensions Recasts

Experimental 3 groups (recast new sentence. control)

Varies. 32-40 mo

27 MLU (in words). no. elements/ noun phrase.

Receptive, longest utterance, MLU: negative Others: ns Typdtoken: negative Others: ns Receptive, longest utterance. MLU, comprehensibility, age: negative Typeltoken: ns

Cross ( 1977)

All ns

Newport, Gleitman, & Gleitnian (1977) Harkness ( 1977)

Mother-infant: negative Older child-infant: ns MLU and adjusted MLU All: negative All ns except imitation: modeling and control equivalent; expansions lowest Ns

Cross ( 1977) Cazden ( 1965)

Feldman (1971)

Cross (1977) Receptive: positive (with pronominal extensions) Others: ns Recast significantly higher Nelson. than controls on all measures Carskaddon, & but MLU and noun phrase Bonvillian ( 1973) recast versus new sentences

TABLE 111 (continued) Categorical adult variable

Specific adult variable

Experimental method

Age of child

intervention for I 1 weeks. 20 min/session twice weekly (middle class)

P

Linguistic teaching devices

Experimental: 2 groups both without complex questions or complex verb phrases, groups given recasts of one type or the other Sampled over 15 min: due to low frequency. a composite measure was devised. including expan-

2.5 yr

N

Child measures

no. verb elementsherb construction. auxiliaries/ verb construction. sentence imitation I2 Complex verb phrase and complex question usage

Significance-direction

Studied by

approaches significance on verb and auxiliaries; new sentences and controls not significantly higher on any measure

Children receiving verb recasts produced significantly more complex verbs than "question" children and vice versa

Nelson ( 1977)

28 mo maternal Pentz (1975) with child 28 mo MLU: positive Number of utterances, questions. and response contingency: positive 36 rno maternal with 36 mo

sions, recasts, corrections, occasional questions, prompts and checks

Function of adult utterances

imitation and response contingency: positive Others: ns

Semantically new utterances Novel, isolated utterances Continuous dialogue

All ns

cross ( 1 977)

All but typehoken: positive Typeitoken: ns

Cross ( 1 977)

Mother-infant and older child-infant: both ns

Harkness (1977)

Direct request

Mother-child both at 36 mo Pentz ( 1 975) tendency to respond verbally to mother: negative Others: ns All ns

$ Indirect request Elicit information

Focus attention

Labeling

All ns Mother-infant and older Harkness (1977) child-infant: negative MLU and adjusted MLU Mother-child both at 36 mo Pentz (1975) number of child utterances and tendency to respond: negative Others: ns Pentz ( 1975) All ns (corititiued)

TABLE 111 (conrinued) Categorical adult variable

c P n

Age

Specific adult variable Property of object Instruction in object use Reinforce utterance Reinforce act Verbalize child act

Verbalize child feeling Verbalize maternal act

Verbalize

Experimental method

of child

N

Child measures

Significance-direction All ns

All ns All ns

All ns MLU at 36 mo: positive with 36 mo maternal verbalization of either child or maternal act (combined measure) Others: ns All ns All ns MLU at 36 mo: positive with 36 mo maternal verbalization of either child or maternal act (combined measure) All ns

Studied by

maternal feeling Play idea Engage in play Prohibition Permission Score of referential field

All ns All ns All ns All ns

Immediate references

All ns

Nonimmediate events

Receptive. longest utterance, MLU, and age: positive Others: ns Comprehensibility: positive Others: ns

Persons/ objects present Child events

Mother events Child and mother events “Adapted from McNew (1981a). *Details of each experiment are given the first time it appears in the table

Receptive, longest utterance. and comprehensibility: negative. Others: ns All ns Age: negative Others: ns

Cross ( 1977)

48

Elizabeth Bates el al.

have attempted to assess the relationship between adult input and child language development. The correlational and experimental findings are summarized in Table 111, adapted from McNew (in preparation, a). Only a handful of correlational studies exist, and these vary considerably in method and range of measures. Compared with the preverbal results in Tables l and 11, more significant, positive findings in studies of motherese can be observed. However, as we shall point out in more detail later, these results are often very difficult to interpret. Critical methodological and statistical controls are sometimes missing, and the available methods usually do not allow causal inferences. One conclusion is clear from the studies in Table 111: more languuge input from adults is related to more and better language in children. For example, Clarke-Stewart ( 1973) reported significant positive correlations between the total amount and variety of language stimulation provided by the mother, and several measures of linguistic competence in the child at the same session. Pentz (1975) reported significant relationships between frequency of maternal stimulation at early sessions, and measures of child comprehension 8 months later. Finally, as reviewed in Bates ( 1975), children whose linguistic input comes primarily from adults (as opposed to peers and older children) are at an advantage in language learning. However, these quantitative relationships do not prove the structural, “teaching” claims of the motherese literature that simplification, repetition, and exaggeration clarify the language learning process. To establish such claims, relationships must be shown between specijk q p e s of language input and spec $ ~outcomes in acquisition. To summarize, although the social-causal theories are at least as convincing on conceptual grounds as the cognitive-causal theories discussed earlier, their empirical grounding is less clear-cut. Shall we conclude that social inputs to language are minimal? It may be that the preverbal and verbal interaction phenomena uncovered in the last few years exist primarily in the service of socialization itself; variations in language acquisition may not be affected beyond the minimal requirement that children be exposed to some language input. We think, however, that this conclusion is premature. To evaluate the findings summarized in Tables 1-111 in more detail, we need to consider some specific conceptual and methodological difficulties that may be resolvable in future work.

VI. Conceptual and Methodological Confounds in Social-Causal Theories of Language Development Thus far, we have assumed a distinction between social and cognitive influences on language which could be defined as follows: 1 . By “social causes” we refer to the effects of interactions with and knowledge of animate beings.

Social Bases of Language Developmenl

49

2 . By “cognitive causes” we refer to the effects of interactions with and knowledge of inanimate beings. However, this animate-inanimate distinction is not all that separates the two lines of research reviewed so far. Many studies of social factors in language development also involve the following conceptual and methodological confounds: 1 . an epistemology emphasizing “internal” as opposed to “external” causal-

ity; 2 . a failure to distinguish structural and motivational contributions; 3 . the question of direction of effects between mother and child; 4. a confound between genetic and environmental variance; 5 . the issue of “threshold effects.” We shall consider these issues one at a time, discussing some possible ways around each one. A.

INTERNAL VERSUS EXTERNAL CAUSES

Most of the research on cognitive bases of language assumes an epistemological model in which the child plays an active role in scanning the environment and assimilating language to processes and structures that transcend the particulars of language proper. The environment must furnish language data; however, cognitive theorists place little emphasis on the role of the environment in selecting or highlighting particular structures. It is, of course, true that the child’s cognitive clues to language are ultimately derived from interactions with the external world (as Piaget has argued for decades). From the very beginning, however, the child’s knowledge of external objects results from her o r his own actions. The causal flow is from “inside” to “outside,” a kind of reverse S-R model in which the child stimulates and the environment responds. In social-causal research, in contrast, some of that emphasis on active, childdriven processes has been lost. Shatz (1981) has noted that social-causal theorists “focus less on the child and more on the role of external elements as important factors in acquisition” (p. 2). Hence, some of this literature represents a twofold break with Chomskian psycholinguistics: ( 1 ) a rejection of the innateness hypothesis for language-specific structures, and (2) a move away from the emphasis in the 1960s on active selective processes that are initiated and executed by the child. In contrast, cognitive-causal theorists have broken only with the innateness principle. The new empiricism is particularly strong in the motherese literature, which emphasizes an environmentally driven process in which caretakers “preprocess ” language for the child by eliminating difficult forms and highlighting and exaggerating those forms for which the child is ready. The flow of causality here

50

Elizubeth Butes et al.

is from “outside” to “inside,” an approach that is epistemologically much closer to the kinds of S-R models that were attacked by psycholinguists in the I 960s. The identification of “social ” with “external ” is not logically necessary. What we really have are two logically independent dimensions: child-driven versus environmentally driven learning mechanisms, interacting with social versus nonsocial objects of knowledge. The apparently greater empirical success of cognitive versus social-causal models may have nothing to do with the relative contribution of nonsocial and social knowledge to language learning. Instead, the difference between the two lines of research may revolve around their relative emphasis on internal causality. When we correlate the child’s progress in language with factors in the child, we may obtain better results than we do when progress in language is correlated with factors in the environment. The way out of this problem for social-causal studies may be to focus less on the correlation between the behavior ofsocial objects and child language, and begin focusing on the correlation between the child’s knowledge of social objects and related changes in language. We shall return to this point later. B . STRUCTURE VERSUS MOTIVATION

A further difficulty in research on social bases of language involves a failure to distinguish between structure and motivation, that is, between what the child knows and is mble to do (structure) and what the same child wants and is willing to do (motivation). This problem applies both to cognitive-causal and socialcausal research, but it is particularly salient in the social domain, in which emotional factors are of paramount importance. The contrast between structure and motivation is valid first at the level of data collection in particular contexts. For example, timid children may perform badly in experiments and in naturalistic observations when strangers are involved. In our own recent longitudinal study, we used both observational data and maternal interviews to measure aspects of language development such as vocabulary and the use of multiword speech. In addition, we asked mothers to fill out an infant temperament inventory (Rowe & Plomin, 1977), which included a “sociability” scale. From the viewpoint of method, the temperament questionnaire and the language interview should have shared the largest amount of ‘irrelevant” task variance (the quality of observation and the tendency of mothers to answer questions quickly, thoroughly, conservatively, etc.). In other words, we might have expected stronger correlations between “sociability’’ and language as reported by the mother than between “sociability” and the language observations. In fact, the opposite was true. The sociability scale bore no relationship at all to the language measures obtained in the interview but was associated with a number of language observations, with the reportedly “more sociable” children

performing better (Bretherton, McNew, Snyder, & Bates, 1980). This finding suggests that language competence measures obtained through observations are confounded by motivational factors to a greater extent than language competence estimates from diaries and maternal report. Hence, it may be useful in some studies to remove irrelevant temperament effects statistically, through partial correlations. The motivation issue is also valid at a more general level, with effects across many contexts. That is, a given mother-child couple may “look bad” in a language experiment not because of a specifically linguistic difference in their interactions, but because they get along badly at every level. In other words, the child’s apparent incompetence at language, and/or the mother’s apparent “linguistic insensitivity, ” may be epiphenomena1 of a generally bad relationship. For example, in the study by Tulkin and Covitz (1975) summarized in Table 11, a significant negative correlation was observed between “prohibition ratio” at I0 months of age and performance on the Peabody Picture Vocabulary Test at 6 years. Do we really want to argue that saying “no” to children decreases their vocabulary? It is likely that the link is much more indirect: a relationship characterized by high rates of saying “no” may be one in which (because of the child’s character, the mother’s, or both) very little decent conversation ever gets under way. The motivational confound may be superficial; that is, if we measured the child with his or her other parent, babysitter, or some other more compatible partner, we might obtain higher estimates of language competence. However, the motivational confound might have interacted to reduce “true” competence; because the relationship was so bad, the child’s progress in language has suffered across the board. We have no way of knowing which interpretation is correct from performance during a “one-shot ” set of observations. We can bring the structure-motivation distinction to bear on the motherese literature and on the multifaceted proposals by Bruner, Schaffer, Trevarthen, and others concerning the potential relationship between preverbal interaction and language development. At least four different kinds of causal hypotheses emerge.

I . Bctsic motivution to interact. When a warm relationship is established between parent and child, the child is more likely to observe, imitate, and above all interact with adults. In other words, motivational variables insure a willingness to participate in social exchange in the first place. From that point on, more direct causal influences can take place. 2 . Attention-direc.tinS functions. Once the child is engaged in interaction, mothers may provide a “scaffolding” (Wood, Bruner, & Ross, 1976) for the optimal use of infant attention. Through joint activities, the child comes to understand maternal gestures (e.g., pointing, direction of gaze toward referents) and intonation patterns (e.g., the tone of voice used in requests for action versus

52

Elizabeth Bares et al.

commands to stop action). The child’s attention then can be directed more easily toward the particular objects and events underlying adult speech. In a sense, this form of facilitation is a mechanism halfway along the structure-motivation dimension that we have drawn here, that is, motivation of interest and direction of attention toward structure. 3. Internalization of language by unulogy. Bruner, in particular, has argued that a literal structural resemblance between aspects of preverbal exchange and aspects of language can be observed. These analogies apply at the levels of pragmatics, semantics, and syntax alike. For example, semantic relations expressed in language (agent-action-receiver-patient) are implicit in games of object exchange, and syntactic operations like the embedding of clauses in complex sentence structures are directly prepared by the embedding of one action sequence in another during mother-child play. Because such analogies exist, the child can derive certain basic linguistic structures directly by internalizing the rules of preverbal exchange. In other words, the child goes from “doing” to “having” basic structures (see also Bates, 1976, Chapter 1). 4. Direct provision of structure. Internalization by analogy is an indirect process, one that involves very little “teaching” by the mother. In addition, claims have been offered in both the preverbal and the motherese literature that mothers preprocess and time linguistic structures directly so that the child will receive them when she or he is ready to understand them. Not that mothers supposedly read texts on grammar to carry out this work; rather, in the process of trying to get a conversation going with the child, mothers necessarily “scale down” their utterances to the point where the child can just understand them. As a result, language input is timed to be just one step ahead of the child’s productive abilities. What we have here, then, is a gradation from indirect to direct causal relationships between the social environment and language learning. At the indirect end, motivation plays a larger role. It is not always an easy matter, however, to keep these factors empirically distinct within a given study. In the prohibition-vocabulary relationship just mentioned, it seems obvious that the intervening causal connection must have more to do with emotional-motivational factors than with a specific effect of one kind of language structure on another. Shatz (1981) and Newport et al. (1977) have argued that direct linguistic effects are more easily established if a facevalid connection between maternal and child variables exists. For example, Newport and co-workers reported significant positive correlations between the number of auxiliary fronted questions used by the mother at Time 1 (e.g., “Are you making a tower?”), and use of auxiliary verbs by the child at later sessions. This is such a specific relationship that it seems reasonable to interpret it as a modeling effect. However, consider the following motivational interpretation.

Sociul Buses of L a n p u g e Development

53

I . In an harmonious relationship, in which partners are actively interested in each other’s affairs, more conversation about mutual activities is likely. 2. In conversations about each other’s activities, a high proportion of progressive verb forms is likely to occur (“What are you doing?”-‘‘I am doing . . . ”). 3. Hence, the correlation between auxiliary use in mother and child could be epiphenomena1 of the conversational topics selected in different kinds of relationships. Both the modeling hypothesis and the conversational topic hypothesis are equally plausible when all we have are correlational data, unless we have some way to partial out those motivational effects that d o not interest us in a given study. To summarize thus far, we have delineated three logically independent dimensions that have been confounded in some of the literature on social bases of language: social versus nonsocial knowledge, internal versus external causality, and structural versus motivational influences. A hypothetical space comprising these three dimensions is shown in Fig. 1. A given instance of language learning undoubtedly involves a complex interaction of all three. It is possible, however, at least in principle, to determine the degree to which different factors affect the course of acquisition, and hence to locate a given instance at some point in this STRUCTURE

INTERNAL CAUSALITY

SOCIAL OBJECTS

EXTERNAL CAUSALITY

NONSOCIAL OBJECTS

MOTIVATION/ ATTENTION

Fig. I . The sociul bases of Iunguage: three independent dimensions

54

Elizcibeth Bares et al

three-dimensional space. If we cannot find some way to separate these influences empirically, we run the risk of generating conceptually uninterpretable results.

VII.

Direction of Effects

One of the most serious problems in interpreting interaction data is the problem of direction of effects. Implicit in some of the studies summarized in Tables I and I11 is the belief that positive correlations reflect a causal effect of the mother on the child. However, we have just as much reason to believe that such correlations reflect an effect of the child on the adult. Let us consider how the issue of bidirectional effects applies to the literature on motherese. Ferguson (1976) has noted that simplifications, repetitions, and exaggerations occur in the speech of adults to any listener whose ability to comprehend is in question: foreigners with poor control of English, adults with hearing difficulties, retarded adults, and normal listeners in a noisy environment. The purpose of these adjustments is to make ourselves understood. The decision to use “motherese” then rests on some kind of cue indicating comprehension difficulties in the listener. Insofar as the use of speech adjustments is a function of the listener’s inability to comprehend, motherese is essentiully an indirect tneusure of childfnilure. In fact, the “causal flow” creating a significant correlation between child competence and maternal speech could even be unidirectional, from child to parent. At least in principle, we could obtain such correlations even if parents had no effect on their children at all. In this case, however, the correlations should be in a negative direction, that is, more motherese, less language tluency in the child. For instance, Bohannon and Marquis (1977) report two studies in which adults responded to naturally occurring or fabricated “incompetence” in children. In each case the adults reduced the lengths of their utterances when children signaled lack of comprehension. Thus, with less comprehension there was more “motherese.” In some studies, however, positive relationships have been reported. Here, too, the direction of effects is an issue. A precocious child is capable of understanding more complex speech, by definition. Hence, if the precocious child does receive more complex speech from her or his parents, the direction of effects could have been one in which the child elicits a type of language from the adult, and not vice versa. The direction of effects problem may also create false negative, that is, nonsignificant, correlations, when in reality a reciprocal causal influence is the case. While motherese may be an indirect measure of child failure, it may also be an aid to comprehension. Thus, any positive influence thut mothrrese may have on the child is canceled out by the original negutive relationship between motherese crnd childfiiilure. This confound was made particularly clear to our research team in a study by Carlson-Luden ( 1979) of causal understanding and maternal teach-

ing style with 10-month-old infants. Carlson-Luden’s infants were given a series of three causal tasks to solve, involving a lever that set some kind of dynamic effect into motion. Mothers were free to demonstrate the toys for the child at their own discretion. We had assumed in the usual fashion that “good teachers” would produce “good learning.” To our surprise, a large set of significant negative correlations was observed between various maternal interventions and measures of success in the 48 children who participated in the study. Does this mean that maternal teaching impedes learning? That is unlikely. The most sensible interpretation is that mothers intervene to the degree that their children fail to understand the task. Hence, this nonverbal form of motherese was actually an indirect measure of infant incompetence, resulting in negative correlations. It might be that good teachers do ultimately get their points across to children. However, if the positive effect of intervention cancels out the negative effect of responding to child failure, then we have nothing more than a zero correlation. From this viewpoint, the facilitative effect of mother and child would be difficult to detect in any correlational design. The direction of effects problem is particularly serious if measures of mother and child are taken at the same point in time, when the dyads are interacting. For example, Cross (1977) reported a large number of significant positivc correlations between maternal and child speech, including mean length of utterance (MLU). However, she failed to control for both differences in age and differences in overall ability across children. Hence, positive correlations could reflect the fact that older and/or more precocious children elicit more complex speech from their mothers-rather than any kind of causal effect of the mother’s speech on the child. In contrast, Furrow (1979) reported a significant negative correlation between maternal MLU at 18 months and child MLU at 27 months. This result was interpreted to mean that mothers who simplify their speech at the earlier stages produce children who are more precocious later on. However, Furrow did not control for linguistic differences among children at the earlier sessions. Newport et al. (1977) partialed out the child’s initial level of competence at Time l in a set of analyses similar to Furrow’s and found no significant relationship between maternal and child MLU when this control was used. This procedure seems a useful one in that it avoids the problem of the relationship of the child to herself over time. However, if the mothers’ own use of motherese at Time 1 was a reaction to the child’s level of competence at Time 1 , and that reaction provides the best input for the child’s next steps, then equating children for initial language level may result in removing the variance that we are interested in (McNew, 1981b). When a couple is locked into sequences of bidirectional effects, partial correlations removing any link in the chain may result in removing everything else of interest from there on. One method that has been offered as a solution to the bidirectional effects issue is the cross-lagged correlation. Both mother and child are assessed at Time 1 and Time 2. Hence, every correlational comparison involves correlations among all

56

Elizaberh Bates et a1

M = mother C = child

1 =Time 1 2=Time 2

Fig, 2 . Illustration of 11 cross-lugged correlutional design.

four poles, as shown in Fig. 2. We assume first that correlations between measures taken in the same session are uninterpretable in terms of direction of effects ( M , - C, ; M, - C2), Second, we also assume that mothers and children can be consistent with themselves over time even if they have no effect at all on one another. Hence, the two horizontal correlations (MI - M 2 ; C , - C,)are not relevant to causal effects between mother and child. To assess direction of effects in an interaction over time, we are interested in the two diagonal correlations: M I - C, and C, - M 2 .If the mother’s influence on the child is greater than the child’s influence on the mother, then the two diagonal correlations should differ significantly and the correlation should be the larger. Although the logic of the cross-lagged method is appealing, some serious problems still remain. First, many researchers fail to meet all the statistical assumptions of the model (Rogosa, 1981). For example, to test the difference between the two diagonals, there should not be a significant difference between the other two interactive correlations (i.e., mother with child within sessions). Second, the cross-lagged panel method requires very large samples for sufficient power to test the differences between correlations. Many investigators are apparently unaware that the sample size needed to create significant individual correlations is far smaller than the sample that is usually needed for two individual correlations to be significantly different from one another. The sample sizes in the studies reviewed in Tables 1-111 rarely exceeded 30-50.The difference between correlations would have to be very large (around .20) to be interpretable. Finally, the cross-lagged method does not get us out of the “canceling out” problem described above. If motherese is an indirect measure of child failure, but also a facilitative measure, then the diagonal correlations in a cross-lagged design might end up around the zero level even though “real” causal influences were going on. In his review, Rogosa suggested that no current methods are adequate to the task of determining predominant cause from correlation. The most we can establish is that two phenomena vary together in a meaningful and nontrivial way.

Social Bmes of Langucigc Development

57

Two other alternatives to the direction of effects problem are available. One approach is illustrated by Zukow, Reilly, and Greenfield (1981) in a study of how children understand offers. In a detailed microanalysis of behavior sequences in mother-child interaction, these investigators isolated every naturally occurring instance of successful or unsuccessful offers, examining the specific adult behaviors that preceded and followed each instance. Hence, they were able to identify “packages” of adult behaviors (establishing eye contact before an offer is made, calling the child’s name, etc.) at different points in development, demonstrating a reduction in the number of elements that were necessary for successful comprehension at each point. These sequential analyses preserve the real-time causal flow, including the number of child failures that preceded adult interventions. This study strengthens the case that a mother’s use of vocal and gestural “supports” is an indirect measure of how much support the child needs, that is, of child failure. However, if the details of behavioral sequencing are preserved, then it should be possible to separate out supports that occur a f e r child.failure from “free” supports or additional supports above and beyond child failure. That is, we could remove the variance from “failure-induced motherese” and assess the facilitative effect of additional support independently. Unfortunately, the kind of exhaustive microanalysis carried out by Zukow and colleagues is very time consuming, making i t difficult to obtain the number of cases that would be needed for correlational analyses. To establish causal relationships, the best alternative is to abandon naturalistic designs for an experimental approach in which adult interventions are manipulated directly as independent variables. McNew (1981b), reported a study of the effect of maternal gestures on child comprehension, using a method that stands midway between naturalistic and experimental approaches. Infants 20 months old were given standardized presentations of pairs of culturally familiar and unusual commands (e.g., “Kiss the baby” versus “Kiss the ball”). After the standardized presentation of the noncanonical commands by the experimenter, the infants’ mothers were asked to “Get her/him to kiss the ball any way that you would normally get her/him to understand.” In this situation of up-to-themoment information about the child’s comprehension, maternal use of demonstrations was negatively related to previous compliance. Thus, the child had an effect on the mother. McNew also analyzed the likelihood that these maternal demonstrations facilitated later success, partialing out child performance on the standardized trial. In other words, two kinds of variance were removed: ( 1 ) the effect of the child on the mother in “provoking” demonstrations; and (2) the likelihood that children who succeed once can succeed twice. In this second analysis, maternal demonstrations had a significant positive correlation with the child’s eventual success. In other words, gestures do facilitate comprehension. Belsky, Goode, and Most (in press) have also demonstrated a facilitative effect of maternal intervention on infant play and exploratory competence, using an experimental technique to increase maternal interventions beyond “baseline”

58

Elizubrth Bures et al.

levels. In an initial session, measures of maternal intervention and child exploration were taken for both control and experimental couples observed in routine household interactions. Then, in three succeeding weekly sessions, mothers in the experimental group received positive feedback whenever they gave their children explicit and helpful stimulation. This stimulation included tendencies to point to or reposition objects, demonstrations and support for the child throughout an action, verbal supports of instructing and questioning, highlighting object properties, and naming objects. Note that the mothers were not trained to carry out any particular kind of intervention. Rather, they were reinforced for aspects of their own spontaneous behavior. Control mothers received no such feedback. One week after the experimental-control sessions, Belsky et ul. observed both groups again, and found a significant increase in the use of facilitative interventions by the experimental mothers compared with control mothers. Finally, in a follow-up two months later, they reported that children of mothers in the experimental group engaged in significantly more exploratory play and investigations of the unique properties of objects. These results reached significance at the .05 level with a one-tailed test. This is not a robust effect, but it was obtained after quite minimal intervention, and the effects appeared after a rather long lag time. Clearly, mothers can increase their interventions beyond spontaneous or “natural” levels. Furthermore, such increases appear to have positive effects on their children. Perhaps the strongest experimental evidence supporting a causal effect of motherese has been provided by Nelson, Carskaddon, and Bonvillian (1973) and by Nelson (1977). An earlier study by Cazden (1965) had shown that “expansions” by adult experimenters in a preschool setting had no effect on language development. An expansion is a restatement of a child’s limited utterance, in a correct and more complete syntactic form, for example, “He making a picture,” “Oh, he’s making a picture.” Nelson and co-workers camed out a similar experiment, using a different type of expansion which they called a “recast.” A recast sentence preserves the basic semantic relations in the child’s previous utterance, but adds further syntacric information, for example, “I want some milk,” “You want some milk, don’t you?” (see Berko-Gleason & Weintraub, 1978, for a further discussion of the Nelson et al. technique). Cazden’s expansions were essentially corrections of the child’s utterances; the Nelson et al. recasts provided syntactic variation and elaboration on a semantic theme selected by the child. Nelson and co-workers found that 3-year-olds who received recast input for weeks obtained higher scores than children in a control group on five different posttest measures of syntactic development. The original recast experiment showed that syntactically varied feedback can have a generully positive effect on syntactic development. In a sense, this is an experimental version of a fact that we already knew from the correlational literature: mvre linguistic input is related to mvre language development. We still do

not know whether any kind of a direct causal link exists between specific structures. In this light, the follow-up study by Nelson (1977) is particularly important. Nelson selected a group of children 2% years old who failed to produce either complex questions or complex verb phrases in a pretest speech sample. The children were then randomly assigned to one of two training groups, with experimental sessions spanning 2 months: half of the children received recast sentences involving complex questions, while the other half received recasts involving complex verb forms. By the end of the training, children in the verb group produced significantly more complex verbs and significantly fewer complex questions than children in the question group. In other words, the respective experimental interventions had very specific structural effects. The Nelson studies are important on two grounds. First, they support the kinds of specific correlational relationships reported by Newport and co-workers, but with stronger evidence for direction of effects from parent to child. Second, the learning process in the recast experiments is essentially child driven. That is, the child selects the semantic theme, and the adult merely provides variations on that theme. This was also true of Cazden’s expansions. However, her expansions corrected the child’s syntax, and Nelson’s recasts respected the child’s formula while providing further information on how the same theme could be expressed. Two problems affect the generalizability of experimental findings like these. First, these are short-term effects that may not correspond to long-term processes in the real world. Second, we do not know how these artificial changes correspond to the spontaneous behavior of caretakers in natural settings. Nevertheless, specific changes in adult speech can clearly have a positive effect-if the adult gives the child what she or he is looking for. If we put together converging evidence from naturalistic, correlational studies and experimental manipulations, we may obtain the best possible estimate of how mothers and children affect one another in language development.

VIII.

Genetic Confounds

Thus far, we have considered whether a correlation can be interpreted as an effect of the mother on her child, or an effect of the child on her or his mother. There is yet another way to interpret significant correlations between maternal behavior and child language: Similarities within mother-child puirs may reflect nothing more thon shared genetic variance. If verbal fluency is genetically based, and/or if verbal measures are influenced by general intelligence, then positive correlations between mother and child on a variety of verbal measures may have nothing to do with causal influences of one partner on the other. Rather, they may simply mean that smarter mothers have smarter babies. One solution to the genetic confound is to obtain a measure of maternal

60

Elizaheih Bates et al.

intelligence, and remove that variance through partial correlation. However, no one really knows what variables “intelligence” comprised as it is measured by standardized tests, and therefore we run the risk of overkill, wiping out interesting and real causal effects in our efforts to remove an uninteresting one. Consider the following causal chain:

1. The real factor influencing the child’s language level is not the genetic variance in IQ inherited from parents, but the quality and sensitive timing of input that the child receives. 2. However, smarter mothers are more sensitive to the child’s needs and hence more likely to provide this high-quality input. 3. We measure maternal sensitivity and child language across a given time span, and carry out partial correlations between maternal and child variables, removing the variance due to maternal intelligence. 4. Because we have removed variation in intelligence, we also remove variation in maternal sensitivity. 5 . We obtain no correlation between maternal speech and child language, and conclude incorrectly that infant competence is unrelated to maternal input. A better alternative to the problem of genetic confounds is a behavior-genetic design, in particular, an adoption study in which the genetic contribution of the biological parents and the environmental contribution of the adoptive parents are clearly separated (Plomin, DeFries, & McClearn, 1980). We are aware of only one study of this type using the kinds of maternal and child measures relevant to current social-causal theories of language development. Hardy-Brown, Plomin, & DeFries (1981) have examined several aspects of language and communicative development at 1 year of age, in a sample of 50 infants who were adopted at birth. Measures of the child included a word diary collected by the adoptive parent, and several observational measures of frequency and quality of speech and gestural communication (taken from videotapes). Correlations of the child measures with one another were high, justifying the use of a child communicative competence factor (the first principle component of a factor analysis of child measures), in addition to the individual measures, in correlations with adult variables. For the biological mother, the measures included a battery of standardized cognitive tests (visual memory, verbal fluency, and the like), which yielded four specific cognitive abilities factors and one general factor (first principle component). The biological fathers were usually not available for testing. For the adoptive parents, the same battery of standardized tests were administered to both father and mother. In addition, 15 measures of maternal style and sensitivity (similar to many of the measures in Tables 1-111) were taken from videotaped interactions between the child and adoptive mother. Finally, scores for socioeconomic and educational status of both adoptive parents were included.

Socrirl Buses of Language Di~veloptnent

61

All of these measures were entered into correlational analyses with the child variables. Because there were so many more measures for the adoptive parents, the case is clearly biased toward uncovering environmental effects. Nevertheless, Hardy-Brown et a l . ’s results provide much more support for biological effects on language and communication at I2 months of age. Direct comparisons between genetic and environmental influences on rate of communicative development in the first year are provided by relating cognitive abilities of the birth and adoptive mothers with the communicative ability of the adopted infant. Of 84 correlational relationships between cognitive assessments of the infant 19% were significant. In comparison, 4% of these relationships were significant between the infant and adoptive mother, and 6% were significant with regard to the adoptive father. This average of 5% for both adoptive parents can be expected on the basis of chance alone, and suggests that brighter parents may not environmentally enhance the communicative development of their children in the first year. Since biological mothers contribute only half the genetic input to the child, the total amount of genetic input to child competence at I year of age is probably even greater than indicated by the above results. Furthermore, of the positive results obtained with adoptive parent measures, almost all came from the videotaped interaction variables. Considering the points we raised earlier about direction of effects, the few correlations that were obtained with adoptive parents could actually reflect an effect of the child on the parent and not vice versa. Hardy-Brown et a / . noted carefully that environmental effects may take longer than 1 year to establish. Communicative developments at 12 months of age may be tied to a child-driven, maturational schedule; similar measures at 2 and 3 years of age may be more strongly influenced by the environmental contribution of adoptive parents. Nevertheless, in view of the large number of theoretical claims that have been made about “scaffolding” effects in the first year of life, HardyBrown’s findings strike a cautionary note.

IX. Threshold Effects One further problem arises in correlational and experimental studies alike, with and without behavior-genetic strategies to remove genetic confounds: the problem of threshold effects. Almost all the studies that we have reviewed thus far assume a linear, cumulative relationship between input and output: if some maternal input is good, then more is better. This may not be the case at all, however. Perhaps the amount and type of social input that is necessary for normal language development is a minimal threshold amount that every normal child receives. Increases in such input beyond the threshold amount may have no effect at all, so that correlational and experimental measures of input variation in

62

Elizubeth Bates et al.

the normal range yield largely nonsignificant results. This is the kind of difficulty that Pentz (1975) and Kaye (1980) alluded to in discussing the “buffering” of language development, and the optimal “fit” of mother to child across our species. To illustrate the threshold problem, let us start with an extreme example. Virtually no individual variation exists in the tendency for human beings to be born with two arms and two legs. Hence, if we were to take a measure of two armedness in parents and their children, we would find no individual variation in the measures. If no variation exists, no correlation is observed. Could we conclude from the lack of correlation that parental armedness and child armedness are unrelated? Bates (1979) offered another extreme example that pertains to two measures taken within the child. It is certainly the case that having a functioning heart and liver is a prerequisite to the acquisition of language. Any child born without a heart and liver will not live to acquire language. However, if we take a sample of 12-month-olds and measure “degree of organ ownership” and correlate that measure with language, we will obtain a zero correlation. Because no variance exists in the organ measure, no statistical link to the variance in language at 12 months is found. Certainly we would not want to conclude that the two are unrelated. It is simply the case that the threshold amount of heart and liver necessary for language is usually established in children by 12 months of age. The threshold problem is obvious in the case of measures that show no variance at all. It is less obvious when two measures (e.g., maternal sensitivity and child language) d o show a large amount of normally distributed variation at the point in development that interests us. It may still be the case that the relevant variation in maternal sensitivity is very small, so that no one in the sample falls below threshold. Increases beyond that amount are irrelevant, and vary independently from individual differences in child competence. In fact, evidence f o r child effects on caretakers leads to the hypothesis that human children seize and create for themselves some minimal amount of communicutive interaction. Some findings by Newport et al. (1977) are particularly relevant in this regard. As noted earlier, they reported only a few significant correlations between maternal and child speech, after the variance due to differences in child ability at Time 1 were removed. However, the relationships that did reach significance formed a coherent pattern: (1) they were all very specific connections between similar aspects of English grammar, for example, auxiliary-fronted questions by the mother versus development of auxiliary verbs in the child; (2) such specific correlations were found only for aspects of grammar that are peculiar to English. This pattern led Newport and colleagues to argue that universal aspects of language are ‘‘environmentally insensitive, ” while language-specific structures may be “environmentally sensitive. ” They concluded that, “the mother has

little latitude to teach her child about the nature of language; but she can at least improve his English” (Newport et a/., 1977, p. 147). The concept of environmental sensitivity can be related to the threshold issue as follows. English-specific structures such as the different forms of auxiliary verbs carry very little semantic content-at least, not content that is critical to a 2-year-old. These structures may be environmentally sensitive because children are not looking for them. Indeed, if they are going to notice these forms at all in the early stages, it will have to be because parents are hitting them over the head with them. Hence, we could call these “high-threshold” forms. In contrast, if environmentally insensitive structures are the ones that encode universal meanings, then children may acquire them with no more than some minimal threshold amount because they art actively scanning the environment for precisely those forms. Hence, the distinction between high- and low-threshold structures, and environmentally sensitive and insensitive forms, rests on the distinction made earlier between passive and active processes in language acquisition. The threshold issue in language acquisition also brings to mind Harlow’s studies of social development in nonhuman primates. Harlow and Harlow (1969) have shown that infant monkeys that are deprived of normal parenting and/or social experience with peers fail to function normally as adults in sexual encounters and in behavior toward their own offspring. In other words, social experience in infancy is crucial to normal social development. However, this requirement becomes obvious only in extreme cases in which threshold amounts of early social experience are denied. The same message may apply to the study of low-threshold, environmentally insensitive aspects of human language. If we focus exclusively on variation in the normal range, we may nor detect some critical relationships. Fortunately, cases of extreme deprivation Li ki Harlow are rare in our species. We do know that “wild children” who are badly abused or neglected in childhood fail to develop language properly (e.g., Curtiss, 1977). However, these children are so badly impaired across the board that particular effects on language are difficult to separate from general social and mental impairment. However, some less extreme “natural experiments” do permit us to assess certain kinds of social inputs separately from others. For example, blind infants are unable to respond to a wide array of social signals that are available to the sighted child-no matter how skilled or well intentioned the child’s mother might be. Studies by Urwin ( I 978), among others, demonstrate that this deprivation leads to marked delays in the earliest stages of language acquisition, when the “idea” of intentional communication through symbols is first established. Similarly, institutionalized children in the 1930s and 1940s were deprived of many aspects of normal social input, even though their own senses were intact. Spitz (1965) has argued that such deprivation often leads to depression and even marasmus if the deprivation is too prolonged. Under these circumstances language has little chance. Fortunately for our species, children usually see to it that adults provide the

64

Elizabeth Bares et al.

essential daily requirements for adequate communicative development. If more and better input is forthcoming, certain aspects of language may indeed improve. The child may ultimately move from conversationalist to poet in response to a verbally enriched environment. However, supermothers are not essential for the species to go on talking.

X. Conclusion We have concluded that the case for externally driven, structural effects on language development is not very good. Research on social factors in language acquisition has concentrated primarily on these kinds of effects, and as a result, social-causal theories have not yet obtained adequate empirical support. We do not want to argue that such approaches are uninteresting or wrong, and that the social environment plays no role in language development beyond passive provision of minimal language data. Instead, we have tried to point out why social effects are so difficult to demonstrate. Several partial solutions have been offered. First, observational and correlational studies should be supplemented with converging evidence from experiments in which types of adult input are manipulated systematically. Second, it may occasionally be possible to disentangle issues in direction of effects, and/or motivational confounds, through partial correlations from which the variance due to confounding factors has been removed (e.g., “sociability” in infants as it relates to observational measures of language). Third, behavior-genetic designs are useful in removing genetic confounds in correlational data. Fourth, the linear view that “if some input is good, then more is better” may apply only to certain kinds of high-threshold, environmentally sensitive language structures. When the threshold amount of input required for normal development is low, we may have to abandon studies of the normal populations in favor of research in clinical settings. That is, cases of extreme deprivation (blind infants who cannot see their mothers’ cues, psychotic mothers who d o not give their children normal cues) may tell us more about what happens when critical amounts of social input are missing. However, the major recommendation that we would like to make is for social-causal theories that reverse the causal flow, an organismically driven approach emphasizing the role of social factors in the child on language development in general, and in particular on those aspects of language that encode uniquely social meanings. We can offer five specific recommendations for areas of research in which social developments in the child contribute to the acquisition of language. 1 . Imitation as a species-specific motive. Imitation was not a popular topic ~ because it was identified with developmental research during the 1 9 6 0 ~perhaps the kinds of passive, mechanistic, environmentally driven processes that charac-

Social Boses of Language Development

65

terized social learning theory in the 1950s (e.g., Mowrer, 1960). However, as we have discussed in more detail elsewhere (Bates, 1979, Chapter 7), imitation is actually one of the most mysterious processes in all of cognitive development. For example, Meltzoff and Moore (1977) have provided some evidence (albeit controversial) that human neonates are capable of imitating facial movements. Chevalier-Skolnikoff ( 1977), in a comparison of human cognitive development with related achievements in other primates, concluded that imitation separates human from infrahuman more than any other area of sensorimotor development in the first two years of life. We are much better at “aping” than the apes. We know less than we should about the role of socially motivated imitation in the acquisition of languages (for reviews, see Bloom, Hood, & Lightbown, 1974; and R. Clark, 1977). Our own longitudinal studies have yielded significant correlations between rate and quality of imitation in infancy, and many aspects of early gestural communication and language (Bates et af ., 1979). At later stages in development, R. Clark (1977) and Fillmore (1976) have demonstrated how imitation can be used by some children to acquire and use certain linguistic forms long before those forms have been analyzed into a set of productive rules. Ochs (1979) has pointed out a wide variety of social functions that are served by repetition of prior discourse, including holding one’s place in a conversation and confirming the speaker’s point. Finally, considerable individual variation exists in the degree to which children use imitation as a communicative tool. Bates (1979), Kempler (1980), Bloom et al. (1974), and Horgan (1979) have all suggested that this variation in the uniquely social activity of imitating others might be related to a variety of other individual differences in the acquisition of grammar, including relative use of pronouns versus nouns and the degree to which children acquire complex syntax through idioms or formulas. 2. Shared reference as a species-specific motive. Bates er d. (1379) and Hardy-Brown et al. (1981) have shown strong correlational relationships between language development at 13 months and several aspects of preverbal communication: giving, pointing, showing, ritualized requesting (e.g., requesting by means of an abbreviated open-and-shut hand gesture). Bates, Carnaioni, and Volterra (1975) have suggested that such gestures may be a kind of “protodeclarative, ” exercising the social function of shared attention and reference to some object or event long before language is available for encoding reference. Scaife and Bruner ( 1 975) have shown that this reference-sharing function in humans, when children follow the line of adult visual regard toward some “third party,” can occasionally be observed as early as 3 months of age and becomes the norm around 9 months. Although the social motivation to share reference is not directly linguistic ( i s . , children do it months before they acquire speech), such a motive would have important indirect consequences for the language acquisition process. As Savage-Rumbaugh, Rumbaugh, and Boysen (1978) noted, one of the peculiarities of the chimpanzee language protocols is the high concentration of imperatives.

66

E/izuhe/li Eutcs et al

They are definitely not interested in conversation for its own sake, nor do they engage in nonverbal communication with humans strictly for the purpose of sharing attention to events in the world (e.g., pointing and labeling without any ulterior motive). What better way to equip a species to acquire large vocabularies and complex grammar than to give it an insatiable lust for small talk? This desire to understand one another, and to see things the same way, means that language acquisition does not have to be motivated extrinsically (e.g., as a means for obtaining bananas). Imitation and shared reference involve affective/motivational inputs to language development, via some very general processes of shared social exchange. The next three points refer to areas in which language is used to encode specifically social content, like the relationship between object permanence and “all gone” in our reviews of cognitive bases to language. 3. lnientionality and human action. We have made much of the “actoraction-recipient-patient” relationship as an example of a semantic structure that serves as a clue to the acquisition of grammar. What we have not yet stressed is that the concept of animacy itself requires considerable social knowledge, an ability to distinguish between actions that obey Newtonian mechanics versus the less predictable actions of willful beings. Comparative studies of a wide variety of languages have led some linguists to suggest that all human grammars have rules built around an “animacy hierarchy” (Cooper and Ross, 1975; Dik, 1978; Kuno, 1976; Li & Thompson, 1976). Consider, for example, the grammatical notion of “subject of the sentence. ” All languages seem to have something like a subject, corresponding to the point of view or topic of a piece of discourse. Furthermore, it seems to be a universal of human conversation that we talk predominantly about ourselves. Hence, the topic is likely to be a “speaker-like’’ element. Languages respect this high-probability relationship between topic and human agency by developing rules that assign subject roles (first position in the sentence, agreement with the verb in person and number, nominative case marking, etc.) to a hierarchy of elements ranging from “most like the speaker” to “least human, animate, willful, etc.” For example, in Dutch it is not permissible to make inanimate instruments the subject of a sentence (as in “The knife cut the salami”). In Navajo, neutral declarative sentences must be ordered from left to right to reflect a complete power hierarchy: chiefs before peasants, adults before children, humans before animals, and so on. Children manage to map such grammatical rules onto knowledge of human action with relatively little trouble. However, an intelligent Martian with tremendous analytic capacities and full knowledge of the physics and chemistry of nonsocial earth, might have much more difficulty acquiring grammars based on these inherently social categories. The relationship between grammar and a theory of human action holds not o n l y for such grammatical forms as word order and subject-verb agreement, but also for verb morphology. For example, English modal verbs (can, would, should, etc.)

require considerable understanding of such ontological concepts as doubt vs certainty, ability vs intention. The same is true for mastery of verb tense (e.g., past vs present) and verb aspect (e.g., completed vs uncompleted action). An important arena for social-causal theories of language acquisition should be the mystery of how children map their developing “theory of human action” onto all these disparate aspects of grammar. (For reviews of animacy and causal understanding in early childhood and infancy, see Ammon, 1980; Bloom, Hood, & Lightbown, 1974; Bretherton & Bates, 1979; Golinkoff, 1975 ) 4 . Social feelings and social roles. Andersen (1977) has shown that children between the ages of 4 and 6 years have a detailed knowledge of the linguistic stereotypes associated with such social roles as male vs female, parent vs child, doctor vs patient, and doctor vs nurse. This knowledge cuts across a heterogeneous set of linguistic devices including intonation, type of lexical items used (e.g., “tough” words are reserved for males), and tag questions (e.g., women tend to say “isn’t i t ” at the end of sentences). Cremona and Bates (1977) have reported that Italian children before the age of 6 years know that their rural dialect is considered socially inappropriate and inferior compared with the standard dialect. Bates (1976) reported that children as young as I % - 3 years of age have a basic understanding of the concept of politeness. This concept in turn affects the acquisition of several aspects of grammar, including interrogative vs declarative sentence types, verb morphology (e.g., past and conditional are more polite than present tense in requests), and noun morphology (e.g., diminutive endings on nouns create “nicer” forms for use in requests). Here, too, an extraterrestrial being with no knowledge of social roles or the regulation of status through language would be unable to decipher a wide range of systematic variation in the grammar and the lexicon. By contrast, very young children apparently do understand at least the rudiments of these social concepts enough to acquire a rather complete, albeit caricatured, system of sociolinguistic variations between 2 and 6 years of age. Ervin-Tripp and Mitchell-Kernan (1977) and Ochs and Schieffelin (1979) have edited volumes summarizing many aspects of sociolinguistic development from preschool onward. Bretherton and her colleagues (e.g., Bretherton & Bates, 1979; Bretherton & Beeghly-Smith, 1981; Bretherton. McNew, & Beeghly-Smith, 1981) have discussed how these social structures and categories are prepared even earlier, in mother-infant interaction from 9 through 28 months of age. Bretherton et a l . argued that early language protocols demonstrate much richer understanding of how others feel and think than is evidenced in many nonverbal studies of role-taking (e.g., Flavell, Botkin, & Fry, 1968; Selman, 1971). When a 28-month-old boy says to his mother, ‘‘I hurt your feelings ’cause I was mean to you” (Bretherton & Beeghly-Smith, 1981), he shows an appreciation of interpersonal power and responsibility that is certainly adequate for the task of acquiring many socially based aspects of language acquisition. 5 . Discourse and syntax: learning to “stage” utterunces for a Listener. A par-

68

Elizabeth Ba/es et al.

ticularly active area of linguistic research in the last decade has been “text grammar” or “discourse grammar” (e.g., van Dijk, 1980). These grammars go beyond the level of individual sentence structure to describe how native speakers use grammatical devices to create cohesion across long passages of discourse. For example, rules that regulate the use of adjectives and other modifiers involve planning at the discourse level, insuring that the listener knows which topic is being talked about. We say “the red dog” instead of “the dog” because there is some danger of confusion between canine referents. When we use a relative clause, as in “The man whom I told you about the other day called this morning ,” we are using some rather intricate knowledge of our past exchanges with the listener and his current state of awareness to “set up” the subject of the sentence in a recognizable way. The same kind of discourse-level function applies to nouns vs pronouns, definite vs indefinite articles, conjunctions and adverbs (e.g., “but” vs “and,” “then” vs “now”), and for that matter virtually every aspect of the grammar (see Bates & MacWhinney, 1979, for reviews). As discussed by Greenfield and Smith (1976), among others, even the earliest rules of child grammar respect a division between “proposed” and “presupposed information. To make that division properly, the child will have to be able to make a vast set of rapid calculations of listener knowledge. This is inherendy social knowledge, not derivable from any other uspects of cognition. Bretherton, McNew, and Beeghly-Smith ( 198I ) have argued that language acquisition requires the development of a “theory of mind,” defined by Premack and Woodruff (1978) as a system of inferences which enables an individual to impute intentions and other mental states to self and other. As Bretherton et al. (1981) suggested, part of our task in studies of language acquisition is to understand how the child maps this “theory of mind” onto the structures of her or his language. We noted at the beginning of this paper that there is hope for a reconciliation in the “marriage” between psychology and linguistic theory. In the 1970s, while searching for cognitive and social inputs to the acquisition process, many of us set aside the problem of grammar and how it is acquired. As Susan ErvinTripp reminded us in 1977, in her keynote address to the Stanford Child Language Forum, we never did solve the syntax problem. It is time for us to gather together the knowledge that has been gained about cognitive and social inputs to language, and tie that knowledge to the very difficult problem of how formal grammatical knowledge is acquired. We know a lot more now than we did in 1957. And they tell us that love is better the second time around. ”

REFERENCES Ainsworth, M . D. S . The development of infant-mother attachment. In B . M . Caldwell & H . N . Ricciuti (Eds.), Review of child developmenr research (Vol. 3). Chicago: Univ. of Chicago Press, 1973.

Social Bases of Language Development

69

Ainsworth, M. D. S . , Blehar, M. C . , Waters, E . , &Wall, S . Potterns ofuttuchment: Apsychological study of the strange situation. Hillsdale, N.J.: Erlbaum, 1978. Amrnon, M. Develupment in the linguisric expression rfcuusul relutions: Comprehension qffeatures uflexirul and periphrostic cuusatives. Unpublished doctoral dissertation, University of California, Berkeley, 1980. Ammon, P. R. Cognitive development & early childhood education: Piagetian and neo-Piagetian theories. In H. L. Horn, Jr. & P. A. Robinson (Eds.), Psychological pruresses in eurly edui~ation.New York Academic Press, 1977. Anderson, E. Learning to speak with style: A study uj‘the sociolinguistic skills of young children. Unpublished doctoral dissertation, Stanford University, 1977. Bakeman, R., & Brown, 1. Early interaction: Consequences for social and mental development at three years. Child Development. 1980, 51,437-447. Bates, E. Peer relations and the acquisition of language. In M. Lewis & L. Rosenblum (Eds.), Friendship and peer relations. New York: Wiley, 1975. Bates, E. Language and context: The ucyuisition ofpragmatics. New York: Academic Press, 1976. Bates, E. Intentions, conventions, and symbols. In E. Bates, L. Benigni, I. Bretherton, L. Camaioni, & V. Volterra, The emergence of symbols: Cognition and communicarion in infancy. New York: Academic Press, 1979. Bates, E., Benigni, L.. Bretherton, I., Camaioni, L., & Volterra, V. From gesture to the first word: On cognitive and social prerequisites. In M. Lewis & L. Rosenblum (Eds.), Interaction. conversutiun and the development of Ianguuge. New York: Wiley, 1977. Bates, E., Benigni, L., Bretherton, I., Camaioni, L., & Volterra, V. Cognition and communication from 9-13 months: Correlational findings. In E. Bates, L. Benigni, 1. Bretherton, L. Camaioni, and V. Volterra, The emergence of symbols: Cognition and communication in infancy. New York: Academic Press, 1979. Bates, E., Camaioni, L., & Volterra, V. The acquisition of perfonnatives prior to speech. MerrillPalmer Quarterly, 1975, 21, 205-226. Bates, E., & MacWhinney, B. A functionalist approach to the acquisition of grammar. In E. Ochs & B, Schieffelin (Eds.), Developmental pragmatic..^. New York: Academic Press, 1979. Bates, E., & MacWhinney, B. Functionalist approaches to grammars. In L. Gleitman & E. Wanner (Eds.), Language acquisition: The state of the art. London and New York: Cambridge Univ. Press, 1981. Bates, E . , & Snyder, L. The cognitive hypothesis in language development. In I. Uzgiris & J. McV. Hunt (Eds.), Research with scales ofpsychological development in infancy. Urbana-Champaign: Univ. of Illinois, in press. Beckwith, L., & Thompson, S . Recognition of verbal labels of pictures, objects and events by 17-to-30-month-old infants. Journal of Speech and Hearing Research, 1976, 47, 579-587. Bell, S. M. The development of the concept of the object as related to infant mother attachment. Child Development. 1970, 41, 291-313. Bell, S . M . Cognitive development and mother-child inreructiun in the ,first three years qf life. Unpublished manuscript, 1978. (Some findings are cited i n Ainsworth et a l . , 1978.) Bell, S . M.,& Ainsworth, M.D. S . Infant crying and maternal responsiveness. Child Development, 1972, 43, 1171-1 190. Belsky, J., Goode, M. K.,& Most, R. Maternal stimulation and infant exploratory competence: Cross-sectional, correlational, and experimental analyses. Child Development, 198 I , in press. Berko, J . The child’s learning of English morphology. Word. 1958, 14, 150-177. Berko-Gleason, J., & Weintraub, S . Input language and the acquisition of communicative competence. In K. Nelson (Ed.), Children’s kinguage (Vol. 1). New York: Gardner, 1978. Block, E.. & Kessel, F. Determinants of the acquisition order of grammatical morphemes: A reanalysis and reinterpretation. Journal ofchild Languuge. 1980, 7, 181-189.

70

Elizabeth Bares et al

Bloom, L. One wordat rr time: The use oj‘single wurd utterances hefure syntux. The Hague: Mouton, 1973. Bloom, L., Hood, L., & Lightbown, P. Imitation in language development: If, when and why. Cognitive Psychology, 1974, 6 , 380-420. Bohannon, J., & Marquis, A. Children’s control of adult speech. Child Dewlopnimt, 1977, 48, 1002. Bowerman, M. Eurly syntadc develupment: A c,rosslinguistic study with spei,ia/ reference ro Finnish. London and New York: Cambridge Univ. Press, 1973. Bowerman, M. Learning the striwture of causutive verbs: A study in the relrtionship of cognitive, senurntic trnd syntcrctic. development (Papers and Reports on Child Language). Stanford, Calif.: Stanford University, 1974. Bowerman, M.F. Semantic factors in the acquisition of rules for word use and sentence construction. In D. M. Morehead & A . E. Morehead (Eds.), Normal a n d d & + v z ti,hildlanguoge. Baltimore: Univ. Park Press, 1976. Bowlby, J. Attuchment and Iuss (Vol. 1). Attirchmmt. New York: Basic Books. 1969. Bowlby, J . Attachment und loss (Vol. 2). Sepurarion. New York: Basic Books, 1973. Brdine. M. D.S . Children’s first word combinations. Monographs qf rhe Society fiw Reseirrch in Child Developmen/, 1976, 41, (1, Serial No. 164). Branigan, G. Some reasons why successive single word utterances are not. Journal ofChild L a n gur/gi’, 1979, 6, 41 1-421. Brazelton, T. B., Koslowski, B., & Main, M. The origins of reciprocity: The early mother-infant interaction. in M. Lewis & L. A. Rosenblum (Eds.), The eflect of the infurir on irs caregiver. New York: Wiley, 1974. Bretherton, I . , & Bates, E. The emergence of intentional communication. New Directionsfor Child Development, 1979, 4, 8 1-100. Bretherton, I., Bates, E., Benigni, L., Camaioni, L., & Volterra, V. Relationships between cognition. communication and quality of attachment. In Bates ct a / . (Eds.) The emergence of synhols: Cognitiun ond commirniccrtiun in infirncy. New York: Academic Press, 1979. Bretherton, I., & Beeghly-Smith, M. Talking crbout internal states: The acquisition of an explicit “theory ofmind.” Paper presented at the biennial meeting of the Society for Research in Child Development, Boston, April 1981. Bretherton. I., McNew, S., & Beeghly-Smith, M. Early person knowledge as expressed in verbal and gestural communication: When do infants acquire a “theory of mind”? In M. E. Lamb & L. R. Sherrod (Eds .), Infirnt six-ial cognition: E m p i r i d und rheureticirl considerations. Hillsdale, N.J.: Erlbaum, 1981. Bretherton, I . , McNew, S . , Snyder, L., & Bates, E. Individual differences at 20 months: Analytic and holistic strategies in language acquisition. Journal qf Chill Languuge, in press. Brown, R. A first language. Cambridge, Mass.: Harvard Univ. Press, 1973. Bruner, J. From communication to language: A psychological perspective. Cognition, 1975, 3, 255-287. (a) Bruner, J . S . The ontogenesis of speech acts. Journrrl ojChild Languugr, 1975, 2 , 1-19. (b) Bruner, J . Early social interaction and language acquisition. In H . R. Schaffer (Ed.), Studies in mother-infirtit intercrcrion. New York: Academic Press, 1977. Bruner, J . . & Ninio, A . The achievement of antecedents of labeling. ./ournu/ of Child Languuge, 1978, 5, 1-19. Bruner, J.. Olver, R., & Greenfield, P. M. Studies in ccignitiwc~gruwth. New York: Wiley, 1966. Bruner, J. S., & Sherwood, V. Early rule structure: The case of peekaboo. In J. S. Bruner, A. Jolly, & K . Sylva (Eds.), P l r y : Its role in developniiw and evolutiun. New York: Basic Books, 1976. Burchinal, P., & Farran, D. H o w importcmt i s eirrly mother-infunr infertrction fiw development during infancy? Paper presented at the International Conference on Infant Studies, New Haven, April 1980.

Caldwell. B. Instruction munuril inisentonfor infiitits (Honre ohsar~wiori,fi)r t n ~ u ~ ~ r r of’~ ~the n~~nt eriivronnimt). Unpublished manuscript, 1970. (Available from Center for Early Development and Education, 8 14 Sherman, Little Rock, Arkansas, 72202.) Carlson-Luden, V . Cnustil understunding in the 10-month-old. Unpublished doctoral dissertation, University of Colorado, 1979. Case, R. Intellectual development from birth to adulthood: A neo-Pagetian interpretation. In R. Siegler (Ed.), Children’s thinking: Whnt rlevelopsl Hillsdale, N.J.: Erlbaum, 1978. Case, R. Intellectual development in ir$aticy: A neo-Piugetian interpretation. Paper presented at the International Conference for Infant Studies, New Haven, April 1980. Cazden, C. Environmental nssisrance to the child’s acqitisitinri of grummur. Unpublished doctoral dissertation, Harvard University. 1965. Chevalier-Skolnikoff. S. The ontogeny of primate intelligence. Implications for communicative potential. In S . Harnad, H. Steklis, & J. Lancaster (Eds.), Origins of 1angrtuge (itid speech. New York: New York Academy of Sciences. 1977. Chomsky. N. Synttrc,tic, structures. The Hague: Mouton, 1957. Chomsky, N. Aspects o f u rheor? of syntax. Cambridge, Mass.: MIT Press, 1965. i New York: Pantheon Books, 1975. Chomsky, N. Refler~tionso ~ lringiroge. Clark, R. What’s the use of imitation’? Journal of Child Lmguoge. 1977, 4, 341-358. Clarke-Stewart, K . A. Interactions between mothers and their young children: Characteristics and consequences. Monogruphs of the Soriety for Reseurch in Child Development, 1973, 38 (6-7 Serial No. 153). Cohen, S., & Beckwith, L. Preterm infant interaction with the caregiver in the first year of life and competence at age two. Child Developtnent, 1979, 50, 767-776. Connell, D. B. Individual differences in tittuchment behcivior: Long-term .stubility und relutionships to lunguuge development. Unpublished doctoral dissertation, Syracuse University, 1977. Cooper, W . , & Ross, J. World order. In R. Grossman. L. San, & T . Vance (Eds.), Poprr.s,fr.om the purczsession on functionn/ism. Chicago: Chicago Linguistic Society, 1975. Corrigan. R. Language development as related to stage 6 object permanence development. Journul of’ Child Language. 1978, 5 , 173-189. Cremona, C . , & Bates, E. The development of attitudes toward dialect in Italian children. Journul of’ Psycholinguistic Reseurch. 1977, 6 , 223-232. Cross, T . Mothers’ speech adjustments: The contributions of selected child listener variables. In C. Snow & C. Ferguson (Eds.), Talking to children: Lutigiruge input and acquisition. London and New York: Cambridge Univ. Press, 1977. Curtiss, S. Genie: A psycholinguistic study of u modern-cloy “wild child.” New York: Academic Press, 1977. Curtiss. S . , & Yamada, J. Lungircige und cognition. Paper presented at the Third Annual Boston University Conference on Language Development, Boston, September 1978. Dale, P. Languuge development: Strrrcrure undfitnction (2nd ed.). New York: Holt, 1976. Dik, S. Funcrional grummur. Amsterdam: North-Holland Publ. 1978. Edwards, D. Sensory-motor intelligence and semantic relations in early child grammar. Cognition, 1973, 2, 395-434. Ervin-Tripp, S., & Mitchell-Kernan, C. (Eds.). Child discourse. New York: Academic Press, 1977. Farran, D., & Ramey, C. Social class differences in dyadic involvement during infancy. Child Development, 1980. 51, 254-257. Feldman, C. The effects o / vtirioirs types ofudidt responses on the syntuctic ucyuisition of two to three veur olds. Unpublished manuscript, Department of Psychology, University of Chicago, 1971. Fenson. L., & Ramsey, D. S. Decentration and integration of the child’s play in the second year. Child Development. 1980, 51, 171-178. Ferguson, C. Baby talk as a simplified register. In C. Snow & C. Ferguson (Eds.), Tulking t o

72

Elizabeth Butes et al.

children: Language input and acquisition . London and New York: Cambridge Univ. Press, 1976. Fillmore, L. The second time around: Cognitive and social strutegies in second language ucyvisition. Unpublished doctoral dissertation, Stanford University, 1976. Fischer, K. W. A theory of cognitive development: The control and construction of hierarchies of skills. Psychological Review, 1980, 87, 477-531. Flavell, J . H. The developmentul theory of Jeuh Piuget. New York: Van Nostrand-Reinhold, 1963. Flavell, J . , Botkin, P., & Fry, C. The development ($role-taking and communicution skills in young children. New York: Wiley, 196R. New York: McGraw-Hill, Fodor, J. A., Bever, T. G., &Garrett, M. F. Theps~~chologyoflanguage. 1974. Freedle, R., & Lewis, M. Prelinguistic conversations. In M. Lewis & L. Rosenblum (Eds.), Inreruction,conversation, und the development of languuge. New York: Wiley, 1977. Furrow, D. The benejh if motherese. Paper presented at the biennial meeting of the Society for Research in Child Development, San Francisco, March 1979. Furth, H. G. Thinking without lunguage: Psychologicd implicutions of deufness. New York: Free Press, 1966. Garcia, J . , & Koelling, R. Relation of cue to consequence in avoidance learning. Psychonomic. Science. 1966. 4, 123-124. Golinkoff, R . Semantic development in infants. The concepts of agent and recipient. Merrill-Palmer Quurterly, 1975, 21, 181-193. Greenfield, P. M., & Smith, J. H. The structure o f communication in early languuge development. New York: Academic Press. 1976. Halpern, E., & Avieter, L. Psycholinguistic skills and sensory motor development within Piaget's theoreticalfr.umework. Paper presented at the 2 I st International Congress of Psychology, Paris, July 1976. Hardy-Brown, K., Plomin, R., & DeFries, J. C. Generic and environmental influences on rate of communicative development in the first year of life. Developmental Psychology, 1981, 17, 704-717. Harkness, S. Aspects of social environment and first language acquisition in rural Africa. In C. Snow & C. Ferguson (Eds.), Talking to childreti. London and New York: Cambridge Univ. Press, 1977. Harlow, H. F., & Harlow, M. K. Effects of various mother-infant relationships on rhesus monkey behaviors. In B. M. Foss (Ed.), Determincmts ($infant behavior (Vol. 4). London: Methuen. 1969. Harmon, R. J., Suwalsky, J. D., & Klein, R. P. Infants' preferential response for mother versus an unfamiliar adult. Journul if the American Academy ofchild Psychiatry, 1979, 18, 437-449. Hock, E. Alternative approaches f a child reuring and their eflects on the mother-inflint relationship. Final Report to the Office of Child Development, 1976. Horgan, D. Nouns: Low 'em or leave 'em. Address to the New York Academy of Sciences, May 7 , 1979. Johnston, J. The language disordered child. In N. Lass, J . Norther, D. Yoder, & L. McReynolds (Eds.), Speech. language. and learning. Philadelphia: Saunders, 1981. in press. Kaye, K . Toward the origin of dialogue. In H. R. Schaffer (Ed.), Studies in mother-infant inferoction. New York: Academic Press. 1977. Kaye, K . The soc~ic~l conlext o f infant developtnetir (Final Report to the Spencer Foundation, June 1974-November 1979). 1980. Kaye, K. The menml untl social life of babies. Chicago: Univ. of Chicago Press, 1981. Kempler, D. Varititiow in langucige ucquisition (UCLA Working Papers in Cognitive Linguistics). Los Angeles: Univ. of California, 1980.

Social Bases of Language Development

73

Kopp, C., Sigman, M., & Parmelee, A . Longitudinal study of sensorimotor development. Developmental Psychology, 1974, 10, 687-695. Kuhn, T . S . The structure of scientific revolutions (2nd ed.). Chicago: Univ. of Chicago Press, 1970.

Kuno, S . Subject, theme and the speaker’s empathy: A re-examination of relation phenomena. In C . Li (Ed.), Subject and topic. New York: Academic Press, 1976. Lashley, K. S. The problem of serial order in behavior. In L. A. Jeffres (Ed.), Cerebral mechunisms in behavior. New York: Wiley, 1951. Leonard, L. Language impairment in children. Merri/l-Pidmer Quirrterly, 1979, 25, 205-232. Lewis, M.,& Rosenblum, L. A. Interaction, conversdon, and the development of lunguage. New York: Wiley. 1977. Li, C.. & Thompson, S. Subject and topic: A new typology of language. In C. Li (Ed.), Subject and topic. New York: Academic Press, 1976. Lock, A. (Ed.). Action, gesture and symbol. New York: Academic Press, 1978. Lock, A. The guided reinvention of liinguage. New York: Academic Press, 1980. Lorenz, K. King Solomon’s ring. New York Crowell, 1965. MacNamara, J. The cognitive basis of language learning in infants. Psychological Review, 1972, 79, 1-13. Main, M . E.rp/ormtion. play iznd ccignitive funi.tioning as related to child-morher attachment. Unpublished doctoral dissertation. Johns Hopkins University, 1973. Matas, L., Arend, R . A . , & Sroufe, L. A. Continuity of adaptation in the second year: The relationship between attachment and later competence. Child Developmenr, 1978,49,547-556. McCall, R. B., Eichorn, D. H., & Hogarty, P. S.Transitions in early mental development. Monographs of the Society for Research in Chill Development, 1977, 42 (3, Serial No. 171). McCarthy, D. Language development in children. In L. Carmichael (Ed.), Munual of Child Psychology (2nd ed.). New York: Wiley, 1954. McNeill, D. The development of language. In P. H. Mussen (Ed.), Carmichtrel’s Manuid of Child Ps-vchology (3rd ed.). New York: Wiley, 1970. McNeill, D. The conceptual basis of lunguage. Hillsdale, N.J.: Erlbaum, 1979. McNew, S. Language and gesture: Co-occurrences and contingencies in speech to young children. Manuscript in preparation, 1981. (a). McNew, S. Muternal gestures: An e.rploration of contingency. Manuscript in preparation, 1981. (b). Meltzoff, A , , & Moore, M. K. Imitations of facial and manual gestures by human neonates. Science, 1977, 198,75-78.

Miller, G. A , , Galanter, E. H., & Pribram, K. H. Pion und the .wwtiire of behavior. New York: Holt, 1960. Miller, J . , Chapman, R., & Bedrosian, J . Detining developmentcilly disubled subjectsfor research: The relationships between etiology. cognitive development rrnd language iind communicative pcvj?ormimce.Paper presented at the Second Annual Boston University Conference on Language Development, Boston, October 1977. Moerk. E. L. Verbal interactions between children and their mothers during the preschool years. Developmental Psychology. 1975, 1 I , 788-794. Mowrer, 0. Leczrning theor-y und behuvior. New York: Wiley, 1960. Nelson. K. Concept, word and sentence: Interrelations in acquisition and development. Psychologiciz/ Review, 1974, 81, 267-285. Nelson, K. Facilitating children’s syntax acquisition. Deve/opmentrd Psy:yc,hology. 1977, 13, 101107.

Nelson, K., Causkaddon. G., & Bonvillian, M. Syntax acquisition: Impact of experimental variation in adult verbal interaction with the child. Child Deve/opmetir. 1973, 44, 497-504. Newmeyer, F. Linguistic theov in Americu. New York: Academic Press, 1980.

74

Elizciheth Bates et al

Newport, E . , Gleitman, L., & Gleitman, H. Mother I’d ruther do it niysey: Some efect.7 rind nun-effects of motherese. In C. Ferguson & C. Snow ( a s . ) , Tulking 10 children. London and New York: Cambridge Univ. Press, 1977. Pp. 109-149. Newson, J. Dialogue and development. In A. Lock (Ed.), Action. gesture. und symbol: The eniergence of’ liitigutige. New York: Academic Press, 1978. Nicolich, L. Beyond sensorimotor intelligence: Assessment of symbolic maturity through analysis of pretend play. Merrill-Pultner Quurterly. 1977, 23, 89-99. Ochs, E. Introduction: What child language can contribute to pragmatics. In E. Ochs & B. Schicffelin (Eds.), Developnicntul prugmmatics. New York: Academic Press, 1979. Ochs, E., & Schieffelin, 8 . (Eds.). Df~vc~/o/~mentrr/ prrigmarics. New York: Academic Press, 1979. Pentz, T. Facilitation ofktnguuge ucquisition: The role of the mother. Unpublished doctoral dissertation, Johns Hopkins University, 1975. Piaget, J. Pluy, dreums rind imittition in child/iood. New York: Norton, 1962. Pinker, S. Formal models of language learning. Cognition, 1979, I , 217-283. Pinker, S . A theory of the acquisition of lexical-interpretive grammars. In J. Bresnan (Ed.), The nienfal representation uf grummciricul relutions. Cambridge, Mass.: MIT Press, 1980. Plomin, R., DeFries, J. C., & McClearn, G. E. Behuviorulgenetics. San Francisco: Freeman, 1980. Premack, D., & Woodruff, G. Does the chimpanzee have a theory of mind? Behrrviorul arid Bruin Sciences, 1978, 1, 515-526. Ramey, C., Farran, D., & Campbell, F. Predicting IQ from mother-infant interactions. Child Development, 1979, 50, 804-814. Ratner, N., & Bruner, J. Games, social exchange, and the acquisition of language. Journul of’Child Ltirigituge. 1978, 5, 39 1-401. Rogosa, D. A critique of cross-lagged correlation. Psychological Bulletin, 1981, 88, 245-321. Rowe, D. C., & Plomin, R. Temperament in early childhood. Juurriril of Personality Assessmcnt, 1977, 41, 150-156. Ryan, J. Early language development. In M. P. M. Richards (Ed.), The integrution ofthe child into a sociul world. London and New York: Cambridge Univ. Press, 1974. Sachs, J . , & Devin, J . Young children’s use of age-appropriate speech styles in social interaction and role-playing. Journul uf Child Lunguuge. 1976. 3, 8 1-98. Sander, L. W. The regulation of exchange in the infant-caregiver system and some aspects of the context-content relationship. In M. Lewis & L. A. Rosenblum (Eds.), /n/errrction, conversution, rind the rlevc~lopmentof lunpuge. New York: Wiley, 1977. Savage-Rumbaugh, E. S., Rumbaugh, D. M., & Boysen, S. Linguistically-mediated tool use and exchange by chimpanzees (Pan Trogludytes). Behuviorul und Bruin Sciences, 1978.4.539-5.54. Scaife, M . , & Bruner, J. S. The capacity for joint visual attention in the infant.’Nuture, (London). 1975, 253, 265-266. Schaffer, H. R. (Ed.). Studies in mothcr-infant interaction. New York: Academic Press, 1977. Schaffer, H. R., Collis, G . M . , & Parsons, G. Vocal interchange and visual regard in verbal and pre-verbal children. In H. R. Schaffer (Ed.), Studies in mother-infbnt intermtion. New York: Academic Press, 1977. Pp. 291-324. Schlesinger, I. M. Relational concepts underlying language. In R. Scheifelbusch & L. Lloyd (Eds.), Lunguuge perspectives: Acyuisitiun. retardurion und intervention. Baltimore: Univ. Park Press, 1974. Schwartz, R. Numes, uctions rrnd ol!jec/s. Unpublished doctoral dissertation, Memphis State University, 1978. Selman, R. Taking another’s perspective: Role-taking development in early childhood. Child Development. 1971, 43, 1721-1734. Shatz, M . On mechanisms of language acquisition: Can features of the communicative environment account for development? In L. Gleitman & E. Wanner (Eds.), Lunguuge acquisition: The, sfute of the urt. London and New York: Cambridge Univ. Press, 1981, in press.

Socitrl Btrses of' Ltrnguugr Development

75

Shatz. M., & Gelman, R. The development of communication skills: Modifications in the speech of young children as a function of the listener. Monogruphs of rho So(.icty./or Resecrr~hin Child Development, 1973, 38 (3, No. 152). Shettleworth, S . J . Constraints on learning. A d w z c ~ r sin the Study o j B r h ~ v i o r ,1972, 4, 1-68. Shore, C. S . . O'Connell, B.,& Bates, E. First sentenw.s in lungutrge tmd symholicpluv. Manuscript in preparation. 198 I . Slobin, D. I. Cognitive prerequisites for grammar. In C. Ferguson & D. Slobin (Eds.), Studies in child lungirugc, d e ~ d o p m e n t .New York: Holt, 1973. Slobin. D. 1. Psycholinguistics (2nd ed.). Glenwood, 111.: Scott, Foreman, 1979. Snow, C.. & Ferguson, C. (Eds.). Tulking to diildren: Lungurrge inpui und cicquisition. London and New York: Cambridge Univ. Press, 1977. Snyder, L. Communicative and cognitive abilities and disabilities in the sensorimotor period. Merrill-Palmer Quurirrly. 1978, 24, 161-180. Spitz, R. A . Thefirst yeor of life. New York: International Universities Press, 196.5. Stern, D. N., Beebe, B., Jaffe, J., & Bennett. S. L. The infant's stimulus world during social interaction. In H. R. Schaffer (Ed.), Studies in mother-infimr iwieruction. New York: Academic Press, 1977. Tinbergen, N. The srrrdv of instinct. London: Oxford Univ. Press, 1951. Tolman, E. C. P ~ r r p s i v rhehuvior oj'unimuls and men. New York: Century, 1932. Trevarthen, C., & Hubley, P. Secondary intersubjectivity: Confidence, confiding and acts of meaning in the first year. In A. Lock (Ed.), Action, gesture und symbol. New York: Academic Press, 1978. Pp. 183-230. Tulkin. S. & Covitz, F. Mother-it$h inrerurrion u r r d intellectual jiuncrioninR at age six. Paper presented at the biennial meeting of the Society for Research in Child Development, Denver, April 197.5. Turiel. E. The development of concepts of social structure. In J. Click & K . A. Clarke-Stewart (Eds.), The development of social utufersrunding. New York: Gardner, 1978. Urwin. C. The development of communication between blind infants and their parents. In A. Lock (Ed.). Ariion. gesture, and symbol: The emergenre of language. New York: Academic Press, 1978. Valian, V. The wherefores and therefores of the competence-performance distinction. In W. Cooper & E. Walker (Eds.), Sentence processing. Hillsdale, N.J.: Erlbaum, 1979. Pp. 1-27. van Dijk, T. Marrosinrrrures. Hillsdale, N.J.: Erlbaum, 1980. Vygotsky, L. S . Thought and lunguage. Cambridge, Mass.: MIT Press, 1962. Werner, H.,& Kaplan, B. Symbolf;mncrtion. New York: Wiley, 1963. Wexler, K.,& Culicover. P. Fvrmul princ.iples c$ lrorguuge trquisition. Cambridge, Mass.: MIT Press, 1980. Williams, C. Personal communication, 1979. Wood, D.. Bruner, J. S., & Ross. G. The role of tutoring in problem solving. Joirrntrl of Child Psychology cmd Psychiurrv, 1976, 17, 89- 100. Zukow, P.. Reilly, J., & Greenfield, P. Making the absent present: Facilitating the transition from sensorimotor to linguistic communication. In K . Nelson (Ed.), Children's ltrngrctrge (Vol. 3). New York: Gardner, 1981, in press.

This Page Intentionally Left Blank

PERCEPTUAL ANISOTROPIES IN INFANCY: ONTOGENETIC ORIGINS AND IMPLICATIONS OF INEQUALITIES IN SPATIAL VISION

Marc H . Bornstein DEPARTMENT OF PSYCHOLOGY

NEW YORK UNIVERSITY

NEW YORK. NEW YORK

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

77

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

79 79 82

11. TWO CLASSES OF PERCEPTUAL ANISOTROPY

A. PERCEPTUAL SALIENCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . PERCEPTUAL EQUIVALENCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. TWO CLASSES OF PERCEPTUAL ANISOTROPY IN INFANCY..

...........

A. PERCEPTUAL SALIENCE IN INFANTS .............................. B. PERCEPTUAL EQUIVALENCE IN INFANTS.. . . . . . . . . . . . . . . . . . . . . . . . . IV. DISCUSSION AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. REVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. SOME INTERRELATIONS BETWEEN PERCEPTUAL ANISOTROPIES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. PERCEPTUAL-COGNITIVE-SOCIAL DEVELOPMENT: ANISOTROPY AND THE FACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. SOME IMPLICATIONS FOR COGNITIVE DEVELOPMENT . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

85 85 97 108

108 109 1I1 I 1I

115

After all, who can fail to be interested in how things grow, mature, and die? -B.

I.

R. MCCANDLESS A N D M. F. GEIS (1975, p. 7)

Introduction

Physical space extends outward from the central ego equally in all directions, and the identification and discrimination of orientation, direction, and location in space are critical to perception and to perceptual development in virtually all 77 ADVANCES IN CHILD DEVELOPMEN7 AND BEHAVIOR, VOL. 16

Copyright @ 1982 by Academic Press. h c . All right6 of reproduction in any form reserved. ISBN 0-12-009716-8

78

Murc H . Bornstein

visual vertebrates. Yet, psychological perceptions of space are skewed. Certain orientations are favored: For example, vertical and horizontal hold many advantages psychologically, while other orientations are at a disadvantage. Likewise, certain directions and locations in space are favored: For example, up and down are salient relative to left and right. A variation in spatial perception or conceptualization with regard to axis or location or direction is termed an anisorropy. ' What are the principal anisotropies in humans? When and how do they manifest themselves psychologically? What perceptual functions do they serve? How general are they phylogenetically? What are their physical, physiological, psychological, or cultural origins? When do they first occur ontogenetically? Do they develop continuously or discontinuously? What is their developmental role? In this article I attempt to address these questions and provide their answers as best we know them. In the first part of this article, I define two common classes of perceptual anisotropy . They are perceptual sulience of the main orthogonuls, meaning the orientation hierarchy of vertical-horizontal-oblique, and perceptual equivalence on the laterul. meaning the similarity of left-right mirror-image sections or enantiomorphs of a pattern. Here, I also briefly review the circumstances under which these perceptual anisotropies emerge in different species, particularly in humans. My discussion is circumscribed to visual perception, since the visual modality is superior to others in the perception of space (e.g.. Geldard, 1970; J. J. Gibson, 1966). In the second and main part of this article, I review and interpret psychological evidence that points to the existence of perceptual anisotropies among newborns and infants in the first year of life. Like older human children and adults and like other infrabuman species, human infants are not equally sensitive to all spatial orientations or to all transformations of patterns that occur in the perceptual world. The two common classes of anisotropy are especially prominent in infancy. Infants consistently find the main orthogonals salient, as shown in a variety of detection, acuity, information-processing, and preference measures. Infants also see and treat as equivalent lateral mirror-image components of visual patterns. This review of the infant literature is not intended to be exhaustive. It purports, rather, to meet three goals: first, to demonstrate that perceptual anisotropies are present in infants; second, to assess how alike or unalike one another immature and mature perceptual anisotropies may be; and, third, to examine some contemporary interpretations of the origins of perceptual anisotropies in infants. Do infants manifest perceptual anisotropies? Are they Anlagen of adult anisotropies? What are their sources? 'Throughout this article orientation will be specified in degrees rotation clockwise from vertical = 0". Orthogonol will specify one or both of the principal orientations in space, viz. vertical (0")or horizontal (90'); and, unless the context specifies otherwise, oblique will specify the principal nonorthogonals, viz. 45" left and right rotations from vertical.

In the third and final part of this article, I discuss some theoretical interrelations between the two classes of anisotropy, the special role of the face in perceptual anisotropies, and some likely implications of infant anisotropies for the development of perception, cognition, and language in childhood.

11.

Two Classes of Perceptual Anisotropy A.

PERCEPTUAL SALIENCE

Although spatial orientation varies infinitely in all directions, the main orthogonals of vertical and horizontal are principal perceptual orientations in visual space for humans as well as for a large number of other animals. Oblique orientations give rise to distinct perceptual disadvantages. The data that construct these generalizations derive from a wide variety of sources (for some general reviews, see Appelle, 1972; Arnheim, 1974; Essock, 1980; J. J. Gibson, 1934, 1966; Howard & Templeton, 1966; Pick, Yonas, & Rieser, 1979; Rock, 1973). Appelle ( 1 972) termed the general disadvantage for obliques in spatial perception the “oblique effect.” Both psychophysical and perceptual studies in adult humans, for example, have shown repeatedly that detection, discrimination, comparison, and assessment are easiest and that identification is most rapid and accurate for stimuli aligned along the principal orthogonals as opposed to any oblique. Essock (1980) reviewed this literature and distinguished two classes of oblique effect, one obtained on tasks that measure the basic functioning of the visual system and the second obtained on tasks that measure differential capabilities of orientation information processing at higher perceptual levels. Class 1 oblique effects derive from psychophysical studies which indicate that visual acuity or sensitivity for a target is greatest when it is oriented along the vertical or horizontal meridian (e.g., Attneave & Olson, 1967; Berkley, Kitterle, & Watkins, 1975; Camisa, Blake, & Lema, 1977; Campbell, Kulikowski, & Levinson, 1966; Corwin, Moskowitz-Cook, & Green, 1977; Emsley, 1925; Higgins & Stultz, 1950; Lennie, 1974; Taylor, 1963; Timney & Muir, 1976) and that perceptual stability is greatest for vertically and horizontally oriented stimuli (Cosgrove, Schmidt, Fulgham, & Brown, 1972; Craig & Lichtenstein, 1953; McFarland, 1968). This class of effect has been linked to characteristics of the neurophysiological substrate of vision (see later). Class 2 oblique effects derive from perceptual-learning and cognitive studies which demonstrate that orthogonally oriented stimuli are at an advantage in tasks that require discrimination, identification, matching, and recognition of targets. Thus, for example, adults judge verticality and horizontality with greater accuracy than obliquity (Keene, 1963); they match and reproduce verticals and horizontals better than obliques

80

Marc. H . Bornsrein

(Lechelt, Eliuk, & Tanne, 1976); and they frequently perceive and reproduce tipped visual forms as closer to the principal orthogonals (Bornstein, 1974, n. 4; Bouma & Andriessen, 1968; J. J. Gibson, 1934, 1937; Olson & Hildyard, 1977). Obliques also take longer than the main orthogonals to name (Attneave & Olson, 1967) and to recognize and identify or differentiate (Olson & Hildyard, 1977). This class of oblique effect has been linked to encoding and memory processes. Essock ( 1980) also investigated adults ' perceptions of differently oriented stimuli in a series of integrated studies that examined the Class 2 oblique effect. In the first study, Essock asked eight observers simply to report the presence or absence of single lines (2.5" of visual angle) presented tachistoscopically (100 milliseconds). Observers were instructed to disregard line orientation (O", 45", 90°, or 135")or position (at a short distance from the fixation point and at one of eight equally spaced locations around the clock). Essock found that when observers were asked simply to detect a stimulus, all orientations were detected equally quickly. In the second study, Essock required eight observers to make a unique response for each of the four orientations used previously. Here observers took significantly longer to identify obliques than the main orthogonals. In the third study, Essock required eight observers simply to classify lines (regardless of orientation) as either obliques or orthogonals. As in his first study, Essock found no oblique effect. In short, "unique identification of stimulus orientation was necessary to produce an oblique effect, whereas tasks requiring detection or classification of the stimuli did not demonstrate an oblique effect" (Essock, 1980, p. 40). Another good example of the role of this anisotropy in spatial vision may be found in the literature of symmetry perception. Moreover, studies of symmetry point up a second important aspect of this anisotropy, namely that between the two principal orthogonals (vertical and horizontal) vertical is especially salient. Symmetry is perceptually special, but not all orientations of symmetry provoke the same effective perceptual advantages; symmetry about the vertical axis is distinctive. The salience of vertical symmetry was early noted by Mach (1906, p. 107), and it has been repeatedly commented on since (e.g., Goldmeier, 1937; Julesz, 1971; Rock & Leaman, 1963). When pitted against matched horizontal or oblique symmetries, repetitions, and asymmetries, vertical symmetries are preferred (e.g., Mach, 1886/1959; Szilagyi & Baird, 1977); vertical symmetries are detected more easily, identified more accurately, and sorted more quickly (e.g., Corballis & Roldan, 1975; Fitts & Simon, 1952; Fitts, Weinstein, Rappaport, Anderson, & Leonard, 1956; Fox, 1975; Julesz, 1971; Palmer & Hemenway, 1978), even when visual presentation hovers around threshold (Johnson & Uhlarik, 1977); and vertical symmetries are remembered better in tasks that may rely on recognition (Fitts et a / . , 1956) or on reproduction (Boswell, 1976; Deregowski, 197 1).

Symmetries about horizontal or oblique axes are usually preferred, processed, and remembered better than asymmetries, but are less salient than vertical symmetries. Some (unresolved) controversy surrounds the relative perceptual salience or advantage for horizontal versus oblique. One position derives from the concept of the “oblique effect” (Appelle, 1972); it argues that vertical and horizontal symmetries are more salient than are oblique ones (Attneave & Curlee, 1977; Gamer, 1970; Goldmeier, 1972; Palmer & Hemenway, 1978). The alternative position derives from the view that symmetry processing involves first mentally rotating nonvertical patterns to the vertical (Shepard & Metzler, 1971); it argues that the salience of nonvertical symmetries is directly related to the degree to which they are displaced from the vertical (e.g.. Corballis & Roldan, 1975; Johnson & Uhlarik, 1977; Mach, 1886/1959). Vertical is not just special in symmetry. Vertical, in contrast to horizontal, is related to a physical universal, a constant, and a natural referent of directiongravity. Vertical divides equal halves of space (left-right) as opposed to unequal ones (up-down). Vertical is perceptually and cognitively salient: A vertical line appears to be slightly longer than an objectively equal horizontal line; velocity along the vertical appears appreciably enhanced over an objectively equal velocity along the horizontal; location decisions on the vertical are easier than on the horizontal; memory for the location of items in a visual scene is more durable on the vertical than on the horizontal; etc. (e.g., Howard & Templeton, 1966; Rock, 1973). The main orthogonals are as perceptually salient for children as they are for adults. Braine (1978a, 1978b) has proposed a theory of perceptual development that posits three distinct stages in the child’s perception and understanding of orientation. She suggests that children first distinguish upright (and vertical) from diagonal, horizontal, and upside down, which are all part of a single class of nonuprights. Second, they distinguish among these nonupright orientations. Third, they distinguish left from right (see Section 11,B). In other words, Braine argues, vertical and upright are special for children. Braine, Lerner, and Relyea (1980) provide clear evidence for a developmental sequence between Braine’s first two stages, the primary focus of this discussion of perceptual salience. Braine and her colleagues gave four groups of 3- and 4-year-olds each one kind of two-alternative discrimination-learning problem that involved pairs of line drawings of familiar figures oriented upright, sideways, or upside down. For upright-nonupright problems, that is, upright-upside down and uprightsideways, 85% of 3-year-olds and 95% of 4-year-olds reached a preestablished learning criterion, whereas for the nonupright problem, that is, upside downsideways and vice versa, only 30% of 3-year-olds and 75% of 4-year-olds reached criterion. By the time upside down is distinguished from sideways, the vertical is providing key information about shape orientation. [Interestingly,

82

Mtrw H . Romstein

Braine rt id. (1980) note that adults take longer to discriminate upside down from sideways than to discriminate upright from upside down or sideways; according to this analysis, levels of orientation processing in adults correspond to developmental stages of coding orientation in children.] Olson (1970) has extensively studied and confirmed perceptuocognitive developmental lags for the diagonal in young children, and Fisher (1980, 1981) has argued that children can code orthogonal lines absolutely, and hence these directions are resistant to change or to decay on account of configuration, external visual cues, presentation conditions, etc., whereas children code oblique lines relationally and in less resistant ways. Bryant (1973), Corballis and Zalik (1977), and Williamson and McKenzie (1979) have all shown that young children have difficulty matching a remembered oblique. In general, perceptual development seems to be at an advantage with respect to vertical and horizontal perhaps because “the frame of reference of vertical and horizontal [is] characteristic of the environment” (E. J . Gibson, 1969, p. 377). The perceptual advantage for the main orthogonals is also widespread phylogenetically. A large number of different species, including octopuses, goldfish, pigeons, rabbits, rats, squirrels, cats, monkeys, and chimpanzees, show one or another class of “oblique effect” (for general reviews, see Appelle, 1972; Bauer, Owens, Thomas, & Held, 1979). One qualification to these generalizations about the main orthogonals is important to underscore. For each individual at any time, there may exist two verticals. “Environmental vertical” is defined by gravity, is fixed, and is perpendicular to the earth; “retinal vertical” bisects the body independent of the body’s rotation in space (Essock, 1980; J. J. Gibson & Mowrer, 1938; Howard & Templeton, 1966; Rock, 1973). Normally, environmental and retinal verticals are congruent. In this article, I refer mainly to verticals as experienced by the organism. 8. PERCEPTUAL EQUIVALENCE

Clearly, the discrimination of orientation is critical to spatial perception. Visual organisms are highly sensitive to orientation specificity and change, and detection of orientation appears to be a relatively early stage of visual processing by the nervous system (e.g., Hubel & Wiesel, 1968). It is curious, therefore, that discriminations of stimuli from their reflections 180” around the main orthogonal axes (as opposed to oblique ones) represent extremely difficult problems for a great variety of animals, including human children and adults (for some general reviews, see Bornstein, 1981; Bradshaw, Bradley, & Patterson, 1976; Corballis & Beale, 1976; Sutherland, 1961; Tee & Riesen, 1974). Reflections about the

vertical axis are called lateral or left-right mirror images, while reflections about the horizontal axis are called vertical or up-down mirror images. It is anecdotal and legend that children and adults frequently experience difficulty discriminating between left and right. In the Analysis ofSensations, Mach (188611959, p. 110) noted that “children constantly confound the letters h and d, p and q . Adults, too, do not readily notice a change from left to right.. . .” Farrell (1979) recently investigated adults’ coding and discrimination of left and right in a series of integrated studies. In his first study, Farrell contrasted orienting with discriminating left, right, up, and down; his design separated the subjects’ simply copying spatial information in the stimulus (manual orienting) from naming (vocal discrimination), which (because the responses themselves are not mirror images in the latter case) constitutes the best evidence for the ability to tell left from right (Corballis & Beale, 1976). Twelve young adults judged the direction of arrows (1” x I ” of visual angle) presented tachistoscopically (100 milliseconds) by moving a lever in the matching direction or by naming the direction; reaction times were measured. Farrell found that telling left from right in a simple perceptual judgment is significantly harder than telling up from down, but there were no dimensional differences in orienting. In his second and third studies, Farrell showed that the left-right disadvantage does not result from difficulty in accessing linguistic codes for direction, and, in a fourth study, he showed that left-right discriminations are confusing for positional as well as directional information. In short, adults find it more difficult to discriminate left-right than up-down. Sekuler and Houlihan (1968) provided another simple and clear experimental demonstration of this phenomenon. They asked 24 young adults to identify as “same” or “different” with regard to orientation pairs of C shapes (approximately 10” square) that were aligned vertically or horizontally and that were repetitions or mirror images of one another. Reaction time was again the principal measure. Considering pairs oriented horizontally, Sekuler and Houlihan found that subjects took longest to identify left-right mirror images (C 7)while they identified up-down mirror images (n U) and repetitions (C C and fl fl) in significantly shorter amounts of time. (The rates for “same” and “different” responses in this experiment were equivalent.) Data of several other investigators experimentally confirm the difficulty or confusability of left-right mirror images in adults (e.g., Bradshaw ef ul., 1976; Butler, 1964; Pomerantz, Sager, & Stoever, 1977; Rock, 1973; Wolff, 1971). Of course, as Mach and innumerable teachers and parents before and since have observed, young children have special difficulties with left and right. The classic demonstration of left-right mirror-image equivalence in children was that of Rude1 and Teuber ( 1 963). They used a two-choice discrimination-learning paradigm in which the two stimuli were simultaneously presented and horizon-

tally aligned. Children were required to identify the “correct” stimulus. The lateral positions of the stimuli were randomized from trial to trial, and the children were told whether they had chosen the correct or incorrect stimulus after each trial. Rudel and Teuber found that children between 4 and 9 years of age have great difficulty in learning to discriminate left-right mirror-image obliques (/ versus \) and mirror-image C shapes (C versus 7)but readily learned to discriminate horizontal from vertical lines (- versus I ) and a U shape from its inversion or up-down mirror image (U versus n). Rudel and Teuber’s results have been repeatedly confirmed in both Western and non-Western cultures (e.g., Huttenlocher, I967b; Over & Over, 1967; Sekuler & Rosenblith, 1964; Serpell, 1971). Left-right confusions are also widespread among infrahuman species. Different animals, including octopuses, fishes, rats, pigeons, cats, and monkeys, perceptually equate left-right mirror images (for some general reviews, see Bornstein, 198I ; Bradshaw cr ul., 1976; Corballis & Beale, 1976; Tee & Riesen, 1974). Several qualifications of these generalizations about mirror-image equivalences need to be made. First, left-right differentiation is usually easier under conditions of simultaneous discrimination than successive discrimination (Aaron & Malatesha, 1974; Bradshaw ef al., 1976; Bryant, 1973; P. L. Harris, Le Tendre, & Bishop, 1974; Over & Over, 1967; Sidman & Kirk, 1974; Stein & Mandler, 1974; Tee & Riesen, 1974; Wohlwill & Wiener, 1964). Thus, coding in memory is implicated in the difficulty of discriminating mirror images; temporal separation of enantiomorphs significantly impairs discriminability of their orientation. Second, the spatial alignment of stimuli is also a significant factor in mirror-image discrimination: Vertical mirror images have been found to be confusing in animals, human children, and human adults (Butler, 1964; Huttenlocher, 1967a, 1967b; Lashley, 1938; Sekuler & Houlihan, 1968; Sekuler & Pierce, 1973; Sekuler & Rosenblith, 1964; Sutherland, 1961; Wohlwill & Wiener, 1964; Wolff, 1971 ), but vertical mirror images have usually been found to be somewhat less confusing than lateral ones (e.g., Bradshaw et al., 1976; Butler, 1964; Huttenlocher, 1967a, 1967b; Sekuler & Rosenblith, 1964). Significantly, mirror images about oblique axes-C are U are mirror images about the /-are not confusing at all. (In the absence of cues associated with stimulus alignment, however, left-right and up-down problems are solved equally well-Fisher, 1979; Fisher & Braine, 1981.) Third, alignment affects discrimination of mirror images. Thus, lateral mirror images horizontally aligned (C 7) are the most difficult to discriminate, followed by vertical mirror images vertically aligned ( u), n followed by lateral mirror images vertically aligned ),( C and vertical mirror images horizontally aligned (U (e.g., Huttenlocher, 1967b; Sekuler & Houlihan, 1968).

n)

85

111. Two Classes of Perceptual Anisotropy in Infancy A.

PERCEPTUAL SALIENCE IN INFANTS

Studies of infant spatial vision suggest that babies see as perceptually salient principal orthogonal-especially vertical-orientations or alignments of patterns. Studies that tap detection and discrimination, processing, and preference in infants all converge on this generalization, and it seems to hold for simple and complex artificial geometric forms as well as for more meaningful patterns like the human face.

I.

Vertical Is Perceptually Salient, Horizotital May Be, and Obliques Are Least So

a . Detection and Discrimitiation . The extreme perceptual disadvantage of obliques relative to the mainorthogonals, the so-called "oblique effect" identified by Appelle (1972), is usually discussed in the context of sensitivity differentials at a postretinal level and is usually studied (and usually emerges) at threshold levels of stimulus contrast. Studies of infant visual detection and acuity that follow in this tradition show that at threshold babies also discern verticals and horizontals better than they do obliques. In addition, infant orientation discrimination has been studied with suprathreshold stimuli. Infants have been shown to discriminate easily between the orthogonals, between orthogonals and obliques, and even between some obliques, though doubtlessly they fail at increasingly fine orientation discriminations (Fellows, 1968). Leehey, Moskowitz-Cook, Brill, and Held (1975) provided the classic demonstration of oblique effects at threshold in young babies. They studied 24 infants between 1.5 and 12 months of age. Leehey and her colleagues used a twoalternative preferential-looking task with 1 1" square-wave gratings displayed at five spatial frequencies (0.75, 1.5, 3.0, 6.0, 12.0 cycles per degree) that were aligned along the vertical, horizontal, or two 45" obliques. Over 10-second stimulus exposures in which an orthogonal was always paired with an oblique of the same spatial frequency, babies at all ages looked at verticals and horizontals equivalently, at obliques equivalently, but at the main orthogonals more than at the obliques. For each age, the maximum difference in frequency of looking at the main orthogonals versus obliques occurred at or near the acuity threshold spatial frequency. Thus, across the first year of life, vertical and horizontal gratings are easier to detect than are otherwise identical oblique gratings. Gwiazda, Brill, Mohindra, and Held ( 1 978) replicated these results and also found that acuity for vertical gratings increases more rapidly over the first year of life than does acuity for oblique gratings.

86

Murc H . Bornstein

Suprathreshold discriminations of orientation are much more common in the infant literature. Although Ling (1941) early on concluded that 6-month-old infants do not discriminate orientation differences in perceiving form, several investigations since have determined that infants facilely discriminate among many changes in the orientation of a stimulus. Orthogonal, that is, vertical-horizontal, discriminations have been the most prominent and the easiest to demonstrate (e.g., Bornstein, Gross, & Wolf, 1978; Essock & Siqueland, 1981; Fagan & Shepherd, 1979; Gross & Bornstein, 1978a; McKenzie & Day, 1971; Moffett, 1969). For example, Gross and Bornstein ( 1978a) demonstrated infants' simultaneous discrimination between orthogonals in a relatively simple and clear way. Seven 4-month-old babies were familiarized over a series of trials with a pair of identical geometric stimuli (1C ) that were approximately 2 1" on a side. On every other trial after familiarization, either a 90" rotation (u)or a 180" rotation ( 3 )of the original familiarization stimulus was substituted for either the left or right member of the familiarization pair. If infants looked more at the probe than at the familiar stimulus. it would indicate that they discriminated the probe from

75

-PREFERENCE

FOR

I I P R E F E R E N C E FOR

u 1

65

55 Oh

45

PREFERENCE

35 25

15 5 N=7 Fig. I . Foll~~witig,fuinilicrrizuti~~~~ with a pair of C shapes. itgonts prefir u 90" orthogonul rottition {!A)of the stimulus to LI 180"niirror-irnuge rotmion (3) of the .stinrulus. 111this r.rprvirtrent, preference indictrtes discrimincrtion (NS.not signijicmr).

the familiar stimulus. (Of course, failure to show a preference would not necessarily give evidence of inability to discriminate.) The results of the u probes are pertinent here. (Results of the left-right mirror-image probe, 7, will be discussed in Section 111,s.)On probe trials, babies looked more at u than at the original L , that is, they discriminated the 90" rotation (Fig. 1). Orthogonal-oblique and oblique-oblique discriminations-even those above normal limits of resolution-are thought to present increasingly difficult problems for young children. Reasonably fine orthogonal-oblique discriminations have been studied in three infant experiments; both geometric and realistic stimuli have been employed. Using a habituation-test paradigm, Bornstein er ul. ( I 978) showed that 4-month-old babies discriminate vertical from 45". Twenty babies habituated to a 45"-right oblique line (1.5" x 12.5") over 10 successive 10-second trials. On nine test trials afterwards, the babies were shown the habituation stimulus, a vertical version of it, and its 45"-left oblique version. (Results for the mirror image are discussed in Section 111,B.) In the test, the babies dishabituated to the vertical relative to the habituation oblique (Fig. 2). Using a similar habituation-test paradigm, Weiner and Kagan ( 1976) examined orientation discrimination between a horizontal line and one rotated 35" from the horizontal. Their 5-month-old babies successfully discriminated between the orthogonal and the oblique. Finally, using a paired novelty-preference paradigm, Fagan and Shepherd ( 1979) studied orthogonal-oblique discriminations with faces. (Here it must be borne in mind that faces are special perceptual stimuli.) Fagan and Shepherd found that infants between 5 and 6 months readily discriminated a 45O-diagonal from a vertical but failed to distinguish diagonal from either horizontal or from upside down. Thus, the upright vertical face is special and uniquely discriminable from diagonal, while horizontal and diagonal are treated similarly (see Braine, 1978a. 1978b). Faces, like symmetry, represent a clear case in which vertical possesses a higher status in the orientation hierarchy than does horizontal. In short, babies in their first half-year discriminate at least 35"-45" rotations of a stimulus from an orthogonal, especially vertical. Finer discriminations have yet to be tested. Oblique-oblique discriminations (like all linear discriminations) are in fact discriminations between mirror images since any pair of obliques falls equally on two sides of a main orthogonal or a midoblique. The few extant studies of infant oblique-oblique discriminations therefore can be placed into two categories. One category encompasses discriminations between obliques that are mirror images about an ohliqite. Vertical-horizontal and orthogonal-oblique discriminations also fall into this category. (The second category encompasses discriminations between obliques that are mirror images about an orthogoizal; a more complete discussion of them will be presented in Section 111,B.) One empirical demonstration of the infant's ability to discriminate between two obliques was provided by

88

c

35t 0 4

301 ......

1':I

:.:, .........,....., .........._.... ..............., ........,...... .........,...... ........,...... ................ ...............

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

L

0

I

\

45ORIGHT

VERTICAL

45" LEFT

Fig. 2 . Following hubituutioti to (I 4S"-right oblique, injiu7t.s muintuin habituation to the 4 5 " - r i ~ h t oblique. they dishubituute to vertical, und they generulize hnbitucition to (I 4S"-lefi oblique, the lutcrril mirror imugr ($the hubituution stimulus. I n this experini~nt,dishabitucition indicures disc.riminc~tion.(A,ftrr Bornstoin, Gross. & Wolf. 1978, Experiment I I . Copyright Elsevier Sequoici S . A . , used by pertnis.sicin. )

Bornstein et (I/. (1978). We tested 10 4-month-olds for their ability to discriminate two rightward-tilting oblique lines-one tilted 20" right from vertical and the other tilted 70"-that were mirror images about the 45" oblique. Babies were habituated to the 20" line ( I " x 12") over 10 successive 10-second trials and immediately tested twice with the line rotated 70". The babies recovered looking from the end of habituation to the test (Fig. 3), that is, they discriminated 50" of rotation between two obliques in the same Cartesian quadrant. This is the only infant study to my knowledge in this category*; to define early capacity further, ZEssock and Siqueland (1981) found that 2-month-olds failed at a successive oblique-oblique discrimination between gratings that were mirror images about an oblique oriented 22.5" left of vertical; however, their experiment is flawed as a study of orientation discrimination since their stimuli, oriented 22.5" right of vertical and 67.5" left of vertical, confound a left-right discrimination with a pure 90" oblique-oblique orientation discrimination. The experiment probably falls into the category of confusable left-right mirror images (see Section 111,B).

89

50 w 5

-

I-

g40 Y

0

s

5 30

W 0

K W

a

z 20 a

W

H

:

.

..

J

I

4

20° RIGHT

70' RIGHT

Fig. 3. From the end of huhirrcution to (1 20"-right obliquc. infunrs dishubituute to a trst 70°-right oblique. I N rliis experiment. clishabitiiation indicates di,srriminution. (After Bornstein, Gross, & Wolf; 1978. Experiment 111. Copyright Elsevier Seyuoiu S . A . , used by permission.)

future research in orientation discrimination ought to be concerned with the development in infancy of visual resolution of a variety of small orientation differences. Finally, it is worthwhile to consider the role of stimulus complexity or organization in a discussion of infant orientation discrimination. Fisher, Ferdinandsen, McCall, Kennedy, and Appelbaum ( 1977) also assessed oblique-oblique orientation discriminations in 2.5-month-olds in a study of the discrepancy hypothesis. Their stimulus was a multicolored shaft with arrowheads at both ends and was oriented o", 33", 66". and 90". Infants' discriminations among all possible pairs in both directions from vertical were studied in a habituation-test design, which included two orthogonal (90")discriminations. two magnitudes (33" and 66")of orthogonaloblique discriminations (vis-a-vis both vertical and horizontal), and two oblique-oblique (33") discriminations. Unfortunately, the design of the experiment and analysis of the results did not permit separate assessment of orientation discrimination: "We caution against attempting to interpret the unanalyzed mean fixation times. , . . Tne major point of the design is that such means are thoroughly confounded-and therefore not interpretable. . ." (McCall rt d , , 1977, p. 779).

Mtrrc H. Bornstein

90

and Bornstein (198 1 ) compared,symmetry and orientation discrimination in four groups of 12 4-month-old babies. This discrimination experiment used a singlelook infant-control procedure (Horowitz, 1974). Babies were habituated with one symmetry pattern (vertical, horizontal, or asymmetrical) to a fixed criterion; afterward, they were tested with a different symmetry pattern. All possible pairings were tested. The patterns were closed symmetrical or asymmetrical polygons that were equated for contour, perimeter, area (of approximately 20” square), and number of turns. Babies discriminated vertical symmetry from asymmetry and vertical symmetry from horizontal symmetry (and vice versa), but they did not discriminate horizontal symmetry from asymmetry or asymmetry from asymmetry. This pattern of results suggests, first, that the global organization embodied in vertical symmetry promotes perceptual discrimination and, second, that the infant’s perceptual advantage for vertical symmetry reflects an interaction between orientation and the unique qualities of symmetry. Infants seem not to be sensitive to the structural organization in symmetry if that organization is aligned about the horizontal axis (rather than the vertical). Vertical symmetry is special, and vertical is special. In summary, babies discern orthogonals best, that is, they show an “oblique effect, at threshold, but they can discriminate between orthogonals, orthogonals and obliques, and obliques and obliques if above-threshold levels of stimulation are used and the stimuli are simple. Though still untested, the resolution of angular disparity by the infant visual system is probably comparatively deficient relative to adults; the child literature (see later) suggests that it ought to be increasingly deficient away from the orthogonals. When more complex stimuli are used, vertical separates from horizontal and shows a perceptual advantage. This review of detection and discrimination studies suggests that for infants the main orthogonals-particularly vertical-are perceptually special next to obliques. ”

b. Processing. The perceptual salience of vertical, or of the two main orthogonals generally, has also been studied by submitting to experimental scrutiny infants ’ long-term visual processing of stimuli that are constant in form but varying in orientation. Presumably, differences in infants ’ long-term patterns of attention, for example, their rates of habituation, to the same stimulus oriented in different directions indicate something about the relative salience of different orientations for them. When E. J. Gibson, Owsley, and Johnston (1978) habituated infants to two types of rigid motion, they found that the babies’ rate of looking descended to a constant habituation criterion faster for a stimulus rotated around a vertical axis than for one rotated around a horizontal axis. As part of a larger study on the perception of symmetry in infants, Bornstein, Ferdinandsen, and Gross ( 198 1) investigated rate and amount of habituation to

91

Perc~eptualAnisotropies in Injkncy

visual stimuli that varied in symmetry or in orientation of symmetry. Eighteen 4-month-old infants habituated to six-element red geometric patterns that were symmetrical about the vertical or horizontal axis or that were asymmetrical; the stimuli were otherwise equivalent in terms of perimeter, contour, and area (approximately 21" on a side). Rate of habituation was the principal measure; rate was assessed by first dividing the 240-second habituation period into 24 consecutive 10-second segments, then using average looking in the first three segments as a baseline for each child, and finally calculating the number of segments for each stimulus type that each child required to reach a fixed habituation criterion (e.g., three consecutive segments

State Answer

Count Objects In Representotion State Lost Number As Answer

-

Fig. 7 . Model oj'strutegy choice in uddition. (Note: The probabilities in parentheses represent the percentage of trials on which children used no visible strategy and answered correctly. They can be thought of as associative strengths.)

298

Robert S. Siegler and Mitchell Robinson

criterion, they give it as the answer. Otherwise, they augment their representations of the numbers indicated by augend and addend, either externally, by putting up their fingers, or internally, by forming some type of imaginal representation. If their confidence in an answer at this point exceeds the criterion, they give it; otherwise, they count their fingers or the imaged objects and state the last number of the count as their result. This model gives rise to the four approaches that we observed. If children answer at the first “recall” point, they will not have used any visible straiegy. If they put up their fingers but answer without counting them, they will have used the finger strategy. If they image objects corresponding to augend and addend and then count aloud, they would be classified as having used the counting strategy. Finally, if they put up their fingers, count them, and answer after counting, they would be classified as having used the counting fingers strategy. The model also suggests explanations for the relative solution times and accuracy rates of the four strategies. It predicts straightforwardly that the no visible strategy approach should be the fastest, the fingers approach the next fastest, and the counting and counting fingers approaches the slowest of the four strategies; all of the processing steps necessary to execute each of the faster strategies are included within the steps necessary to execute the slower ones. Predictions of relative accuracy also can be derived, albeit not quite as directly. Both the overall low accuracy and the bimodal distribution of accuracies of the no visible strategy approach follow from the view that some children used this approach because they set loose criteria for deciding when they knew the answer, and others used it because they did not need any external aids to retrieve the correct answer. The high accuracy of the counting fingers strategy would have been expected since 4and 5-year-olds have been shown to be very adept at counting the 2-9 objects required by the problems (Gelman & Gallistel, 1978). Several considerations may have contributed to the high accuracy of the fingers approach: pattern recognition of the number of fingers that were put up, kinesthetic cues associated with putting up particular sets of fingers, and longer search time than that typical of the no visible strategy approach.’O Finally, the relative inaccuracy of the accounted for 50%. Another possibly relevant source of data was that on the counting and counting fingers strategy trials, on which the 4- and 5-year-olds were heard counting aloud, their counts always started from one. Thus, it was difficult to tell on the no visible strategy trials whether children generally used the min approach, whether they generally used the sum approach (like the min model but counts start at one), whether some children used one and some the other, or whether the same children sometimes used one and sometimes the other. ‘“Tounderstand why kinesthetic cues associated with raising one’s fingers might be helpful for recalling a sum, consider what would happen if your eyes were closed and someone lifted one or more fingers on each of your hands. With at least some combinations, the number of fingers raised might well “feel” like 2,4, or 10 (for example). Whether children are helped by such information is an open question, but the kinesthetic cues offered at least one explanation for the superior accuracy of the fingers approach.

Numerical Ur~ciersrcirulings

299

counting approach was predicted by Kosslyn’s (1978) finding that children have difficulty maintaining a clear image for as long as it took to execute this strategy, 9 seconds. Perhaps the most important feature of the model is that it allows us to account for the correlation between problem difficulty and strategy use. No explicit knowledge about problem difficulty is required to produce the relationship. Instead, at each step in the solution process, the child considers whether his or her confidence in an answer exceeds the level demanded by his or her criterion. The more difficult the problem, the less likely it is that this will occur. Thus, at least those children whose criteria are relatively high are led to take increasingly effortful steps to solve the more difficult problems. In a sense, these children use internalized strategies when they can and externalized ones when they must. Such flexible strategy selection has obvious advantages. It minimizes effort while maximizing the probability of a correct answer. By simply adjusting the confidence criterion, children can adapt to situations in which accuracy is the critical consideration or to situations in which speed or lack of effort is. Adults have been shown to possess similar propensities to avoid cognitive effort and to adjust strategies to the difficulty of particular problems. Siegler and Atlas (reported in Siegler & Klahr, in press) found that adults computed quantitative solutions to balance scale problems only when the problems could not be solved by simpler qualitative comparisons. Glushko and Cooper (1978) found that even in simple sentence verification situations, adults varied their approach depending upon the task demands. These and the findings presented here suggest that from early in childhood, two systemic principles may govern the construction of information-processing routines: minimize the effort needed to accomplish any particular goal, and maximize each routine’s flexibility to adjust to different task environments. These principles would have obvious adaptive value and would be in keeping with the flexible strategy use that has been so frequently observed. The implication for future research is that examining the ways in which people choose among alternative strategies for solving problems may be at least as informative as focusing on how they execute any given strategy.

VI.

Conclusions: The Development of Numerical Knowledge

At the outset of this article, we proposed to examine several aspects of young children’s knowledge of numbers, to devise models of their knowledge within each task domain, and eventually to formulate one or more comprehensive models, including the information within each of the specific ones. This last goal, the formulation of models that stretch across task domains, has been given

300

Rohcrt S. Siegler and Mitchell Robinson

considerable homage in the abstract by developmental psychologists, but few such accounts have been stated at a sufficiently precise level to be meaningfully evaluated. We believe that the formulation of detailed but encompassing models is crucial to understanding cognitive growth. Therefore, in this last section, we shall focus on the issue of what type of larger system might produce the numerical skills of preschoolers that we observed in each of the particular areas. Newel1 (1973), in his article, “You Can’t Play 20 Questions With Nature and Win, ” eloquently argued the case for building large-scale integrative models. He contended that although cognitive psychologists have succeeded in identifying robust phenomena and in accounting for performance in particular situations, the research has failed to cumulate. Among Newell’s suggested means of escape from this dilemma was devising a single model capable of producing performance on many tasks. This suggestion has been influential in motivating large-scale computer simulations of thought, language, imagery, and memory (Anderson, 1976; Kintsch & van Dijk, 1978; Kosslyn, 1978; LNR, 1975). The models that we shall present of young children’s knowledge of numbers differ from these others in being much less ambitious in scope, in not yet being specified at the level of running computer simulations (programs in OPS5 are currently being written), and in being primarily concerned with development. However, the motivation for building them was the same. Figure 8 outlines our current understanding of preschoolers’ knowledge of numbers. The three models within Fig. 8 are ordered from the least to the most advanced, and correspond to the knowledge that we hypothesize is most often possessed by 3-, 4-, and 5-year-olds, respectively.” A cursory examination of the models reveals two features: they are quite forbidding looking, and they appear rather similar to each other. Because of the models’ forbidding appearance, we will describe one of them, Model 11, at some length. Because of the similarities among the three models, we will characterize the depictions of the least and the “The models in Fig. 8 are intended to provide descriptions of the modal tendencies among 3-, 4-, and 5-year-olds, but should not be taken to imply a lockstep progression among the three skills or a perfect correlation between age and skill. Empirically, each skill possessed a moderately high correlation with age and also a moderately high correlation with the other two skills. Age correlated r = .64 with percentage of correct answers on the numerical comparison problems, r = .68 with percentage of correct answers on the addition problems, and r = .49 with how high children counted. Across the three age groups, the percentage of correct answers on the addition task correlated r = .80 with the percentage of correct answers on the numerical comparison task, the percentage of correct answers on the comparison task correlated I’ = .60 with the highest number counted on the counting task, and the percentage of correct answers on the addition task correlated r = .65 with the highest number counted on the counting task. The correlations between pairs of tasks were also fairly high within age groups, averaging r = .54 for the 9 within-age correlations. Overall, 27 of the 39 children (70%) who performed all three tasks would have been assigned to the same aggregate model by virtue of their performance on each of the three individual tasks.

most advanced knowledge in terms of differences between them and the formulation of intermediate level knowledge. When we examine Model 11, the model of 4-year-olds’ knowledge, a basic hierarchical form becomes evident. Numbers as a class are at the top of the hierarchy, then categories of numbers (e.g., small numbers), and finally individual numbers (e.g., 6). There are connections both across and within levels of the hierarchy. At the top, numbers as a class can be operated upon by a number of processes: they can be counted, their magnitudes can be compared, and they can be added and (presumably) subtracted. That children treat numbers as a class distinct from other classes was evident in what they did not do as much as in what they did. No child ever gave a nonnumeric answer to an addition problem or used any nonnumeric term in counting. Children at times did use nonstandard numbers in their counting strings, but the nonstandard numbers were always combinations of standard ones. Therefcre, we believe that several processes that can operate upon numbers are attached to numbers as a class rather than to particular groups of numbers or to individual numbers. (The details of these processes are omitted from the diagrams in Fig. 8 only because of considerations of space; they are shown earlier in the article, in the figures indicated.) At the next lower level of the hierarchy are categories of numbers, ordered in terms of magnitudes. Both the number conservation data reported by Siegler (1981a) and tne magnitude comparison data reported in the present investigation suggest that these categories possess psychological reality for young children. Illiistratively , the conservation operators that are applied to small numbers differ from those that are applied to large ones; as shown in Fig. 1 B, children apply the correct transformational rules to small numbers of objects, but with larger groups judge the longer row to have more. The numerical categories occupy an intermediate position within the hierarchy. Each category is linked both upward to the class of numbers and downward to individual numbers. The particular probabilities linking the categories to the individual numbers are based on those that appeared in the Fig. 5B representation-the empirically derived probabilities that 4-year-olds assigned each label to each number. (In order to make Fig. 8 relatively readable, all probabilities have been rounded to the nearest tenth and probabilities of 10% or less are not shown.) The lowest level of the hierarchy involves individual numbers. In addition to being tied to the category labels with varying probabilities, the numbers are at times tied to each other by “next” connections. Some numbers are also labeled as members of the digit repetition and rule applicability lists. The smaller ones are involved in specific addition facts that the children are more or less confident of knowing. Although our experiments did not tap other information about indi-

A MODEL

COUNTING PROCESSISHOWN IN FIGURE 3 A ) MAGNITUDE COMPARISON PROCESSISHOWN IN FIGURE 5 A 1 AOOlTlON PROCESS SUBTRACTION PROCESS ON PROCESSISHOWN IN FIGURE IA)

I

w N 0

0 *I72

*,;3

+2=3 Usually sfarl

Number of

covnllng with

WCL.

orms.

Less thon

Number of

Number of

c o n . rides ond stdes ond leqr comers on cornera on

on verson

lrsongle

Nurnbcr of

NUMBER OF FINGERS

Imperr on hond

square

m y other

number

Fig. 8 . Integrative models of preschoolers' knowledge of numbers: ( A ) Model I .

29

30 31

39

100. 1000

6. MODEL

IT NUMBERS &UPON

W

0 w

+1=2(FJ6) +2=3(6B) 3 =4( 45) * 4 = 5 ( 35) 6 5=61491

*I = 3(75) *2=4(69) +3.5(49) +4=6(2 8 ) +5=7(26)

+1=4(45) .2=5(40) +3=6(35) t4=7( 09) +5=B( 13)

usually start countmg

Numbn 01 Of eyes. ears, arms. legs on

Number 01 sides ond

wlh

.I = 5 ( 35) * 2 =6(23)

+3=7(071 + 4 = B (36) .5=9( IS) Numbtr 01 sides ond

corners on

corners on

lr4a"gle

sawre

+I =6149) + 2 =7( 52) t 3 = a( 341 + 4 = 9( O W +5=10(38) Number 01 Impas on hand

CAN BE OPERATED+ BY

COUNTING PROCESS (SHOWN IN FIGURE 38) MAGNITUDE COMPARISON PROCESS(SH0WN IN FIGURE 581 ADDITION PROCESS(SH0WN IN FIGURE 7 ) SUBTRACTION PROCESS

Number of lingers

Old

Number o f loes

person

Less than any olher number N u m b n of heads. bodies. noses. mouths m person

*=

DIGIT R E P E T I T I O N LIST

Doddy 15 30 years

* *=

RULE A P P L I C A B I L I T Y L I S T

Fig. 8 . ( B ) Model I I . Model III on following p a g e .

C MODEL III (COUNTING PROCESSISHOWN IN FIGURE 3 C I

CATEGORICAL ORGANIZATIONl PROBABIL. I T I

A

5

L

I

2.3

4.5

6.7.8.9

B

2

I

I

2

3.4

1

I

2.3.4

5.6

5.6.7.8.9 7.8.9

D

I

I

2.3

4.5.6.7

8.9

1

E

I

I

2.3.4.5

6.7.8

9

I

c

*:DIGIT

REPETITION LIST

**=RULE

APPLICABILITY L I S T

*t*= HUNDREDS

LIST

Fig. 8. ( C ) Model 111.

I I

J

Nicmericul Understundings

305

vidual numbers, informal discussions with preschoolers suggest that they know many other facts about them. For example, a 4-year-old told us that 1 is the number that she starts counting with, that it is the number of heads, bodies, noses, and mouths on a person, and that it is the smallest number. The Model I1 depiction of moderately skilled performance provides a vantage point for considering the more advanced knowledge depicted in Model 111 and the less advanced knowledge depicted within Model I. First, let us consider some properties that are hypothesized not to change within this age and skill range. As mentioned above, the basic hierarchical form of the representation is constant across the three models. The children have knowledge about numbers in general, about categories of numbers, and about particular numbers. Also relatively constant across the models are many of the particular connections within and across levels of the hierarchy: even in Model I, the larger numbers are more often assigned to the larger categories; even in Model I, some of the “next” connections between digits are present; even in Model I, some facts linking individual numbers to other semantic properties (e.g., people have two hands) are known. Development in these (though not all) aspects of the representations appears to be a gradual, incremental process. At the other extreme, development can be seen in sharpest relief in the processes that children apply to their representations. These processes change greatly in all three task domains between Models I and 11, and the processes for counting and comparing undergo large changes between Models I1 and 111 as well. The counting process changes from using only “next” connections in Model I to also using rule applicability and digit repetition lists in Model I1 to using all of the above information and also the hundreds list in Model 111. The magnitude comparison process changes from guessing in Model I to comparing labels attached to individual numbers in Model I1 to comparing labels derived from categorical organizations in Model 111. The addition process changes from sole reliance on memorized facts in Model I to supplementary use of reconstructive strategies such as putting up fingers and counting fingers in Models I1 and 111. The pattern is reminiscent of the often-expressed speculation that development entails at least as great a growth in what children can do with information as in the amount of information that they possess (Bruner, 1973; Piaget, 1972; Simon, 1972). The issue of intertask relationships is addressed implicitly in all three models. The models suggest that preschoolers ’ understandings of counting, comparing, conserving, and adding are linked in some ways, but not in all of the ways that they could be. All of the processes operate upon a common representation, and this seems to produce some commonalities. Most dramatically, both conservation and magnitude comparison processes utilize the categorizations of numbers, and all processes except magnitude comparison make use of the links among the individual numbers. Other intertask connections that could have been present

306

Ruhcw S . Sieglrr

(inti

Mitchell Rohinsr~ri

were not, however. Preschoolers could have used their knowledge of counting to compare numerical magnitudes but they did not seem to. They could have used their knowledge of comparing to add numbers, as in the Groen and Parkman min model. but again they did not seem to. This last finding especially suggests that early mathematical skills may develop in relative isolation from one another; once children are proficient in the individual skills, they may make greater use of the potential interconnections among them. How can we evaluate the quality of these models? Empirically, they predict in detail preschoolers’ counting, comparing, conserving, and adding. Model I1 can again be used to illustrate. When counting, the model will stop at “9s,” will skip and repeat entire decades, will introduce nonstandard numbers if the boundary of its digit list is too high, and will count on at least to the ne3t “9” from points within or beyond its spontaneous counting range. When comparing the magnitudes of numbers, it will err most often on problems with large minima and small splits, will assign labels that correspond reasonably well to the relative magnitudes of the numbers, and will learn from instruction that adds an overall categorical organization to the existing connections between individual numbers and categories. When performing number conservation problems, it will judge small number problems in terms of the type of transformation but will judge large number problems in terms of the relative lengths of the rows. When adding, it will usually recall the answers to the easiest problems without using visible strategies, and will more frequently use such visible strategies as the problem increases in difficulty. Thus, the model mimics a considerable range of preschoolers’ behaviors in manipulating numbers. A second virtue of the models is the quality that Klahr and Wallace (1976) termed developmental tractability. For most of the changes between models, we can easily imagine how the more advanced form could grow out of the less advanced one. In counting, children first learn the “next” connections that are the only relations that bind the first numbers they encounter; then they add to this knowledge information about the cyclical patterns inherent in the next higher numbers they learn; eventually they extend the list membership notion to include the much larger numbers that they encounter yet later. In learning about numeiical magnitudes, children first obtain a rough sense of magnitudes that allows them to assign individual numbers to categories having some correspondence to the sizes of the numbers; then they learn how to use the categorical information to compare magnitudes; finally they impose overall categorical organizations that subsume the connections between individual numbers and categories but avert errors. In learning about number conservation, children first rely on the type of transformation only in limited situations, and gradually expand that reliance to encompass all three transformations and all set sizes. In adding, children first memorize solutions to specific problems, and then learn supplementary reconstructive strategies to use on problems where they cannot retrieve the answer.

Numerical iJnder.s:andings

307

Thus, development in this age range and content area involves few false starts; children build on what they already know to construct increasingly successful approaches. The separation between representations and processes in the model proved useful for specifying the source of this developmental tractability. Modeling approaches that focus solely on processes, such as the rule assessment approach, might have revealed as much about developmental changes in the preschool period, but probably would not have revealed the developmental constancies that also were present. The representation-process distinction also provided a basis for hypothesizing why early error patterns often foreshadow later reaction time patterns: increasingly powerful processes operating upon fundamentally similar representations. A final strength of the models is that they should be easy for us and other investigators to build upon. They can be expanded both outward, to encompass additional aspects of preschoolers’ knowledge of numbers, and upward, to include the more advanced knowledge of school age children. One early sign of this intellectual “developmental tractability” was the ease with which we could integrate the new information that 4-year-olds in the magnitude comparison training experiment were taught with the model of their existing knowledge (i.e., the extension of the Fig. 5B model to produce the one represented in Fig. 5D). We anticipate that it also will be relatively straightforward to add to the present models information about preschoolers’ ability to subtract, to count objects, to subitize, and to estimate the numerosity of large collections. Other reasonable goals include expanding the models upward so as to include more complex addition and subtraction skills, the relationships of addition and subtraction to multiplication and division, and the extension of arithmetic operations to the rational numbers. This is not to underestimate the difficulty of achieving these objectives, but rather to affirm that the present hierarchical models provide a base that is far from closed. All models have weaknesses as well as strengths. The two greatest weaknesses of the present models seem to be a lack of detail concerning how children choose to use a particular process in a particular situation and a lack of flexibility to cope with novel situations. With regard to the first point, the model of each process is introduced by the rather opaque test, “conditions for process x met?” Even the 3-year-olds counted when we asked them to count, compared when we asked them to compare, and added when we asked them to add. We do not understand, however, how they knew to do so. Recent efforts to discover how children interpret arithmetic word problems (Carpenter & Moser, 1981; Greeno et NI., 1981 ; Nesher, I98 1) represent a first step toward modeling how children understand instructions. Without further research on how language understanding occurs, however, this part of the model must remain a black box. The second weakness of the models is their lack of flexibility for adapting to

Robert S . Siegler and Mitchell Robinson

308

novel situations. This weakness applies most directly to the portrayal of magnitude comparison. The links between categories and individual numbers are presented as fixed in all three models. This may be a realistic depiction of the long-term memory contents of preschoolers, but almost certainly would not continue to be a realistic depiction in older children and adults. The lack of even a poorly developed mechanism for taking into account the effects of context is a weak point in the general developmental tractability of the models. In addition, just as children may not apply the same process to solving each addition problem, they may not apply the same process to solving every magnitude comparison problem. They may directly retrieve some pairs, may judge others relative to some common reference point, and may in general use any number of idiosyncratic judgment techniques in the comparison process. Illustratively, when one of us recently asked his 5-year-old son whether 16 or 33 was the bigger number, the child counted out loud from 1 to 16 and then said that 33 was bigger. Our error data, Sekuler and Mierkiewicz’s (1977) reaction time data, and previous observations of this child indicate that the counting strategy is far from the rule in this age range or even for this individual; nonetheless, children may use it, and probably many other approaches, sometimes. Both the strengths and the weaknesses of our models converge on two final points. First, conceptual development is far too complex for us to assess children’s understanding by examining performance on a single task. No one age is the age at which a concept is understood, and there is little meaning to saying that one concept is understood before, after, or simultaneously with another. Conceptual understanding has many facets, and only by investigating a concept both broadly and deeply do we have any hope of discovering what people know about it. Second, it is possible and desirable to build integrative models of children’s knowledge across different tasks corresponding to a single concept. These models help us to realize which aspects of children’s understandings we have accounted for, to notice the aspects that we have not yet addressed, and to face those implications that we did not intend and would like to change. In short, such models can help the work cumulate.

REFERENCES Aiken, L. R . , & Williams, E. N. Three variables related to reaction time to compare digits. Perceptual and Motor Skills, 1968, 27, 199-206. Anderson, 1. R . Language, memory, and thought. Hillsdale, N.J.:Erlbaum, 1976. Anderson, J R. Arguments concerning representations for mental imagery. Psychological Review, 1978, 85, 249-277. Ashcraft, M. H., & Battaglia, J. Cognitive arithmetic: Evidence for retrieval and decision processes in mental addition. Journal of Experimental Psychology: Human Learning and Memory, 1978, 4, 527-538. I

Numerical Understundings

309

Banks, W. P. Encoding and processing of symbolic information in comparative judgments. The Psychology ofleurning and Motivation, 1977, 11, 101-159. Banks, W. P., Fujii, M., & Kayra-Stuart, F. Semantic congruity effects in comparativejudgments of magnitude of digits. Journal of Experimental Psychology: Human Perception and Performunce, 1976, 2, 435-447. Banks, W. P., & Hill, D. K. The apparent magnitude of number scaled by random production. Jourtzul ojExperimeniol Psychology, 1974, 102, 353-376. (Monograph) Beilin, H. Cognitive capacities of young children: A replication. Science. 1968, 162, 920-921. Bever, T. G., Mehler, J., & Epstein, J. What children do in spite of what they know. Science, 1968, 162, 921-924. Braine, M. D. S. The ontogeny of certain logical operations: Piaget's formulation examined by nonverbal methods. Psychological Monographs. 1959, 73 (Whole No. 475). Brainerd, C. J. The origins of the number concept. New York: Praeger, 1979. Brown, A. L. The construction of temporal succession by preoperational children. In A. D. Pick (Ed.), Minnesota Symposium on Child Psychology (Vol. 10). Minneapolis: Univ. of Minnesota, 1976.

Brown, A. L. Knowing when, where, and how to remember: A problem of metacognition. In R. Glaser (Ed.), Advances in instructional psychology. Hillsdale, N.J.: Erlbaum, 1978. Bruner, J. S. Beyond the informution given: Studies in the psychology of knowing. New York: Norton, 1973. Bryant, P. E. Perception and understanding in young children. New York: Basic Books, 1974. Carpenter, T. P., & Moser, J . M. The development of addition and subtraction problem solving skills. In T. P. Carpenter, J . M. Moser, & T. A. Romberg (Eds.), Addition and subtraction: A cognitive perspective. Hillsdale, N.J.: Erlbaum, 1981, in press. Chi, M. T., & Klahr, D. Span and rate of apprehension in children and adults. Journal of Experimenid Child Psychology, 1975, 19,434-439.

Cooper, R., Starkey, P., Blevins, B., Goth, P., & Leitner, E. Number development: Addition and subtraction. Paper presented at the meeting of the Jean Piaget Society, Philadelphia, May 1978. Estes, W. K. The problem of inference form curves based on group data. Psychologicul Bulletin. 1956, 53, 134-139. Fairbank, B. A. Experiments on the temporal uspects of number perception. Unpublished doctoral dissertation, University of Arizona, 1969. Flavell, J . H. Stage-related properties of cognitive development. Cognitive Psychology. 1971, 2 , 42 1-453.

Fuson, K. C., & Richards, J. Children's construction of the counting numbers: From a spew to a bidirectional chain. Paper presented at the meetings of the American Education Research Association, 1980, Boston. Fuson, K. C., & Richards, J. The acquisition anb elaboration of the number sequence. In C. Brainerd (Ed.), Progress in cogniriw development.Vol. I . Berlin and New York: Springer-Verlag. 1982, in press. Gelman, R. The nature and development of early number concepts. Advunres in Child Development und Behuvior 1972, 7 , 115-167.

Gelman, R. Cognitive development. Annual Review oj't'sychology. 1978, 29, 297-332. Gelman, R . , & Gallistel, C. R. The child's understanding qfnumber. Cambridge, Mass.: Harvard Univ. Press, 1978. Gelman, R., & Starkey, P. Development of addition and subtraction abilities prior to formal schooling in arithmetic. In T. P. Carpenter; J. M. Moser, & T. A. Romberg (Eds.), Addifion und subtraction: A cognitive perspective. Hillsdale, N.J.:Erlbaum, 1981, in press. Ginsburg, H. Children's urithmetic: The learning process. New York: Van Nostrand, 1977.

310

Robert S . Siegler orid Mitchell Robinson

Glushko, R. J . , & Cooper, L. A. Spatial comprehension and comparison processes in verification tasks. Cognitive Psychology, 1978, 10, 391 -421. Greeno, J. G., Riley, M. S., & Gelman, R. Young children’s counting and understanding of principles. Unpublished manuscript. Groen, G . J . , & Parkman, J . M. A chronometric analysis of simple addition. Psychological Review, 1972, 79, 329-343. Groen, G . J . , & Resnick, L. B. Can preschool children invent addition algorithms? Journal of Educational Psychology. 1977, 69,645-652. Hebbeler, K. The development ofchildren’s problem-solving skills in addition. Unpublished doctoral dissertation, Cornell University, 1976. Ilg. F., & Ames, L. B. Developmental trends in arithmetic. Journd of Genetic Psychology. 195 1, 79, 3-28. Johnson, S . C. Hierarchical clustering schemes. Psychometrika. 1967, 32, 241 -254. Kintsch, W . , & van Dijk, T . A. Toward a model of text comprehension and production. Psychological Review, 1978, 85, 363-394. Klahr, D., & Wallace, J. G. Cognitive development: An infurmaticin-processing view. Hillsdale, N.J.: Erlbaum, 1976. Knight, F. B., & Behrens, M. S. The learning of the 100 oddiiion combinutions and the 100 subtraciion combinations. New York Longmans, Green, 1928. Kosslyn, S. M. Imagery and cognitive development: A teleological approach. In R. S. Siegler (Ed.), Children’s thinking: Whaf develops? Hillsdale, N.J.: Erlbaum, 1978. Kosslyn, S . M., Murphy, G. L . , Bemesderfer, M. E., & Feinstein, J. J. Category and continuum in mental comparisons. Journal of Experimental Psychology: General. 1977, 106, 341 -375. Kruskal, J . B., & Wish, M. Multidimensional scaling. Beverly Hills, Calif.: Sage Univ. Press, 1978. Lehman, H. Intrrduciion to the philosophy of mathematics. Totowa, N.J.: Rowman & Littlefield, 1979. LNR. Explurutions in cognition. San Francisco: Freeman, 1975. Maki, R. H. Categorization and distance effects with spatial linear orders. Journul of Experimental Psychology: Human Learning and Memory. 1981, I , 15-32. Manis, F. R., Keating, D. P., & Morrison, F. J. Developmental differences in the allocation of processing capacity. Journal of Experimental Child Psychology, 1980, 29, 156-169. Mehler, J . , & Bever, T. G. Cognitive capacity of very young children. Science, 1967, 158, 141142. Miller, S . A. Nonverbal assessment of conservation of number. Child Development, 1976, 47, 722-728. Moyer. R. S., & Bayer, R . H. Mental comparison and the symbolic distance effect. Cognitive PSychology, 1976, 8, 228-246. Moyer, R. S . , & Dumais, S. T . Mental comparison. The psychology of learning and motivution, 1978, 12, 117-155. Moyer, R. S . , & Landauer, T. K. The time required for judgments of numerical inequality. Nature (London). 1967, 215, 1519-1520. Neisser. V . General, academic, and artificial intelligence. In L. B. Resnick (Ed.), The nuture qf intelligence. Hillsdale, N.J.: Erlbaum, 1976. Nesher, P. Levels of description in the analysis of addition and subtraction. In T. P. Carpenter, J. M. Moser, & T. A. Romberg (Eds.), Addition and subtraction: A cognitive perspective. Hillsdale, N.J.: Erlbaum, 1981, in press. Newell. A. A note on process-structure distinctions in developmental psychology. In S. FarnhamDiggory (Ed.), Information processing in children. New York Academic Press, 1972.

Numericul Understandings

31 I

Newell, A . You can’t play 20 questions with nature and win: Projective comments on the papers of this symposium. In W . G. Chase (Ed.), Visuul information prowusing. New York: Academic Press, 1973. Parkman, J. M. Temporal aspects of digit and letter inequality judgments. Journal of Experimentcil P ~ j ~ h ( ~ l o g1971, y . 91, 191-205. Peano, G . Formulaire de mathematic1ues (Vols. 1-5). Turin: Broca, 1894-1908. Piaget, J . The child’s concept ofnumber. New York: Norton, 1952. Piaget, J. lntellectual evolution from adolescence to adulthood. Humcin Development. 1972, 15, 1-12. Pliske, R. M., & Smith, K. H. Semantic categorization in a linear order problem. Memory & Cognition, 1979, 7, 297-302. Pollio, H. R., & Reinhardt, D. Rules and counting behavior. Cognitive Psychology. 1970, 1, 388-402. Pollio, H. R., & Whitacre, J . Some observations on the use of natural numbers by preschool children. Perceptual and Motor Skills, 1970, 30, 167-174. Rothenberg, B. B., & Courtney, R. G . Conservation of number in very young children. Developmental Psychology, I 969. 1, 493-502. Rule, S. J. Equal discriminability scale of number. Journal ofExperimental Ps.ychology. 1969, 79, 35-39. Schaeffer, B., Eggleston, V. H . , & Scott, J. L. Number development in young children. Cognitive Ps.ycho1og.V. 1974, 6 , 357-379. Sekuler, R., Armstrong, R., & Rubin, E. Processing numerical information: A choice time analysis. Journul of Experimental Psychology, 1971, 90, 75-80. Sekuler, R., & Mierkiewicz, D. Children’s judgments of numerical inequality. Child Development, 1977, 48, 630-633. Shepard, R. N., Kilpatric, D. W., & Cunningham, J. P. The internal representation of numbers. Cognitive Psychology. 1975, 6 , 82-138. Shepard. R. N.. & Podgorny, P. Cognitive processes that resemble perceptual processes. In W . K . Estes (Ed.), Handbook of learning and cognitive processes. Hillsdale, N.J.: Erlbaum, 1978. Siegler, R . S. Defining the locus of developmental differences in children’s causal reasoning. Journal of Experimental Child Psychology, 1975, 20, 512-525. Siegler, R. S. Three aspects of cognitive development. Cognitive Psvcho1og.y. 1976, 8, 481-520. Siegler. R. S. The origins of scientific reasoning. In R. S. Siegler (Ed.), Children’s thinking: What develops? Hillsdale, N.J.: Erlbauni, 1978. Siegler, R. S. Developmental sequences within and between concepts. Monographs of the Sociery,for Research in Child Development. 1981, 46 (Whole No. 189). (a) Siegler, R. S. Information processing approaches to development. In W. Kessen (Ed.), Manual C~ Child Psychology: History, theories und methods. New York: Wiley, 1981, in press. (b) Siegler. R. S., & Klahr, D. When do children learn: The relationship between existing knowledge and the ability to acquire new knowledge. In R. Glaser (Ed.), Advcinces in instructionul psychology. Hillsdale, N.J.:Erlbaum, 1981, in press. Siegler, R. S., & Robinson, M. Preschoolers’ knowledge of very large numericul magnitudes. Manuscript in preparation. I98 I . Simon, H. A. On the development of the processor. I n S . Farnham-Diggory (Ed.), Infmnution processing in children. New York: Academic Press, 1972. Smedslund, J. Microanalysis of concrete reasoning. 1. The difficulty of some combinations of addition and subtraction of one unit. Scnndinmian Journal ofPsychology, 1966, 7 , 145-156. Sternberg. R. J. Component processes in analogical reasoning. Psyc‘holo~icdRevirw, 1977, 84, 353-378.

312

Robert S . Siegler and Mitchell Robinson

Strauss, S., & Levin, 1. A commentary on Siegler, R. S. Developmental sequences within and between concepts. Monographs of the Society for Research in Child Development. 1981, 46, (Whole No. 189). Svenson, 0. Analysis of time required by children for simple additions. Act0 Psychologica, 1975, 39, 289-302. Svenson, 0.. & Broquist, S. Strategies for solving simple addition problems. Scandinavian Journal of Psychology, 1975, 16, 143-151. Trabasso, T.The role of memory as a system in making transitive inferences. In R. V. Kail, Jr. & J. W . Hagen (Eds.), Perspectives on the developmeni of memory and cognition. Hillsdale, N.J.: Erlbaum, 1977. Trabasso, T. How do children solve class inclusion problems? In R. S. Siegler (Ed.), Children's thinking: What develops? Hillsdale, N.J.: Erlbaum, 1978. Trabasso, T., Riley, C. A,, & Wilson, E. G . The representation of linear order and spatial strategies in reasoning: A developmental study. In R. J. Falmagne (Ed.), Reasoning: Representation and process. Hillsdale, N.J.: Erlbaum, 1975. Yoshimura, T. Strategies for addition among young children. Paper presented at the 16th annual convention of the Japanese Association of Educational Psychology, 1974.

AUTHOR INDEX Numbers in italics refer to the pages on which the complete references are listed.

A

Bates,E., 8, 1 1 , 12, 13, 14, 15, 16, 17, 18, 19, 2 0 , 2 1 , 4 8 , 5 1 , 5 2 , 6 2 , 6 5 , 6 7 , 6 8 , 69, 70. 7 / , 75 Battaglia, J., 289, 290, 308 Batter, B. S., 226, 237 Bauer, J. A , , 82, 115 Bayer, R. H., 268, 310 Bayley, N . , 215,216,217,218,222,225,234, 237 Beale, 1. L., 82, 83, 84, 106, 117 Beckwith, L., 24, 27, 28, 29, 30, 31, 33, 69, 71. 225, 237 Bedrosian, J . , 13, 14, 73 Beebe, B., 22, 75 Beeghly-Smith, M., 12, 16. 67, 68, 70 Behrens, M. S., 287, 292, 295, 310 Beilin, H . , 242, 309 Bell, S . M., 19, 20, 69. 112, 115 Belsky, J., 69 Bemesderfer, M. E., 286, 310 Benigni, L., 13, 16, 17, 18, 19,20,21,65, 69, 70 Bennett, S . L.. 22, 7.5 Benton, A. L., 114, 115 Berkley, M. A,, 79, 115 Berko, J . , 9, 69 Berko-Gleason, J., 35, 58, 69 Berti, F. B., 113, 114, 119 Best, C. T., 106. 119 Bever, T. G., 10, 72, 243, 309, 310 Beverley, K. I., 96, 121, 122 Birch, H. G., 106, 123 Birns, B., 223, 237 Bishop, A., 84, I19 Bishop, P. O., 95, I 2 f Blake, R.. 79, 116 Blakemore, C., 95, 122 Blehar, M. C . , 17, 69, 234, 237 Blevins, B., 288, 309 Block, E., 15, 69 Block, J . , 173, 208

Aaron, P. G., 84, 113, 115 Abelson, W. D., 225, 239 Abramovitch, R., 180, 209 Adams, J. E., 95, I19 Aiken, L. R., 268, 308 Ainswonh, M. D. S., 17, 19, 20, 68, 69, 214, 233, 234, 237 Allen, T. W., 110, 119 Allman, J . M., 106, 115 Ames, L. B., 289, 3 / 0 Ammon, M., 67, 69 Ammon, P. R.,69 Anderson, E., 67, 69 Anderson, J. R., 128, 129, 164. 244,249,300, 308 Anderson, N., 80, 118 Andriessen, J. J., 80, 95, 116 Anglin, J. M., 139, 151, 164 Annis, R. C., 96, 11.5 Appelbaum, M. I., 89, 121 Appelle, S., 79, 81, 82, 85, 115 Arend, R. A , , 19, 20, 73 Armstrong, R., 268, 311 Arnheim, R . , 79, 115 Ashcraft, M. H., 289, 290, 308 Atkinson, J . , 95, 100, I I I , 115 Attneave, F., 79, 80, 81, 115 Auk, R. L., 133, 164 Austin, G., 131, 164 Aviezer, L., 13, 72

B Baird, J. C., 80, 122 Bakeman, R.,27, 28, 32, 33, 69 Banks, M. S.,I l l , 115 Banks, W. P., 268, 269, 270, 275, 284, 3OY Bartlett, F. C . , 102, I15 Bassili, J., 172, 208

313

3 I4

Author Index

Bloom, L., 12, 65, 67, 70 Boden, M . , 158, 164 Bohannon, J., 54, 70 Bonvillian, M.. 43, 58, 73 Borke, H., 179, 208 Bornstein, M. H., 80, 82, 84, 86, 87, 88, 89, 90,91,92,93,94,95,96,97,98,99, 100, 101, 102, 103, 104, 105, 106, 110, 113, 114, 116. 118. 119 Boswell, S. L., 80, 92, 116 Botkin, P., 67, 72 Bouma, H., 80, 95, 116 Bourne, L. E., 131, 164 Bowennan, M. F., 1 I , 12, 15, 70 Bowlby, J . , 17, 70, 214, 233, 237 Boysen, S., 65, 74 Braddick, F., 11 I , 115 Braddick, O., 111, 115 Bradley, D., 82, 83, 84, 100, 116 Bradshaw, J . , 82, 83, 84, 100, 116 Braine, L. G . , 81, 82, 84, 87, 116, 118 Braine, M. D. S., 15, 70. 243, 309 Brainerd, C. J . , 242, 309 Branigan, G., 14, 70 Bransford, J . D., 138, 141, 142, 143, 165 Brazelton, T. B., 22, 70, 186, 208 Bresson, F., 106, 116 Bretherton, I . , 12, 13, 16, 17, 18, 19, 20, 21, 51, 65, 67, 68, 69, 70 Brill, S., 8 5 , 95, 109, I I Y , 121 Broadbent, D. E., 111, 116 Bronson, G . , 102, 116 Brooks, B., 106, 116 Brooks, J . , 174, 178, 179, 210 Brooks, L., 146, 147, 148, 149, 151, 153, 157, I64 Broquist, S . , 288, 312 Brown, A. L., 243, 250, 309 Brown, D. R., 79, 117 Brown, J., 27, 28, 32, 33, 69 Brown, R., 9, 15, 70 Bruner, J . S . , 9, 1 I . 12,22,51,65, 70, 74, 75, 131, 159, 164, 305, 309 Bryant, P. E., 82, 84, 116. 243, 309 Buck, R., 172, 173, 174, 181, 182, 183, 184, 206, 207, 208, 208, 209 Burchinal, P., 27, 29, 32, 33, 70 Butler, J . , 83, 84, 106, 116 Butterfield, E. C., 127, 165

C

Cairns, N. U . , 113, 116 Caldwell, B., 33, 71 Caldwell, R. A., 104, 116 Camaioni, L., 12, 13, 16, 17, 18, 19, 20, 21, 65, 6 9 , 70 Camisa, J . M . , 79, 116 Campbell, F., 24,25,26,28,30,31,32,33, 74 Campbell, F. A , , 223, 238 Campbell, F. W . , 79, 95, 116. 119 Campos, J . J . , 172, 174, 177, 178, 205, ZOY, 210 Caplan, P. J., 106, 116 Carey, S.,139, 164, 180, 209 Carlson-Luden, V., 54, 71 Caron, A. J., 104, 116, 117 Caron, R. F., 104, 116, 117 Carpenter, P. A , , 114, I20 Carpenter, T. P., 307, 309 Case, R . , 13, 14, 71, 150, 164 Casler, L., 233, 237 Caul, F., 172, 173, 182, 209 Caul, W., 172, 182, 209 Causkaddon, G . , 43, 58, 73 Cazden, C., 43, 58, 71 Chang, J.-J., 111, 119 Chapman, R . , 13, 14, 73 Charlesworth, W. R., 170, 174, 209 Chase-Lansdale, P. L., 234, 238 Chevalier-Skolnikoff, S . , 65, 71 Chi, M. T. H., 150, 165. 286, 309 Chikvishvili, L., 172, 209 Chomsky, N., 8, 9, 10, 71, 140, 165 Clark, E. V., 114, 117, 152, 165 Clark, H. H., 114, 117 Clark, R., 65, 71, 106, 123 Clarke-Stewart, K. A., 20, 21, 36, 37, 48, 71, 224, 225, 227, 234, 237 Cohen, D.,185, 204, 205, 208, 209 Cohen, L. B., 96, 117 Cohen, S . , 24, 27, 28, 29, 30, 31, 33, 71 Cohen, S. E., 225, 237 Collins, A. M.,161, 165 Collis, G. M., 12, 74 Conezio, J . , 144, 166 Connell, D. B., 20, 21, 71 Cooper, L. A., 156, 165. 299, 310 Cooper, R.,288, 309

315

Author Index

Cooper, W., 66, 71 Corballis, M. C., 80, 81, 82, 83, 84, 96, 106, 1 1 I , 117 Cornell, E. H . , 98, 117 Comgan, R., 13, 71 Corwin, T. R . , 79, 117 Cosgrove, M. P., 79, 117 Courtney. R. G . , 242, 243, 311 Covitz, F., 24, 25, 26, 28, 29, 30, 31, 32, 51, 75 Craig, E. A , , 79, 117 Crano, W. D., 217, 222. 237 Cremona, C., 67, 71 Critchley, M., 106, 117 Cross, T., 38,39,40,41,42,43,45,47,55, 71 Culicover, P., 8, 75 Cullen, J. K . , Jr., 95, 96, 121 Cunningham, J. P., 268, 311 Curlee, T. E., 81, 115 Curtiss, S . , 15, 63, 71

D Dale, P.,9, 71 Daly. E., 180, 209 Darwin, C., 171, 209 Davidson, C. V.. 237 Davidson. H. P. A., 110, 113, 117 Davis, A. E., 106, 117 Day, R. H . , 86, 95, 121 Decarie, G. T., 174, 178, 210 De Frank, R. S.,211 De Fries, J. C., 60, 65, 72. 74 De Loache, J. S., 96, 117 Deregowski, J. B., 80, 117 Derryberry, D., 232, 238 de Schonen, S . , 106. 116 Devin, J., 35, 74 Diamond, R . , 180, 209 Dik, S., 66, 71 Dolan, A. B., 216, 217, 223, 233, 237, 2-38 Dougherty, L. M . . 172. 177, 210 Dumais, S. T., 268, 269, 310

E Edsterbrooks, M. A., 235, 237 Edwards, D . , 1 1 , 12, 71 Egglesron, V. H., 269, 270, 311

Eichorn, D. H., 14, 73, 228, 238 Ekman, P., 171, 172, 175, 176, 179, 185, 2 0 ~ . 210 Eliuk, J., 80, 119 Ellsworth, P., 172, 20Y Emde, R. N., 172, 174, 176, 177, 178, 205, 20Y. 210 Emsley, H. H . , 79, 117 Essock, E. A . , 79, 80, 82, 86, 88, 97, 117 Epstein, J . , 243, 30Y Eriksen, C. W . , 129, 165 Ervin-Tripp, S., 67, 68, 71 Estes, D. A . , 219, 235, 23Y Estes, W. K., 290. 309 Eysenck, H. J., 184, 185, 189, 206, 208, 209

F Fagan, J. F., 86, 87, 102, 104, 105, 117 Fairbank, B. A , , 268, 309 Fantz, R. L., 98, 117 Farah, M. J . , 165 Fman, D., 24, 25, 26, 27, 28, 29, 30, 31. 32, 33, 70, 71, 74 Farrell, W. S . , 83, 106, I18 Feinstein, J . J., 286, 310 Feldman, C., 43, 71 Fellows, B. J., 85, 118 Fenson, L., 14, 71 Ferdinandsen, K . , 84, 89, 90, 91, 92, 93, 94, 116. 118 Ferguson, C., 35, 54, 71, 72, 75 Feshbach, N., 180, 209 Field, T., 185, 193, 204, 205, 208, 20Y, 211 Fillmore, L . , 65, 72 Finke, R. A . . 165 Finlay, D. C., 94, I I Y Fischer, K. W.,13, 72 Fisher, C. B., 82, 84, 89, 90, 92, 113, I18 Fitts, P. M.,80, 92, 118 Flavell, J . H . , 9, 67, 72, 133, 154, 165, 243, 309 Fodor, J . A . , 10, 72, 127, 128, 133. 165 Fox, J . , 80, 118 Franks, J. J., 138, 141, 142, 143, 165 Freedle, R . , 25, 28, 72 Freedman, D. G., 231, 233, 237 Freeman. R . B., Jr., 106, 121 Freeman, R. D.. 95, 118

316

Author Index

Fregnac, Y., 96, 118 French, J., 95, 100, 115 Friedman, S . L., 104, 117 Friesen, W . V., 171, 172, 179, 209 Frith. U., 113, 118 Frodi, A. M . , 235, 238 Frodi, M.,235, 238 Frost, B. J., 95, 96, 115, 118 Fry, C . , 67. 72 Fujii, M., 269, 275, 284, 309 Fulgham, D. D., 79, 117 Furrow, D., 38, 39, 40, 41, 55, 72 Furth, H. G., 9, 72 Fuson, K. C., 309

G Gaensbauer, T., 174, 178, 209 Galanter, E. H., 9, 73 Gallistel, C. R., 250, 270, 298, 309 Garcia, J., 72 Gardner, J . , 106, 118 Garn, S. M., 217, 224, 228, 237 Garner, W. R . , 81, 118, 129, 157, 165 Garrett, M. F., 10, 72 Geis, M. F., 77, 121 Geldard, F. A., 78, 118 Gelman, R., 35, 74, 243, 250, 270, 287, 288, 298, 309, 310 Gibson, E. J . , 82, 90, 96, 107, 110, 112, 113, I I8 Gibson, J . J., 78, 79, 80, 82, 112, 118 Ginsburg, A. P., 1 1 1 , 118, 119 Ginsburg, H.,250, 251, 270, 309 Gitter, G . , 179, 209 Glanville, B. B., 106, 119 Glass, A. L., 150, 160, 161, 165 Gleitman, H.,35, 37, 39,40,41, 42, 43, 52, 55, 62, 63, 73 Gleitman, L., 35,37, 39,40,41,42,43,52,55, 62, 63, 73 Glushko, R. J . , 299, 310 Golden, M., 223, 237 Goldmeier, E., 80, 81, 105, 119 Goldsmith, H. H., 231, 232, 237 Goldstein, A. J . , 101, 119 Golinkoff. R.. 72 Goode, M. K., 69 Goodman, N., 131, 165 Goodnow, J., 131, 164

Gordon, E. W., 106, 123 Goren, C. C., 186, 209 Goth, P., 288, 309 Gottesman, I. I . , 231, 232, 237 Goyen, J., 114, 119 Grandy, C. A., 114, 119 Gray, J . A., 184, 210 Green, M. A , , 79, 117 Greenberg, R . , 185, 204, 205, 208, 209 Greenfield, P. M., 9, 11, 57, 68, 70. 72. 75, 159, 164 Greeno, J. G., 310 Groen, G. J., 250,288,289,290,292,295,310 Gross, C. G., 84,86,87,88,89,90,91,93,94, 97, 98.99, 100, 101, 102, 103, 104, 105, 106, 110, 113, 114. 116, 118, 119 Guillaume, P., 210 Gwiazda, J., 85, 95, 109, 119, 121

H Haber, R. N., 144, 166 Hack, E., 20, 72 Haith, M. M., 92, 96, 119 Hake, H. W.,129, 165 Hall, J. A., 211 Halpern, E., 13, 72 Hamilton, C. R.. 106, 119 Hamm, A , , 106, 123 Hardy-Brown, K., 60, 65, 72 Harkness, S.,38, 40, 41, 42, 43, 45, 72 Harlow, H. F., 63, 72 Harlow, M. K . , 63, 72 Harmon, L. D., 101, 111, 119 Harmon, R. J . , 19, 72 Harris, D. S., 113, 114, 119 Harris, L. J . , 110, 119 Harris, P. L., 84, 119 Haupe, G., 114, 119 Haviland, J., 174, 178, 179, 210 Hayes-Roth, B., 134, 142, 153, 165 Hayes-Roth, F., 128, 134, 142, 153, 165 Hebbeler, K., 289. 310 Held,R.,82, 85, 95, 109, 115, 119, / 2 / Heldmeyer, K. H., 150, 160, 165 Hemenway, K., 80, 81, 119 Henderson, C., 174, 178, 209 Herzka, H. S., 175, 204, 210 Heubner, R. R., 172, 177, 210 Hiatt, S. W., 172, 177, 205, 210

317

Author Index

Higgins, G. C., 79, / / Y Hildyard, A., 80, 115, 121 Hill, D. K., 268, 309 Hill, L., 181, 205, 211 Hirsch, H . V. B., 96, 119 Hogarty, P. S . , 14, 73, 222, 227, 228, 238 Hogben, J. H., 92, 119 Holyoak, K . J., 161, 165 Hood, L., 65, 67, 70 Horgan, D.,72 Horowitz, F. D., 90, 119, 176, 186, 211 Houlihan, K., 83, 84, 106, 122 Howard, I. P., 79, 81, 82, 112, 1IY Hubel, D. H., 82, 95, 119 Hubley, P., 12, 15, 75 Hunt, J. McV., 224, 23Y Hurlburt, N . , 222. 227, 228, 238 Huttenlocher, J., 84, 119 Hwang, C.-P., 235, 238

I Irnbert, M., 96, 118 Inhelder, B., 166 Ivinskis, A . , 94, 119 Iwahara, S . , 96, 123 1zard.C. E., 171, 172, 175, 176,177, 179, 181 182, 205, 210

J Jaffe, J., 22, 75 Jastrow, J.. 112, IIY Johnson, D. L., 238 Johnson, R. M.,80, 81, 119 Johnson, S. C., 310 Johnston, J . , 13, 72, 90, 118 Jonckheere, A. R., 105. 119 Jones, H. E., 173, 174, 178, 184, 206, 207, 208, 210 Julesz, B., 80, 92, 96, 111, I I Y Jung, R., 106, 116 Just, M.A . , 114, I20

K Kaas, J. H . , 106, 115 Kagan, J., 87, 123 Kaminer, J. J., 95, 118 Kaplan, B . , 16, 75

Karmel, B. Z., 96, 120 Kato, H., 95, 121 Kaye, K . , 22, 25, 26, 27, 28, 31, 34, 62, 72 Kayra-Stuart, F., 269, 275, 284, 309 Keating, D. P., 286, 310 Keating, M. T., 217, 224, 228, 237 Kee-le, S. W., 135, 136, 138, 155, 156, 166 Keene, G.C., 79, 119 Keller, B., 231, 233, 237 Kempler, D., 65, 72 Kendler, H. H., 150, 165 Kendler, T. S . , 150, 165 Kennedy, C. B., 89, 121 Kessel. F., 15, 6Y Kessen, W., 92, 96, 98, 116. 119, 122 Kilbride, J., 181, 205, 211 Kilpatric, D. W., 268, 311 King, W. L . . 223, 238 Kinskume, M., 106, 116 Kintsch, W., 300, 310 Kirk, B., 84, 122 Kitterle, F., 79, 115 Klahr, D., 286, 306, 30Y, 310, 311 Kleck, R., 173, 210 Klein, R. P., 19, 72 Kligman, D. H., 176. 209 Knight, F. B., 287, 292, 295, 3 / 0 Koelling, R . , 72 Koivumaki, J., 172, 211 Kopp, C., 33, 72 Kopp, C. B . , 225, 237 Koslowski, B., 22, 70 Kossan, N., 150, 151, 152, 165 Kosslyn, S. M., 128, 129, 150, 154, 155, 156, 159, 160, /65, 286, 299, 300, 310 Kreutzer, M. A., 170. 174, 209 Kruskal, J. B., 273, 310 Kuhn, T. S., 9, 73 Kulikowski, J. J., 79, 116 Kuno, S . , 66, 73

L La Barbera, J. D., 175, 205, 210 Lachman, R.. 127, 165 Lamb, M.E., 214, 217, 218, 219, 224, 225, 227, 228, 232, 234, 235, 237; 238, 23Y Landauer, T. K., 267, 268, 310 Lanzetta, J., 173, 210 Lashley, K. S . , 9, 73. 84, IIY

318

Aiirhor Index

Lasky, R. E., 140, 143, I65 Lam, M., 172, 209 Leaman, R., 80, 92, 122 Leavitt, L. A . , 176, 205, 2 / 0 Lechelt, E. G., 80, / I 9 Leehey, S. C., 85, 119 Lehrnan, H., 270, 310 Lehmim, R. A. W., 106, 1IY Leitner, E., 288, 30Y Lema, S . , 79, 116 Lernond, C., 181, 206, 210 Lennie, P., 79, 119 Leonard, J. A., 80, I18 Leonard, L., 13, 73 Lepez, R., 113, / / Y Lerner, C., 81, 82, 116 Lesk, A. B., 101, I I Y Le Tendre, J. B., 84, 119 Leventhal, A. G.,96, l / Y Levin, H., 107, 110, 112, 113, 118 Levin, I . , 244, 312 Levine, M., 131, 165 Levinson, J . , 79, / I 6 Levinson, R., 106, I19 Lewis, A , , 20Y Lewis, M., 22, 25, 28, 72. 73. 174, 178, 179, 210, 227, 238 Lewkowicz, D., 106, 118 Li, C., 66, 73 Liberman, 1. Y., 113, 114, 119, 122 Lichtenstein, M., 79, 117 Lightbown, P., 65, 67, 70 Ling, B. C., 86, / I 9 Lipets, M., 172, 2 / 1 Llg, F., 289, 3 / 0 Lock, A , , 12, 16, 22, 73 Lorenz, K.,9, 73 Lyle, J . G., 114, IIY

M Mach, E., 80, 81, 83, 96, 106, I I Y MacNamara, J., 11, 12, 73 MacWhinney, B., 8, 15, 68, 6 Y Maffei, L., 95, I19 Main, M., 19, 20, 21, 22, 70, 73, 227, 238 Maisel, E. B . , 96, I19 Maki, R. H . , 114, //Y. 286, 3 / 0 Maxwell, S . E . , 238 Malatesha, R. N., 84, 113, / / 5

Mandler, J . M., 84, 122 Manis, F. R . , 286, 3 / 0 Mansfield, R . J. W., 95, 102, / I 9 Marcy, T. G., 225,237 Marg, E., 95, 119 Marquis, A , , 54, 70 Martello. M., 102, 119 Matas, L., 19, 20, 73, 178, 211 Matheny, A. P., 216,217, 223, 231,232, 233, 237, 238 Maurer, D., 102, 119 Maury, L., 106, 116 May, J. G.,95, 96, 121 Mayer, M. J., 109, 12/ McCall, R. B., 14, 73, 89, 121, 222,227,228, 238 McCandless, 9.R., 77, 121 McCarter, R., 170, 171, 206, 2 / / McCarthy, D., 9, 73 McClearn, G. E., 60, 74 McFarland, J . H., 79, / 2 / McGinnes, G. C., 172, 177, 210 McGowan, R . J . , 238 McGurk, H., 103, 104, I21 McKenzie, B., 82, 86, 95, 121, 123 McNeill, D., 9, 73 McNew, S., 12,47, 51, 5 5 , 57,67,68, 70, 73 Medin, D. L., 144, 145, 146, 152, 153, 163, I66 Mehler, J . , 243, 30Y, 310 Meltzoff, A. N., 65, 73, 186, 2 / 0 Mervis, C. 9..136, 137, 138, 155, 166 Metzler, J . , 8 1 , 122 Mierkiewicz, D., 269, 273, 274, 285, 308, 3 / / Miller, A., 106, 117 Miller, E. R., 172, 173, 182, 209 Miller, G. A., 9, 73 Miller, J., 13, 14, 73 Miller, R. S., 166, 172, 182, 209 Miller, S . A., 242, 310 Millward, R. B., 132, 166 Minichiello, M. D., 104, 1 / 7 Mistler-Lachman, J., 127, 165 Mitchell-Kernan, C., 67, 68, 71 Mow, K . , 1 1 1 , 115 Moerk, E. L., 35, 73 Moffett, A , , 86, 95, I21 Mohindra, I . , 85, 95, 109, //Y, 121 Molfese, D. L., 106, 12/ Money, J., 114, 12/

Author Index

Moore, M. K., 65, 73, 186, 210 Morgan, M. J . , 106, 117 Morrison, F. J., 286, 310 Morse, P. A . , 176, 205, 210 Moser, J. M.,307, 30Y Moskowitz-Cook, A , , 79, 85,95, 96, 117, I I Y . 121

Most, R., 6Y Mostofsky, D., 179, 20Y Mowrer, 0. H., 65, 73, 82, I18 Moyer, R. S.,267, 268, 269, 310 Muir, D. W., 79, 96, 123 Murphy, G . L., 286, 310

N Nagamura, N . , 96, 123 Neisser, V.,244, 310 Nelson, C. A . , 176, 205, 210 Nelson, J . I . , 95, 121 Nelson, K., 12, 43, 44, 58, 59, 73 Nesher, P., 307, 310 Neumann, P. G., 138, 143, 166 Newell, A . , 127, 166, 287, 290, 300, 310, 311

Newmeyer, F., 1 I , 73 Newport, E., 35,37,39,40,41,42,43,52,62, 63, 73 Newson, J . , 23, 74 Nicolich, L . , 14, 74 Ninio, A , , 22, 70 Noble, J . , 106, f21 Norman, D. A . , 158, 166

319

P Palermo, D. S . , 106, 121 Palmer, S. E., 80, 81, 121 Parisi, S . A., 175, 205, 210 Parkman, J . M., 250, 268, 288, 289, 290, 292, 295, 310, 311 Parmelee, A. H., 33, 72, 225, 237 Parsons, G . , 12, 74 Pascual-Leone, J., 150, 166 Patterson, K., 82, 83, 84, 100, I16 Pawlby, S., 205, 210 Peano, G., 311 Pederson, J . A., 225, 227, 234, 237 Pentz, T., 18, 20, 21, 34, 36, 39, 40, 41, 42, 45, 48, 62, 74 Piaget, J . , 12, 13, 74, 112, 121, 166, 224,228, 238, 242, 305, 311 Pick, H. L., 79, 112, l l Y Pieraut-Le Bonniec, G., 106, 116 Pierce, S., 84, 122 Pinker, S . , 8, 15, 74 Pliner, P., 180, 20Y Pliske, R. M., 286, 311 Plomin, R., 50, 60, 65, 72. 74. 232, 233, 238 Podgorny, P., 268, 311 Pollio, H. R., 250, 255, 311 Pomerantz, J . R., 83, 106, 121 Posnansky, C. J . , 143, 166 Posner, M. I . , 135, 136, 138, 155, 156, 166 Premack, D., 68, 74 Pribram, K. H., 9, 73 Provost, A . M., 174, 178, 210 Pyiyshyn, 2. W., 128, 166

0 Ochs, E., 65, 67, 74 O’Connell, B., 14, 75 Odom, R . , 181, 206, 210 Olson, D. R., 80, 82, 115. I I Y , 133, 166 Olson, R. K . , 79, 80, 115 Olver, R., 9, 70, 159, 164 Orlando, C . , 113, 114, 119 Onon, S. T., 106, 113, 114, l l y Oster. H., 175, 185, 210 Over, J., 84, 100, 121 Over, R., 84, 100, 121 Owen, M. T., 234, 23X Owens, D. A., 82, 115 Owsley, C. J . , 90, I18

Q Quillian, M. R., 161, 165 Quincy, A , , 179, 20Y

R Ramey, C., 24, 25. 26, 27, 28, 30. 31, 32, 33, 71, 74, 223, 238 Ramsey, D. S., 14, 71 Rappapon, M., 80, 118 Ratner, N., 22, 74 Rced,S. K . , 135, 136, 140, 141, 144, 155, 156, 157, 166 Reese, H. W . , 155, 166

320

Author Index

Regan, D., 95, 96, 121, 122 Reich, J . H., 176, 209 Reilly, J., 57, 75 Reinhardt, D., 255, 311 Relyea, L., 81, 82, 116 Resnick, L. B., 288, 310 Rheingold, H. L., 233, 238 Richards, J . , 309 Riesen, A. H., 82, 84, 100, 123 Rieser, J.. 79, 112, 121 Riley, C. A., 286, 312 Riley, M. S., 310 Rips, L. J., 160, 161, 166 Risser, D., 172, 177, 210 Robinson, M.,269, 311 Rock, I., 79, 80, 81, 82, 83, 92, 105, 122 Roe, K.,180, 209 Rogosa, D., 56, 74 Roldan, C. E., 80, 81, 96, 1 1 1 , 117 Ronner, S. F., 95, 102, 119 Rosch,E., 112, 122, 136, 137, 138, 140, 152, 155, 157, 166 Rose, D., 95, 122 Rosenblith, J . F., 84, 122 Rosenblum, L. A , , 22, 73, 227, 238 Rosenfeld, H. M.,176, 186, 211 Rosenthal, R., 172, 211 Ross, G.,5 1 , 66, 71, 75, 92, 119 Rothbard, M. K . , 232, 238 Rothenberg, B. B., 242, 243, 311 Rowe, D. C., 50, 74, 232, 233, 238 Rubin, E., 268, 311 Rudel, R. G., 83, 114, 122 Rule, S. J., 268, 311 Rumbaugh, D. M., 65, 74 Rumelhart, D. E., 158, 166 Rutkin, B., 95, 119 Ryan, J., 12, 74 S

Sachs, J., 35, 74 Sager, L. C., 83, 106, 121 Salapatek, P.,92, 96, 111, 115, 119, 122 Sandeman, D. R., 106, 123 Sander, L. W . , 22, 74 Sarty, M., 186, 20Y Savage-Rumbaugh, E. S., 65, 74 Savin, J . V., 172, 173, 182, 209

Scaife, M., 65, 74 Scan, S . , 232, 238 Schachter, S., 173, 210 Schaefer, E. S . , 217, 222, 225, 234, 237 Schaeffer, B., 269, 270, 311 Schaffer, H. R., 12, 22, 74 Schaffer, M., 144, 145, 146, 152, 153, 163, I66 Schieffelin, B., 67, 74 Schlesinger, I. M., 15, 74 Schmidt, M. J., 79, 117 Schneider, W., 158, 166 Schwartz, R., 9, 74 Scott, J . L., 269, 270, 311 Seegmiller, B. R., 223, 238 Seitz, V., 225, 239 Sekuler, R. W., 83, 84, 106, 114, 122, 268, 269, 273, 274, 285, 308, 311 Selman, R., 67, 74 Serpell, R., 84, 114, 122 Shankweiler, D. P., 113, 114, 119, 122 Shantz, C. U., 214, 238 Shapiro, M., 114, 122 Shatz, M., 35, 49, 52, 74 Shepard, R. N., 81, 122, 156, 165, 268, 311 Shepherd, P. A., 86, 87, 102, 117 Sherman, M., 174, 210 Sherrod, L. R., 111, 122, 214, 238 Sherwood, V., 12, 70 Shettleworth, S . J., 9, 75 Shiffrin, R. M., 158, 166 Shoben, E. J., 160, 161, 166 Shore, C. S . , 14, 75 Sidman, M., 84, 122 Siegfried, J. B., 95, 96, 121 Siegler, R. S., 243, 244, 245, 269, 270, 287, 301, 311 Sigman, M., 33, 72, 225, 237 Simon, C. W., 80, 92, 118 Simon, H. A., 127, 166. 305, 311 Simpson, C., 166 Siqueland, E. R., 86, 88, 97, 117 Slater, A., 95, 122 Slobin, D. I., 9, 10, 12, 15, 75 Smedslund, J . , 287, 311 Smith, C., 139, 164 ' Smith, E. E., 160, 161, 166 Smith, J. H., 1 1 , 68, 72. 150, 166 Smith, K. H . , 286, 311

321

Author Index

Snow, C . , 35, 75 Snow, M. E., 225, 227, 234, 237 Snyder, L., 13, 51, 69, 70, 75 Spencer, D. D., 106, 119 Spitz, R. A,, 63, 75 Sroufe, L. A , , 19, 20, 73, 177. 178, 210, 2 1 1 , 226, 239 Staller, J., 114, 122 Standing, L., 144, 166 Starkey, P., 287, 288, 309 Stein, N. L., 84, 122 Steiner, J. E., 175, 185, 204, 210, 211 Stem, D. N., 22, 75 Sternberg, R. J . , 286, 311 Stevenson, M. B., 214, 217, 218, 219, 224, 225, 226, 227, 234, 239 Steward, M. S . , 113, 116 Stoever, R. I., 83, 106, I21 Storandt, M . , 106, 122 Strauss, M. S., 96, 117 Strauss, S., 244, 312 Stultz, K.,79, 119 Sutherland, N. S.,82, 84, 122 Suwalsky, J. D., 19, 72 Svenson, O., 288, 289, 290, 292, 312 Sykes, M.,95, 122 Szilagyi, P. G., 80, 122

T Tanne, G., 80, 119 Taylor, M. M.,79, 122 Tee, K. S . , 82, 84, 100, 123 Templeton, W. B., 79, 81, 82, 112, 119 Templin, M . , 133, 166 Teuber, H. L., 83, 114, 122 Thibos, L. N., 95, 118 Thomas, J., 82, I15 Thompson, J . , 172, 211 Thompson, R. A,, 217, 219, 232, 232, 239 Thompson, S . , 33, 66, 69, 73 Tieman, S. B., 106, I I Y Tighe, L. S., 150, 166 Tighe. T. J., 150, 166 Timney, B. N . , 79, 96, 123 Tinbergen, N., 9, 75 Tolman, E. C., 9, 75 Tomkins, S. S., 170, 171, 206, 211 Trabasso, T., 243, 275, 286, 312

Trevarthen, C., 12, 15, 75 Tulkin, S., 24, 25, 26, 28, 29, 30, 31, 32, 51, 75 Turiel, E., 15, 75 Turkewitz, G., 106, 118. 123

U Uhlarik, J., 80, 81, 119 Umeh, B. J., 225, 227, 234, 237 Urwin, C., 63, 75 Uzgins, I. C . , 224, 239

V Valian, V., 68, 75 Valsiner, J., 172, 209 van Dijk, T. A , , 68, 75, 300, 310 Vellutino, F. R . , 114, 123 Vietze, P., 175, 205, 210 Volterra, V., 12, 13, 16, 17, 18, 19,20,21,65, 69, 70 Vygotsky, L. S., 16, 75, 131, 166

W Wada, J. A., 106, 117. 123 Wade, T. D., 176, 209 Walden, T . , 193, 209. 211 Wall, S., 17, 69, 234, 237 Wallace, J . G., 306, 310 Waters, E., 17, 69. 178, 211, 225, 234, 237, 239 Watkins, D. W., 79, 115 Watson, J . S . , 95, 103, 104, 110, 123 Weiner, K . , 87, 123 Weiner, M., 84, 100, 123 Weinstein, M., 80, 118 Weintraub, S . , 35, 58, 69 Weiskopf, S . , 98, 116 Weiss, S . J . , 104, 116, 117 Werner, H . , 16, 75, 167 Wertheimer, M., 112, 123 Wexler, K., 8, 75 Whitacre, J., 250, 311 White, S. H., 150, 167 Wiesel, T. N . , 82, 95, 119 Williams, C., 35, 75 Williams, E. N., 268, 308

322

Author Index

Williamson, A. M., 82, 123 Wilson, E. G., 286, 312 Wilson, R. S.,216, 217, 223, 233, 237. 238 Winston, P. H . , 155, 158, 167 Wish, M., 273, 310 Wittgenstein, L., 167 Wohlwill, J . F., 84, 100, 123 Wolf, J. Z . , 86, 87, 88, 89, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 110, 116 Wolff, P . , 83, 84, 106, 123, 174, 189, 211 Wood, D.,5 1 , 75 Woodruff, G . , 68, 74 Woods, B . , 180, 209 Woodson, R . , 185, 204, 205, 208, 2UY Wu, P. Y. K . , 186, 209 Wunsch, J . P . , 177, 210

Y Yamada, J . , 15, 71 Yarczower, M . , 181, 205, 211 Yin, R . K . , 105, 123 Yonas, A., 79, 112, 121 Yoshida, S . , 96, 123 Yoshimura, T . , 289, 312 Young-Browne, G . , 176, 186, 211 2

Zalik, M. C., 82, 117 Zeki, S . M . , 106, 123 Zigler, E . , 225, 239 Zuckerman, M . , 172, 211 Zukow, P . , 57, 75

SUBJECT INDEX

A

organization of concept representations and, 160- 16 1 Counting from a point beyond one, 256-259 models of three levels of expertise in, 259267 from one, 250-256 Cue set representations, 140-142 evidence that children use. 142-143

Addition existing research on children, 287-290 model of strategy choice in, 296-299 preschoolers’ strategies for, 290-296 Attachment. language development and, 17-18

C D Causes, internal vs external, language development and, 48-54 Children, facial expressions and physiological responsivity and, 183-184 production or encoding of, I8 1-1 83, 191 -204 recognition or decoding of, 179-181, 191204 Cognitive development, perceptual anisotropies and, 1 1 1-1 I 5 Cognitive inputs, to language, 12-15 Cognitive performance, sociability and, see Sociability Computer analogy, concept development and, 127-129 Concept development, 125-127 contents of concept representations and, 130 exemplar-based, 143- I52 multiple interpretations of data and, 152153 non-rule-based, 134-143 rule-based, 130- 134 format of concept representations and, 153I55 images and descriptions, 155-158 representational-development hypothesis and, 158-160 information-processing theories of, 127 computer analogy and, 127-129 mental representation and, 129- 130

Decoding, of facial expressions in adults, 172-174 in children, 179-181, 191-204 in infants, 175-176 Descriptions, as concept representations, 155I56 Discrimination, of facial expressions, in infants, 175-176

E Encoding, of facial expressions in adults, 172-174 in children, 181-183, 191-204 in infants, 189-191 Environment, sociability and, 233-235 Equivalence, perceptual, 82-84 in infancy, 97-108 Exemplar-based concept representations, 143150

evidence that children use, 150-152 Expressions, see Facial expressions

F Face, perceptual anisotropy and, 100-103, 104-105, I I I

323

324

Subject Index

Facial expressions, 170-171 adult, I71 cataloguing, 171-172 relationships between encoding-decoding and physiological measures, 172-1 74 children and production or encoding by, 181 - I83 recognition or decoding by, 179-181 relationships between encoding and physiological responsivity in, 183-184 infants and discrimination or decoding by, 175-176, 185- 189 neonatal, 174-175 production by, 176-177, 185-189 relationships between expressions and physiological responsivity of, 178- 179

G Genetic confounds, language development and, 59-61

social bases of, 8 attachment and, 17-18 historical aspects, 8-12 internal vs external causes and, 49-50 preverbal interaction and, 18-34 structure vs motivation and, 50-54 verbal interaction and, 35-48 threshold effects and, 6 1-64 cognitive inputs and, 12-15 social inputs and, 15-34

M Magnitudes, see Numerical understandings Measurement, of sociability, 2 15-222 Mental representation, concept development and, 129-130 Mirror images, perceptual equivalence and, 97-108 Motivation, structure vs. language development and, 50-54

N I Images, as concept representations, 155-156 Individual differences, in sociability biogenetic influences on, 23 1 -233 environmental influences on, 233-235 Infants, facial expressions and discrimination or decoding of, 175-176, 189-191 encoding of, 189-191 in neonates, 174-175 physiological responsivity and, 178- 179 production of, 176-177 Information processing, concept development and, 127 computer analogy and, 127-129 mental representation and, 129-130 Interaction preverbal, language development and, 18-34 verbal, language development and, 35-48

L

Non-rule-based concept representations, 134143 Numerical understandings, 242-244 addition and existing research on children, 287-290 model of strategy choice in, 296-299 preschoolers’ strategies for, 290-296 counting and from a point beyond one, 256-259 models of three levels of expertise in, 259-267 from one, 250-256 development of, 299-308 number conservation and, 244-250 numerical magnitudes and effects of teaching a labeling strategy on, 278-280 preschoolers’ comparisons, 271 -275 research on adults and children, 267-271 models of magnitude comparison, 280-287 verbal labeling of numbers and, 275-278

P Labeling, numerical magnitudes and, 275-280 Language development direction of effects and, 54-59 genetic confounds and, 59-61

Perceptual anisotropies, 77-79 equivalence, 82-84 in infancy, 97-108

Subject Index

implications for cognitive development, 11 1-115 interrelations between, 109-1 11 perceptual-cognitive-social development and, 111 salience, 79-82 in infancy, 85-97 Physiological responsivity , facial expressions and in adults, 172-174 in children, 183-184 in infants, 178-179 Preference, perceptual salience and, 92-95 Preverbal interaction, language development and, 18-34 Processing, perceptual salience and, 90-92 Prototypes, as concept representations, 135138 evidence that children use, 138-140

325

Sociability, 213-215 cognitive performance and in older children, 225-226 relationship between, 226-230 outside test situation, 224-225 test sociability and, 222-224 individual differences in, 230-231 biogenetic influences on, 23 1-233 environmental influences on, 233-235 measuring, 215-222 Social-causal theories, of language develop ment, 48-49 internal vs external causes and, 49-50 structure vs motivation and, 50-54 Spatial vision, see Perceptual anisotropies Strategies addition, 290-2% model of strategy choice in, 296-299 labeling, 278-280 Structure, motivation vs, language development and, 50-54

R

T Representational-development hypothesis, concept representations and, 158-160 Rule-based concept representations, 130- 134

S Salience, perceptual, 79-82 in infancy, 85-97

Test sociability, cognitive performance and, 222-224

V Verbal interaction, language development and, 35-48 Vision, spatial, see Perceptual anisotropies

This Page Intentionally Left Blank

Contents of Previous Volumes Volumc 1 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. Lipsirt Some Methodological Contributions from a Functional Analysis of Child Development Sidney W . Bijou and Donald M . Baer The Hypothesisof 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 Betty 1. House and David Zeaman AUTHOR INDEX-SUBJECT

INDEX

Selected Anatomic Variables A d y d for lnterage Relationships of the Size-Size, Size-Gain, and Gain-Gain Varieties Howard V . Meredirh AUTHOR INDEXSUBJECT INDEX

Volume 3 Infant Sucking Behavior and Its Modification Herbert K a y The Study of Brain Electrical Activity in Infants Robert J . Ellingson Selective Auditory Attention in Children Eleanor E. Uaccobv 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 Herberr L. Pick, Jr.. Anne D . Pick, and Robert E. Klein Component Process Latencies in Reaction Times of Children and Adults Raymond H. Hohle AUTHOR INDEX-SUBJECT

INDEX

Volume 2 Volume 4 The Paired-Associates Method in the Study of Conflict Aljred Castaneda Transfer of Stimulus Retraining to Motor PairedAssociate and Discrimination Learning Tasks Joan H. Conror The Role of the Distance Receptors in the Development of Social Responsiveness Richard H. Walrers 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 Hierarchical Arrangement of Learning Rocesses Sheldon H. White

321

Developmental Studies of Figurative Perception David Elkind The Relations of Short-Term Memory to Development and Intelligence John M. Belmont and Earl C . Butterfield Learning, Developmental Research. and Individual Differences Frances Degen Horowitz Psychophysiological Studies in Newborn Infants S. J . Hurt, H.G . Lennrd, and H. F . R. Prechll Development of the Sensory Analyzers during Infancy Yvonne Brackbill and Hiram E. Firzgerald The h b 1 e m of Imitation Justin Aronfreed AUTHOR INDEX-SUBJECT

INDEX

328

Contents of Previous Volumes

vohnnc 5

Volame 8

The Development of Human F e d Activity and Its Relation to Postnatal Behavior T?ypheM Humphrey Arousal Systems and Infant Heut Rate Responses Frances K. Graham and Jan C. Jackson Specific and Diversive Exploration Corinne Hut2 Developmental Studies of Mediated Memory John H . Flavell Development and Choice Behavior in Probabilistic and ProbIem-Solving Tasks L. R. Goukt and Kathryn S. Goodwin

Elaboration and Learning in Childhood and Adolescence Williarn D.Rohwer, Jr. Exploratory Behavior and Human Development Jum C. Nunnally and L. Charles Lomnd Operant Conditioning of Infant Behavior: A Review Robert C. Hulsebus Birth Order and h n t a l Experience in Monkeys and Man G. Mitchell and L. Schroers Fear of the Snanger: A Critical Exmination Harriet L. Rheingold and Carol 0.Eckennan Applications of HuU-Spence Theory to the Transfer of Discrimination W i n g in Children Charles C . Spiker and Joan H . Cantor

AUTHOR INDEX-SUBJECT INDEX Volume 6 Incentives and Learning in Children Sam L. Wit@ Habituation in the Human Infant Wendell E. Jefley and Leslie 8. 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 H o w d V. Meredith Imitation and Language Development James A. Sherman Conditional Responding as a Paradigm for Observational. Imitative Learning and Vicarious-Reinforcement Jacob L. Gewinz AUTHOR INDEX-SUBIECT INDEX Volumc 7

Superstitious Behavior in Children: An Experimental Analysis Michael D.Zeiler Learning Strategies in Children from Different SocioceonomicLevels Jean L. Bresnahan 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 Con-

=F

Rock1 Gelman Laming and Adaptation in Infawy: A Comparison of MOdClS Arnold J . Sameroff

AUTHOR INDEX-SUBJECT INDEX

AUTHOR INDEX-SUBJECT INDEX

Volume 9 Children’s Discrimination k i n g Based on Identity or Difference Betty 1. House, Ann L.Brown, and Marcia S.Scan Two Aspects of Experience in Ontogeny: Development and Learning Ham G . Furth The Effccts of Contextual Changes and Deof Component Mastery on Transfer of Training Joseph C. Campiane and Ann L. Brown Psychophysiological Functioning, AmusaI. Attention. and Learning during the First Year of Life Richard Hirschman and Edward S.Katkin Self-ReinforcementProcesses 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 Representationsof LargeScale 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

Contents of Previous Volumes The Development of Selective Attention: From Perceptual Exploration to Logical Search John C. Wright and Alice G . Vlietstra

AUTHOR INDEX-SUBJECT INDEX

Vdumr 11 The Hyperactive Child: Characteristics,Treatment, and Evaluation of Research Design Gladys B. Baxley and Judith M. LeBlanc Peripheral and Newchemical Parallels of Psychopathology: A Psychophysiological Model Relating Autonomic Imbalance to Hypcnaivity, Psychopathy. and Autism Stephen W. Porges

Constructing Cognitive Operations Linguistically Harry Beilin

Operant Acquisition of Social Behaviors in Infancy: 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 Developmend Psychology for Child Development: Introductory Remarks Paul B. Boltes

Theory and Method in Life-Span Developmental Psychology: Implications for Child Development Aletha Huston-Stein and Paul B. B a l m 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 pnd Life-Span Approaches to the Study of Child Development Michael 1.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 Volume I2 Research between 1%0 and 1970 on the Standing Height of Young Children in Different Pons of the World Howard V. Meredith

The Representation of Children's Knowledge David KIahr and Robert S.Siegler

Chromatic Vision in Infancy Marc H . Bornstein

329

Developmental Memory Theories: Baldwin and Piaget Bruce M. Ross and Stephen M . Kerst Child Discipline and the husuit of Self An Historical Interpretation Howard Gadlin

Development of Time Concepts in Children William 1.Friedman AUTHOR INDEX-SUBJECT INDEX Volume 13 Coding of Spatial and Temporal Information in Episodic Memory Daniel 6 . Berch A Developmental Model of Human Leanzing Barry Gholson and Harry Beilin The Development of Discrimination l a m i n g : A Levels-of-Functioning Explanation Tracy S.Kendler

The Kendler Levels-of-Functioning Theory: Comments and an Alternative Schema Charles C. Spiker and Joan H. Cantor Commentary on Kendler's Paper: An Alternative Perspective Barry Gholson and Therese Schuepfer Reply to Commentaries Tracy S. Kendler On the DeveIopment of Speech Perception: Mechanisms and Andogjes Peter D.Eimas and Vivien C. Tanter

The Economics of Infancy: A Review of Conjugate Reinforcement Carolyn Kent Rovee-Collier and Marcy J. Gekoski

Human Facial Expressions in Response to Taste and Smell Stimulation Jacob E . Steiner AUTHOR INDEXSUBJECr INDEX Volume 14 Development of Visual Memory in Infants John S. Werner and Marion Perhurter Sibship-Constellation Effccts on Psychosocial Development, Creativity, and Health Mazie Eark Wagner, Herman J. P . Schubert. and Daniel

S.P. Schuben

The Development of Understanding of the Spatial Terms Front and Back Louren Julius Harris and Elkn A. Strommen Thc Organization and Contml of Infant Sucking C. K . Crook Neurological Plasticity, Recovery from Brain Insult, and Child Development Ian Sf. James-Roberts

AUTHOR INDEXAUBJECT INDEX

330

Contents of Previous Volumes

Volume IS

Validating Theories of Intelligence Earl C . Burrofield, Dennis Siladi, and John M. Belmonf

Visual Development in Ontogenesis: Some Reeva,uations Juri AIIik and Joan Valsiner Binocular Vision in Infants: A Review and a Theoretical Framework Rirhard N . A s h and Susan T.Dumais

Cognitive Differentiation and Developmental Learning William Fowier Children's Clinical Syndromes and Generalized ExpeeFred Rorhbaum AUTHOR INDEXSUBJECT INDEX