Download the full version of the ebook at

https://ebookfinal.com

Adaptive Perspectives on Human Technology

Interaction Methods and Models for Cognitive
Engineering and Human Computer Interaction
1st Edition Alex Kirlik
https://ebookfinal.com/download/adaptive-
perspectives-on-human-technology-interaction-
methods-and-models-for-cognitive-engineering-and-
human-computer-interaction-1st-edition-alex-
kirlik/

Explore and download more ebook at https://ebookfinal.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Human robot Interaction A Special Double Issue of human

computer Interaction 1st Edition Sara Kiesler (Editor)

https://ebookfinal.com/download/human-robot-interaction-a-special-
double-issue-of-human-computer-interaction-1st-edition-sara-kiesler-
editor/
ebookfinal.com

Issues and Trends in Technology and Human Interaction

Bernd Carsten Stahl

https://ebookfinal.com/download/issues-and-trends-in-technology-and-
human-interaction-bernd-carsten-stahl/

ebookfinal.com

Building interactive systems principles for human computer

interaction 1st Edition Dan R . Olsen

https://ebookfinal.com/download/building-interactive-systems-
principles-for-human-computer-interaction-1st-edition-dan-r-olsen/

ebookfinal.com

Paradoxes of Interactivity Perspectives for Media Theory

Human Computer Interaction and Artistic Investigations 1.
Aufl. Edition Uwe Seifert (Editor)
https://ebookfinal.com/download/paradoxes-of-interactivity-
perspectives-for-media-theory-human-computer-interaction-and-artistic-
investigations-1-aufl-edition-uwe-seifert-editor/
ebookfinal.com
The Human Computer Interaction Handbook Fundamentals
Evolving 2nd Edition Andrew Sears

https://ebookfinal.com/download/the-human-computer-interaction-
handbook-fundamentals-evolving-2nd-edition-andrew-sears/

ebookfinal.com

Multimodal Human Computer Interaction and Pervasive

Services Premier Reference Source 1st Edition Patrizia
Grifoni
https://ebookfinal.com/download/multimodal-human-computer-interaction-
and-pervasive-services-premier-reference-source-1st-edition-patrizia-
grifoni/
ebookfinal.com

Human Computer Interaction and Innovation in Handheld

Mobile and Wearable Technologies 1st Edition Joanna
Lumsden
https://ebookfinal.com/download/human-computer-interaction-and-
innovation-in-handheld-mobile-and-wearable-technologies-1st-edition-
joanna-lumsden/
ebookfinal.com

Advanced Geotechnical Engineering Soil Structure

Interaction using Computer and Material Models 1st Edition
Chandrakant S. Desai (Author)
https://ebookfinal.com/download/advanced-geotechnical-engineering-
soil-structure-interaction-using-computer-and-material-models-1st-
edition-chandrakant-s-desai-author/
ebookfinal.com

Affective computing and interaction psychological

cognitive and neuroscientific perspectives 1st Edition
Gülsen Yildirim
https://ebookfinal.com/download/affective-computing-and-interaction-
psychological-cognitive-and-neuroscientific-perspectives-1st-edition-
gulsen-yildirim/
ebookfinal.com
Adaptive Perspectives on Human Technology Interaction
Methods and Models for Cognitive Engineering and
Human Computer Interaction 1st Edition Alex Kirlik
Digital Instant Download
Author(s): Alex Kirlik
ISBN(s): 9780195171822, 0195171829
Edition: 1
File Details: PDF, 2.53 MB
Year: 2006
Language: english
 Adaptive Perspectives on
Human–Technology
Interaction:
Methods and Models for
Cognitive Engineering and
Human–Computer Interaction

Alex Kirlik,
Editor

OXFORD UNIVERSITY PRESS

ADAPTIVE

PERSPECTIVES o n Human–Technology Interaction

SERIES IN HUMAN–TECHNOLOGY INTERACTION

S E R I E S E D I T O R

Alex Kirlik

Adaptive Perspectives on Human–Technology Interaction: Methods and Models for Cognitive

Engineering and Human–Computer Interaction
Edited by Alex Kirlik
PCs, Phones, and the Internet: The Social Impact of Information Technology
Edited by Robert Kraut, Malcolm Brynin, and Sara Kiesler
ADAPTIVE
PERSPECTIVES

on

Human–Technology
Interaction
Methods and Models for Cognitive Engineering
and Human–Computer Interaction

E D I T E D B Y

Alex Kirlik

1
2006
1
Oxford University Press, Inc., publishes works that further
Oxford University’s objective of excellence
in research, scholarship, and education.
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam

Copyright © 2006 by Oxford University Press, Inc.

Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016
www.oup.com
Oxford is a registered trademark of Oxford University Press
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.

Library of Congress Cataloging-in-Publication Data

Adaptive perspectives on human-technology interaction : methods and models for cognitive
engineering and human-computer interaction / edited by Alex Kirlik.
p. cm. — (Human-technology interaction)
Includes bibliographical references and indexes.
ISBN-13 978-0-19-517182-2
ISBN 0-19-517182-9
1. Human-computer interaction. 2. Human-machine systems. I. Kirlik, Alex. II. Series.
QA76.9.H85A34 2005
004'.01'9—dc22 2005009304

9 8 7 6 5 4 3 2 1
Printed in the United States of America
on acid-free paper
To Egon Brunswik and Kenneth R. Hammond
This page intentionally left blank
 Kenneth R. Hammond

Foreword

This book will no doubt stand as an advance for variables could be employed (as in factorial design),
cognitive engineering, but it will also stand as an they were required, and research blossomed, along
affirmation of Egon Brunswik’s claims for the with scientific prestige—well, a little, anyway. And
significant change he claimed was necessary for you better not try to publish in a major journal
the advancement of psychology. His claims—put without a prominent use of ANOVA. Yet it was this
forward in very scholarly yet unusually bold terms very technique—this goose that was laying the
—were that behaviorism built on narrow, determin- golden egg of scientific respectability—and research
istic, stimulus-response theory, and its accompany- money—that Brunswik was trying to kill. Of course,
ing methodology (the rule of one variable) derived his challenge didn’t stand a chance, and it didn’t get
from a physicalistic theme that should be given up one.
in favor a theme almost exactly opposite to that. It In 1941, however, Brunswik got his chance to
has taken over a half century for change in that di- go head to head with Clark Hull and Kurt Lewin,
rection to reach this point, but as the many con- the leaders of the conventional approaches to psy-
tributors to this volume show, a firm step in the chology. In his presentation Brunswik made this
direction Brunswik advocated has now been taken. statement:
Although Brunswik’s book Perception and the
Representative Design of Psychological Experiments The point I should like to emphasize is . . .
(1956) presented his arguments in a coherent and the necessary imperfection, inflicted upon
substantive fashion, it could not have appeared at achievements . . . by the ambiguity in the
a worse time for his thesis to be considered. The causal texture of the environment. . . . Because
methods of analysis of variance (ANOVA) intro- of this environmental ambiguity, no matter
duced by Fisher some 30 years earlier had by then how smoothly the organismic instruments and
been discovered by psychologists and found to be mechanisms may function, relationships cannot
an answer to their dreams. No longer would they be foolproof, at least as far as those connecting
be restricted to Woodworth’s 1938 dictum about with the vitally relevant more remote distal
the “rule of one variable” that exhausted the experi- regions of the environment are concerned.
mental methodology of the day, and once multiple (Hammond & Stewart, 2001, p. 59)
viii Foreword

Now, to be frank, in 1941 no psychologist but Brunswik was fully aware of the hard road he
Brunswik spoke like this: “causal texture?” “Envi- faced in trying to direct the attention of the psy-
ronmental ambiguity?” “Vitally relevant more re- chologists of the mid-twentieth century to the role
mote distal regions of the environment?” None of of the causal texture of the environment, and in
these terms were part of a psychologist’s vocabu- 1955 he made his final effort with the publication
lary in the 1940s. As a result, many simply refused of Perception and the Representative Design of Psycho-
to try to understand what he was saying and re- logical Experiments (published posthumously in
jected everything with scornful remarks about 1956). In the preface of that book he stated: “This
Brunswik’s inability to write. Although it is prob- book has been written with two major purposes in
ably true that some readers even today won’t be mind. One is the exposition of the more complex
familiar with those words and what they signify, attainments of perception, those attainments that
they will come much closer than Hull and Lewin help stabilize our grasp of the relevant features of
did to understanding that Brunswik was saying a the physical and social environment. The other
great deal, albeit in unfamiliar language applicable purpose is the development of the only methodol-
to a bold new conception of psychology. ogy by which the [aforesaid goal] can be reached,
Today’s cognitive engineers, however, won’t be that is, representative design” (Brunswik, 1956,
afraid of that sentence because the ideas in it— p. vii). Those complex attainments will be exactly
wholly mystifying to mid-century learning theo- those that interest cognitive engineers, and the
rists—are now common. Brunswik was pointing methodology of representative design will be one
out that human beings were going to make inaccu- that they will struggle with for some time; their
rate empirical judgments (“the necessary imperfec- results need to generalize to the environment of
tion, inflicted upon achievements”) and these will interest. In short, it is essential to their purposes.
occur not through any fault of their own (“no mat- The reader will see some of those struggles in
ter how smoothly the organismic instruments and the following chapters, for it is no accident that
mechanisms may function”) but instead because Kirlik chose to collate these chapters within the
of “environmental ambiguity,” and that this was Brunswikian framework. He chose them because no
particularly true of the more important (“vitally paradigm is more conducive to the goals of cogni-
relevant more remote distal regions of the environ- tive engineering than the Brunswikian one, and that
ment”) judgments we make, about other people, is because no other paradigm so clearly differenti-
for example. Thus he prepared the way for differ- ates proximal and distal environmental material
entiating among various cognitive goals (ranging both theoretically and methodologically (even his-
from accurate proximal judgments close to the skin torically). Whereas Brunswik focused his work on
along a continuum to accurate judgments about the natural environment to speak to the academic
covert distal personal or meteorological variables). psychologists of the day, he was mindful of the
Communication failed not because he was poor strong implications his paradigm held for the arti-
writer (he was highly precise), but because in call- ficial, or “engineered, ” environment. Those impli-
ing attention to environmental ambiguity, no one cations can be reduced to one: The environment
knew or was interested in what he was talking toward which the researcher intends to generalize
about, and that was because no one was then giv- should be specified in advance of the design of the
ing any consideration to the environment. The or- experiment. That (a priori) specification should
ganism and its “instruments and mechanisms” include theory as well as method. Conventional
dominated everything; all that was needed was a academic psychology ignored both requirements
stimulus to get it going and produce a response. But throughout the twentieth century and suffered the
not for cognitive engineers: Given the goals of their consequences of producing floating results; “float-
profession—and these include the design of the ing” because the simple logic of generalization was
environment, and the design of technical displays— applied only to subject populations; environmen-
cognitive engineers know exactly what Brunswik tal generalization was ignored; consequently, the
was saying to psychologists, and that was: Consider implications of the results were left unanchored.
the informational characteristics of the environment The studies Kirlik chose to include here may not
and how these affect the judgments of individuals. meet Brunswikian ecological criteria in every way
 Foreword ix

in every case, yet they will illustrate the need chapters in this volume show (see especially chap-
for meeting them and will bring us closer to our ter 12 and chapter 18).
goal of understanding the problems of cognitive If the cognitive engineering of information is
engineering. now indispensable, it is also highly varied and com-
It would be hard to find a field that will evoke plex. A glance at the table of contents of this book
a greater fit for the Brunswikian paradigm than will be sufficient to grasp that it involves an aston-
cognitive engineering, for here the distinctions ishingly wide range of topics. This can also be seen
between natural environments and engineered en- in the editor’s introductions to the various parts;
vironments arise immediately. An organism in a they offer an education in the struggle to make psy-
natural environment is required to cope with what chological knowledge useful. This broad vision
Brunswik called an “uncertainty-geared” environ- works to the reader’s advantage, for it removes cog-
ment; probabilism is at its core. (This was an idea nitive engineering from mere application of knowl-
neither Hull and Lewin could stomach in that sym- edge already at hand to the forefront of knowledge
posium in 1941, and they made their revulsion acquisition, which was surely the editor’s intention.
known to Brunswik.) But accepting that contrast This sophisticated consideration of theory and
between uncertainty-geared and certainty-geared method together with the pursuit of utilization led
environments made clear exactly what the goal of to the conclusion that representing the environment
cognitive engineering would be, namely, replacing was essential; theory and method would have to be
the uncertainty-geared natural environment with a adjusted accordingly. Thus cognitive engineering is
certainty-geared environment; the optimal replace- changing not only applied psychology but psychol-
ment created by cognitive engineers. ogy itself.
Indeed, one might say that such replacement Therefore this book does more than just affirm,
defines the field of cognitive engineering. Why? it points to the future. The down-to-Earth charac-
Because uncertainty in the environment means er- teristics of its contents will show how the recogni-
ror in judgment, errors of judgment can be extraor- tion of the duality of error increased the research
dinarily costly, therefore in situations in which sophistication of its authors, thanks largely to the
errors are costly, cognitive engineers should drive fruitfulness of the application of such techniques
out or at least reduce to a minimum environmen- as signal detection theory (SDT) and the Taylor-
tal uncertainty. Russell (T-R) diagram. Although introduced roughly
Possibly the best and most successful example a decade earlier than SDT, it was not applied in
of meeting the challenge to cognitive engineering the field of cognitive science until much later (see
to reduce environmental uncertainty is illustrated Hammond, 1966, for a brief history of both). Both
in aviation psychology. Aviation is an example of a techniques are directed toward the idea of separat-
trade or profession that began with its practitio- ing false positive errors and false negative errors
ners—the pilots—utterly dependent on informa- and that separation immediately made apparent
tion provided by perceptual (including kinesthetic) the critical role of cost, benefits, and, most impor-
cues afforded by the uncertainty-geared natural tant, values and trade-offs among them. SDT and
environment. But aviation engineering has now the T-R diagram offered valuable quantitative means
moved to the point where information from this of clarification of these relationships in a manner
environment is ignored in favor of information from not seen before. The future will bring these ideas
an artificial certainty-geared, wholly engineered to a new prominence, and in doing so new dis-
environment. Uncertainty has been driven out of tinctions will appear that will advance theory and
the flight deck to a degree unimaginable when avia- research. For example, it will become obviously
tion began. That change made commercial airline necessary to distinguish between environments of
travel practical and saved countless lives. Exactly reducible and irreducible uncertainty both theoreti-
how much the study of cognition contributed to cally and methodologically.
that engineering achievement is unclear, but history Environments that permit reducible uncertainty
is not likely to give it much credit. We are now at a will be targets of opportunity for cognitive engi-
point, however, where cognitive engineering of in- neers—provided other engineers haven’t already
formation for the pilot is indispensable, as several exploited them. The aviation industry was already
x Foreword

exploited because the essentially uncluttered nature irreducible uncertainty. Such “territories” offer the
of the sky invited exploitation, and the necessary best example of requiring humans to cope with
technology fit with the rapid development of elec- environments of irreducible uncertainty. Here is the
tronics. The navigational (and traffic) uncertainties strong future challenge to cognitive engineers; they
of the uncluttered sky were steadily reducible by need not and should not restrict themselves to situ-
electronic means of measurement and communica- ations involving gauges and electronics and me-
tion. Allow me a brief anecdote to show how recent chanical artifacts. Their knowledge and skills are
that was. In 1941 I was an observer in the Weather badly needed in areas where disputes remain largely
Bureau at the San Francisco Airport. On the night at the level of primitive people past and still in
shift I often talked with the janitor, a bent-over old modern times regularly leading to mayhem, mur-
man who had been a sky-writer pilot in the early der, and wholesale slaughter, not to mention deg-
days of such stunts. His stories were fascinating; he radation and poverty. But cognitive engineers are
often mentioned lack of instruments (“we had al- accustomed to work at abstract systems levels;
most nothing”) and what that meant (attacks of the therefore their theories and methods should be
“bends”) due to a too-rapid descent. Thus, in less applicable, albeit at a level of complexity that will
than a century aviation engineering went from al- demand innovation in theory and method and
most nothing to the glass cockpit in which infor- thought because of the shift to environments char-
mation from inside the cockpit means more than acterized by irreducible uncertainty.
information from without. That uncluttered envi- At the level of irreducible uncertainty the
ronment was successfully exploited by conven- Brunswikian approach will be of considerable as-
tional, largely electronic engineering and remains sistance to cognitive engineers because it will allow
now only to be tidied up by cognitive engineers. the broadest range of theories and methods to be
There remain many other environments that offer included. Domains of reducible uncertainty will be
reducible uncertainty that cognitive engineers will conquered with and without the aid of cognitive
exploit to the benefit of all of us. engineers, as indeed the domain of aviation psy-
Fortunately, Brunswikian theory and method chology already has. But the domains of irreduc-
provides a big tent; it is inclusive rather than ex- ible uncertainty will demand all the knowledge
clusive. By virtue of its demand that the study and skills and ingenuity of the modern cognitive
must be designed to justify generalization to the engineer to cope with the consequences of the
environment of interest, it permits the use of any duality of error that follow from irreducible un-
design that meets that criterion. That means that certainty. That demand will surely include the
when generalization can be met by factorial or newfound knowledge of Brunswikian psychology
other forms of ANOVA, then these designs will be that includes cognitive theory and the methodol-
appropriate. When, however, the situation toward ogy appropriate to it.
which the generalization is intended involves in-
terdependent variables and other features not rep-
resented by factorial and similar designs, then they References
should not be used because the generalization will
Brunswik, E. (1956) Perception and the representative
not be justified. Representation, whatever its form,
design of psychological experiments. Berkeley, CA:
is key to generalization.
University of California Press.
But contrast that virgin—“blue sky”—territory Hammond, K. R. & Stewart, T. R. (2001) The essential
with its reducible uncertainty with the murky ter- Brunswik: Beginnings, explications, applications.
ritories, such as social policy formation, that entail New York: Oxford University Press.
Contents

Foreword vii
K. R. Hammond

Contributors xv

I Background and Motivation

1 Cognitive Engineering: Toward a Workable Concept of Mind 3
Alex Kirlik

2 Introduction to Brunswikian Theory and Method 10

William M. Goldstein

II Technological Interfaces
Introduction 27
Alex Kirlik

3 Knowledge versus Execution in Dynamic Judgment Tasks 29

Ann M. Bisantz, Alex Kirlik, Neff Walker, Arthur D. Fisk. Paul Gay, and Donita Phipps

4 Understanding the Effects of Computer Displays and Time Pressure on the

Performance of Distributed Teams 43
Leonard Adelman, Cedric Yeo, and Sheryl L. Miller

5 Supporting Situation Assessment through Attention Guidance and

Diagnostic Aiding: The Benefits and Costs of Display Enhancement
on Judgment Skill 55
William J. Horrey, Christopher D. Wickens, Richard Strauss, Alex Kirlik,
and Thomas R. Stewart
xii Contents

6 Applying the Multivariate Lens Model to Fault Diagnosis 71

Pratik D. Jha and Ann M. Bisantz

III Automation and Decision Aiding

Introduction 89
Alex Kirlik

7 Measuring the Fit between Human Judgments and Alerting Systems:

A Study of Collision Detection in Aviation 91
Amy R. Pritchett and Ann M. Bisantz

8 Trust, Automation, and Feedback: An Integrated Approach 105

Younho Seong, Ann M. Bisantz, and Gordon J. Gattie

9 Human–Automated Judgment Learning: Enhancing Interaction with

Automated Judgment Systems 114
Ellen J. Bass and Amy R. Pritchett

IV Alternatives to Compensatory Modeling

Introduction 129
Alex Kirlik

10 Inferring Fast and Frugal Heuristics from Human Judgment Data 131
Ling Rothrock and Alex Kirlik

11 Viewing Training through a Fuzzy Lens 149

Gwendolyn E. Campbell, Wendi L. Van Buskirk, and Amy E. Bolton

12 Achieving Coherence: Meeting New Cognitive Demands

in Technological Systems 163
Kathleen L. Mosier and Shane T. McCauley

V Into the Field: Vicarious Functioning in Action

Introduction 177
Alex Kirlik

13 What Makes Vicarious Functioning Work? Exploring the Geometry

of Human–Technology Interaction 179
Asaf Degani, Michael Shafto, and Alex Kirlik

14 Understanding the Determinants of Adaptive Behavior in a Modern

Airline Cockpit 197
Stephen M. Casner

15 Abstracting Situated Action: Implications for Cognitive Modeling

and Interface Design 212
Alex Kirlik

VI Ecological Analysis Meets Computational Cognitive Modeling

Introduction 227
Alex Kirlik
 Contents xiii

16 The Emerging Rapprochement between Cognitive

and Ecological Analyses 230
Wayne D. Gray

17 The Use of Proximal Information Scent to Forage for Distal Content

on the World Wide Web 247
Peter Pirolli

18 Kilograms Matter: Rational Analysis, Ecological Rationality, and

Closed-Loop Modeling of Interactive Cognition and Behavior 267
Michael D. Byrne, Alex Kirlik, and Chris S. Fick

VII Reflections and Future Directions

19 Reflections from a Judgment and Decision Making Perspective 287
Terry Connolly

20 Reflections from a Cognitive Engineering and Human

Factors Perspective 292
Kim J. Vicente

Name Index 297

Subject Index 303
This page intentionally left blank
Contributors

Leonard Adelman Stephen M. Casner

Department of Systems Engineering and Human Factors Research and Technology
Operations Research Division
George Mason University NASA Ames Research Center

Terry Connolly
Ellen J. Bass
Department of Management and Policy
Department of Systems and Information
University of Arizona
Engineering
University of Virginia
Asaf Degani
Computational Sciences Division
Ann M. Bisantz NASA Ames Research Center
Department of Industrial Engineering
University at Buffalo, State University of Chris S. Fick
New York Department of Psychology
Rice University
Amy E. Bolton
Training Systems Division Arthur D. Fisk
U.S. Naval Air Warfare Center School of Psychology
Georgia Institute of Technology

Michael D. Byrne Gordon J. Gattie

Psychology Department Department of Industrial Engineering
Rice University University at Buffalo, State University
of New York
Gwendolyn E. Campbell
Training Systems Division Paul Gay
U.S. Naval Air Warfare Center Delta Airlines

xv
xvi Contributors

William M. Goldstein Ling Rothrock

Department of Psychology The Harold and Inge Marcus Department of
University of Chicago Industrial and Manufacturing Engineering
Pennsylvania State University
Wayne D. Gray
Department of Cognitive Science Michael G. Shafto
Rensselaer Polytechnic Institute Exploration Systems Mission Directorate
NASA Headquarters
Kenneth R. Hammond
Department of Psychology Younho Seong
University of Colorado Department of Industrial & Systems
Engineering
William J. Horrey North Carolina A&T State University
Liberty Mutual Research Institute for Safety
Thomas R. Stewart
Pratik D. Jha
Center for Policy Research
Titan Corporation
University at Albany, State University of
New York
Alex Kirlik
Human Factors Division and Beckman
Richard Strauss
Institute
Fatwire Software
University of Illinois at Urbana-Champaign

Wendi L. Van Buskirk

Shane T. McCauley
Training Systems Division
Department of Psychology
U.S. Naval Air Warfare Center
San Francisco State University

Sheryl L. Miller Kim J. Vicente

Department of Systems Engineering and Department of Mechanical and Industrial
Operations Research Engineering
George Mason University University of Toronto

Kathleen L. Mosier Neff Walker

Department of Psychology UNAIDS
San Francisco State University
Christopher D. Wickens
Donita Phipps Human Factors Division and Department of
Department of Psychology Psychology
Georgia Institute of Technology University of Illinois at Urbana-Champaign

Peter Pirolli Cedric Yeo

User Interface Research Department of Systems Engineering and
PARC Operations Research
George Mason University
Amy R. Pritchett
School of Industrial and Systems
Engineering
Georgia Institute of Technology
I
Background and Motivation
This page intentionally left blank
1 Alex Kirlik

Cognitive Engineering: Toward a

Workable Concept of Mind
It seems plain to me now that the “cognitive revolution” . . . was a response to the technologi-
cal demands of the Post-Industrial Revolution. You cannot properly conceive of managing a
complex world of information without a workable concept of mind.
—Bruner (1983, p. 63)

Perhaps no one has understood the depth to which provide methods and models that can be fruitfully
the ever-increasing technological nature of the applied to solving practically relevant problems in
human ecology has shaped psychological theory human–technology interaction. These problems in-
better than Jerome Bruner. In his memoir In Search clude designing and evaluating technological inter-
of Mind (1983), Bruner shared his reflections on faces, decision aids, alerting systems, and training
the origins of the cognitive revolution. Although technology, as well as supporting human–automa-
a great many factors may have played a role (e.g., tion interaction and human–computer interaction.
Chomsky, 1959; Miller, 1956; Newell & Simon, In short, the aim of this book is to provide practi-
1972), Bruner turns much conventional thinking cal resources for addressing the menagerie of prob-
on its head, implying that scientists had to invent a lems making up cognitive engineering (Hollnagel &
theory of mind in response to the practical demands Woods, 1983; Kirlik & Bisantz, 1999; Norman,
of finding coherent ways of understanding and 1986; Rasmussen, 1986). Along the way, many con-
coordinating a largely invented world of people tributors to this volume also present insights and
engaged with post–Industrial Revolution technolo- approaches that may shed light on fundamental
gies. The seeds of this scientific revolution, it seems, problems in the science of adaptive cognition and
were not so much “in the air” as in the digital cir- behavior. This may be especially true when it comes
cuitry and in the need to understand and manage to the challenge of understanding and formally ar-
“a complex world of information.” ticulating the role of the environment in cognitive
theory.
Six themes unite the contributors’ orientation
A Workable Concept of Mind toward developing a concept of mind that is both
workable and valuable from a cognitive engineer-
The purpose of this book is to take additional steps ing perspective. These themes are illustrated in the
toward building what Bruner referred to as a “work- selection of research problems, methods, and analy-
able concept of mind.” Special emphasis is given sis and modeling techniques presented in the fol-
here to the word workable. The central goal is to lowing chapters.

3
4 Background and Motivation

1. An Ecological or and “cognitive analysis” to highlight the study of the

Systems Perspective internal or mental activities of their human com-
ponents, the intended meaning of each term will
Pioneers of the cognitive engineering discipline, be clear from context. At this point, suffice it to say
such as Rasmussen (1990) and Woods (1995), have that the perspective taken in this book is that a psy-
emphasized the essential ecological character of the chological theory or model, especially one capable
cognitive engineering enterprise. What does this of providing cognitive engineering with a workable
mean, and why is it the case? The answers to these concept of mind, should include a description of
questions lie in the nature of the systems studied the “whole” human–technology–environment sys-
and the practical problems associated with their tem and not merely internal cognition.
analysis and design. Cognitive engineering research This systems or ecological orientation toward
is largely concerned with the analysis and design of cognitive engineering has long roots in human fac-
human–technology systems and the role of technol- tors, and especially within the field of “human–
ogy as a mediator between humans and the external machine systems.” As noted by Sheridan (2002,
objects and events comprising a work domain. As p. 5), shortly after World War II human factors re-
such, cognitive engineering is fundamentally con- searchers increasingly began to borrow modeling
cerned with systems composed of humans, mediat- techniques from engineering, such as mathemati-
ing technologies and task environments, as well as cal theories of estimation, information, decision,
the interactions among these system components. Of and control. They did so “to look at information,
all the schools or approaches to psychology relevant control, and decision making as a continuous pro-
to cognitive engineering (and there are many), both cess within a closed loop that also included physi-
the theoretical orientation and unit of analysis pre- cal subsystems—more than just sets of independent
sumed by ecological psychology are perhaps best stimulus-response relations” (Sheridan, 2002, p. 4).
aligned with the task of understanding and support- The rationale for including a description of physi-
ing the mediated coupling between humans and cal subsystems within such models, that is, descrip-
their environments. By reviewing the Instructions tions of the task environment and any mediating
for Authors appearing in each issue of Ecological technology, was well expressed by Baron, a human–
Psychology, one learns that the focus of the disci- machine systems engineer: “Human behavior, either
pline is a broad range of psychological problems (per- cognitive or psychomotor, is too diverse to model
ception, cognition, communication, development, unless it is sufficiently constrained by the situation
learning, etc.), but with one noteworthy constraint. or environment; however, when these environmen-
In particular, ecological psychology is concerned with tal constraints exist, to model behavior adequately,
those problems only “to the extent that those prob- one must include a model for that environment”
lems derive from a consideration of whole animal- (Baron, 1984, p. 6).
environmental systems, rather than animals or Exactly this point, the need for a psychological
environments in isolation from each other.” model to include an environmental model, was one
This broad and inclusive perspective on what of the defining features of the theory and method
it means to take an ecological approach to cogni- developed by one of ecological psychology’s pio-
tive engineering is the one presumed in this book. neers, Egon Brunswik (1903–1955). As will be
As such, the approach is much in the spirit of the discussed in greater detail by Goldstein in the fol-
Joint Cognitive Systems tradition in cognitive sys- lowing chapter, Brunswik is perhaps best known
tems engineering (Hollnagel, Mancini, & Woods, for his lens model, whose ecological nature is di-
1986). It is inherently a systems approach in which rectly apparent in its symmetrical arrangement as
each element of the human–technology–environ- a pair of joined human and environmental mod-
ment unit of analysis receives attention and treat- els (Brunswik, 1952, 1956; also see Hammond &
ment during analysis and design and where each Stewart, 2001). The pervasive influence of Bruns-
element is considered in light of its functional role wik’s theory of probabilistic functionalism and
within the overall human–technology–environment experimental methodology of representative design
system. Although some authors in this book use the is evident throughout this volume.
phrase “ecological analysis” to highlight the study This is not to imply, of course, that only Bruns-
of the environmental components of these systems wik’s ecological theory is capable of giving rise to
 Cognitive Engineering: Toward a Workable Concept of Mind 5

workable resources for cognitive engineering. For gramming a VCR or the flight control automation
example, Vicente (1999) has drawn heavily and in a modern glass cockpit airliner (Degani, Shafto,
profitably on the ecological theory of James J. & Kirlik, 1999; Sarter & Woods, 1992).
Gibson (1979/1986) in the development of his One reason the authors of this volume have
Cognitive Work Analysis and Ecological Interface become attracted to Brunswik’s functionalist theory
Design (also see Burns & Hajdukiewicz, 2004) cog- of cognition and behavior is its grounding in ex-
nitive engineering techniques. As both Vicente actly these gulfs. In particular, and as explained in
(2003) and Kirlik (1995, 2001) have observed, it greater detail by Goldstein in chapter 2, Brunswik’s
is possible to view the ecological theories of Bruns- theory is founded in an examination of the proxi-
wik and Gibson to be complementary rather than mal–distal relationships characterizing a person’s
conflicting, despite what much of the psychologi- encounter with the world (also see Tolman &
cal research conducted in each of the two traditions Brunswik, 1935). A technological interface provides
may lead one to believe. As such, one should view us with both proximal, or directly available, infor-
the research presented in the current volume, mation sources and proximal opportunities for
largely grounded in Brunswikian theory, and the action. The intended target of our interaction, how-
research program of Vicente and his colleagues, ever, is all too often distal, or not so directly avail-
influenced by Gibsonian theory, as similarly com- able to us: a goal state to be achieved by taking a
plementary rather than conflicting. Neither sub- correct sequence of proximal actions, and an un-
sumes the other with respect to the central derstanding of whether we have achieved our goal,
problems addressed or the techniques provided. which can often be gained only by correctly inte-
The same could also be said of the distributed cog- grating proximally available information.
nition approach (e.g., Hollan, Hutchins, & Kirsh, In 1972, Newell and Simon began their semi-
2000; Hutchins, 1995) and any other framework nal book, General Problem Solving, with an expres-
embodying the ecological notion that the unit of sion of debt to Tolman & Brunswik’s (1935)
psychological analysis and modeling must span the Psychological Review article emphasizing the neces-
human–environment boundary if it is to provide sity of a detailed analysis of these proximal–distal
cognitive engineering with a workable concept of relations to understand goal-directed behavior.
mind (also see Clark, 2003; Dourish, 2001; Kirlik, Newell and Simon appreciated the insight that such
1998; Kirsh, 1996; Olson & Olson, 1991; Zhang behavior is typically directed toward distal objects
& Norman, 1997). or ends that can be achieved only by the adaptive
use of proximal information and action resources.
One result was Newell and Simon’s characterization
2. An Adaptive, of problem solving as a search through a “problem
Functional Perspective space” to find a series of proximal actions that would
lead to distal goals. A problem space is a model of a
Anyone who has ever read (or even better, taught problem solver’s environment.
from) Donald Norman’s (1988) insightful book The Every chapter in this volume deals in one way
Psychology of Everyday Things will almost certainly or another with an examination of the proximal–
recognize his characterization of the problems faced distal relations characterizing or mediating one’s
by a technology user in terms of bridging a “gulf of encounter with the environment and a parallel ex-
execution” and a “gulf of evaluation.” How do I get amination of the degree to which humans are well
it to work? What is it doing? These are questions adapted to these relations. This is what is meant
we find ourselves asking all to often in our interac- here by an adaptive, functional perspective on cog-
tions with technology. As one who has graded over nition and behavior. There is no initial assumption
a thousand students’ answers to “define and give an that people are either well or poorly adapted to the
everyday example” exam questions about these demands and opportunities of any particular situ-
gulfs, I have found that if a student is going to get ation. Instead, the approach taken in the following
anything correct on an exam it is likely to be these chapters is to perform detailed functional analyses
questions. These concepts are immediately intuitive of task environments and then empirically measure
to anyone who has ever experienced difficulty or the degree of adaptivity attained in light of both the
frustration when using technology, whether in pro- cognitive and environmental resources available.
6 Background and Motivation

Various chapters in parts IV, V, and VI of this book goal of “conducting one’s study with representative
also examine the relationship between Brunswik’s samples of subjects, tasks, and contexts to which
original approach to these problems and more re- one wishes to generalize” (Lipshitz et al., 2001b,
cent yet related adaptive approaches to cognition p. 386). This is illustrated in this volume by the com-
such as Anderson’s (1990) Rational Analysis and plete reliance on either field observation or the use
ACT-R model (Anderson & Lebiere, 1998), and of dynamic and interactive simulations modeled
the Ecological Rationality approach developed by after the target context of scientific generalization.
Gigerenzer, Todd, and their colleagues (Gigerenzer, As will be described by Goldstein in more detail in
Todd, & the ABC Research Group 1999). chapter 2, this methodological commitment is con-
sistent with Brunswik’s methodology of represen-
tative design. However, as will be illustrated in
3. Embracing Uncertainty nearly all of the chapters that follow, this orienta-
tion does not preclude the use of systematically
Another aspect of Brunswik’s thinking adopted by designed interventions in representatively designed
the authors in this volume is the idea that the re- experiments. This book is filled with examples of
lationship between the human and environment investigators using hybrid representative/systematic
must often be characterized in probabilistic terms. experimental designs to both foster the generaliza-
Note, however, that this does not reflect an a priori tion of results to a target context and also to test
commitment to probabilism but instead the need various hypotheses regarding the efficacy of design,
to have conceptual and technical resources avail- training, or aiding interventions and to examine
able for measuring and modeling uncertainty where how adaptivity may be influenced by factors such
it is found to exist. Having techniques available to as time stress.
represent the possible probabilistic structure both
within a task environment and within the opera-
tions of inner cognitive processes is especially im- 5. A Formal Perspective
portant for the purpose of evaluating the adaptivity
of behavior, and also when motivating interventions The research presented in this volume displays a
aimed at enhancing it. Why? commitment to abstraction and formalization in the
First, environmental uncertainty places a ceiling creation of modeling and measurement techniques
on the accuracy of adaptive behavior in any given (also see Byrne & Gray, 2003). The contributors to
instance. As Lipshitz et al. (2001a) have (qualitatively) this volume, as illustrated by their demonstrated
put it, “Uncertainty is intimately linked with error: the commitment to study cognition and behavior in
greater the uncertainty, the greater the probability of context and to perform formal (mathematical or
making an error” (p. 339). As such, it is important to computational) modeling, agree with Todd and
recognize the existence of environmental uncertainty Gigerenzer (2001) in noting that the alternatives
from a forensic perspective because human “error” for describing context-sensitive, adaptive cogni-
must always be expected when people are perform- tion “are not context-free formal modeling versus
ing in environments with irreducible uncertainty context-bound informal modeling” (p. 382).
(Hammond, 1996). Second, the possible presence of Instead, I hope that the chapters that follow
uncertainty suggests that questions about adaptive illustrate that it is quite possible to have a deep
cognition be addressed and answered at the level of appreciation for the role of the environmental con-
how well tailored or calibrated a performer’s judg- text in cognition and behavior, yet also to have an
ments or actions are to the environment on average, appreciation for and ability to formally model the
or at a distributional level of analysis, rather than on essential aspects of this context. As I have pointed
an instance-by-instance basis. out elsewhere (Kirlik, 2003), research in fields such
as human factors and cognitive engineering nearly
always begins (or should begin) with a qualitative,
4. Embracing Representativeness naturalistic phase to identify and distill the central
features of a target problem to be solved or phenom-
The research presented in this book shares with the ena to be investigated. Yet if attention then turns
Naturalistic Decision Making (NDM) paradigm the directly to creating an intuitive solution or qualita-
 Cognitive Engineering: Toward a Workable Concept of Mind 7

tive account (regardless of how well received by models presented here appreciate that these tech-
stakeholders), without bringing these central fea- niques are all works in progress, and additional
tures to an abstract level, it is often impossible to progress depends on extending and elaborating
know the conditions in which that same solution these tools and techniques. Only then will cogni-
will prove useful. As such, each cognitive engineer- tive engineering come to possess the kind of diverse
ing problem will have to be solved largely from and reusable toolbox of measurement, analysis, and
scratch. A workable concept of mind useful for cog- modeling techniques characteristic of other engi-
nitive engineering should be one that is fertile in neering disciplines.
giving rise to a toolbox of formal analysis and mod-
eling techniques, as scientific generalization rides
chiefly on the winds of abstraction. Conclusion

After the immediately following chapter on theory

6. A Problem-Solving Perspective and method, this volume consists of a set of chap-
ters presenting research adhering to each of the six
The problem-solving orientation displayed in the general themes or perspectives outlined in this
research presented herein will be immediately ap- chapter. Although it is traditional in a volume such
parent. What might be less obvious in these pages as this to provide an introductory overview and
is that in virtually every case researchers had to roadmap of the chapters that follow, this is unnec-
work with the behavioral situation given to them, essary for two reasons. First, in chapter 2 Goldstein
and this required them to invent a wide variety of provides an introduction and tutorial that explic-
novel extensions and improvisations on the general itly situates each chapter by name within its intel-
theoretical perspective uniting these studies within lectual and historical place in Brunswik’s thought
a common intellectual and historical perspective. and scholarship. Second, I have provided introduc-
It is likely that the more important contributions tory overviews prefacing each of the volume’s subse-
within these chapters lie in the development of these quent sections. Should one desire to gain a quick
extensions to current theory and method, rather than overview of the substantive contents of the volume,
in any particular empirical findings presented. It may reading Goldstein’s chapter and/or reading through
be useful to highlight this aspect of the book at the each of these brief section introductions should
outset. serve this purpose well.
To explain, the study of human–technology in-
teraction and the disciplines of cognitive engineering,
human factors, and human–computer interaction are References
changing so rapidly that the practically relevant life-
Anderson, J. R. (1990). The adaptive character of
time of any particular empirical effect or finding is
thought. Hillsdale, N.J.: Erlbaum.
likely to be quite short. In a world continually under- Anderson, J. R., & Lebiere, C. (1998). The atomic
going reinvention, the specific barriers to adaptive components of thought. Mahwah, N.J.: Lawrence
cognition and behavior are likely to change as rap- Erlbaum.
idly as the ecology itself (Kirlik, 2005). As is often the Baron, S. (1984). A control theoretic approach to
case with the workers and performers cognitive en- modeling human supervisory control of dynamic
gineers intend to support, our own work as research- systems. In W. B. Rouse (Ed.), Advances in
ers, practitioners, and students of human–technology human-machine systems research, vol. 1. (pp. 1–
interaction is one of almost continuous adaptation to 48). Greenwich, Conn: JAI Press.
novel problems and opportunities. As such, a useful Bruner, J. (1983). In search of mind. New York: Harper
& Row.
method, model, or measurement technique is likely
Brunswik, E. (1952). The conceptual framework of
to have a longer shelf life, as measured by its dura-
psychology. In International Encyclopedia of
tion of practical relevance, than any particular experi- Unified Science, vol. 1, no. 10. (pp. 1–102)
mental result or field observation. Chicago: University of Chicago Press.
As such, one can only hope that those who Brunswik, E. (1956). Perception and the representative
might embrace the perspectives on which this book design of psychological experiments. Berkeley:
is based and those who might use the methods and University of California Press.
8 Background and Motivation

Burns, C. M., & Hajdukiewicz, J. R. (2004). Ecological K. R. Hammond and T. R. Stewart (Eds.), The
interface design. Boca Raton, Fla.: CRC Press. essential Brunswik (pp. 238–242). New York:
Byrne, M. D., & Gray, W. D. (2003). Returning Oxford University Press.
human factors to an engineering discipline: Kirlik, A. (2003). Human factors distributes its
Expanding the science base through a new workload. Review of E. Salas (Ed.), Advances in
generation of quantitative methods—preface to human performance and cognitive engineering
the special section. Human Factors, 45, 1–4. research. Contemporary Psychology, 48(6).
Chomsky, N. (1959). A review of Skinner’s Verbal Kirlik, A. (2005). Work in progress: Reinventing
Behavior. Language, 35, 26–38. intelligence for a reinvented world. In R. J.
Clark, A. (2003). Natural-born cyborgs. New York: Sternberg & D. Preiss (Eds.), Intelligence and
Cambridge University Press. technology (pp. 105–134). Mahwah, N.J.: Erlbaum.
Degani, A., Shafto, M., & Kirlik, A. (1999). Modes in Kirlik, A., & Bisantz, A. M. (1999). Cognition in
human-machine systems: Review, classification, human-machine systems: Experiential and
and application. International Journal of Aviation environmental aspects of adaptation. In P. A.
Psychology, 9(2), 125–138. Hancock (Ed.), Handbook of perception and cognition
Dourish, P. (2001). Where the action is: The foundations (2nd ed.): Human performance and ergonomics
of embodied interaction. Cambridge, Mass.: MIT (pp. 47–68). New York: Academic Press.
Press. Kirsh, D. (1996). Adapting the environment instead
Gibson, J. J. (1979/1986). The ecological approach to of oneself. Adaptive Behavior, 4(3/4), 415–452.
visual perception. Hillsdale, NJ: Erlbaum. Lipshitz, R., Klein, G., Orasanu, J., and Salas, E.
Gigerenzer, G., Todd, P. M., & the ABC Research (2001a). Focus article: Taking stock of naturalis-
Group. (1999). Simple heuristics that make us tic decision making. Journal of Behavioral Decision
smart. New York: Oxford University Press. Making, 14, 331–352.
Hammond, K. R. (1996). Human judgment and social Lipshitz, R., Klein, G., Orasanu, J., and Salas, E.
policy: Irreducible uncertainty, inevitable error, (2001b) Rejoinder: A welcome dialogue—and the
unavoidable injustice. New York: Oxford Univer- need to continue. Journal of Behavioral Decision
sity Press. Making, 14, 385–389.
Hammond, K. R., & Stewart, T. R. (Eds.). (2001). The Miller, G. A. (1956). The magical number seven, plus
essential Brunswik. New York: Oxford University or minus two: Some limits on our capacity for
Press. processing information. Psychological Review, 63,
Hollan, J., Hutchins, E., & Kirsh, D. (2000). 81–97.
Distributed cognition: Toward a new foundation Newell, A., & Simon, H. A. (1972). Human problem
for human-computer interaction research. ACM solving. New York: Prentice Hall.
Transactions on Human-Computer Interaction, 7(2), Norman, D. A. (1986). Cognitive engineering. In D. A.
174–196. Norman & S. W. Draper (Eds.), User centered
Hollnagel, E., Mancini, G., & Woods, D. D. (1986). system design: New perspectives in human-computer
Intelligent decision support in process environments. interaction (pp. 31–61). Mawhah, N.J.: LEA.
New York: Springer-Verlag. Norman, D. A. (1988). The psychology of everyday
Hollnagel, E., & Woods, D. D. (1983). Cognitive things. New York: Basic Books.
systems engineering: New wine in new bottles. Olson, G. M., & Olson, J. S. (1991). User-centered
International Journal of Man-Machine Studies, 18, design of collaboration technology. Journal of
583–600. Organizational Computing, 1, 61–83.
Hutchins, E. (1995). Cognition in the wild. Cambridge, Rasmussen, J. (1986). Information processing and
Mass.: MIT Press. human-machine interaction: An approach to
Kirlik, A. (1995). Requirements for psychological cognitive engineering. New York: North-Holland.
models to support design: Toward ecological task Rasmussen, J. (1990). Cognitive engineering, a new
analysis. In J. Flach, P. Hancock, J. Caird, and profession? In L. P. Goodstein, S. E. Olsen, &
K. J. Vicente (Eds.), Global perspectives on the H. B. Anderson (Eds.), Tasks, errors and mental
ecology of human-machine systems (pp. 68–120). models (pp. 325–334). New York: Taylor &
Mawhah, N.J.: LEA. Francis.
Kirlik, A. (1998). The design of everyday life Sarter, N., & Woods, D. D. (1992). Pilot interaction
environments. In W. Bechtel and G. Graham with cockpit automation. I. Operational
(Eds.), A companion to cognitive science.(pp. 702– experiences with the Flight Management System.
712). Oxford: Blackwell. International Journal of Aviation Psychology, 2,
Kirlik, A. (2001). On Gibson’s review of Brunswik. In 303–321.
 Cognitive Engineering: Toward a Workable Concept of Mind 9

Sheridan, T. B. (2002). Humans and automation: direct perception: Toward a broader view of
system design and research issues. Santa Monica, ecological psychology. Ecological Psychology,
Calif.: Human Factors and Ergonomics Society 15(3), 241–267.
and John Wiley & Sons. Woods, D. D. (1995). Toward a theoretical base
Todd, P., & Gigerenzer, G. (2001). Putting naturalis- for representation design in the computer
tic decision making into the adaptive toolbox. medium: Ecological perception and aiding
Journal of Behavioral Decision Making, 14, 381– human cognition. In J. Flach, P. Hancock,
382. J. Caird, and K. J. Vicente (Eds.), Global
Tolman, E.C., & Brunswik, E. (1935). The organism perspectives on the ecology of human-machine
and the causal texture of the environment. systems (pp. 157–188). Mawhah, N.J.:
Psychological Review, 42, 43–77. LEA.
Vicente, K. J. (1999). Cognitive work analysis. Zhang, J., & Norman, D. A. (1997). The nature of
Mahwah, N.J.: Erlbaum. external representations in problem solving.
Vicente, K. J. (2003). Beyond the lens model and Cognitive Science, 21(2), 179–217.
10 Background and Motivation

2 William M. Goldstein

Introduction to Brunswikian
Theory and Method

Egon Brunswik (1903–1955) was a perceptual psy- butions it has facilitated in the remaining chapters
chologist with deep interests in the history and phi- of this book.
losophy of science. His work on perception led him It might help to introduce Brunswik’s thinking
to develop a general vision for psychology called by outlining the kind of research problem that in-
probabilistic functionalism (Brunswik, 1952, 1956; spired it. Brunswik’s early research, conducted
Hammond, 1966; Hammond & Stewart, 2001). during the 1920s and 1930s when he was a mem-
Brunswik’s thinking had a systemic integrity with ber of the Vienna Psychological Institute, concen-
interwoven perspectives on the nature of psychol- trated on various perceptual constancies. Size
ogy, its definitive problems, and proper methodol- constancy, for example, is demonstrated by the
ogy. Unfortunately, some of Brunswik’s positions fact that the apparent (e.g., estimated) size of an
were out of step with the mainstream psychology of object tends to remain more or less constant at
his day, and his work was not well received at the various distances from an observer, even though the
time (Gigerenzer, 1987; Kurz & Tweney, 1997; object’s projection on the retina differs. Other fea-
Leary, 1987). Even today, although a number of tures, such as shape and color, similarly tend to be
Brunswik’s ideas are widely employed by psy- perceived as relatively constant despite changes in
chologists who may be unaware of their origins, the proximal information available to the observer.
Brunswikian thinking is represented mainly by a The issue of perceptual constancy, stated this way,
vocal minority of researchers who study judgment contains the seeds of some enduring concerns for
and decision making (Goldstein, 2004). However, Brunswik: (1) the relationship between the proxi-
Brunswik’s probabilistic functionalism has been mal information available to the observer and the
receiving increasing attention from researchers in distal state of affairs, emphasizing the ambiguous
cognitive engineering (see chapter 1), and one or even misleading implications of individual items
purpose of this book is to announce, explain, and of proximal information taken in isolation; (2) the
promote that trend. The purpose of this chapter establishment of a stable percept despite (or per-
is to acquaint readers with the basics of probabi- haps because of) variability and interrelationships
listic functionalism so that they can see its appeal among the items of proximal information; and (3)
for cognitive engineering and evaluate the contri- the accuracy of the stabilized percept. Briefly,

10
 Introduction to Brunswikian Theory and Method 11

Brunswik highlighted the importance of the ind- including aviation (piloting and pilot–automation
ividual’s adjustment to the world (i.e., not just in- interaction: Bass & Pritchett, chapter 9; Byrne, Kirlik,
ner perceptual processes), and he portrayed the & Fick, chapter 18; Casner, chapter 14; Degani,
individual’s understanding of the world as mediated Shafto, & Kirlik, chapter 13; Mosier & McCauley,
by proximal information that could be unreliable. chapter 12; Pritchett & Bisantz, chapter 7), military
Although Brunswik continued to emphasize per- applications (threat assessment, resource allocation,
ceptual examples, he extended his thinking to overt identification of unknown craft, and other aspects
action after a visit from Edward Chase Tolman re- of command and control: Bisantz, Kirlik, Walker,
sulted in collaboration (Tolman & Brunswik, Fisk, Gay & Phipps, chapter 3; Campbell, Van
1935). For overt action, Brunswik again empha- Buskirk , & Bolton, chapter 11; Gray, chapter 16;
sized the individual’s adjustment to the world (i.e., Horrey, Wickens, Strauss, Kirlik, & Stewart, chap-
the ability to bring about distal states of affairs that ter 5; Seong, Bisantz, & Gattie, chapter 8; Rothrock
match one’s internal desires) and the unreliable & Kirlik, chapter 10), dynamic system or process
mediation of that adjustment (i.e., by proximal ac- control (fault diagnosis, worker-tool-environment
tions that may or may not have the desired distal interaction: Jha & Bisantz, chapter 6; Kirlik, chap-
effects). ter 15), team performance (Adelman, Yeo, & Miller,
The highlighting of (1) adjustment to the world chapter 4), and human-computer interaction (Gray,
and (2) the mediated nature of that adjustment are chapter 16; Pirolli, chapter 17).
exactly the features of probabilistic functionalism In this chapter, I introduce Brunswik’s ideas in
that make it attractive to cognitive engineering. As a way that will illuminate their use in these many
stressed by Alex Kirlik (chapter 1), cognitive engi- applications. In the first section, I review some of
neering is a problem-centered field where the goal the core principles of Brunswik’s probabilistic func-
is to develop and use technology that will help tionalism, concentrating on: (1) functionalism, (2)
people accomplish tasks more effectively. It is not vicarious functioning, (3) probabilism, and (4) rep-
a field that can be content to study only the inner resentative design. In the second section, I outline
psychological processes underlying people’s inter- some theoretical extensions and lines of empirical
actions with technology; the ultimate goal is to research that have led to social judgment theory, a
improve people’s adjustment to the world. More- Brunswikian approach to the study of judgment. In
over, that adjustment is increasingly mediated by the third section, I draw connections between chap-
complex technology whose connection to the world ters in this book and aspects either of Brunswik’s
may be less than obvious to the user. Technology original theory or its subsequent extension in the
augments and/or replaces components of a percep- form of social judgment theory.
tion- or evaluation-like process when it is used (1)
to sense and collect information, (2) to display in-
formation, or (3) to suggest inferences. Technology Fundamentals of Brunswik’s
augments and/or replaces components of overt ac- Probabilistic Functionalism
tion when it is used to respond with great speed,
strength, or precision; across great distances; or Functionalism
with any type of performance that humans them-
selves cannot execute. Technology may also com- Adjustment to the World
bine the processes of perception/evaluation and The term functionalism refers to too many schools
overt action in various forms of automation (e.g., and movements in psychology (and other disci-
flying on autopilot). A source of theoretical and plines) for it to be meaningful without elaboration.
methodological ideas for studying such technologi- Brunswik was a functionalist in at least two senses.
cally mediated relationships with the world is surely First, he considered psychology to be “concerned
worth the consideration of cognitive engineers. with the interrelationships between organism and
This is where probabilistic functionalism and environment” (Brunswik, 1957, p. 5), with special
its extensions come in. The contributors to this emphasis on the interrelationships by which organ-
volume apply or extend Brunswik’s ideas to the isms manage to perform important tasks (i.e., to
technological mediation of perception/evaluation “function”) in their environments. This sense of
and overt action in many substantive domains, functionalism has Darwinian connotations and can
12 Background and Motivation

be linked to the functionalist movement of early mic and environmental systems. Brunswik explic-
twentieth-century American psychology (e.g., Angell, itly recognized that whatever degree of central–
1907). Second, Brunswik also had ties to the nine- distal correspondence the organism managed to
teenth-century European functionalism (also known achieve, it brought it about by the mediation of
as act psychology) of Franz Brentano (1874/1973). proximal events and processes. Moreover, these
Act psychology was known for its insistence that mediating processes drew a great deal of Bruns-
mental states have the quality of being “about” wik’s theoretical attention. However, as we will
something (e.g., about the outside world). Mental see shortly, Brunswik thought they should be re-
states are said to be “intentional,” in the technical garded as secondary in empirical investigations.
sense that philosophers give to this term. That is,
they point to or “intend toward” something outside Vicarious Functioning
themselves. Both senses of functionalism direct at-
tention to people’s adjustment to the world, and Stabilization of Central–Distal Relationships
both are reflected in Brunswik’s emphasis on cen- via Adaptable Mediation
tral–distal (cognition–world) relationships. “Vicarious functioning” is a phrase that Brunswik
borrowed from Hunter (1932) to refer to a critical
Central–Distal Correspondence property of mediating processes. Hunter was at-
Brunswik (1957, p. 5) considered the organism and tempting to distinguish the subject matter of psy-
environment to be distinct systems, each with its chology from that of physiology. He indicated that
own “surface and depth, or overt and covert re- whereas the physiological functions of one organ
gions,” and he argued that the “coming to terms” are rarely carried out by another organ, it is typical
of these “equal partners” most crucially involved the of the behavior studied by psychologists that if the
development of “rapport between the central, co- parts of the body normally used in a performance
vert layers of the two systems.” The central, covert are impaired, other parts of the body can function
layer of the environment contains the distal vari- “vicariously” to perform the behavior. Generalizing
ables the organism cares about, all of which are from Hunter’s examples, Brunswik used the phrase
objects, events, or properties that are remote from “vicarious functioning” to refer more broadly to
the organism in space and/or time. The organism’s exchangeability of means to an end. He emphasized
own central, covert variables include especially its that through the flexible adaptation of means to an
perceptions, understandings, and motivational states. end the observable hallmark of purposive behavior
(In what follows, the term “central” will be restricted is achieved: the “ ‘stabilization’ of the end stage” and
somewhat to refer to the covert layer of the organ- the “diversity of preceding stages” (Brunswik, 1952,
ism only.) Brunswik’s emphasis on the correspon- p. 17). That is, overt behavior mediates the relation-
dence between covert distal and central variables ship between organisms’ (central) desires and the
followed from his functionalist orientation, in both (distal) state of affairs, and vicarious functioning in
of the senses discussed earlier. Adequate adjust- their overt behavior enables organisms to stabilize
ment to the world requires accurate perception and this relationship, helping them reach their goals
effective actions, and these are matters of central– even if different means must be employed on dif-
distal correspondence: (1) bringing one’s (central) ferent occasions. (In the tradition of Gibsonian,
perceptions into line with (distal) objects and (2) rather than Brunswikian, ecological psychology,
bringing about (distal) states of affairs that coincide vicarious functioning is more commonly discussed
with one’s (central) desires. Brunswik referred to in terms of the “context-conditioned variability” of
central–distal correspondence as “achievement,” behavior, e.g., Turvey, Fitch, & Tuller, 1982.)
“attainment,” or “functional validity.” In today’s Brunswik extended this line of thought by ap-
parlance, this correspondence is perhaps most fre- plying a similar analysis to perception. In this case
quently discussed in terms of the “adaptive” charac- the “end stage” refers to the formation of a (central)
ter of behavior or cognition (e.g., Anderson, 1990). percept rather than the reaching of a goal, and the
It is important to note that by emphasizing the (mediating) “preceding stages” refer to the orga-
rapport between the central and distal layers, nism’s collection and use of sensory information
Brunswik was deemphasizing the proximal/periph- rather than its overt behavior. It is not obvious that
eral layer that is the interface between the organis- perception requires the same adaptable mediation
 Introduction to Brunswikian Theory and Method 13

Functional arc Vicarious mediation

as goal pursuit. However, Brunswik saw (1) that (probabilistic (family-hierarchy of
stabilization, cues, habits)
percepts must be inferred or constructed from achievement) Feedback

proximal sensory “cues” to the distal object’s prop-

erties and (2) that proximal cues must be selected
and used differently to perceive objects under dif- Process Process
detail detail
ferent environmental conditions. For example, con- Initial Terminal
focal focal
cerning proximal cues to the distance between variable variable
oneself and an object (e.g., binocular parallax, con- Stray
causes
Stray
effects
vergence of the eye axes, etc.), Brunswik wrote:
Figure 2.1. The lens model: Composite picture of
[N]one of these proximal variables can be the functional unit of behavior. Reprinted by
considered to be the distance cue in the sense permission of University of Chicago Press from
of an effect which would be present without Brunswik, 1952, p. 20.
exception whenever the distal condition
should obtain. Some of the cues will more
initial focal variable, and the scatter of rays repre-
often, others less often, be present, depending
sents a set of proximal means among which the
on circumstances, and occasionally all of them
organism selects, all of which lead to the same dis-
may be cut off (so that the fact of a certain
tal end shown as the terminal focal variable. In dif-
distance relationship must remain unrecog-
ferent contexts, organisms select different subsets
nized by the organism in question). (Brunswik,
of cues or means to achieve similar ends, and the
1943, p. 256, emphasis in original)
composite picture shows the collection of cues or
Thus, adaptable mediation again enables organisms means that are used over a large number of contexts.
to stabilize the relationship between their (central)
percepts and (distal) objects, thereby producing On Research in Psychology
the perceptual constancies that initially attracted When Brunswik referred to particular variables as
Brunswik’s attention. Brunswik referred to multi- focal, in Figure 2.1, he was not merely alluding to
plicity and flexibility in the use of both cues and the picture’s resemblance to a convex lens. Rather,
means to an end as vicarious functioning, and he he was distinguishing between variables that can
took vicarious functioning (broadly construed to and cannot be brought into stable relationships with
include multiplicity, flexibility, intersubstitutability, each other. Relationships between focal variables
and combination of both cues and means) to be are stabilized by the adaptable mediation of vicari-
the core feature that makes behavior interesting to ous functioning. The fact that it is central and dis-
psychologists. tal variables that turn out to be focal, on Brunswik’s
analysis, reinforced his emphasis on functional ac-
The Lens Model hievement (i.e., central–distal correspondence) as
Brunswik’s lens model derives its name from the the proper subject matter of psychology. As argued
fact that a pictorial representation of vicarious func- earlier, functionalism identifies central–distal rela-
tioning resembles a diagram of a convex lens. Fig- tionships as the ones of primary importance to the
ure 2.1 shows the lens model as it appeared in organism. In addition, the lens model indicates that
Brunswik’s (1952, p. 20) book, where he referred these are the relationships that will be stable enough
to it as a “composite picture of the functional unit to bear scientific investigation.
of behavior.” The figure is shown in a general form By contrast with his support for studying
that is applicable to both perception and overt be- central–distal relationships, Brunswik was wary of
havior. Interpreted in terms of perception, the im- research on mediating processes, especially the me-
age is one in which a distal stimulus at one focus diating internal processes that occupy the majority
(i.e., the “initial focal variable”) emits a scatter of of attention of today’s cognitive scientists. Vicarious
rays, representing proximal cues, from among functioning is a stabilizing process (i.e., it produces
which the organism selects a subset to be recom- stable relationships between central and distal vari-
bined into the central perception at the other focus ables) but is itself unstable (i.e., it exhibits flex-
(i.e., the “terminal focal variable”). For overt behav- ibility and variability), producing “relative chaos
ior, a central motivational state is pictured as the in the regions intervening between focal variables”
14 Background and Motivation

(Brunswik, 1952, p. 20). Therefore, Brunswik terms of perception (with a comparable statement
thought mediating processes to be a difficult and applying to action), this means the following. First,
unpromising area for research. In his early writings, the environment must be studied to determine what
he suggested that a classification of the kinds of Brunswik called the “ecological texture.” That is,
distal objects “attained” by an organism would pro- one must (1) identify the features that correlate with
vide a description of its abilities and performance, the distal object (and are therefore candidates for
and he favored the pursuit of such inventories over being used by the organism as proximal/peripheral
the investigation of mediating processes. Brunswik’s cues), (2) ascertain the strengths of the relationships
later writings, on the other hand, did acknowledge between the distal object and the (potential) proxi-
the study of mediating processes as important, but mal/peripheral cues (i.e., the “ecological validities”
his enthusiasm for the subject remained restrained, of the cues), and (3) assess the interrelationships
and he urged psychologists to keep research prob- among the cues (for this is part of what permits the
lems in perspective. Specifically, he warned against cues to function vicariously for one another). Sec-
replacing the functionalist study of organisms’ ond, one must study the organismic system to de-
adjustment to the world with a focus on quasi- termine which of the (potential) proximal/peripheral
solipsistic “encapsulated” organisms in isolation. In- cues are actually used and in what strengths. Finally,
deed, Brunswik implied that such a trend had already the environmental and organismic systems must be
gone too far when he wrote, “Psychology has forgot- compared to see if the organism’s utilization of cues
ten that it is a science of organism-environment is appropriate, where “[w]e may call it appropriate
relationships, and has become a science of the or- in a generalized sense if the strength of utilization of
ganism” (Brunswik, 1957, p. 6). In this comment a cue is in line with the degree of its ecological va-
he presaged the critique of cognitive science to be lidity” (Brunswik, 1957, p. 11).
leveled some 35 years later by Donald Norman In sum, Brunswik’s theoretical ideas about func-
(1993), who caricatured its nearly sole focus on tionalism and vicarious processing lead to a general
internal processes as resulting in a theory of cogni- outline for investigations of human–environment
tion as “disembodied” and cut off from the world relations, an outline that continues to apply when
in which it is embedded. these relations involve mediation by technologies of
Consequently, Brunswik addressed the study various sorts. This is part of Brunswik’s appeal for
of mediating processes with caution and circum- cognitive engineers, and further details that elabo-
spection. In particular, he distinguished between rate this outline will be given in what follows. First,
studies of “macromediation” and “micromediation.” however, some of Brunswik’s other theoretical ideas
The former examine “the gross characteristics or that also have implications for methodological prac-
macrostructure of the pattern of proximal and pe- tice will be discussed in the next two subsections.
ripheral mediation between the distal and central
foci,” and thereby address “the problem of the grand Probabilism
strategy of mediation” (Brunswik, 1957, p. 8). The
latter “attempt to break down the cognitive process Although in principle the relationship between
further into its component parts” and thus concern proximal/peripheral and distal variables must be
“mediational tactics” (p. 9). As he (p. 9) put it, determined in studies of ecological texture, Bruns-
“Achievement and its strategy are molar problems; wik was convinced that these relationships would
tactics is a molecular problem,” and clearly he felt virtually always prove to be somewhat unreliable,
the former problems should precede and inform the ambiguous, or “equivocal.” At least, this was the
latter. argument for individual cues and means. The ex-
The lens model helps clarify what is involved in tent to which an organism could combine cues and
elucidating the (macromediational) “grand strategy” select means to overcome their limitations when
by which an organism achieves a stable central–distal taken singly is, of course, the problem of achieve-
relationship. Specifically, beyond assessing the de- ment—the main subject of research in Brunswik’s
gree of achievement that the organism actually functionalist view.
reaches, the organismic (for the purposes of this The equivocality of means may be more readily
book, human) and environmental systems must be apparent than that of cues. A proverb reminds us
analyzed separately and compared. Expressed in that the best-laid plans of mice and men oft go
 Introduction to Brunswikian Theory and Method 15

awry, and it is easy to see that failure is possible even make a posit, or wager” (Brunswik, 1943, p. 259),
in simple cases. Regarding perception, Brunswik gave and wagers are occasionally lost (e.g., see Byrne,
numerous examples of proximal cues that were Kirlik & Fick, chapter 18; Rothrock & Kirlik, chap-
ambiguous, and he suggested that this would be ter 10). Brunswik (1943, p. 270) concluded that
typical. A trapezoidally shaped retinal image may “there can be no truly molar psychology dealing
be due to a rectangular object seen at an angle, or with the physical relationships of the organism with
the object may actually be trapezoidal. Of two ob- its environment unless it gives up the nomothetic
jects, the one with the larger retinal projection may [universal, lawlike] ideal in favor of a thoroughly
be the closer, or its larger image may be due to its statistical conception.”
great physical size. These examples are telling be- An implication for cognitive engineers is that
cause they indicate a pattern: In both cases the cue they should expect technological aids to produce
is ambiguous not because of an inherently proba- no more than a probabilistic relationship between
bilistic environment but because of incompleteness a person and the environment (also see Kirlik, 2005;
of the information. Without challenging determin- Vicente, 1999). Even when technology is designed
ism, Brunswik argued that organisms usually have specifically to reduce the uncertainty in this rela-
access to sets of proximal cues that are incomplete tionship (e.g., by responding with greater precision
or otherwise inadequate for applying the laws that than people are capable of), a person’s central goals
govern distal objects and distal–proximal relation- or judgments might fail to correspond with the dis-
ships. He wrote, “The universal lawfulness of the tal environment because of conditions the technol-
world is of limited comfort to the perceiver or ogy was not designed to handle. Although these
behaver not in a position to apply these laws. . . ,” situations don’t reflect a failure of the technologi-
and that “ordinarily organisms must behave as if in cal aid, they nonetheless render the central–distal
a semierratic ecology” (Brunswik, 1955, p. 209). relationship uncertain.
That is, the environment must appear to organisms
to be probabilistic, even if in a philosophical sense Representative Design
it is deterministic. To emphasize the organism’s pre-
dicament, Brunswik coined the metaphor that the The last aspect of Brunswikian theory to be
perceptual system must act as an “intuitive statisti- discussed here was and remains one of the most
cian” (e.g., Brunswik, 1956, p. 80). controversial. By the 1950s, it was commonly ap-
From these observations, Brunswik drew an preciated that participants in psychological inves-
important implication for both theory and method. tigations must be sampled at random from a
Specifically, certain parts of the project to explicate specified population for the findings to generalize
achievement and its strategy must be expressed in beyond the participants themselves. Brunswik ar-
probabilistic terms. First, as just argued, the envi- gued similarly that for findings to generalize beyond
ronment or that part of it accessible to an organ- the particular stimuli and conditions employed in
ism must appear to it to be probabilistic. Therefore, a study, they also must be sampled to be represen-
studies of ecological texture, which aim to describe tative of the ecology. Such experiments had what
the relationships between distal variables and the Brunswik called a “representative design,” in con-
accessible proximal/peripheral cues and means, trast to the more typical “systematic design” where
must describe a set of probabilistic relationships. experimenters manipulate the stimuli and condi-
For example, although the relationship between tions to produce orthogonal independent variables
size of retinal projection and size of distal object is that are convenient to analyze yet often uncharac-
deterministic for objects at a fixed distance from the teristic of the ecology.
organism, the relationship is probabilistic when it Recall that Brunswik’s project to study achieve-
must be taken over objects at the various distances ment and its strategy decomposed into four parts:
the organism encounters in its environment. Sec- (1) the study of achievement itself, (2) the study of
ond, in addition to ecological texture, achievement ecological texture, (3) the study of the organism’s
itself must be described in probabilistic terms. The utilization of cues, and (4) a comparison of the eco-
reason is that, faced with ambiguous, apparently logical and organismic systems. It is relatively easy
probabilistic cues, “[a]ll a finite, sub-divine indi- to see the importance of representative design for
vidual can do when acting [or perceiving] is . . . to the first two of these four parts. Clearly achievement
16 Background and Motivation

can be made excellent or poor by the use of stimuli one might reply that sensitivity to task environment
that in the appropriate sense are easy or hard. Little is an empirical question. The frequency with which
is revealed about achievement in a target environ- cognitive engineers have heard experienced per-
ment unless the stimuli and conditions are represen- formers (e.g., pilots) complain about even minor
tative of those that occur in that target environment. issues associated with the lack of complete repre-
Even more clearly, the ecological texture of relation- sentativeness of a research environment (e.g., a
ships among distal and proximal/peripheral variables flight simulator) should bring this point home.
cannot be studied adequately in a set of stimuli that As indicated, the need for representative design
distorts the relationships to be studied. The real con- remains a controversial issue, turning in large part
troversy surrounding the use of representative de- on empirical questions about people’s sensitivity to
sign centers on its application to studies of cue task environments (see Goldstein, 2004, for further
utilization, and only a sketch of the issues can be comments). At the very least, an implication for
given here, stated in terms of the sampling of stimuli. cognitive engineers is that they should be wary of
One argument for representative design, even in assessing technological aids in test environments
studies of cue utilization, holds that the logic of sta- that depart from the settings in which they are in-
tistical inference requires the probability sampling of tended to be used. Such a sentiment is reflected in
units from populations. Generalizing beyond the attempts to study or simulate field conditions rather
specific stimuli employed in the study requires this than use abstract laboratory tasks (as demonstrated
no less than generalizing beyond the specific par- in nearly all of the chapters of this volume).
ticipants. Some might find this argument uncon-
vincing, either because there are other, nonstatistical
bases for generalization (e.g., theory-guided infer- Development of Social
ence) or because they are not convinced that or- Judgment Theory
ganisms are sufficiently sensitive to their task
environments for the orthogonalization of stimulus As mentioned earlier, Brunswikian research in
properties to affect their cue utilization. One might psychology has been pursued most vigorously by
reply, however, that sensitivity to the task environ- a group of researchers who study judgment and
ment is itself a matter for empirical investigation. decision making (Goldstein, 2004). The point of
A second argument for representative design departure for these researchers was a paper by
meets a similar objection, and one might give a simi- Hammond (1955), in which he drew an analogy
lar reply. Specifically, the heart of cue utilization— between Brunswik’s analysis of perception and the
and from a Brunswikian perspective the main problem of clinical judgment. Specifically, just as
interest in a study of cue utilization—is vicarious people must infer or construct a percept from a
functioning. Therefore, organisms should be given collection of sensory cues that provide only incom-
full scope to display the vicarious functioning they plete and fallible information, Hammond argued that
normally employ in their environment. Moreover, a patient’s behaviors, expressions, and test scores
vicarious functioning includes the flexible selection provide ambiguous cues to the patient’s personality
of cues, that is, allowing some cues on occasion to and diagnosis. In both cases the clinician/perceiver
substitute for other cues, on the basis of prior learn- must use multiple cues and indicators to infer some-
ing of cue intercorrelations in the environment. thing that goes beyond the cues themselves.
Deliberately destroying these cue intercorrelations Subsequent work by Hammond and colleagues
in an orthogonal design could easily produce “inter- produced a voluminous literature of empirical find-
combinations of variates [that] may be incompatible ings, as well as theoretical and methodological ex-
in nature or otherwise grossly unrealistic” (Brunswik, tensions of Brunswik’s approach. The work evolved
1955, p. 205). Encountering such strange stimuli into an approach called social judgment theory
might confuse the participants, prevent them from (Brehmer, 1988; Brehmer & Joyce, 1988; Doherty,
taking the task seriously, or at least make them wary 1996; Hammond et al., 1975, 1977; see also Ham-
of relying on the intersubstitutability of cues that mond, 1996, 2000; Hammond & Joyce, 1975; Ham-
would normally characterize their vicarious function- mond & Stewart, 2001; Hammond & Wascoe, 1980;
ing. As before, a critic might suggest that participants Juslin & Montgomery, 1999; Rappoport & Sum-
are not so sensitive to their task environments, and mers, 1973). In this section, I selectively review
 Introduction to Brunswikian Theory and Method 17

some of the developments that are of particular Tucker, 1964). The lens model equation (LME) is
relevance to cognitive engineers and/or connect to a formula for decomposing the achievement coef-
the contributions in this book. ficient, that is, the correlation between criterion and
judgment, for a given set of proximal cues:
The Lens Model Equation
ra = GReRs + C√(1 – R2e)√(1 – R2s),
Figure 2.2 shows the lens model as it was adapted
where (see Figure 2.2):
(Hammond, 1955; Hammond et al. 1975) for the
study of judgment. Modeled on the case of percep- ra is the achievement coefficient, that is, the
tion, rather than action, the to-be-judged criterion correlation between the criterion variable Ye
variable Ye plays the role of the distal stimulus at the and the judgment variable Ys;
initial focus, whereas the person’s judgments Ys play Re is the multiple correlation of the criterion
the role of central perceptions at the terminal focus. variable with the proximal cues;
The proximal cues X1, X2, . . . , Xn constitute the in- Rs is the multiple correlation of the judgments
formation available to the judge and are related to the with the proximal cues;
criterion by ecological validities (i.e., cue–criterion G is the correlation between the linear
correlations) and to the judgments by cue utilization components of the criterion and judgment
coefficients (i.e., cue–judgment correlations). Judg- variables, that is, the correlation between the
mental accuracy or achievement is assessed by the values Y'e that are predicted by linear
correlation between judgment and criterion values. regression of the criterion variable on the
At the time of Brunswik’s death, the environ- proximal cues and the values Y's that are
mental and organismic systems could be modeled predicted by linear regression of the
separately (e.g., by linear regression), but only a judgments on the proximal cues; and
qualitative and impressionistic comparison of the C is the correlation between the nonlinear
two systems could be offered. That situation was components of the criterion and judgment
changed by the development of the lens model variables, that is, the correlation between the
equation (Hursch, Hammond, & Hursch, 1964; residuals Ye – Y'e and the residuals Ys – Y's.

Achievement (ra)

X1

re,1 rs,1
X2
re,2 rs,2
ri,j
Criterion Ye Ys Judgment
re,3 rs,3
X3

.
re,n . rs,n
.
Ecological Cue Utilization
Validities (re,i) Xn
Coefficients (rs,i)

Y’e Proximal Cues (Xi) Y’s

Figure 2.2. Brunswik’s lens model as modified for social judgment theory, shown together with compo-
nents of the lens model equation. See text for details.
18 Background and Motivation

The LME has become a standard tool of Bruns- Phipps, where it is found to explain individual dif-
wikian research because it allows the investigator ferences in performance in a combat information
to decompose performance in a meaningful way center environment, and in chapter 4 by Adelman,
and to see which component(s) of performance are Yeo, and Miller, where it is used to explain group
affected by various manipulations (e.g., different achievement with different computer displays un-
kinds of feedback about one’s task performance, der time pressure. A number of chapters (to be dis-
introduction of a technological aid). The indices Re cussed shortly in a somewhat different context)
and Rs indicate the linear predictability of the two suggest alternatives to the LME’s implicit linear and
systems, environmental (criterion) and organismic static modeling of both the environment and the
(judgment), respectively, based on a particular set organism. Still others consider various mathemati-
of proximal cues. The indices G and C indicate the cal extensions or embeddings of the LME in other
extent to which the judge has succeeded in match- decompositions of performance (chapter 5; chap-
ing his or her own systematic use of the proximal ter 6; also see Stewart, 2001).
cues (linear/additive and nonlinear/configural, re-
spectively) with the criterion variable’s systematic Individual Learning
dependence (linear/additive and nonlinear/con-
figural) on the proximal cues. In a sense, then, G Brunswik portrayed organisms as having to adapt
and C assess the judge’s knowledge of the task. to environments whose distal properties were in-
It is relatively common to find that neither the dicated only probabilistically by proximal cues. This
criterion variable nor the judge has any consistent raises the question of how organisms learn proba-
nonlinear or configural dependence on the proximal bilistic relationships, and Brunswikian researchers
cues, and this situation simplifies both the LME it- have produced a large body of literature on the
self and the interpretations of the indices. In this case, subject using a paradigm called multiple-cue prob-
the LME reduces to the following approximation: ability learning (MCPL). On each trial of an MCPL
study, the respondent examines a profile of cue
ra ≈GReRs
values and predicts the value of a criterion variable.
because in the absence of systematic nonlinear vari- In a study using outcome feedback, the respondent
ance the expected value of C equals zero. The indi- is then told the correct criterion value and must
ces Re and Rs, then, indicate the total predictability learn the cue–criterion relationships over many tri-
of the environmental and organismic systems, re- als (for a review, see Klayman, 1988). In general,
spectively (i.e., linear/additive predictability is all learning from outcome feedback tends to be sur-
there is, apart from sampling fluctuations). The prisingly slow and limited. Todd and Hammond
index Rs indicates the consistency with which the (1965) introduced the notion of cognitive feedback,
judge executes the (only) systematic component of where respondents are shown statistical properties
his or her judgment, and therefore the index is over blocks of trials rather than the outcomes of
called cognitive control (Hammond & Summers, single trials. The result is that learning is generally
1972). The index G indicates the extent to which faster and better than with outcome feedback. (See
the (only) systematic component of the judge’s Balzer, Doherty, & O’Connor, 1989, for a decom-
performance is related to the (only) systematic com- position of cognitive feedback into components
ponent of the task environment and therefore is that are differentially responsible for the improved
called knowledge. The LME reveals the fact that performance.)
even when knowledge is perfect (G = 1), achieve- A number of the chapters in this book touch
ment is limited by the consistency with which this on issues of feedback and training, sometimes in
knowledge is executed (Rs) and the consistency of the context of other concerns. For example, Seong,
the task environment (Re). Bisantz, and Gattie (chapter 8) discuss the use of
Many of the chapters in this volume either cognitive feedback, although their main interest has
employ the LME as a tool for analyzing the impact to do with human trust of automated systems, and
of technological aids or address the LME itself by Bass and Pritchett (chapter 9) provide a feedback-
proposing modifications and extensions. For ex- based technique for improving the correspon-
ample, cognitive control figures importantly in dence between human and automated judgments.
chapter 3 by Bisantz, Kirlik, Walker, Fisk, Gay, and Chapters by Rothrock and Kirlik (chapter 10) and
 Introduction to Brunswikian Theory and Method 19

Campbell, Van Buskirk, and Bolton (chapter 11) examined in investigations of interpersonal learn-
also address issues of training, although they are ing and conflict. Hammond et al. (1975) review the
mainly concerned with using nonlinear models to effects of task consistency, ecological validity of
represent human performance. cues, cue-criterion function forms, and cue inter-
correlations. (For other reviews and discussions,
Interpersonal Conflict and including extensions to negotiations and small-
Interpersonal Learning group processes, see Adelman, Henderson, & Miller,
2001; Brehmer, 1976; Cooksey, 1996; Mumpower,
The LME was described earlier as applying to the 2001; Rohrbaugh, 2001.)
correlation between judgments and criterion values. The relevance of these studies and these sorts
However, the LME applies to the correlation between of analyses for cognitive engineering is that the vari-
any two variables that are each in turn related to the ous judges in an N-system lens model may be pro-
same set of cues. Therefore, one can use it to decom- vided with different kinds of displays of the cues,
pose the correlation between two people’s judg- so that (dis)agreement between judges reflects the
ments. Indeed, one can decompose the correlation impact of display conditions. Pritchett and Bisantz
between any pair of variables in a group of variables. (chapter 7) present just such an analysis of judges
In such a case, the model is referred to as the N-sys- employing different cockpit display designs in a
tem lens model. If one of the N systems is a criterion simulated conflict detection task. They also apply
variable and there are N – 1 judges, then N – 1 of the N-system lens model to assess the agreement
the correlations will refer to judges’ achievement and between human judges and automated systems that
the remaining (N – 1)(N – 2)/2 correlations will re- are designed to produce suggested judgments. Com-
fer to agreement between pairs of judges. Changes parable issues are examined by Horrey, Wickens,
in the LME components of each correlation, over Strauss, Kirlik, and Stewart (chapter 5) in a simu-
blocks of trials, indicate the results of the interactions lation of military threat assessments and resource
taking place among the people in addition to any allocation decisions. These investigators evaluate
other feedback that might be available. people’s performance with and without display
Social judgment theory has generated a large conditions that highlight the most threatening en-
number of experiments that examine the resolution emy units and with and without the use of an au-
of conflicts between people who must reach con- tomated display enhancement that suggests an
sensus judgments despite their (initial) differences appropriate allocation of resources. The contri-
in the use of proximal cues (Hammond, 1965, butions by Seong, Bisantz, and Gattie (chapter 8)
1973). In addition, social judgment theory has ad- and by Bass and Pritchett (chapter 9) further exam-
dressed two different kinds of interpersonal learn- ine the interactions between humans and automated
ing. First, there is the case in which one person systems for suggesting judgments and decisions, the
learns from another about the skillful performance latter chapter explicitly drawing on social judgment
of the task (Earle, 1973). A second kind of inter- theory research on interpersonal learning.
personal learning can be studied in a procedure that
adds a phase to the experiment, after the interac-
tion between people, where the participants are Further Applications and Extensions
separated and each is asked to judge new cases as of Brunswikian Principles
well as to predict the other participant’s judgments
of these cases. Using this procedure (Hammond, The areas of research reviewed in the previous sec-
Wilkins, & Todd, 1966), one can study what par- tion are the classic ones that contributed directly
ticipants learn about each other (i.e., what they learn to the development of social judgment theory.
about the judgmental systems of their partners) as However, this brief review cannot begin to convey
well as what they learn from each other (i.e., how the range of theoretical issues and domains of ap-
their own judgmental systems are changed due to plication where Brunswikian-inspired research on
exposure to their partners). (It is the social nature judgment has been fruitful. In this section, I touch
of these lines of research that gave social judgment on some of the applications and extensions of so-
theory its name.) Many of the same factors that have cial judgment theory to which the chapters in this
been manipulated in MCPL studies have also been volume make contributions.
20 Background and Motivation

Dynamic Tasks and Time Stress to describe (1) criterion–cue relationships and (2)
cue–judgment relationships. Studies that focus on the
Because many tasks involve time as a key element latter type of model are often referred to in social judg-
in one form or another, a number of researchers ment theory as efforts in “policy capturing” (referring
have been interested in tasks that go beyond that to the judge’s “policy” for using the cues; see Brehmer
of static judgment or the one-shot choice of an ac- & Brehmer, 1988). Moreover, in keeping with Bruns-
tion. For example, information may be accrued over wik’s advice to study macromediation before micro-
time, judgments or actions may be required under mediation (see previous discussion), researchers
time pressure, actions may need to be adjusted as frequently fit models to people’s judgments without
their results unfold over time, and so on. Research- much concern for the model’s fidelity to the “process
ers working within the context of individual learn- detail” (using Brunswik’s expression as it is repro-
ing (see previous discussion) have studied dynamic duced in Figure 2.1) by which people are “actually”
environments in which the validity of cues can using the cues. Models that describe criterion–cue re-
change (Dudycha, Dumoff, & Dudycha, 1973; Peter- lationships while remaining uncommitted about the
son, Hammond, & Summers, 1965; Summers, 1969) detailed time course of the underlying cognitive pro-
or where the value of the criterion variable is affected cesses are often called paramorphic (Hoffman, 1960;
by the respondent’s predictions or actions (Camerer, see Doherty & Brehmer, 1997, for a different view of
1981; Kleinmuntz, 1985; Kleinmuntz & Thomas, paramorphic models).
1987; Mackinnon & Wearing, 1985; for research on Some of the contributions to this volume in-
learning in even more complex dynamic environ- volve efforts in this spirit. That is, models are fit to
ments, see Brehmer, 1992, 1995, 1996). capture the structure and predictability of people’s
Several of the contributions to this volume behavior, but the authors refrain from making
share this concern with dynamic aspects of task strong claims that the models describe underlying
environments and performance (chapter 4; chap- processes. All of the chapters presenting applica-
ter 18; chapter 16; chapter 6; chapter 10). The chap- tions of the linear-additive formulation of the lens
ter by Bisantz, Kirlik, Walker, Fisk, Gay and Phipps model are of this type. In addition, the canonical
(chapter 3) in particular extends the lens model correlation analysis by Degani, Shafto, and Kirlik
approach by developing time-stamped, idiographic (chapter 13) and Casner’s analysis of adaptation
(individualized) models of the task environment, (chapter 14) are concerned with determining the
which is necessary for examining judgment in dy- task demands and environmental conditions to
namic tasks. The chapter by Kirlik (chapter 15) which aircraft pilots are sensitive in deciding where
provides a technique for formalizing both percep- to set the aircraft’s mode of control on a continuum
tion and action cue-criterion relationships in a uni- between fully manual and fully automatic. In both
fied environmental model, which is necessary for cases, the goal of describing the internal details of
studying interactive tasks where the performer has the pilot’s decision process is deferred.
some control over available stimulation. These con- In other chapters, the authors depart more radi-
tributions complement the work described next, in cally from the use of linear models and techniques
which researchers extend the other (internal) “half” (see Dawes, 1979; Dawes & Corrigan, 1974) to
of the lens model by proposing nonlinear models describe people’s judgment policies. Instead, the
of the organism. authors employ models that are thought by some
to be more consistent with cognitive theory and
Macromediation: Gross Characteristics with people’s verbal reports of their processes. In
of the Pattern of Proximal and particular, Rothrock and Kirlik (chapter 10) present
Peripheral Mediation between a method for inferring rule-based heuristics from
Distal and Central Foci people’s judgments. These simplified, noncompen-
satory strategies are more plausible descriptions of
As discussed earlier, Brunswik’s project to illuminate people’s judgment processes, especially when people
achievement and its “grand strategy” requires sepa- are time-stressed or otherwise facing cognitive
rate models of the environment and the organism and limitations. Similarly, Campbell, Van Buskirk, and
a comparison of those models. Phrased in terms of Bolton (chapter 11) investigate the use of fuzzy logic
judgment, that means that separate models are needed for modeling people’s judgments and decisions.
 Introduction to Brunswikian Theory and Method 21

Micromediation: Breaking Down page) to predict the utility of distal pages, and this
the Cognitive Process into predictive activity is modeled with spreading acti-
Its Component Parts vation in a network of nodes that represent the
person’s understanding of the search domain.
Still other chapters are more explicit in attempting Degani, Shafto, and Kirlik (chapter 13) do not
to flesh out the process detail of people’s adapta- so much attempt to contribute to theory of cogni-
tion to the environment. Recall that it was late in tive processes but instead to draw on recent cogni-
Brunswik’s career before he explicitly incorporated tive theory to address fundamental questions in
the study of inner psychological processes into his Brunswik’s probabilistic functionalism. In particu-
framework (but see Goldstein & Wright, 2001). As lar, they ask the question “What makes vicarious
noted earlier, Brunswik was restrained about stud- functioning work?” They are motivated to ask this
ies of cognitive processes for two reasons. He question by their observation that many largely
thought that researchers might tend to study cog- digital, technological interfaces seem to defeat vi-
nitive processes apart from the task environments carious functioning and thus adaptive behavior
in which they occurred, abandoning the study of itself. Degani and colleagues conclude that the dis-
functional adaptation. Also, he thought that the crete, technological ecology often does not support
flexibility of vicarious functioning might produce the approximation and convergence (learning) op-
“relative chaos in the regions intervening between erations supported by continuous ecologies. These
focal variables” (Brunswik, 1952, p. 20), making it learning operations are claimed to be crucial to the
difficult and unrewarding to study cognitive pro- development of robust and adaptive “intuitive”
cesses in detail. Although Brunswik’s concerns on (Hammond, 2000) or “System I” (Kahneman, 2003)
the first point may have been valid, the cognitive cognition.
revolution in psychology has shown the study of
cognitive processes to be eminently feasible and Revisiting the Fundamentals
amply rewarding.
Some of the chapters in this volume have taken A number of researchers have urged that some of
up the challenge of integrating detailed models of the principles of Brunswik’s program be revisited
cognitive processes into Brunswik’s framework. and possibly reinterpreted. Two of these lines of
Whereas organisms’ sensitivity to their task envi- thought will be mentioned here, concerning (1) the
ronments was offered earlier as an argument for appropriate meaning (and measure) of functional
representative design and a matter for empirical adaptation and (2) the adequacy of studying per-
study, Gray (chapter 16) uses a current cognitive ception/judgment and overt action separately.
theory (ACT-R) and related computational tech- Regarding adaptation, Gigerenzer and his col-
niques to consider how even minor changes in task leagues (Gigerenzer & Kurz, 2001; Gigerenzer, Todd,
environments can affect cognitive processes and & the ABC Research Group, 1999) note that Bruns-
influence overt performance. He goes even further wik’s functionalist perspective led him to empha-
to sketch out his vision more generally of a “rap- size the adaptiveness of an organism’s behavior in
prochement” between cognitive and ecological specific environments, but that accuracy of judg-
analyses. In a study of airline pilots’ performance ment is only one facet of its adaptiveness. Speed and
in landing and taxiing to gate under foggy condi- the ability to make use of limited information are
tions, Byrne, Kirlik, and Fick (chapter 18) combine also important features. Similar reasoning has
detailed cognitive models with ecological analyses. prompted Rothrock and Kirlik (chapter 10) and
In particular, they model pilots’ taxi route decisions Byrne, Kirlik, and Fick (chapter 18) to suggest that
as heavily involving the use of two simple heuris- researchers reconsider the task criterion that is ana-
tics (see also chapter 10), and they find that the lyzed, especially in studies of time-stressed tasks.
model is predictive of decision errors in ecologically Both chapters attend to the observation that ap-
atypical situations that short-circuit these heu- proximate yet fast (rather than accurate yet slow)
ristics’ efficacy. Pirolli (chapter 17) investigates judgment and decision strategies may be favored in
people’s navigation among pages on the Web. Part dynamic tasks. Mosier and McCauley (chapter 12)
of his approach involves portraying people as us- elaborate on a distinction emphasized by Ham-
ing proximal cues (provided by the current Web mond (1996) between a focus on correspondence
22 Background and Motivation

with the environment (e.g., accuracy of judgment) Acknowledgments. Parts of this chapter have been
versus coherence of judgments and/or actions (e.g., adapted from Goldstein (2004).
logical consistency). Mosier and McCauley draw
attention to the fact that technology has trans-
formed a number of tasks so that human operators
References
must assess the coherence of information and, if
necessary, take steps to reestablish coherence. The Adelman, L., Henderson, D., & Miller, S. (2001).
cognitive processes underlying these tasks and the Vicarious functioning in teams. In K. R.
standards of adaptiveness for evaluating them must Hammond & T. R. Stewart (Eds.), The essential
be considered carefully. Brunswik: Beginnings, explications, applications
Finally, Kirlik (chapter 15) reconsiders the (pp. 416–423). Oxford: Oxford University Press.
separate application of Brunswik’s lens model to Anderson, J. R. (1990). The adaptive character of
matters of perception or judgment versus matters thought. Hillsdale, N.J.: Lawrence Erlbaum
of overt action. Although Brunswik recognized and Associates.
discussed the applicability of his approach to both Angell, J. R. (1907). The province of functional
psychology. Psychological Review, 14, 61–91.
sets of issues (compare Tolman & Brunswik, 1935),
Balzer, W. K., Doherty, M. E., & O’Connor, R., Jr.
his own substantive interest concerned problems of
(1989). Effects of cognitive feedback on
perception, and most of his followers have been performance. Psychological Bulletin, 106, 410–433.
interested in judgment. The implicit assumption Brehmer, A., & Brehmer, B. (1988). What have we
is that perception/judgment and action can be learned about human judgment from thirty years
studied independently without serious loss. Kirlik of policy capturing? In B. Brehmer & C. R. B.
demonstrates that the simultaneous analysis of Joyce (Eds.), Human judgment: The SJT view
perception/judgment and action in a specific task (pp. 75–114). Amsterdam: Elsevier Science
environment constrains theories of the underlying Publishers B.V. (North-Holland).
cognitive processes much more adequately than Brehmer, B. (1976). Social judgment theory and the
modeling them in isolation. Casner (chapter 14), analysis of interpersonal conflict. Psychological
Bulletin, 83, 985–1003.
Gray (chapter 16), Kirlik, Byrne, and Fick (chapter
Brehmer, B. (1988). The development of social
18), and Pirolli (chapter 17) also demonstrate ben-
judgment theory. In B. Brehmer & C. R. B. Joyce
efits from an integrated analysis of perception/judg- (Eds.), Human judgment: The SJT view (pp. 13–
ment and overt action. 40). Amsterdam: Elsevier Science Publishers B.V.
(North-Holland).
Brehmer, B. (1992). Dynamic decision making:
Human control of complex systems. Acta
Conclusion
Psychologica, 81, 211–241.
Brehmer, B. (1995). Feedback delays in complex
Brunswik’s probabilistic functionalism emphasizes dynamic decision tasks. In P. A. Frensch &
(1) adjustment to the world and (2) the mediated J. Funke (Eds.), Complex problem solving: The
nature of that adjustment. It is becoming increas- European perspective. Hillsdale, N.J.: Erlbaum.
ingly clear that cognitive engineering must do so as Brehmer, B. (1996). Man as a stabiliser of systems:
well. Brunswik’s approach leads to important theo- From static snapshots of judgement processes to
retical and methodological implications, and his dynamic decision making. In M. E. Doherty (Ed.),
original formulation has been amplified and ex- Social judgement theory [Special issue of Thinking
tended by social judgment theory and other lines and Reasoning, 2(2/3), 225–238]. East Sussex,
of research. This volume attempts to explain the UK: Psychology Press.
Brehmer, B., & Joyce, C. R. B. (Eds.). (1988). Human
applicability of Brunswik’s ideas to cognitive en-
judgment: The SJT view. Amsterdam: Elsevier
gineering and demonstrate some of its potential
Science Publishers B.V. (North-Holland).
contributions. Those who would like a more sub- Brentano, F. (1874/1973). Psychology from an
stantial historical presentation of Brunswik’s ideas empirical standpoint. L. L. McAlister, Ed.; A. C.
and the manner in which they have been applied Rancurello, D. B. Terrell, & L. L. McAlister,
and extended should see Hammond and Stewart Trans. London: Routledge.
(2001). Brunswik, E. (1943). Organismic achievement and
 Introduction to Brunswikian Theory and Method 23

environmental probability. Psychological Review, functioning reconsidered: A fast and frugal lens
50, 255–272. model. In K. R. Hammond & T. R. Stewart (Eds.),
Brunswik, E. (1952). The conceptual framework of The essential Brunswik: Beginnings, explications,
psychology. Chicago: University of Chicago Press. applications (pp. 342–347). Oxford: Oxford
(International Encyclopedia of Unified Science, vol. University Press.
I, no. 10.) Gigerenzer, G., Todd, P. M., & the ABC Research
Brunswik, E. (1955). Representative design and Group (Eds). (1999). Simple heuristics that make
probabilistic theory in a functional psychology. us smart. New York: Oxford University Press.
Psychological Review, 62, 193–217. Goldstein, W. M. (2004). Social judgment theory:
Brunswik, E. (1956). Perception and the representative Applying and extending Brunswik’s probabilistic
design of psychological experiments (2nd ed.). functionalism. In D. J. Koehler & N. Harvey
Berkeley: University of California Press. (Eds.), Blackwell handbook of judgment and decision
Brunswik, E. (1957). Scope and aspects of the making (pp. 37–61). Oxford: Blackwell.
cognitive problem. In H. E. Gruber, K. R. Goldstein, W. M., & Wright, J. H. (2001). “Percep-
Hammond, & R. Jessor (Eds.), Contemporary tion” versus “thinking”: Brunswikian thought on
approaches to cognition: A symposium held at the central responses and processes. In K. R.
University of Colorado (pp. 5–31). Cambridge, Hammond & T. R. Stewart (Eds.), The essential
MA: Harvard University Press. Brunswik: Beginnings, explications, applications
Camerer, C. F. (1981). The validity and utility of (pp. 249–256). Oxford: Oxford University Press.
expert judgment. Unpublished doctoral disserta- Hammond, K. R. (1955). Probabilistic functionalism
tion, Graduate School of Business, University of and the clinical method. Psychological Review, 62,
Chicago. 255–262.
Cooksey, R. W. (1996). Judgment analysis: Theory, Hammond, K. R. (1965). New directions in research
methods, and applications. San Diego, CA: on conflict resolution. Journal of Social Issues, 21,
Academic Press. 44–66.
Dawes, R. M. (1979). The robust beauty of improper Hammond, K. R. (Ed.). (1966). The psychology of Egon
linear models in decision making. American Brunswik. New York: Holt, Rinehart & Winston.
Psychologist, 34, 571–582. Hammond, K. R. (1973). The cognitive conflict
Dawes, R. M., & Corrigan, B. (1974). Linear models in paradigm. In L. Rappoport & D. A. Summers,
decision making. Psychological Bulletin, 81, 95–106. (Eds.), Human judgment and social interaction
Doherty, M. E. (Ed.). (1996). Social judgement theory (pp. 188–205). New York: Holt, Rinehart &
[Special issue of Thinking and Reasoning, 2(2/3)]. Winston.
East Sussex, UK: Psychology Press. Hammond, K. R. (1996). Human judgment and social
Doherty, M. E., & Brehmer, B. (1997). The para- policy: Irreducible uncertainty, inevitable error,
morphic representation of clinical judgment: A unavoidable injustice. New York: Oxford University
thirty-year retrospective. In W. M. Goldstein & Press.
R. M. Hogarth (Eds.), Research on judgment and Hammond, K. R. (2000). Judgments under stress. New
decision making: Currents, connections, and York: Oxford University Press.
controversies (pp. 537–551). Cambridge: Hammond, K. R., & Joyce, C. R. B. (Eds.). (1975).
Cambridge University Press. Psychoactive drugs and social judgment: Theory and
Dudycha, A. L., Dumoff, M. G., & Dudycha, L. W. research. New York: Wiley.
(1973). Choice behavior in dynamic environ- Hammond, K. R., Rohrbaugh, J., Mumpower, J., &
ments. Organizational Behavior and Human Adelman, L. (1977). Social judgment theory:
Performance, 9, 323–338. Applications in policy formation. In M. Kaplan &
Earle, T. C. (1973). Interpersonal learning. In S. Schwartz (Eds.), Human judgment and decision
L. Rappoport & D. A. Summers (Eds.), Human processes in applied settings (pp. 1–30). New York:
judgment and social interaction (pp. 240–266). Academic Press.
New York: Holt, Rinehart & Winston. Hammond, K. R., & Stewart, T. R. (Eds.). (2001). The
Gigerenzer, G. (1987). Survival of the fittest probabi- essential Brunswik: Beginnings, explications,
list: Brunswik, Thurstone, and the two disciplines applications. Oxford: Oxford University Press.
of psychology. In L. Kruger, G. Gigerenzer, & Hammond, K. R., Stewart, T. R., Brehmer, B., &
M. S. Morgan (Eds.), The probabilistic revolution: Steinmann, D. (1975). Social judgment theory. In
Volume 2. Ideas in the sciences (pp. 49–72). M. F. Kaplan & S. Schwartz (Eds.), Human
Cambridge, MA: MIT Press. judgment and decision processes (pp. 271–312).
Gigerenzer, G., & Kurz, E. M. (2001). Vicarious New York: Academic Press.
Random documents with unrelated
content Scribd suggests to you:
 "Mikä?" kysyi Sam hämmästyneenä.

"Sanantapa", vastasi mies ja vilkasi tovereihinsa.

"En oikein ymmärrä teitä", sanoi kokki. "Miten joku paikka voi olla
sanantapa?"

"Asia johtuu kokonaan eräästä velikullasta, jonka nimi oli kapteeni

Piggot", selitti kalastaja. "Hän oli kovin hauska veitikka ja erään
kerran oli päänsä tavallista täysinäisempänä. Silloin hän löysi
Etelä-Amerikan! Eikös se niin ollut Dan?"

"Niin kai se oli", vastasi vanha mies. "Hän luuli löytäneensä uuden
maanosan", jatkoi kalastaja "ja laski laivansa erääseen lahteen, veti
Suurbritannian lipun tangon nenään ja kutsui lahden nimensä
mukaan Piggotin lahdeksi. Tämä pääsi ulos miehistöltä kotia
saavuttua. Ja nyt jos joku täällä käyttäytyy hullumaisesti sanotaan
hänen ajavan takaa Piggotin lahtea. Siinä koko juttu."

Hän alkoi uudestaan nauraa, ja Sam, joka surukseen huomasi

miehen suureksi ja vahvarakenteiseksi, kääntyi ympäri ja seurasi jo
mäkeä ylöspäin harppaavaa kokkia. Saavuttuaan kukkulalle
katsoivat miehet taakseen; Stonepen Quay nauroi vielä.

Turvatakseen "koskemattomuuttaan" tekivät miehet ison

kierroksen maallepäin. "Ennen pari penikulmaa toiselle suunnalle
ennenkuin kohdata Dickiä", oli kokin tunnuslause. Puhe ei ottanut
sujuakseen, ja kuivin suin ja särkevin säärin asteltiin alakuloisina
tomuista tietä.

"Merilokkiin" saavuttua kello 7 pyydettiin laivalla yksinään oleva

Henry keittämään teetä.
 "Missä Dick on?" kysyi Sam ikäänkuin sattumalta.

"En ole nähnyt sitten päivällisen", vastasi poika. "Luulin hänen

olleen teidän seurassanne."

Sam pudisti päätään, hörppäsi teensä ja heittäytyi purjepakalle

kokin viereen. Väsyneinä kävelystä ei kumpikaan liikahtanut
paikaltaan ennenkuin yhdeksältä, jolloin noustiin ylös. Odottavan
näköisinä vilkaistiin suunnalle, mistä Dickiä varrottiin, mutta kun ei
mitään näkynyt, kiivettiin kojuihin.

Lamppu sai palaa romaania lukevan Henryn suureksi mielihyväksi.

Kello löi kaupungilla kymmenen — miehet vaihtoivat levottomia
katseita. Kun molemmat nyt makasivat lämpimissä vuoteissaan,
muistui väkistenkin mieleen Dick; oltiin ehkä oltu liian
ajattelemattomia. Puoli tuntia myöhemmin katsoi Henry yht'äkkiä
ylös; jotakin pehmeää kuului pudonneen kannelle ja lähestyvän
tassutellen hyttiä. Seuraavassa silmänräpäyksessä huudahti poika
pyhällä hartaudella.

"Dick! Dick?!”

"Tuki suusi!" ärisi Dick vaipuen merimies-arkulleen. "Oi, Herra

Jumala!
Jos joku ihminen tietäisi mitä minä olen kärsinyt!"

"Ihmettelen teitä", sanoi Henry ja heitti peitteen alastoman

merimiehen hartioille. "Mihin häveliäisyytenne on joutunut Dick?"

"Jos vielä sanallakin pilkkaat, väännän niskasi nurin, kirottu

irvinaama", sanoi Dick suuttuneena. "Ellei minulla häveliäisyyttä olisi,
niin päivällä jo olisin laivaan saapunut. Ei yksikään syntinen ihminen
voi tietää, mitä minä olen kärsinyt."

"Missä teidän vaatteenne ovat?" kysyi Henry. "Mistä helkatista

minä sen tiedän", tiuskasi toinen. "Jätin ne rannalle, kun menin
uimaan, ja palatessani olivat ne poissa. Olen istunut noilla kirotun
kylmillä kivillä aina kello kolmesta iltapäivällä, eikä ristin sielua
kulkenut ohitseni. Se oli ensi kerta, kun etsin kapteeni Gethingiä,
mutta viimeinen se myöskin oli."

"Vai sellaisissa asioissa olette liikkunutkin", sanoi Henry.

"Sanoinhan teille niistä seuraavan vaan ikävyyksiä."

"Olet aivan liian viisas ikäiseksesi", murisi Dick. "Samille ja kokille

et puhu tästä mitään. Ymmärrätkö?"

"Miksikä ei?"

"Sentähden, että minä en tahdo sitä", sanoi Dick ärtyisesti. "Se on

syy."

"Luultavasti he sen jo tietävätkin", sanoi Henry kiusottelevasti.

"Luulen Samin kuuntelevan unissaan."

Dick nousi ylös ja tarkasti varovasti molempia nukkuvia. Toistaen

varoituksensa ja uhaten kauheimmilla rangaistuksilla
tottelemattomuudesta meni hän omaan kojuunsa ja unohti
kärsimyksensä unen helmaan.

Dick säilytti salaisuutensa koko seuraavan päivän, mutta kun hän

tiistai-aamuna kannelle tultuaan huomasi vaatteensa ankkurin
vieressä likaiseen sanomalehtipaperiin käärittynä, paljastui
salaisuuskin. Sekä Samille että kokille kertoi hän kaikki, ja kun
miehet eivät nauraneet, kuten hän kovasti oli pelännyt, nousi heidän
arvonsa hänen silmissään suuresti. Kumpikaan ei edes suutaan
hymyyn vetänyt, kuuntelivat vaan tahdottomina, ja heti kun Dick oli
lopettanut kertomuksensa nukahtivat siihen paikkaan. Sen he
muuten tekivät aina kun vaan saivat siihen tilaisuutta.
VIII LUKU.

"Merilokki" seisoi Cocklemouthissa vielä kolme päivää, jona aikana

Dick, käveltyään kaksitoista penikulmaa sai Piggotin lahdesta
tietoonsa kaikki, mitä tiedettiin. Tämä toinen loukkaus oli saattaa
hänet vuoteen omaksi, mutta silloin juuri pälkähti hänen päähänsä
muuan mainio ajatus.

"Olen tullut ajatelleeksi, Sam", sanoi hän seuraavana aamuna,

"että olen ollut hyvin itsekäs. Tarkoitan tuota osakeyhtiötä. Minun
olisi pitänyt yhtyä siihen."

"Voithan sen vielä tehdä", sanoi Sam.

"Niin, ajattelin näin: parempi myöhään kun ei koskaan", sanoi Dick

ja kääntyi kokkiin. "Toimitan teidät kapteeni Gethingin jälille."

Kokki näytti hämmästyvän.

"Tiedän varmasti hänen asuvan eräässä paikassa, jota kutsutaan

Piggotin lahdeksi", jatkoi Dick. "Jos teitä haluttaa kävellä sinne
jonakuna iltapuhteena ja ottaa selvä kapteenista, saatte kaksi
kultaista' mieheen. Minä puolestani tyydyn yhteen."
 "Oo…", sanoi Sam ja seisoi kuin kivettyneenä moisesta
julkeudesta.

"Ei! — Sehän olisi vahinko sinulle, Dick", sanoi kokki. "Emmekä

me voi nauttia etuja sinun kustannuksellasi. — Ei. — Viisi puntaa
ovat sinun."

"Minä en niistä huoli", sanoi Dick totisena. "Ahneuteni saakoon

rangaistuksensa. Jos te kaksi menette ja otatte selvän hänestä,
pidän sitä oikein rehellisenä ystävyyden merkkinä."

"No! Voimmehan siellä käydä", sanoi kokki teeskennellyllä

ihastuksella.

"Dickillä on sydän oikealla paikallaan", sanoi Sam kokille. "Parasta

kun menemme heti juomaan teemme."

"Anna kun puristan kättäsi, Dick", sanoi kokki kiitollisena.

"Sen tahdon minäkin tehdä", sanoi Sam ja tarttui vuorostaan

Dickin kouraan. "Sinä olet oiva mies, Dick, siitä ei voi olla eri
mielipidettä."

"Oikein kunnon mies", säesti kokki.

"Lähdemme heti teen jälestä, jos sinä viitsit katsoa lipun meille,
Dick", sanoi Sam.

"Lipun?" toisti Dick. "Lipun?"

"Niin, tietystikin! Suurbritannian lipun", sanoi Sam ja katseli miestä

viattomalla kummastuksella. "Eihän Piggotin lahteen voi
Suurbritannian liputta mennä. Etkös sitä tiedä? Kävithän siellä eilen
illalla."

Sam seisoi rauhallisena ja odotti vastausta, katsellen ihmetyksellä

oikein ammattipetturia — ja Dick ryntäsi hyttiin suuttuneena kuin
kimalainen. Ollen luulevinaan Dickin lähteneen etsimään puheessa
ollutta lippua, pisti Sam varovasti päänsä hytin ovesta ja sanoi parin
uimahousujakin kelpaavan, paremman puutteessa — pilapuhe,
minkä hän halusta olisi ottanut takaisin, ennenkuin tapasi Dickin
päivällisellä.

Northfleet'iin uudestaan saavuttua oli koko etsimishalu

näennäisesti kadonnut jokaiselta. Ensiksikään ei ketään haluttanut
antautua alttiiksi Henryn pilapuheille ja toiseksi, kun asiaa punnittiin
järjen kannalta, tuntuivat edellytyksetkin hyvin mitättömiltä.
Kajuutassakin vallitsi alakuloinen mieliala; ainoastaan perämies
hoiteli puhetorvea.

"Sehän on suoraan vastoin parempaa ymmärrystä", sanoi hän

erään kerran, kun asiasta oli sukeutunut pitempi keskustelu. "Te ette
edes voi odottaa löytävänne häntä. Seuratkaa neuvoani: jatketaan
vanhoja vuoroja Brittlesean välillä; tyytykää siihen."

"En, siihen en tyydy", sanoi laivuri itsepintaisesti.

"On aivan synti heidän kanssaan — tarkoitan ukkoja, joilla on

vähänkin ikää ja harmaa parta —, että heitä ajetaan takaa kuin
mitäkin metsän otuksia. Ja sitäpaitsi joutuu miehistökin ennen pitkää
selkkauksiin, siitä olen varma."

"Puhukaa niin paljon kuin teitä haluttaa, se ei minua liikuta", sanoi

laivuri ja hieroi saippuaa kasvoihinsa kovassa puhdistushommassa.
 "Vai niin! Koetatte taas kaunistaa itseänne", sanoi perämies. "Ja
kenen tähden? Toisen miehen kihlatun morsiamen. Niin se kuitenkin
on, vaikka asiaa mihin päin vääntelisitte."

"Kun tarvitsen teidän neuvojanne, niin pyydän niitä", sanoi laivuri

suuttuneena.

Hiljaisuuden vallitessa lopetti hän vaatetuksensa, meni maalle,

käveli ja katseli puotien ikkunoita, lopulta kuitenkin kääntyen
Gravesendiin päin. Ainoat valonsäteet hänen nykyisessä
elämässään olivat säännölliset kouluvuorot, ja jatkaessaan
matkaansa siunasi hän opettajain tarkkaa velvollisuuden tunnetta —
joka ainakin kello neljältä ilmeni.

Mutta tänään riemuitsi hän liian aikaiseen, sillä vaikka lapset jo

aikoja sitten olivat päässeet koulusta, ei vaan näkynyt varjoakaan
Annis Gething'istä. Laivuri asteli edestakaisin katua ja ihmetteli.
Puoli viisi. Hän odotti kello kuuteen ja aikoi juuri lähteä matkaansa,
kun Annis neljännestä yli kuusi ilmestyi koulun rappusille.

"Kas! Tekin olette taas täällä", sanoi hän ja tervehti laivuria.

"Saavuimme juuri äsken."

"Ei mitään tietoja isästäni, otaksun?" kysyi Annis.

"Ei, ikävä kylläkin", vastasi laivuri! "Olitte myöhään koululla tänään.

Eikö totta?"

"No! — Kyllä."

"Näytätte väsyneeltä", sanoi laivuri auliisti.

 "Mutta sitä en ole", vastasi Annis. "Olin miss Grattan luona
juomassa kupin teetä. Äiti lupasi mennä ulos enkä minäkään
sentähden pitänyt mitään kiirettä."

"Onko hän ulkona vieläkin?" kysyi laivuri.

Miss Gething nyökkäsi ja pysähtyi seuraavassa kadun kulmassa.

"En aio vielä mennä kotiin", sanoi hän. "Täytyy vähäsen kävellä
raittiissa ilmassa."

Wilson oli kahdella päällä tarjoutuako seuraan vai ei.

"Hyvästi", sanoi hän lopulta ja ojensi kätensä.

"Hyvästi", vastasi Annis. "Jos haluatte pistäytyä meille ja odottaa,

niin äiti varmaan kotiin saavuttuaan on hyvin iloinen."

"Onko siellä ketään, jotta pääsen sisään", kysyi laivuri.

"Herra Glover, luullakseni", vastasi Annis ja katsoi pitkin katua.

"Sitten — sitten tulen toisella kerralla", änkytti Wilson

hätääntyneenä. "Luulin muuten…"

"Mitä luulitte?"

"En mitään", vastasi laivuri. "Minä… Menettekö pitkällekin?"

"No, enpä juuri. Miksi sitä kysytte?"

"Ajattelin vaan, että kernaammin kai kävelette yksinänne?"

 "Samantekevää", sanoi Annis. "Voittehan tekin tulla mukaan jos
haluatte."

He astuivat äänettöminä katua alas.

"Mitä te äsken aioitte sanoa?" kysyi Annis vihdoin, kun vaitiolo

näytti vaivaavan kumpaakin.

"Koska?"

"Kun sanoin herra Gloverin olevan meillä, niin sanoitte te

luulevanne…?" Tätä sanoessaan katsoi Annis laivuriin silmäyksellä,
jota tämän oli mahdoton käsittää.

"Niin, minä luulin", sanoi laivuri epätoivoisen rohkeudella, "teidän

kernaimmin haluavan kotiin."

"En ymmärrä teitä", sanoi Annis kylmästi. "Olette äärettömän

nenäviisas."

"Pyydän anteeksi", sanoi Wilson nöyrästi. "Olen hyvin — hyvin

pahoillani."

Seurasi uusi hiljaisuus.

"Antaisin vaikka mitä, jos löytäisin teidän isänne", sanoi Wilson

totisena.

"Niin! Toivoisin teidän onnistuvan — toivoisin teidän onnistuvan",

sanoi Annis katsoen olkapäänsä ylitse laivuria.

"Otaksumpa herra Gloverinkin tekevän parhaansa", sanoi Wilson.

 "Minun täytyy saada isäni takaisin", huudahti yht'äkkiä Annis.
"Minun täytyy saada hänet, mutta halusta soisin jonkun toisen, kuin
hra Gloverin, löytävän hänet."

"Tehän juuri menette naimisiin, jos isänne löytyy", sanoi laivuri

ihmeissään.

"Niin, jos herra Glover hänet löytää", sanoi Annis matalalla

äänellä.

"Tarkoitatteko…?" sanoi laivuri ja tarttui tytön käsivarteen.

"Tarkoitatteko ettette mene naimisiin herra Gloverin kanssa, ellei hän
löydä isäänne?"

"Niin", sanoi Annis. "Siten se on päätetty. Äiti haluaa sitä niin — ja

minulle on asia yhdentekevää kun vaan löydän isäni. Siihen olen
antanut lupaukseni."

"Mutta jos joku toinen hänet löytää?" änkytti Wilson.

"Siinä tapauksessa", sanoi Annis ja katsoi veitikkamaisesti

Wilsonia, "en mene naimisiin. Niinkö tarkoitatte?"

"En… En minä juuri sitä tarkoittanut", sanoi laivuri. "Aijoin

sanoa…"

"Katsokaa", sanoi Annis ja pysähtyi samassa, "mikä ihana näköala

tästä on joelle.

"Hurmaava", sanoi laivuri.

"Tämä onkin lempikävelypaikkani", jatkoi Annis.

Wilson painoi sen mieleensä.

 "Erittäinkin silloin kun herra Glover sattuu olemaan kotonanne",
sanoi hän enempää ajattelematta.

"Herra Glover on ollut hyvin kiltti", sanoi Annis totisena. "Hän on

ollut hyvin kiltti äidille ja nähnyt paljon vaivaa isäni etsimisestä."

"Toivon kumminkin, ettei hän onnistu löytämään häntä", virkkoi

Wilson.

Annis kääntyi ja tarkasti laivuria.

"Onko tuo ystävällistä?" sanoi hän kiivaasti.

"Tahdon itse hänet löytää", sanoi laivuri. "Te tiedätte

minkätähden."

"Nyt täytyy minun kiirehtää kotiin", sanoi Annis vastaamatta

laivurin arveluihin.

Wilson tunsi rohkeutensa loppuvan ja yritti kaikin mokomin saada

sanottavansa lausutuksi.

"Toivoisin teidän kohtelevan minua samalla tavalla kuin herra

Gloveriakin."

"Sen teen halusta", vastasi Annis kiireesti. Hän ei voinut estää

huuliensa värisemistä, ja hänen katseensa kääntyi toisaalle.

"Olen rakastanut teitä ensi näkemästä", sanoi laivuri yht'äkkiä

laukaisten.

Aivan valmistumattomana näin pikaiseen hyökkäykseen ei miss

Gething tiennyt mitä vastata. Osat olivat vaihtuneet, ja punastuneena
kääntyi hän toisaalle.
 "Olen viettänyt päiviä käyden edestakaisin koulunne edessä, edes
vilahdukseltakin nähdäkseni teidät", jatkoi Wilson. "Toisinaan olen
ihmetellyt, etteivät lapset ole jo mitään huomanneet."

Miss Gethingin karmiinin punainen poski kääntyi häneen.

"Sen ne kylläkin ovat tehneet", sanoi hän. "Eräs pieni toitottaja sai
seistä tunnin nurkassa oppiakseen pitämään suunsa kiinni."

"Minä en voi sitä auttaa", sanoi Wilson. "Saatte panna vaikka koko
koulun nurkkaan ettekä sittenkään saa minua estetyksi rakastamasta
teitä."

"Ehkä nyt käännymme ympäri", sanoi Annis kylmästi. Molemmat

astuivat ajatuksissaan eikä kumpikaan puhunut sanaakaan. Vähää
ennen kaupunkiin saapumista Wilson pysähtyi ja katsoi pelotta ja
vakavasti tyttöä silmiin. Hän oli päättänyt ajaa asiansa loppuun,
maksoi mitä maksoi. Miss Gething peräytyi ja katsoi maahan.

"Lupaatteko kohdella meitä molempia yhtäläisesti?" kysyi Wilson

matalalla äänellä.

"En", vastasi Annis ja katsoi pelokkaasti hymyillen laivuria.

Ikäänkuin huumaantuneena tarttui tämä tytön käteen ja yritti vetää
hänet luoksensa.

"Ei", sanoi Annis ja vetäytyi kiireesti syrjään. "Ei, se ei sovi."

Wilson pelästyi, huomaten menneensä liian pitkälle, ja pelko sai

hänessä uudestaan vallan.

"Miksikä se ei sovi?" kysyi hän koettaen näyttää viattomalta.

 Lyhyt, mutta onnettomuutta uhkaava kipinä välähti miss Gethingin
silmissä.

"Pyydän anteeksi", sanoi laivuri nöyrästi.

"Mitä?" kysyi miss Gething, ja nyt oli hänen vuoronsa näyttää

viattomalta.

Väsyneinä kiertoteihin, joissa hän aina joutui tappiolle, päätti

Wilson taas käydä suoraan asiaan.

"Kun koetin suudella teitä ja sitten olin olevinani ymmärtämätön",

sanoi hän avomielisesti.

"Kapteeni Wilson… minä… minä en käsitä mitä te tarkoitatte."

"Kyllä te käsitätte", sanoi Wilson levollisena.

Kipinä välähti taas miss Gethingin silmissä, hän puri huultaan ja

kääntyi poispäin, huomattuaan mahdottomaksi rangaista tätä
syntistä olentoa.

"Varmaankin sanotte ensi kerran tällaisia asioita jollekin tytölle",

sanoi hän lopulta. "Niin", sanoi Wilson vakuuttavasti.

"Tahdotte harjotella", sanoi miss Gething halveksivasti.

"Niin, sitä juuri haluan", sanoi Wilson innokkaasti.

Hän teki taas liikkeen lähestyäkseen tyttöä, mutta tämän katse

pidätti hänet.

"Ei kuitenkaan tytön kanssa, joka jo on puoliksi kihloissa", sanoi

Annis ja katsoi laivuria lempeän moittivasti. "Se ei ole oikein."
 "Tietääkö hän asian oikean laidan?" kysyi Wilson, tietystikin
tarkoittaen herra Gloveria. Miss Gething nyökkäsi.

"Silloin katson asian olevan aivan oikein ja paikallaan", sanoi

Wilson.

"Mutta sitä en tee minä", sanoi Annis ja ojensi kätensä. "Hyvästi

nyt!" sanoi hän varmalla äänellä. "En halua tavata teitä ennen isäni
löytymistä. Ja, jos herra Glover hänet löytää, en halua tavata teitä
koskaan. Voikaa hyvin."

Laivuri tarttui ojennettuun käteen ja ihmetellen rohkeuttaan, veti

hän vastusta tekevän tytön luoksensa. Sitten taivutti hän päänsä ja
miss Gethingin suosiollisella avulla suuteli hänen hattunsa
koristuksia. Tyttö riuhtasi itsensä irti, juoksi aina kadun toiseen
päähän ja katosi nurkan taakse. Laivuri vilkasi vihaisesti aidan
takana katselevaa hevosta, tapauksen ainoaa elollista näkijää, ja
lähti kiireesti kohti Northfleetia, kehottaakseen rohkeutensa
menettänyttä miehistöään uusiin ponnistuksiin.
IX LUKU.

Laivurin ihmeeksi ja harmiksi piti Annis sanansa. Hän tiesi hyvin,

ettei voinut estää laivuria kohtaamasta itseään kadulla, ja sentähden
otti hän mukaansa pari koulutyttöä. Kun ei tämäkään näyttänyt
auttavan, lisättiin lukua, ja sen päivän perästä, jolloin Anniksen
mukana kulki viisi nauravaa pienokaista, ei laivuria enää näkynyt.

Etsimistä jatkettiin koko kesä, ja "Merilokin" miehistö sai osakseen

monet naurut ja ihmettelyt nuuskivasta esiintymisestään.

Syyskuun alkupäivinä oltiin Iranbridgessä, pienessä kaupungissa,

Lebbe-joen rannalla. Kuten tavallista, eivät laivurin tiedustelut
johtaneet mihinkään tuloksiin. Iranbridge oli siksi vähäinen
paikkakunta, ettei kukaan siellä voinut piilottautua. Kuitenkin sai
Henry eräänä kauniina päivänä, kun huomattiin hänen
vastenmielisyytensä lastin purkaukseen, luvan mennä maalle
ostamaan tupakkaa kokille ja samalla katselemaan vähän
ympärilleen.

Henry astuskeli hiljalleen ja heitti tuontuostakin katseen taaksensa

tomupilveen, mikä ympäröi "Merilokkia." Paikkakunta oli uusi ja
tuntematon ja hänen mielensä hehkui seikkailuihin.
 Kaupungissa ei löytynyt montakaan näkemisen arvoista paikkaa.
Ennen rautateitä oli se kylläkin ollut vilkas satamakaupunki; nyt olivat
sen kadut tyhjät ja ranta melkein autio. Paitsi "Merilokkia", löytyi
vaan pari jahtia, joiden lastia miehet koreilla kantoivat maalle.

Hiljaisuus vaikutti painostavasti Henryyn; juotuaan kohtuullisen

puolikkaan olutta sytytti hän piippunsa ja asteli Isoakatua kädet
housuntaskuissa.

"Ei! Kyllä Lontoo on sentään toista", sanoi hän ja pysähtyi korkean

tiilimuurin viereen katselemaan hedelmäpuita sen takana. "Täällähän
on kaikki ihan kuollutta."

Hän kiipesi muurin harjalle, istui sinne hajareisin ja vihelteli.

Kasvava oluthalunsa oli muuttunut vanhaan rakkauteen omeniin.
Puut peittivät takana olevan rakennuksen, tilaisuus oli mainio, ja
meikein pakosta hyppäsi hän muurin toiselle puolelle, ja alkoi
kiireimmän kautta sulloa omenia taskuihinsa.

Kaikki oli äärettömän hiljaista. Poika matki romaanissa lukemiansa

"punanahkain" hiipiviä liikkeitä tällaisissa tapauksissa, niin hyvin kuin
kaksi tai kolme omenakiloa taskuissa sallivat, lähestyi sitten tutkien
isoa huvihuonetta, ja kurkisti ovesta sisään. Se oli tyhjä, paitsi
pöytää ja paria puupenkkiä. Henry istuutui ja maistoi omenaa.

Mutta samassa huomasi hän tilansa vaarallisuuden. Kuului

lähestyviä askeleita, joiden päämääränä epäilemättä oli huvihuone.
Äänettömästi ja kiireesti kuin vanha tottunutkin murtovaras ryömi
Henry pöydän alle.

"Pysy tuolla, hävytön penikka", kuului samassa naisääni. "Ja

ennen et pääse ulos, kun osaat joet aivan kuin vettä."
 Joku työnnettiin huvihuoneeseen, ovi suljettiin ja väännettiin
lukkoon. Askeleet poistuivat ja hämmästynyt ryöväri käsitti nyt tilansa
koko vaarallisuuden: elämänsä riippui pienestä tytöstä.

"Minä en välitä teidän saarnoistanne", kuului vastustava ääni.

"Minä en välitä — en! Syljen teidän jokiinne — syljen teidän jokiinne
— syljen teidän jokiinne."

Äänen omistajatar istuutui pöydälle. Hädissään koetti Henry

parantaa asentoaan, vaan löi vahingossa päänsä kovasti pöytään.

"Ooooh!…" huusi pieni tyttö kauhistuneena. "Älkää peljätkö", sanoi

Henry nöyrästi ja katsoi tyttöön. "En minä tee teille pahaa."

"Hyi! — Hyi", sanoi tyttö. "Poika…!"

Henry ryömi esille ja istuutui arvokkaasti, vaan yskäsi samassa

nolona huomatessaan tytön katseen kiintyvän hänen taskuihinsa.

"Mitä sinulla on taskuissasi?" kysyi tyttö. "Omenia", vastasi Henry

houkuttelevasti. "Ostin niitä kaupungista."

Tyttö otti pari kappaletta ja tarkasti niitä tyystin. "Olet paha ja

kelvoton poika", sanoi hän ja purasi toista omenaa. "Mutta kyllä tulet
saamaankin, sinä, kunhan miss Dimchurch tulee."

"Kuka miss Dimchurch on?" kysyi Henry, anteeksiannettavalla

uteliaisuudella. "Opettajatar."

"Onko tämä koulu?"

Pieni tyttö nyökkäsi suu omenaa täynnä. "Onko täällä miehiäkin

opettamassa?" kysyi Henry, teeskennellyllä välinpitämättömyydellä.
Tyttö pudisti päätään.

"Sinä olet ainoa poika, minkä koskaan olen täällä nähnyt", sanoi
hän iloisesti. "Kyllä sinä saat, sinä, kunhan miss Dimchurch tulee."

Joku painostava levottomuus putosi Henryn sydämeltä ja hän

hymyili tyytyväisenä.

"Vanhoja mamselleja minä en pelkää", sanoi hän rauhallisesti, veti

piipun taskustaan ja täytti sen.

Pienen tytön silmät loistivat ihastuksesta.

"Oi! Jos minäkin olisin poika", sanoi hän kaiholla. "Silloin en

välittäisi mitään miss Dimchurchista. Oletko sinä merimiespoika?"

"Merimies", oikaisi Henry. "Kyllä."

"Minä pidän merimiehistä", sanoi pieni tyttö ystävällisesti. "Saat

kappaleen omenastani, jos haluat."

"En, kiitoksia", sanoi Henry kiireesti. "Minulla on vielä

alkamattomia."

Pikku tyttö nousi pystyyn ja katsoi ylpeästi poikaan, mutta

huomattuaan, ettei tämä vilkaissutkaan häneen, puri hän uudestaan
omenaansa. "Mikä sinun nimesi on?" kysyi tyttö. "Henry Hatkins",
vastasi poika. "Entäs sinun?"

"Gertrude Ursula Florence Harcourt", sanoi tyttö suoristaen

selkäänsä.
"Minä en pidä Hatkins nimestä."
 "Etkö?" sanoi Henry ja koetti peittää loukkaantumistaan. "Minä en
taas pidä Gertrudesta taikka Ursulasta taikka Florencesta, ja
Harcourt minusta on kaikkein hullumpi."

Miss Harcourt siirtyi pari kolme tuumaa etemmäksi ja naputti

hermostuneesti sormenpäillään pöytää.

"Minä en välitä mistä sinä pidät", sanoi hän ylpeästi.

"Minä pidän Gertystä", sanoi Henry tuntijan äänellä ja katsoi pieniä

rusottavia kasvoja. "Gerty on hyvin kaunis, siitä minä pidän."

"Siten minua tavallisesti kutsutaankin", sanoi miss Harcourt

huolimattomasti. "Purjehtiiko sinun laivasi suurilla merilläkin?"
"Kyllä", vastasi poika.

"Ja kuinka monta kertaa", sanoi Gertrude Ursula Florence

Harcourt, "olet taistellut merirosvojen kanssa?"

"En nyt niin varmaan muista — kuusi tai seitsemän kertaa", sanoi
Henry
Hatkins. "Luulen niitä sentään olleen vaan kuusi."

"Kerro minulle niistä kaikista", sanoi miss Harcourt, joka vaivoin

pysyi alallaan.

Henry puri kappaleen omenastaan ja alkoi, romaanejaan

muistellen, juttunsa. Hän valloitti laivoja tavalla, jota eivät edes
amiraalit olisi voineet uneksiakaan, vielä vähemmän merirosvot,
jotka aina tulivat lyödyiksi. Miss Harcourt pidätteli henkeään, ja
hänen pienet (eivät suinkaan liian puhtaat) sormensa puristivat pojan
käsivartta kovemmin ja heikommin, aina kertomuksen kulun mukaan.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookfinal.com