Instant Ebook Access, One Click Away – Begin at ebookgate.

com

Psychology of Reasoning Theoretical and Historical

Perspectives 1st Edition Ken Manktelow

https://ebookgate.com/product/psychology-of-reasoning-
theoretical-and-historical-perspectives-1st-edition-ken-
manktelow/

OR CLICK BUTTON

DOWLOAD EBOOK

Get Instant Ebook Downloads – Browse at https://ebookgate.com

Click here to visit ebookgate.com and download ebook now
Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Thinking and Reasoning An Introduction to the Psychology

of Reason Judgment and Decision Making 1st Edition Ken
Manktelow
https://ebookgate.com/product/thinking-and-reasoning-an-introduction-
to-the-psychology-of-reason-judgment-and-decision-making-1st-edition-
ken-manktelow/
ebookgate.com

The Sociology of Religion Theoretical and Comparative

Perspectives Second Edition Malcolm Hamilton

https://ebookgate.com/product/the-sociology-of-religion-theoretical-
and-comparative-perspectives-second-edition-malcolm-hamilton/

ebookgate.com

Inseparable Nature of Love and Aggression Clinical and

Theoretical Perspectives 1st Edition Otto F. Kernberg

https://ebookgate.com/product/inseparable-nature-of-love-and-
aggression-clinical-and-theoretical-perspectives-1st-edition-otto-f-
kernberg/
ebookgate.com

Historical and conceptual issues in psychology 1st Edition

Marc Brysbaert

https://ebookgate.com/product/historical-and-conceptual-issues-in-
psychology-1st-edition-marc-brysbaert/

ebookgate.com
Minding Time A Philosophical and Theoretical Approach to
the Psychology of Time 1st Edition Carlos Montemayor

https://ebookgate.com/product/minding-time-a-philosophical-and-
theoretical-approach-to-the-psychology-of-time-1st-edition-carlos-
montemayor/
ebookgate.com

Defense Mechanisms Volume 136 Theoretical Research and

Clinical Perspectives 1st Edition Uwe Hentschel

https://ebookgate.com/product/defense-mechanisms-
volume-136-theoretical-research-and-clinical-perspectives-1st-edition-
uwe-hentschel/
ebookgate.com

Theoretical Issues in Psychology An Introduction Second

Edition Sacha Bem

https://ebookgate.com/product/theoretical-issues-in-psychology-an-
introduction-second-edition-sacha-bem/

ebookgate.com

Combinatorial Optimization and Theoretical Computer

Science Interfaces and Perspectives 1st Edition Vangelis
Th. Paschos
https://ebookgate.com/product/combinatorial-optimization-and-
theoretical-computer-science-interfaces-and-perspectives-1st-edition-
vangelis-th-paschos/
ebookgate.com

Irregular Migration and Human Rights Theoretical European

and International Perspectives 1st Edition Barbara Bogusz

https://ebookgate.com/product/irregular-migration-and-human-rights-
theoretical-european-and-international-perspectives-1st-edition-
barbara-bogusz/
ebookgate.com
Psychology of Reasoning
Psychology of Reasoning
Theoretical and Historical Perspectives

edited by
Ken Manktelow & Man Cheung Chung
First published 2004
by Psychology Press
27 Church Road, Hove, East Sussex BN3 2FA
Simultaneously published in the USA and Canada
by Psychology Press
270 Madison Avenue, New York, NY 10016
This edition published in the Taylor & Francis e-Library, 2005.
“To purchase your own copy of this or any of Taylor & Francis or Routledge’s
collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”
Psychology Press is a part of the Taylor & Francis Group
Copyright © 2004 Psychology Press
Cover design by Hybert Design
All rights reserved. No part of this book may be reprinted or
reproduced or utilized in any form or by any electronic, mechanical, or
other means, now known or hereafter invented, including photocopying
and recording, or in any information storage or retrieval system,
without permission in writing from the publishers.
The publisher makes no representation, express or implied, with regard
to the accuracy of the information contained in this book and cannot
accept any legal responsibility or liability for any errors or omissions
that may be made.
This publication has been produced with paper manufactured to strict
environmental standards and with pulp derived from sustainable
forests.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
Psychology of reasoning: theoretical and historical perspectives /
edited by Ken Manktelow and Man Cheung Chung.—1st ed.
p. cm.
Includes bibliographical references and index.
ISBN 1-84169-310-3 (hbk)
1. Reasoning (Psychology) I. Manktelow, K., 1., 1952–
II. Chung, Man Cheung, 1962–

BF442.P79 2004
153.4′3–dc22
2004016579
ISBN 0-203-50693-6 Master e-book ISBN

ISBN 0-203-59521-1 (Adobe eReader Format)

ISBN 1-84169-310-3 (Print Edition)
Peter Wason, the progenitor of the modern psychology of reasoning and
a friend to everyone here, died while this volume was being prepared.
We dedicate this book to his memory.
Contents

List of contributors ix

1 The contextual character of thought: Integrating themes from

the histories and theories of the study of reasoning 1
KEN I. MANKTELOW AND MAN C. CHUNG

2 Reasoning and task environments: The Brunswikian approach 11

MICHAEL E. DOHERTY AND RYAN D. TWENEY

3 Rationality, rational analysis, and human reasoning 43

NICK CHATER AND MIKE OAKSFORD

4 The psychology of conditionals 75

DAVID OVER

5 Violations, lies, broken promises, and just plain mistakes: The

pragmatics of counterexamples, logical semantics, and the
evaluation of conditional assertions, regulations, and promises
in variants of Wason’s selection task 95
DAVID P. O’BRIEN, ANTONIO ROAZZI, MARIA G. DIAS, JOSHUA B.
CANTOR, AND PATRICIA J. BROOKS

6 The natural history of hypotheses about the selection task:

Towards a philosophy of science for investigating human
reasoning 127
KEITH STENNING AND MICHIEL VAN LAMBALGEN

7 Reasoning and rationality: The pure and the practical 157

KEN I. MANKTELOW

8 The history of mental models 179

PHILIP N. JOHNSON-LAIRD
viii Contents
9 Some precursors of current theories of syllogistic reasoning 213
GUY POLITZER

10 History of the dual process theory of reasoning 241

JONATHAN ST. B. T. EVANS

11 Coherence and argumentation 267

DAVID W. GREEN

12 Reasoning about strategic interaction: Solution concepts in

game theory 287
ANDREW M. COLMAN

13 What we reason about and why: How evolution explains reasoning 309
GARY L. BRASE

14 The proof of the pudding is in the eating: Translating Popper’s

philosophy into a model for testing behaviour 333
FENNA H. POLETIEK

15 Constructing science 349

SANDY LOVIE

Author index 371

Subject index 381
Contributors

Gary L. Brase, Department of Psychological Sciences, University of

Missouri–Columbia, 210 McAlester Hall, Columbia, MO 65211, USA.
Patricia J. Brooks, Department of Psychology, College of Staten Island, City
University of New York, 2800 Victory Blvd., 4S-108, Staten Island, NY
10314, USA.
Joshua B. Cantor, Department of Rehabilitation Medicine, Mount Sinai
School of Medicine, 1 Gustave L. Levy Place, New York NY, USA.
Nick Chater, Department of Psychology, University of Warwick, Coventry
CV4 7AL, UK.
Man Cheung Chung, Department of Psychology, University of Plymouth,
Drake Circus, Plymouth PL4 8AA, UK.
Andrew M. Colman, Department of Psychology, Astley Clarke Building,
University of Leicester, University Road, Leicester LE1 7RH, UK.
Maria G. Dias, Universidade Federal de Pernambuco, Av. Acad. Hélio
Ramos, s/n – CFCH, 8° Andar, Recife 50670–901, PE Brazil.
Michael E. Doherty, Bowling Green State University, Bowling Green, Ohio,
USA.
Jonathan St. B. T. Evans, Centre for Thinking and Language, University of
Plymouth, Drake Circus, Plymouth PL4 8AA, UK.
David W. Green, Department of Psychology, University College London,
Gower Street, London WC1E 6BT, UK.
Philip N. Johnson-Laird, Department of Psychology, Princeton University,
Green Hall, Princeton, NJ 08544, USA.
Sandy Lovie, Department of Psychology, University of Liverpool, Eleanor
Rathbone Building, Bedford Street South, Liverpool L69 7ZA, UK.
Ken Manktelow, Division of Psychology, School of Applied Sciences, Uni-
versity of Wolverhampton, Millennium City Building, Wulfruna Street,
Wolverhampton WV1 1SB, UK.
x Contributors
Mike Oaksford, School of Psychology, Cardiff University, P.O. Box 901,
Cardiff CF1 3YG, Wales, UK.
David O’Brien, Baruch College and the Graduate School of the City
University of New York, 55 Lexington Avenue, New York NY, USA.
David Over, Department of Psychology, University of Sunderland, Sunder-
land SR6 0DD, UK.
Fenna H. Poletiek, Unit of Experimental and Theoretical Psychology,
University of Leiden, P.O. Box 9555, 2300 RB Leiden, The Netherlands.
Guy Politzer, CNRS: Laboratoire “Cognition et Activités Finalisées”,
Université de Paris 8, 2 Rue de la Liberté, 93526 Saint-Denis, France.
Antonio Roazzi, Universidade Federal de Pernambuco, Av. Acad. Hélio
Ramos, s/n – CFCH, 8° Andar, Recife 50670–901, PE Brazil.
Keith Stenning, Human Communication Research Centre, Dept. of Psych-
ology, Edinburgh University, 2 Buccleuch Place, Edinburgh EH8 9LW,
UK.
Ryan D. Tweney, Bowling Green State University, Bowling Green, Ohio,
USA.
Michiel van Lambalgen, Department of Philosophy, Amsterdam University,
Plantage Muidergracht 24, 1018 TV Amsterdam, The Netherlands.
1 The contextual character of
thought
Integrating themes from the
histories and theories of the
study of reasoning
K. I. Manktelow and M. C. Chung

The study of reasoning, unlike what was once said about psychology as a
discipline, has both a long past and a long history. Indeed, it would not be too
far-fetched to claim that people have been studying and reporting on their
reasoning abilities for as long as they have been studying anything. So the
shadow cast by this history is probably longer than in any other field of the
cognitive sciences: modern authors still regularly appeal to the writings of
their ancient philosophical forebears, most notably Aristotle, who lived over
2300 years ago. Even within the confines of contemporary empirical psych-
ology, a long line can be traced back: the earliest such reference in this book is
a near century old, from 1908.
All the more curious, then, that present-day students of the psychology of
reasoning can find their attempts to situate the field rather an arduous busi-
ness. While you will often find such material in the introduction sections of
empirical papers, and in the background sections of theoretical reviews, you
can feel rather like an archaeologist piecing together an object from various
fragments you happen upon. One aim of this book is to enable those inter-
ested in the psychology of reasoning to apprehend more of its general histor-
ical base, and hence get a better appreciation of the direction in which the
ideas are flowing, by finding more of these pieces in one place. We will have a
better idea of where we are going if we are clearer about where we have come
from.
In addition to the rich strata of history on which the psychology of reason-
ing stands, there has been an increasing acknowledgement, as it has turned
from a task-based data-gathering enterprise to a deeper and wider theoretical
one, that there are connections to be made both to other aspects of the study
of human thought, and to cognitive science in general. Psychologists study-
ing reasoning have often made appeals for engagement with these other fields
and have increasingly gone out looking for it in recent years. Every recent
major theoretical statement has been made in these terms, as several of the
present chapters portray, although sometimes, as Stenning and van Lambal-
gen argue here, there has been a step in the other direction and a detrimental
2 Manktelow and Chung
split has opened up. A second aim of this book is to present some of these
theoretical extensions and contacts.
The third aim of the book was something of an act of faith on the part of
its editors. We have 14 contributions here, taken from a wide range of areas of
study; some of these areas will be familiar to and expected from people
already interested in the psychology of reasoning, some perhaps less so. We
believed (hoped) that it would be possible to identify emergent themes, from
both the historical and theoretical strands, that would enable both ourselves
and you, the reader, to form the more grounded view of the field that is one of
the benefits of its progress towards being a more integrated science. That faith
has been amply repaid, and in the rest of this introduction we shall set out
some of these themes (you may well find others); they have informed the way
the chapters have been organized.
One of the first true empirical psychologists, Hermann Ebbinghaus,
devoted his life’s work to the attempt to explain memories from their cradles
to their graves, as it were: from their acquisition to their loss. He realized that
in order to accomplish the first of these tasks, the material to be remembered
had to be free of all prior associations, otherwise he could not be sure he was
looking at the very start of an act of memory. Hence the development of his
famous nonsense syllables. Something like this idea has also been applied in
the psychology of reasoning: if one wants to study an inference, it had better
not be contaminated by memory, otherwise any interesting effects could be
attributed to knowledge rather than thought. This was the nub of Mary
Henle’s (1978) famous objection, that she had never found errors that could
unambiguously be attributed to faulty reasoning. The answer seems obvious:
strip out, as Ebbinghaus thought he had done, all possibility of other influ-
ences, and study only reasoning. Thus a research programme grew up based
around the study of responses to various abstract tasks. This programme is
reviewed in greatest detail in older texts such as Wason and Johnson-Laird
(1972) and Evans (1982), although it is also at the core of later ones such as
Evans, Newstead, and Byrne (1993), Garnham and Oakhill (1994), and
Manktelow (1999).
However, the assumption that one can study reasoning directly by using
abstract tasks has been replaced by a realization that, if one strips out con-
tent, one is losing something essential about reasoning (as one is about mem-
ory). This is far from a new idea in itself, as Mike Doherty and Ryan Tweney
relate; it was at the heart of Brunswik’s approach to psychology. (It was also
acknowledged by Wason and Johnson-Laird, in both the title and the sub-
stance of their 1972 book.) Brunswik was guided by the principle of represen-
tative design: that the essential elements of the environment in question must
figure in research design. If this principle is taken on board, certain con-
sequences follow as to the way you go about designing experiments and inter-
preting their findings, as Doherty and Tweney show. Brunswik’s ideas are
probably most familiar to students of reasoning through the work of Gig-
erenzer and colleagues on fast and frugal heuristics and probabilistic mental
 The counterfactual character of thought 3
models (see Gigerenzer, Todd, & the ABC Research Group, 1999), but they
are becoming more influential more widely: a whole issue of the journal
Thinking and Reasoning was devoted to this approach in 1996.
The need to acknowledge the relation between the task and the natural
cognitive ecology of research participants has also been the fundamental
concern of the “rational analysis” approach to cognition, as Nick Chater and
Mike Oaksford demonstrate. They turn Brunswik’s principle around, and ask
of a given reasoning task, what environment is that task reflecting? In com-
mon with Brunswik, rational analysis thus requires a specification of the
cognitive environment. A scientist might design a task assuming that the
participant’s response to it reflects one kind of cognition, but as Porgy said, it
ain’t necessarily so. The participants might be coming at it from quite another
direction, in which case our evaluation of their performance is under serious
question. Chater and Oaksford take the example of tasks ostensibly designed
to assess logical deduction, but on which there is evidence that participants
may be attempting optimal information gain. This approach is one of the
most important recent theoretical and methodological innovations in the
psychology of reasoning, and can only gain in influence in the future. It
chimes with the standpoint of the philosopher Gilbert Harman (e.g., 1999),
who argues that reasoning is all about belief revision, and hence that “deduct-
ive reasoning” is not reasoning, since deduction is concerned only with find-
ing proofs, and so does not increase information. Chater and Oaksford
remind us of this position when they declare that, since much of the experi-
mental study of deductive reasoning does not seem to engage it, the term
“psychology of deductive reasoning” may be a misnomer.
These are fundamental ideas about the way research into reasoning is and
ought to be done, and the sorts of theories that ought to be compiled, but the
acknowledgement of the importance of the contexts of thought can be
cashed out in other ways. Much of our thinking – and almost all of the
thinking that has been studied by psychologists – is verbal, and so the lin-
guistic context comes to the fore. And at the heart of reasoning, as David
Over emphasizes, is the conditional, usually expressed using sentences con-
taining “if”. This is reflected in the experimental study of human reasoning,
where studies of conditional reasoning far outweigh any other kind. Over
points out problems with “logicism”, the idea that the study of conditionals
must be restricted to those that embody statements expressing certainties.
Most of the statements we make about the world are uncertain to some
degree, and so most real-world conditional arguments cannot strictly be
deductively valid. Adherence to a strong logicist approach to conditionals
leads to some long-recognized, jarring paradoxes. Furthermore, many of our
conditional expressions concern counterfactual or deontic situations, and so
are not amenable to a strict deductive analysis in the first place. Over traces
the history of the study of conditionals both in psychology (through the
theories of natural deduction and mental models) and philosophy (focusing
on the work of Ramsey and Stalnaker). He concludes that considerations of
4 Manktelow and Chung
utility and probability are necessary in explaining reasoning with all types of
conditional, with the hitherto “standard” indicative conditional being use-
fully seen as a special case of the deontic conditional. This decision-theoretic
approach to conditional reasoning is shared by Chater and Oaksford,
although they take a radically different theoretical stance.
The semantics and pragmatics of conditionals are also the focus of David
O’Brien, Antonio Roazzi, Maria Dias, Joshua Cantor, and Patricia Brooks;
they reflect on what a close analysis of these factors implies for research using
the most well-known experimental paradigm on conditional reasoning, the
Wason selection task. They concentrate on the indicative–deontic distinction,
which has been the locus of much contemporary theorizing. It has often been
held that when we evaluate an indicative conditional, in the selection task and
elsewhere, we are dealing with its possible falsification, whereas when we
evaluate a deontic conditional, we are dealing with possible violations of it.
So in the former case, the status of the sentence is in question but in the latter
it is not. Using a large range of task scenarios, O’Brien and colleagues show
that it is easy to be simplistic about this, for instance in falsely assuming that
deontic rules cannot be falsified, or in distinguishing lies and mistakes. The
solution is a precise specification of both the semantics and the pragmatics of
the task in question (a Brunswikian moral). Such an analysis leads in their
view to a unified conception of conditional reasoning, where differences in
performance can be explained using pragmatic principles and general-purpose
reasoning mechanisms, rather than domain-specific processes.
The importance of the reasoner’s interpretation of task materials is also
the starting point for Keith Stenning and Michiel van Lambalgen’s analysis.
They also address the selection task, and like Over and O’Brien and col-
leagues, strongly contest the “standard” reading of the conditional as corres-
ponding to the logic of material implication (which states that a conditional
is true when its antecedent is false or its consequent is true). Unlike the more
intuitive analysis of O’Brien and colleagues, they point to developments in
logical semantics as a descriptive rather than normative discipline, and urge a
reunification between psychology and logic, whose divorce, they argue, has
hindered the progress of research. Similarly, selection task research has con-
tained an outdated philosophy of science based on Popper’s falsificationist
doctrine, another area that has moved on in recent times. As with O’Brien
and colleagues and Chater and Oaksford, Stenning and van Lambalgen
argue that researchers should not blithely assume that participants in a selec-
tion task experiment are using an implication conditional and seeking pos-
sible falsifications, and they too go into detail about the indicative–deontic
distinction. Rather, the participants’ task is to discover how the researcher
intends the task and rule to be understood. Without this basic semantic
account, we are in no position to argue about mental processes. Thus,
echoing a sentiment first expressed by Anderson (1978) and urged on the
psychology of reasoning by Evans (1982), theories require an account of
both representation and process.
 The counterfactual character of thought 5
As we have seen above, there has been a lot of attention devoted to the
indicative–deontic distinction in recent times, and this is the concern of the
chapter by Ken Manktelow. He also locates his analysis in research on
the Wason selection task, completing a trilogy of chapters that do so; the
reason for this is that the empirical base of the upsurge in interest in deontic
reasoning lies in work using that problem. It was given its most powerful
stimulus in the mid 1980s, in the widely known studies of Cheng and Holyoak
and Cosmides. One of the offshoots of the direction of attention to deontic
reasoning was the reawakening of a bipolar argument concerning the integ-
rity of reasoning. Broadly speaking, an indicative statement concerns matters
of fact, while a deontic statement prescribes an action. This difference has
been acknowledged since Aristotle, as the difference between pure, or theor-
etical, reasoning and practical reasoning. Are they two sides of the same coin,
or different currencies? Manktelow considers a variety of approaches to this
fundamental question, and raises another: that there may be a further distinc-
tion to be made, as causal reasoning does not fit comfortably under either
heading.
There have been many centuries of argument over whether the intuitive
pure–practical distinction reflects a categorical distinction in human cog-
nition, and recent empirical studies have led to a number of modern authors
espousing that position. Two foregoing chapters here, by Over and by
O’Brien and colleagues, however, come down on the side of a unified
approach to reasoning. That being so, how is reasoning carried out in the
human mind? One answer is alluded to by O’Brien et al. – that people use a
kind of mental logic, allied to the pragmatic principles about which they go
into detail. The major alternative to this proposal is reviewed by Phil
Johnson-Laird, who traces the historical antecedents of the theory of mental
models. This is unquestionably the most influential theoretical proposal in the
field, as reflected in the volume of published research that appeals to the
notion of mental models, and the range of its applications. These began
outside the field of reasoning, and extend ever more widely within and
beyond its conventional boundaries; some flavour of this breadth of applica-
tion can be gained here. Johnson-Laird finds precedents in 19th-century
science, but locates the theory’s true ancestry in the work of the philosopher
C. S. Peirce, the Mozartish psychologist Kenneth Craik, and his contempor-
ary, the “Gestalt behaviourist” E. C. Tolman. Another measure of a theory’s
influence is, of course, the amount of criticism as well as support that it
attracts, and you will find many such statements here, most notably in the
chapters by Over and by Stenning and van Lambalgen.
Before the field’s preoccupation with conditionals, the selection task, and
other such delights, its territory was largely occupied by the classical syllo-
gism. This once again reflects its ancient antecedents, as the syllogism was
first systematized and extensively written about by Aristotle. Once again,
then, we can look for a line of inquiry through a long stretch of time, and this
is what Guy Politzer does. As with Johnson-Laird’s history of mental models,
6 Manktelow and Chung
Politzer finds a rich fossil record and concludes, rather challengingly for pres-
ent-day researchers, that the important psychological observations were
made about a century ago. And, as with mental models, one can go back even
further: students of the syllogism will already be familiar with the 18th cen-
tury creations of Venn and Euler, though perhaps less so, surprisingly, with
Leibniz’s pre-emption of such devices. One can of course go all the way back
to Aristotle, who made psychological as well as logical proposals whose
echoes can still be heard in modern theorizing.
Modern theorizing brings us to the third of these avowedly historical chap-
ters, this time focusing on a theory that has a more recent origin, but which,
like the theory of mental models, has had a wide influence on research into
reasoning. Jonathan Evans gives a personal historical account of the origins
and development of this theory, going back to his work with Peter Wason in
the 1970s. This is particularly apt given that Wason was responsible more
than any other individual for the way that the psychology of reasoning looks
today, and it is fitting that his direct as well as his indirect influence is repre-
sented here (other references to Wason’s own research can be found in the
chapters by Stenning and van Lambalgen and by Poletiek, and a number of
his former associates figure among this book’s contributors). The dual
process theory began life as an explanation for some curious selection task
findings, but has taken on much greater import in the years since then,
most extensively in its revisions resulting from Evans’ collaboration with
Over. Along the way, it has been influenced by ideas from other areas of
cognitive psychology, and interestingly has been paralleled by independent
developments of a similar kind, again in other areas.
Within the psychology of reasoning, dual process ideas inform the stance
of David Green, who reviews an approach to the study of reasoning that,
while sharing some features and interests with the kinds of research we have
been dealing with so far, goes into a different kind of territory: the use of
reasoning. Green’s concern is with argumentation, and as with so many areas
of reasoning research, it has some direct philosophical ancestry, this time in
the work of Stephen Toulmin. Green focuses on the dual representation of
arguments: their structure and the mental models that comprise them (again
showing the scope of influence of the theory of mental models). He also
ventures into terrain that is relatively less explored not only in most research
on reasoning, but in most of cognitive psychology: the interaction between
affect and cognition. He argues that both need to be considered if argumenta-
tion is to be understood. Anyone who has ever been involved in argument, no
matter how “reasoned” – that’s anyone – will testify that it can become
decidedly hot, and it is likely that the emotional context of reasoning will
come increasingly into focus in future.
Green’s chapter is an example of one that takes the psychology of reason-
ing into areas where it does not often go. This was an aim we stated at the
outset: to show that the study of reasoning is an enterprise that connects
with a wide range of fields of research into human mentality. Several of the
 The counterfactual character of thought 7
previous chapters (e.g., Chater and Oaksford, Over, Manktelow) have made
one connection in particular: that between reasoning and decision making.
Some (e.g., Doherty and Tweney, O’Brien et al., Green) have alluded to the
fact that thinking and reasoning often take place in a social context. Andrew
Colman reviews an area where such issues are thrown into sharp relief: psy-
chological game theory. As with the Brunswikian approach, this is a field of
study that does not loom large in much of the mainstream literature on
reasoning, but it should. Game theory is concerned with the kinds of think-
ing that go on between interdependent decision-makers, that is, when the
decisions taken by one party affect the decisions taken by others. As with so
many of the areas under review here, it has historical roots outside psych-
ology, but which have raised important psychological questions. Colman
deals with one in particular: that because of the actions of others, the “play-
ers” in a social interaction have incomplete control over the situation. This
profoundly affects the normative question of what each one should do. Thus,
while game theory poses important psychological questions, psychology has
equally unsettling effects on game theory.
What people should do in situations of social interaction is a many-layered
question, one that has been the central focus of theorists working in an
evolutionary tradition. This approach arrived in the psychology of reasoning
with a bang on the appearance of Leda Cosmides’ famous paper on the
application of evolutionary social contract theory to the explanation of
behaviour on the selection task in 1989, and has had an enormous impact.
Gary Brase reviews the evolutionary approach here, and it is interesting that
even in his title, the normative question is confronted immediately. Evo-
lutionary theorizing is driven by the empirical predictions made on the
assumption that people (and other animals) will tend to behave in ways that
reflect inclusive fitness, i.e., that increase the likelihood that they will survive
and pass on their genes. Brase argues that this perspective leads inevitably to
the idea that reasoning, along with all other parts of cognition, will be
domain-specific and, hence, modular. If one accepts this, there then follows a
debate about how these domains are arranged. Evolutionary theory provides
us with some signposts about this, too. This approach to cognitive theory has
been as controversial as it is influential, as Brase relates in his chapter and as
can be seen not only elsewhere in this book (e.g., Over and O’Brien et al.
reject the strict modularity that many evolutionary reasoning theorists
espouse) but in others, such as the recent volume edited by Over (2003).
Theoretical, historical, and philosophical ideas clearly run through the
study of reasoning like the veins in granite but, as we saw in the case of game
theory, the flow is not all one way. For instance, Stenning and van Lambalgen
base some of their argument on the philosophy of science; and of course
reasoning, particularly conditional reasoning, is a large part of scientific
thought. The influence has, as they contend, been mutual: ideas from the
philosophy of science have found their way into the design and interpretation
of reasoning research. Foremost among these has been the falsificationism of
8 Manktelow and Chung
Karl Popper. The chapter by Fenna Poletiek presents a close examination of
falsificationism, not this time in terms of its use by the selection task coterie,
but as a model for testing behaviour in general, both scientific and everyday.
As always, the closer one looks at something, the more complex it appears.
Falsification and confirmation turn out to be conjoined twins, in that the
success of an attempt at falsification is tied in intimately with how severe the
test is in the first place. If you give your hypothesis a very severe test, you go
in expecting a falsification, so when you achieve it, you have not learned very
much. Falsifications are more useful when confirmations are expected, and
vice versa. Thus hypothesis testing can be seen, as can conditional reasoning
in Over’s scheme, as a decision problem. Two-valued testing is too stringent
for the real world, both in the lab and outside it: one does not simply ask
whether a hypothesis is true or false, one thinks about the usefulness of a test
and the likelihood of its possible outcomes.
Science and life are alike in another complicating aspect as well, of course:
they are both inherently social activities. That much can also be taken from
the Brunswikian approach expounded by Doherty and Tweney, from the
pragmatic perspective of O’Brien and colleagues, from Stenning and van
Lambalgen’s semantically based examination of the progress (or lack of it) in
selection task research, and from the evolutionary and game-theoretic stand-
points on reasoning set out for us by Brase and Colman respectively. Green
explores how reasoning is used in argumentation, and Stenning and van
Lambalgen urge a sociological analysis of how reasoning researchers have
gone about their work. Something along those lines is provided by Sandy
Lovie, although his scope is rather wider; in fact, it is probably safe to say that
the kind of perspective he offers is not common in books on the psychology
of reasoning. Lovie acknowledges from the start the central roles of reason-
ing and argumentation in the doing of science, and directs us to the rhetorical
practice of science as a socially constructed activity. He echoes Poletiek’s
basic outline of the structure of hypothesis-testing situations – hypothesis,
test, interpretation – but rejects the idea of an experimental psychology of
science, arguing instead for psychologically informed but naturalistic studies.
Science, from this viewpoint, is a set of texts of belief and action in context,
with reasoning employed in the service of argumentation, persuasion, and
change in belief. Once again, there are reminders here of philosophical cross-
currents that appear elsewhere in the book, such as Toulmin’s views on the
structure of argument, and Harman’s idea that reasoning is about belief
revision.
Context is the theme that runs through all the contributions to this book.
Reasoning is an activity whose objects cannot be detached from their context,
be they syntactic, semantic, or pragmatic; which takes place in a context, be it
linguistic, interpersonal, biological, or cultural. The psychology of reasoning
itself takes place in context: methodological, theoretical, historical. In turn,
the psychology of reasoning provides findings and ideas that reflect back on
how we view everyday thinking, human rationality, the practice of science
 The counterfactual character of thought 9
and hence human culture in general. We hope that this volume will help you
to better appreciate these matters, and thank all the contributors for their
efforts in this.

REFERENCES

Anderson, J. R. (1978). Arguments concerning representations for mental imagery.

Psychological Review, 85, 249–277.
Cosmides, L. (1989). The logic of social exchange: Has natural selection shaped how
humans reason? Studies with the Wason selection task. Cognition, 31, 187–316.
Evans, J. St. B. T. (1982). The psychology of deductive reasoning. London: Routledge &
Kegan Paul.
Evans, J. St. B. T., Newstead, S. E., & Byrne, R. M. J. (1993). Human reasoning; the
psychology of deduction. Hove, UK: Lawrence Erlbaum Associates Ltd.
Garnham, A., & Oakhill, J. V. (1994). Thinking and reasoning. Oxford: Blackwell.
Gigerenzer, G., Todd, P., & the ABC Research Group. (1999). Simple heuristics that
make us smart. Oxford: Oxford University Press.
Harman, G. (1999). Reasoning, meaning and mind. Oxford: Oxford University Press.
Henle, M. (1978). Foreword. In R. Revlin & R. E. Mayer (Eds.), Human reasoning.
Washington, DC: Winston.
Manktelow, K. I. (1999). Reasoning and thinking. Hove, UK: Psychology Press.
Over, D. E. (Ed.). (2003). Evolution and the psychology of thinking: The debate. Hove,
UK: Psychology Press.
Wason, P. C. & Johnson-Laird, P. N. (1972). Psychology of reasoning: Structure and
content. London: Batsford.
2 Reasoning and task
environments
The Brunswikian approach
Michael E. Doherty and Ryan D.
Tweney

If there is anything that still ails psychology in general, and the psychology of
cognition specifically, it is the neglect of investigation of environmental or
ecological texture in favor of that of the texture of organismic structures and
processes.
(Brunswik, 1957, p. 6)

BRUNSWIK’S INTELLECTUAL HERITAGE

Egon Brunswik (1903–1955), born in Hungary but educated mostly in

Vienna, came of age as a psychologist in the Vienna Psychological Institute
founded by Karl and Charlotte Bühler. Karl Bühler was his primary mentor,
but Moritz Schlick, founder of the Vienna Circle, was also an important
influence. After completing his doctoral dissertation in 1927, with Bühler and
Schlick as his committee, Brunswik became laboratory assistant to Bühler,
assuming responsibility for the perception lab.
Brunswik’s functionalist orientation derived from Bühler and reflects a
very different set of influences than the American functionalism associated
with such figures as William James and John Dewey. Instead of a post-
Darwinian adaptationist perspective, Bühler’s position emphasized the pri-
macy of “intentionality”, a central concept in the act psychology of Franz
Brentano (Kurz & Tweney, 1997; Leary, 1987). Brentano, seeking a criterion
for what constituted mental phenomena, used this term for the referential
character of thought. Bühler, who had participated in the “Imageless
Thought” school of Külpe, Messer, Ach, and others (Humphrey, 1951),
combined Brentano’s intentionality with a phenomenological approach
(based loosely on Husserl; see Kusch, 1995) and argued that psychology must
establish the fundamental units of study in the intentional consciousness. In
spite of his phenomenological emphasis, however, Bühler found himself at
odds with the Gestalt psychologists, disagreeing in particular with their
extension of Gestalt perceptual laws to thinking.
By the time Brunswik arrived as a student, Bühler was emphasizing the
importance of “thoughts” as the fundamental units of experience. For
12 Doherty and Tweney
Bühler, this led to research on the nature of language (Bühler, 1934/1990),
and an emphasis on the social environmental context within which language
and thought developed (Bühler, 1930). Perception became Brunswik’s niche
within the overall Bühlerian framework and, encouraged by his mentor, he
sought to understand the perceptual constancies using an experimental
approach. If mental phenomena are “intentional” and object oriented, and if
the units of mind are “thoughts”, how can the gap be bridged between the
world of objects, the perceptions that “represent” those objects, and the
“thoughts” (or judgements, to use a modern term often associated with
Brunswik) that are the results of the process?
Another Bühler student, Karl Popper, received his PhD in 1928, one year
after Brunswik and also with Bühler and Schlick as his advisors (Kurz, 1996).
Like Popper, the young Brunswik cannot be classified as a “logical positivist”
in spite of the influence of Schlick and the Vienna Circle. For both, Bühler
had posed a fundamental issue that they saw as resolved too simplistically by
the logical positivists. No matter how one sought to put the formulations of
science on a sound logical and empirical footing, there remained the problem
of bridging the gap between the mental world of those formulations and
the world of objects. Brunswik sought the solution in psychology, whereas
Popper decided that psychology was a dead end for understanding science,
turning to inductivist formulations and a falsificationist account of science.
Popper did this only in the context of how scientists justify propositions. How
they discovered those propositions was, for Popper, a problem for “mere”
psychology (Popper, 1974; see also Kurz, 1996 for an account of Popper’s
rejection of the cognitive psychology he learned from Bühler).
For Brunswik in the 1920s and 1930s, the task at hand was to show how the
perceptual constancies could be investigated in a way that did justice to the
object orientation of the perceiver. Even his earliest work rejected the usual
experimental approach to such issues, which manipulated one variable at a
time while observing the effects on a dependent measure (usually a judgement
of magnitude). Instead, to understand size constancy, Brunswik used “multi-
dimensional psychophysics” (a term used by Brunswik, 1956a, to describe his
earlier work); he manipulated a multiplicity of variables simultaneously dur-
ing a size judgement task because isolating one variable in the laboratory
failed to reflect the environments in which organisms actually perceive. For
Brunswik, the most important outcome of this research was the finding that
although distal and proximal stimuli, which he came to call “cues”, were only
weakly related, and proximal cues and the final judgements were also only
weakly related, there was an extremely high correspondence between distal
stimuli and judgements. Brunswik referred to this apparent paradox as the
“perceptual achievement” of the organism, and the development of this
theme animated the rest of his career.
In 1933, Brunswik met a like-minded American who was visiting Vienna,
the young Edward Chace Tolman; each found something of value in the ideas
of the other, and their collaborative paper (Tolman & Brunswik, 1935) reveals
 The Brunswikian approach 13
the overlap between their views. The paper gave the first clear statement
of the causal texture of the environment, and developed a theme important in
the later work of both men, that the environmental cues available to an
organism may be equivocal. The central concepts of the paper, local represen-
tation, means-objects, and – especially – hypotheses, blend Tolman and
Brunswik, and establish the latter as neobehaviourist, like Tolman. Even so,
hindsight permits us to see that there are deep differences as well. Tolman’s
later work focused on central processes and how these mediate behaviour,
whereas Brunswik’s later work emphasized the relation between behaviour
and the distal world (Kurz & Tweney, 1997).
Following a reciprocal visit to Berkeley in 1935–36, Brunswik moved per-
manently to the University of California in 1937, just prior to the 1938 Nazi
takeover of Austria. Else Frenkel also left Vienna, and their marriage in 1938
brought her to Berkeley as well. The move to the US marked a turning point
in Brunswik’s career, and he soon began the process of adapting his views to
what, to him, must have seemed a very different intellectual milieu. In particu-
lar, he began to incorporate statistical views into his work. If behaviour was
mediated by probabilistic cues (see Gigerenzer, 1997), then it made sense to
conceptualize the organism as something of an intuitive statistician
(Gigerenzer & Murray, 1987). It was not long before Brunswik had developed
a first version of the lens model (see below) to capture this in formal terms.
In the late 1930s, American psychology was permeated by statistical
notions, one consequence of the heavy emphasis on predicting and control-
ling the aggregated behaviour of many individuals (Danziger, 1990). For
Brunswik, this meant that new tools were available – in particular, he was able
to use correlation to capture the inexact relation between cues and objects
(Brunswik, 1940). We will show how he did this in subsequent parts of this
chapter, but it is worth noting that, in one respect, Brunswik remained Euro-
pean in his use of statistical representations. By contrast to the prevailing
trend in America, Brunswik’s work never emphasized aggregation of data
across “subjects”, the newly coined term used to distance the researcher from
the object of study.
At Berkeley, Brunswik developed his theoretical system and the empirical
studies that supported it, moving beyond the reliance on perception that
characterized his Vienna research, and establishing his probabilistic func-
tionalism as an important school of psychological thought. Yet the austere
and intellectual Austrian, an uncompromising perfectionist, was often seen as
intimidating and, perhaps as a result, left few followers to carry on his work.
Among his American contemporaries, Brunswik’s use of correlational stat-
istics in an experimental context was novel – too novel, as it turned out.
When, shortly after the war, statistical inference techniques became the
touchstone of American psychology (Rucci & Tweney, 1980), Brunswik’s
approach began to be criticized for its use of correlation, which Hilgard
called an “instrument of the devil” (Hilgard, 1955, p. 228; see also Kurz &
Tweney, 1997). Nor was it merely a question of this or that statistical tool; as
14 Doherty and Tweney
we note below, the post-war emphasis on a methodological canon, first stated
by Woodworth (1938) as the rule of one variable, and the attendant reliance
on experimental designs that required strict separation and balancing of fac-
torial variables, made Brunswik’s emphasis on capturing the causal structure
of probabilistic environments seem wrong-headed to his peers.
It would be wrong to say that Brunswik was ignored by his American
contemporaries; his papers were published in the best journals, his views were
attacked by the most prominent experimentalists (see Hammond, 1966), and
he participated in some seminal conferences, including a 1955 meeting at
Colorado that is often cited as one of the precursors of the cognitive revolu-
tion (Brunswik, 1957). One early follower of Brunswik, Kenneth Hammond,
became a powerful advocate and extender of Brunswik’s ideas. Hammond
was among the first to use Brunswik’s lens model in judgement and decision-
making tasks, was the leading advocate of the Brunswikian approach known
as Social Judgement Theory (Doherty & Kurz, 1996), and the founding
father of the Brunswik Society. In large part because of Hammond’s
advocacy, Brunswik’s memory remains vivid.
Brunswik himself was not wedded to one statistical tool, nor to one meth-
odological approach. In particular, but for his premature death in 1955, we
might have seen him apply the new information theory of Claude Shannon
and the cybernetics of Norbert Wiener to the understanding of the probabil-
istic relations among cues and behaviour. He noted both of these in sympa-
thetic terms (Brunswik, 1952), and it is interesting to speculate on how he
might have used such approaches to capture uncertainty in a formal model.

THE METHODOLOGICAL POSTULATE OF

BEHAVIOUR–RESEARCH ISOMORPHISM

Brunswik had a deep interest in methodological issues. He believed that

empirical research would be wasteful and inefficient unless it was based on
methods appropriate to the investigation of psychological phenomena. The
heritage of European functionalism is reflected in his methodological postu-
late of behaviour–research isomorphism, which asserted that “the ‘order,’ or
pattern, of research ‘ideas,’ or design, should be the same as the pattern of
‘things’ studied, which for psychology is behaviour” (1952, p. 25).
The principle of behaviour–research isomorphism is a hard doctrine. It
says that research should focus wherever behaviour focuses. In espousing it,
Brunswik asserted that creating laboratory situations that lack the funda-
mental features an organism faces in its natural environment precludes
generalization beyond the specific task studied. Brunswik was advocating a
radical position, one antithetical to the Zeitgeist of his time. He saw that it
would entail both conceptual and practical difficulties, in that the features of
entities that were the foci of instrumental or perceptual behaviour were
inevitably correlated with the features of other entities in the environment.
 The Brunswikian approach 15
His emphasis on the correlational structure of the ecology had all too clear
implications for “the rule of one variable” (Woodworth, 1938) and for the
then burgeoning use of factorial design (Rucci & Tweney, 1980).

THE ENVIRONMENT HAS TEXTURE AND DEPTH

Brunswik maintained that psychology must investigate how the organism

adapts to and functions in its normal environment. This was not a unique
view, but Brunswik was virtually alone in his belief that for psychology to
take its proper place in the sciences, it must envision both the organism and
environment as systems in their own right, each with its own properties, each
with surface and depth, each with a rich texture (Brunswik, 1943, 1952). He
lamented what he saw as the encapsulation of psychology inside the head –
the concentration on understanding psychological processes either without
regard to the environment, or with a naive “constancy hypothesis” that pos-
tulated one-to-one relations between environmental events, proximal stimuli,
brain events, and psychological events.1
As an alternative, Brunswik proposed a “psychology without a subject”,
that is, he advocated investigating the properties of the “textural ecology as a
propaedeutic to functional psychology” (1956a, p. 119). Thus, Brunswik and
Kamiya (1953) carried out a study with an “n” of zero, in which they ana-
lyzed photographs to establish whether the environmental conditions for
learning Gestalt-like proximity relations were present. Brunswik and Kamiya
concluded that they were, and that learning was therefore a plausible
alternative to the innate processes postulated by the Gestaltists.
The premise that psychology must address the causal texture of the
environment was highly unusual in the historical context in which Brunswik
worked – and remains so today. The environment of interest to Brunswik was
not, of course, the environment as conceived by the physicist, but rather the
environment confronted by the organism, an environment with a proximal
surface, or “skin”, and extending in depth away from the organism in both
time and space. Yet it is not knowledge of the “skin” of the environment that
is relevant to survival; it is knowledge of things in depth, distal things rather
than proximal cues, that matters. And things in the world cannot, in principle,
be known with certainty because causes in the world scatter their effects
“semi-erratically”. Brunswik thus envisaged psychology in terms of two
characteristics of the environment: (1) “The environment has a causal
texture . . . in which different events are regularly dependent upon each
other” (Tolman & Brunswik, 1935, p. 43), and (2) “Such causal connections
are probably always to some degree equivocal, that is, not connected in simple
one-one, univocal . . . fashion, with the types of entities represented” (p. 44).
This implies that perceptual cues are partially intersubstitutable for one
another, a central concept for Brunswik. It appears throughout his work, in
the lens model and in the idea that vicarious functioning in the organism
16 Doherty and Tweney
mirrors the vicarious mediation in the organism’s ecology (see Brunswik,
1956a, p. 50, 1956b, p. 158).
Two of Brunswik’s papers used manipulations in a fashion especially con-
sistent with his conception that the environment was imperfectly knowable:
Probability as a determiner of rat behaviour (1939) and Probability learning of
perceptual cues in the establishment of a weight illusion (Brunswik & Herma,
1951). Each reflects his belief that the environment is, as far as the organism
can know, probabilistic. The very first sentence of the 1939 paper is “In the
natural environment of a living being, cues, means or pathways to a goal are
usually neither absolutely reliable nor absolutely wrong” (p. 175). Thus, he
varied the probabilities of reward on the two arms of a T-maze, the differ-
ences between the probabilities, the ratios of the probabilities, and which of
the two sides presented ambiguous information. In this, his only rat study and
one of the earliest studies of probability learning, Brunswik found that rats
could not only learn in a probabilistic environment, but could discriminate
between probabilities as well. He concluded that “the strength of response
appears to follow rather a kind of compromise” of tendencies, an idea deeply
entrenched in his thinking about perception, and closely related to the
concept of vicarious functioning (see also Doherty, 2001, p. 15).
Brunswik and Herma (1951) dealt with probability learning in perception.
They set up “an artificial miniature environment in which the relationship
between cue and referent was probabilistic rather than absolute, that is, in
which reinforcement is ‘partial’ . . . rather than of the conventional
unequivocal kind” (1951, p. 281). Björkman, in commenting on this paper,
attributes its significance to “the fact that it represents an early effort (the
experiments were conducted in Vienna in 1937) to construct a task that
mirrored important aspects of the environment” (2001, p. 170).

THE BASIC FACT OF PERCEPTION IS DISTAL FOCUSING

Brunswik maintained that all behaviour, including perceptual behaviour, has

a distal focus. It follows that behaviour is not concerned with (does not focus
on) immediate sensory experiences resulting from proximal stimuli, but
rather on things in the world that must be inferred from arrays of intercorrel-
ated cues. Research that has any chance of illuminating how organisms
achieve survival and success must involve intercorrelated cues to entities in
the environment, and must investigate behavioural repertoires relevant to that
achievement.

THING CONSTANCY MAKES HIGHER LIFE POSSIBLE

Brunswik held that cognitive structures paralleled whatever was important in

the environment. Since distal entities in the world maintain their identities in
 The Brunswikian approach 17
changing circumstances, they must be perceived by the organism as maintain-
ing those identities: “Thing constancy is nothing but the mechanism that
makes the behavioural environment conform to the geographic environment
to a considerable extent, especially in its biologically more distal aspects, thus
making higher life possible” (Brunswik, 1956a, p. 62). Thus, perceptual con-
stancies acted as “stabilization mechanisms”, providing the organism with a
stable external world in the face of constantly shifting proximal stimuli
(1956a, p. 47f). Brunswik saw a close analogy between perceptual constancy
and Piaget’s concept of conservation, in that both constancy and conserva-
tion provided the developing organism with a picture of things in the world as
unchanging in the face of changing appearances (Brunswik, 1944, p. 38; see
also Smedslund, 1966). Thus, whereas the experimental psychology of the
early 1900s focused on the detection or discrimination of proximate sensory
continua, Brunswik devoted much of his early research to the perceptual
constancies. We see a person who comes towards us from 20 feet away to 10
feet away as maintaining a constant height even as the retinal image doubles
in size. Clearly, the organismic response is more highly correlated with the
distal stimulus than with the proximal stimulus, in this case, the retinal image.
It is easy to dismiss perceptual constancy as a curiosity to which a sentence
or two is dedicated in introductory psychology textbooks. But constancy is
a truly fundamental achievement, one of a high order of complexity. Per-
ceptual constancy is an effect “of extreme biological importance to the
organism, since otherwise no orderly and self consistent ‘world’ of remote
manipulable ‘independents’ could be established” (Brunswik, 1937, p. 229).
Perceptual constancy epitomizes distal focusing, since the essence of the
constancies is that the same percept is achieved across different situations.
Under normal viewing circumstances, humans have an extraordinary cap-
acity to see things in their actual sizes. In Brunswik’s early work, in which he
analyzed the data using his “constancy ratio”, he typically found constancy
ratios for size perception near 1.0, indicating nearly perfect constancy. In a
later laboratory study using systematic design, he reported the correlation
between actual size and perceived size to be .97, whereas the correlation
between bodily size and proximal stimuli was only .10 and that between
proximal stimuli and perceptual response was just .26 (1940, p. 72). These
results are represented schematically in Figure 2.1.
Even so, size constancy is not perfect. Perception involves a “compromise
tendency between distal and proximal focusing” (1952, p. 15). This is a
specific instance of the general tendency in Brunswik’s thinking, noted above,
concerning the role of compromise in behaviour. In size constancy, he con-
sidered compromise “. . . as primordial as any theoretical principle in psych-
ology proper is ever apt to become” (1951, p. 215). Following Karl Bühler’s
“duplicity principle” that perception is determined not only by the focal
stimulus but also immediately and inextricably by the context, Brunswik
argued that context was not a “variable” that could be separated from the
organism, event, or focal stimulus. Instead, context was an inextricable part
18 Doherty and Tweney

Figure 2.1 Figure 1. Correlations among distal stimuli, proximal stimuli, and per-
ceptual responses. (After Brunswik, 1940.)

of the organism/environment system, part of the unit of analysis, and not a

separable causal influence on the unit of analysis.
Further, those errors reflected in constancy ratios less than 1 or correlation
coefficients less than 1.0 are better understood as errors towards something
rather than as away from some “correct” value. In discussing his research on
value constancy (with coins of varying values, sizes, and numerosities),
Brunswik noted that “To those of us who see in functionalism mainly an
avenue to the study of veridicality in perception the emphasis is shifted from
the negative aspect of value-induced error to the positive aspect of value
constancy” (1956a, p. 78). Thus, in a size constancy experiment, error is
directed to the proximal focus of intention rather than away from the distal
focus.2 The ability to know that distal entities maintain their identities despite
wildly varying proximal stimuli is a great cognitive achievement. This
achievement can be properly investigated only by varying the conditions of
experiments so that subjects can exercise and display it, which requires aban-
doning many uses of the “isolate a factor” method characteristic of current
psychological research.

VICARIOUS FUNCTIONING IS OF THE VERY ESSENCE

OF BEHAVIOUR

In his 1931 presidential address to the American Psychological Association,

Walter S. Hunter, in defining the subject matter of psychology, asserted that
“Typically, psychology deals with behaviour which although normally per-
formed by certain bodily structures in certain ways need not be so per-
formed” (Hunter, 1932, p. 18). Brunswik cited Hunter’s address several times
(e.g., 1952, p. 17; 1957, p. 9) and considered Hunter’s “vicarious functioning”
as “of the very essence of behaviour” (Brunswik, 1952, p. 92) and “the basic
definiens of all behaviour” (1957, p. 9). In effect, the organism can attain
 The Brunswikian approach 19
perceptual and instrumental goals in a variety of ways. Here, Brunswik found
common ground with his colleague and supporter, Edward Chace Tolman,
and with his critic, Clark Hull and his concept of habit family hierarchy. Note
that if vicarious functioning defines behaviour, then according to the
behaviour–research isomorphy principle, psychological research must allow
subjects to behave vicariously in attaining their goals. This consideration
led Brunswik to the best known of his methodological innovations,
representative design.

REPRESENTATIVE DESIGN IN THE STUDY OF DISTAL

ACHIEVEMENT

If we wish to make principled generalizations to environments that are

tangled webs of partial causes and partial effects, then we must represent the
essential features of those environments in our research. If we wish to under-
stand how people achieve accurate perceptions of the sizes of distal objects,
we must allow the person the opportunity to use the multiple mediation of
multiple proximal cues; we cannot learn how people perceive the sizes of
things by presenting one size cue at a time. Representative design, not factorial
design, is, according to Brunswik (1955a), the way to achieve these goals.
Representative design also requires situation sampling. If we wish to make
principled generalizations concerning how organisms behave towards stable
distal stimuli, the identities of which are mediated by a constant flux of
changes in proximal stimuli, then our research must investigate behaviour
across situations. Brunswik complained of a “double standard” (1956a, p. 39)
in the experimental research of his day. Investigators would assiduously
sample subjects from a carefully described population, present a single situ-
ation, then generalize over situations. The practice is not uncommon today,
even though a two groups t-test with 50 subjects per group may have 98 df for
populational generalization but 0 df for situational generalization.
The principle of behaviour–research isomorphism is a hard doctrine; so
too is representative design. Yet it is required by Brunswik’s conception of
psychological science: for an experiment to be truly representative, experi-
mental situations must reflect the essential features of the environment. But
how does an investigator decide what features are essential? Hochberg
pointed out that “representative sampling is no substitute for a theory of
what to look for” (1966, p. 369). For Brunswik, psychological research had to
begin with an ecological survey (Brunswik & Kamiya, 1953; see also Brehmer,
1984), after which representative environments might be instantiated in the
research. Alternatively, an investigator might randomly sample situations in
which measures of ecological success on the behaviour of interest could be
correlated with measured features of the environment. Sampling situations
randomly from the natural environment maintains both the substantive and
formal features of the environment. That is, the features of the sample of
20 Doherty and Tweney
situations that the subject confronts have about the same means, standard
deviations, and cue intercorrelations, i.e., the same causal texture, as the
population of situations to which the investigator wants to generalize.

“PSYCHOLOGY . . . HAS TO BECOME STATISTICAL

THROUGHOUT”

Brunswik argued that “psychology, as long as it wishes to deal with the vitally
relevant molar aspects of adjustment and achievement, has to become stat-
istical throughout, instead of being statistical where it seems hopeless to be
otherwise . . .” (1943, p. 262). In a 1940 paper, Thing constancy as measured by
correlation coefficients, Brunswik took a radical step, proposing that experi-
mental data should be analyzed by correlating responses with the stimuli. The
correlational approach and the very idea of probabilistic laws were heresies to
his contemporaries; Hilgard, as noted, expressed his animus for the idea by
referring to correlation as an “instrument of the devil”. For Brunswik,
though, there could be no deterministic laws in psychology. Brunswik’s
exhortation that psychology become statistical throughout flew in the face of
the orthodoxy of his (and, to a large extent, our own) day, the orthodoxy
enshrined in Woodworth’s 1938 Experimental psychology, the “Columbia
Bible”, which holds that the control of extraneous variables is a sine qua non
of psychological science.
Note that Brunswik’s position that can be contrasted with that of his other-
wise sympathetic contemporary, James J. Gibson. Gibson shared Brunswik’s
belief that an ecological approach was essential for psychology, but disliked
Brunswik’s emphasis on unreliable cues and on cues that could be discrepant
from one another. Instead, Gibson wanted to “concentrate on the theory of
those spatial perceptions for which the determinants are supplementary to
one another, not discrepant, and for which the stimulus conditions are opti-
mal rather than impoverished or inadequate” (1950, p. 150f). As Kirlik (2001)
noted, Gibson was uncomfortable with Brunswik’s emphasis on singular
objects, rather than field-like arrays, in his constancy experiments. Gibson’s
deterministic emphasis, suggests Kirlik, is the reason contemporary Gibsoni-
ans work on perception and action, whereas Brunswikians concentrate on
social judgement and decision making.
Note that to have the data necessary to compute correlation coefficients
such as those in Figure 2.1, there had to be multiple instances of the distal
stimulus, b, varying in size. This entailed multiple instances of p, the size of
the retinal image, and multiple responses, e, the subject’s estimates of the size
of the distal stimulus. The correlations for the subject’s data in Figure 2.1
illustrate the idea that perception is distally focused, since the correlation
between the distal stimulus and the perceptual response is far higher than the
correlations between distal and proximal stimulus and between proximal
stimulus and perceptual response.
 The Brunswikian approach 21
BRUNSWIK’S DISTINCTION BETWEEN PERCEPTION
AND THINKING

Brunswik sought an objective distinction between “perception” and “explicit

reasoning”. In a striking demonstration experiment, he had subjects judge
size under two conditions. In one, subjects judged the size of an 8 cm object
“in the laboratory, with the normal array of distance cues left intact” (1956a,
p. 91). This condition represented perception, which he conceived as
“uncertainty-geared” and characterized by multiple mediation with its
organic multiplicity of factors involving rivalries and compromises (see also
Brunswik, 1955b). The other condition represented thinking, which he
conceived as “certainty-geared”, and was characterized by single channel
mediation. Here, subjects were given the numerical information necessary to
make a precisely correct calculation of the size of the object. The results are
shown in Figure 2.2.
The perceptual judgements fall in a “compact and fairly normal frequency
distribution . . . with no particularly outstanding frequency of precisely cor-
rect answers” (Brunswik, 1956a, p. 91). The distribution of responses based
on explicit reasoning is radically different. “Almost half of all the answers . . .
are on-the-dot correct.” But look at the distribution of errors! One error
(64 cm) is an order of magnitude too high. Another (590 cm) is about two
orders of magnitude too high, and a third (4800 cm) is almost three orders of
magnitude too high! The secondary mode at 16 cm, which is the value of the
retinal equivalent, “demonstrates one of the typical pitfalls of reasoning,
namely, the going off in the wrong direction by being right about something
else.”
Brunswik argued that the multiple mediation characteristic of perception
“must be seen as chiefly responsible for the above noted relative infrequency
of precision. On the other hand, the organic multiplicity of factors entering
the process constitutes an effective safeguard against drastic error” (1956a,
p. 92). He went on,

So long as we accept, with Hunter (1932) and Tolman (1932), vicarious-

ness as the foremost objective criterion of behavioral purposiveness, per-
ception must appear as the more truly behaviour-like function when
compared with deductive reasoning with its machine-like, precariously
one-tracked, tight-roped modes of procedure. The constantly looming
catastrophes of the intellect would be found more often to develop into
catastrophes of action were it not for the mellowing effect of the darker,
more feeling-like and thus more dramatically convincing primordial
layers of cognitive adjustment.
(1956a, p. 93)

Brunswik also alluded in these pages to the “flash-like speed of perceptual

responses,” which he characterized as “a biologically very valuable feature,
Figure 2.2 Distributions of accuracy and error in perception vs thinking. (After Brunswik, 1956a.) Reproduced by permission of the University
of California Press.
 The Brunswikian approach 23
especially where . . . chances of success depend on quick action” (1956a,
p. 92). Brunswik’s distinction between perception and thinking illustrates his
contention that psychology must become statistical throughout. When it
came to showing the distinction objectively, he turned to an unequivocal
statistical demonstration, one between the two distributions, not just between
the means.

BRUNSWIK’S IDIOGRAPHIC STATISTICAL APPROACH

Brunswik’s 1944 monograph exemplifies his call for psychology to be stat-

istical throughout. At randomly selected times, a single subject estimated the
size of what in her surroundings was in her focus of attention. “The sur-
roundings . . . included scenes on the street and campus, in the laboratory, at
the desk, at home, and in the kitchen. An attempt was made to cover rec-
reation and study, daytime and evening (including periods of artificial lighting)
under conditions and in proportions representative of the daily routine”
(1944, p. 5). Subsequent physical measurements of the distal stimuli provided
the values necessary to assess achievement, and the data were analyzed by
way of correlation coefficients. The subject achieved extraordinarily high
levels of size constancy across a variety of situations, distances, and bodily
sizes. Note that this study had multiple situations but a single subject. It
involved an almost fully representative design that allowed for vicarious func-
tioning, i.e., multiple mediation of size by a variety of cues, but did not
directly assess multiple mediation.
How might Brunswik’s idiographic statistical approach be implemented
today? Consider an idealized study with a single subject. Assume that an
investigator has identified an environment important to the organism’s sur-
vival and success. That environment varies from situation to situation, hence
is a theoretically infinite population of environments. Assume that the
investigator has identified 10 features of the environment that are important
to a perceptual or instrumental goal of the organism. The investigator has
randomly sampled 100 instances of situations from the population of
environments to which he or she wishes to generalize and confronted a sub-
ject with those 100 samples. The subject has responded to each one. Hence,
the investigator has 100 measurements of the 10 features of potential eco-
logical relevance to the organism, of the distal goal to be attained, and of the
response to each situation. How can one make sense of the mass of data that
results, even with a single subject? That is where the lens model comes into play.

THE LENS MODEL

The lens model is a simple, elegant representation that captures the essential
features of representative design, including vicarious mediation, vicarious
24 Doherty and Tweney
functioning, the duplicity principle, cue intercorrelation, and the distal focus-
ing of perception. We will present two versions of the lens model – the
first, Figure 2.3, from Brunswik (1952) and the second, Figure 2.4, a current
version with the typical statistical indices shown.
The initial focal variable is the ecological object of either perception or an
overt behavioural act. We describe the lens model in terms of perception. The
initial focal variable scatters its effects semi-erratically. The cues, or proximal
stimuli, come from the same initial focal variable, so they are intercorrelated
with and intersubstitutable one for another, and can provide useful
redundancy.
The multiple rays from the initial focal variable thus represent vicarious
mediation, the re-collection of those rays into the terminal focal variable
represents vicarious functioning. The heavy, wide arc at the top of the figure
shows the relation between the organismic response and the ecological goal.
That is, it measures achievement. The stray causes and stray effects arrows
reflect the idea that the environment is complex and imperfectly knowable,
and the feedback arrow indicates that the “lens patterns do not stand in
isolation but are apt to reflect back upon the organism in a future state . . .
such as when food is followed by satiation and reinforcement of the preceding
behaviour . . .” (1952, p. 20).

Figure 2.3 The lens model. (After Brunswik, 1952.) Reproduced with the permission
of The University of Chicago Press.
 The Brunswikian approach 25

Figure 2.4 A contemporary lens model showing the typical notation for the indices
that constitute an idiographic-statistical analysis.

A contemporary lens model, due essentially to the elaboration of

Brunswik’s system by Kenneth Hammond, is shown in Figure 2.4. The con-
temporary lens model is the conceptual and statistical framework for what
has become “lens model methodology” (Doherty & Kurz, 1996). Consider
how the lens model might be used to study diagnostic thinking. Suppose we
are interested in assessing the diagnostic skills of a physician. Clearly we
cannot make valid generalizations about his or her competence from a single
case; there may be something unique about that case. We must observe the
physician’s behaviour in multiple cases sampled from the domain of cases to
which we wish to generalize. Consider a physician who has made 50 diag-
nostic judgements (Ys) of the presence or absence of a particular disease on
the 50 cases, each case represented by 7 potential symptoms, or cues, (Xi).
Subsequently, we ascertain the true ecological criterion value (Ye) for each
case, perhaps via post mortems. The physician’s achievement is assessed by ra,
the correlation between the diagnoses and the criterion measures over the 50
cases.
The multiple correlation, Rs, which represents the degree to which the
person’s judgements are predictable, assuming a linear, additive model,
should ideally be 1.0. Typically, linearity and additivity are assumed, though
they need not be. The cue utilization coefficients (ris) ought to match the
ecological validities (rie), both of which are zero-order correlations. (The term
26 Doherty and Tweney
“ecological validity” was, for Brunswik, a technical term meaning the correl-
ation between a cue and distal stimulus, not the degree to which the experi-
ment might be said to match some unspecified reality.) The G coefficient, the
zero-order correlation between the predicted values of the two linear models,
can be interpreted as the validity of the person’s knowledge of the linear
additive components of the environment. The C index, the zero-order correl-
ation between the residuals from the two models, reflects the extent to which
unmodelled aspects of the person’s knowledge match unmodelled aspects
of the environmental side of the lens, and can be partly interpreted as the
validity of the person’s configural knowledge.
All of the lens model indices are conceptually and statistically interrelated.
The degrees to which rie and ris are appropriate measures of relationship are
influenced by whether we have specified the correct function forms and by the
real strength of relationship. The degree to which the rie and ris are ordinally
related determines G. Re and Rs are determined by how predictable the
environment and the person are, and by how well we have specified the true
model of each. Finally, achievement is influenced by all of the above, as
expressed in Tucker’s (1964) lens model equation:

ra = Re * Rs * G + C [(1 − R2e) * (1 − R2s)] ½.

In words, a person’s achievement in predicting the world is determined by

how predictable the world is, given the available data (Re), how consistently
the person uses those data (Rs), and how well the person understands the
relevant causal texture of the environment (G and C).
The statistical indices that derive from the idiographic analyses can be
tested for statistical significance. The domain of generalizability would be the
physician-subject who had been confronted with the sample of situations,
and the generalization would be to the population of situations from which
the sample had been drawn. Of course, an investigation can include multiple
individuals, the idiographic indices can themselves be subjected to nomo-
thetic analyses, and generalizations across people as well as situations can be
made (Schilling & Hogge (2001). Brunswik argued that the study of the
environment is propaedeutic to the study of the organism. We might add
that the study of the individual is propaedeutic to the study of individuals in
general.
The lens model reflects many aspects of Brunswik’s thinking, including the
focus on achievement, vicarious mediation, and vicarious functioning, and
the idea that correlation coefficients can be used to assess not only achieve-
ment but also the micro-mediational aspects of the organism’s interaction
with the environment. Perhaps the aspect of the lens model most directly
reflecting the core of Brunswik’s thinking is its symmetry, with the two sides,
the environment and the person, being mirror images assessed with the same
statistical tools.
 The Brunswikian approach 27
Can a regression equation be considered a process model? Brunswik was
aware of how multiple correlation might be used to explain how multiple cues
of low ecological validities (rie) could be utilized (ris) to arrive at high levels of
functional validity (ra). Referring to Else Frenkel-Brunswik’s use of multiple
correlation in her analysis of “clinical intuition”, he wrote that “Prediction of
response even from low utilization cues could be built up by such a procedure.
And the same procedure could be used symmetrically for ecological validity
so that eventually an understanding of the mechanism of the super ordinate
functional achievement arcs . . . could be obtained” (Brunswik, 1956a, p. 110).
Hammond and Stewart (2001b) interpret this as showing that Brunswik con-
sidered multiple regression equations as “simulating” but not “duplicating”
the form of cognitive activity that Brunswik termed quasirationality. For an
extended discussion, see Doherty and Brehmer (1997).

CONTEMPORARY THEORY IN THE BRUNSWIKIAN

TRADITION

Two main theoretical developments have built upon Brunswik’s contributions.

One is Cognitive Continuum Theory. The other derives from the study of fast
and frugal heuristics, a programme referred to as “the adaptive toolbox”.

Cognitive continuum theory

There is growing consensus that thinking is not a unitary phenomenon, sus-
ceptible of easy definition. There are variations on the theme, but the idea
that there is a difference between analytic and intuitive thinking has been
advanced by a variety of theorists (Doherty, 2003; Evans & Over, 1996). The
long history of the distinction between intuition and analysis is traced by
Hammond (1996a), who explored the behavioural markers of intuition vs
analysis and the task conditions that elicit each. Hammond’s Cognitive Con-
tinuum Theory holds that thinking lies on a continuum from intuition at one
pole to analysis at the other, with most human cognitive activity being
between the two, a type of cognition to which Hammond applied Brunswik’s
term “quasirationality”. Hammond’s postulate that most thinking is quasi-
rational, a compromise between analysis and intuition, recalls Brunswik’s
emphasis on compromise. For an empirical assessment of Cognitive Con-
tinuum Theory and a further explication of the concepts of surface and depth
see Hammond, Hamm, Grassia, and Pearson (1997).
A related contrast (Cooksey, 1996; Hammond, 1996a) is that between cor-
respondence and coherence criteria for success. Correspondence is the same
idea as Brunswik’s achievement, ra in the lens model. It is the functionalist’s
criterion of success or failure in interactions with the world. Coherence, by
contrast, is the degree to which behaviour can be shown to be consistent
with some normative model, such as modus tollens, Bayes’ theorem, or the
28 Doherty and Tweney
conjunction rule in probability theory. Researchers who adopt a coherence
criterion are likely, because of the need to have a normative model against
which to assess the behaviour studied, to investigate tasks that elicit analytic
thinking. Correspondence theorists, especially those of a Brunswikian orien-
tation, are likely to investigate tasks that confront the subject with multiple,
probabilistic cues that elicit quasirational thinking. It is with respect to cor-
respondence vs coherence criteria that Brunswikian approaches to research
differ most sharply from the mainstream of thinking research.

Fast and frugal heuristics: The adaptive toolbox

Gerd Gigerenzer and his colleagues at the Max Planck Institute, (the Adap-
tive Behaviour and Cognition, or ABC, group), have developed a broad
theoretical programme based on the premise that the organism has a set of
simple heuristics that have evolved to achieve success in the environment. A
favoured metaphor of the group is Simon’s (1967/1996) description of
bounded rationality as a pair of scissors, one blade being the cognitive limita-
tions of the person, the other being the structure of the environment. “Minds
with limited time, knowledge, and other resources can be nevertheless suc-
cessful by exploiting structures in the environment” (Gigerenzer & Selten,
2001, p. 2).
In traditional lens model research, the term “intersubstitutability” implies
weighting and averaging, typically implemented with multiple regression
analysis, a compensatory model that processes all available information (but
see Hammond, 1996b). Gigerenzer’s fast and frugal models use a more literal
meaning of intersubstitutability, assuming that people draw inferences in
many circumstances by using a single cue at a time. The ABC approach
follows Brunswik in positing that the environment has a structure such that
multiple cues are probabilistically related to the distal criterion, and con-
ceptualizes that uncertainty in terms of conditional probabilities. Their
research programme has focused, in part, on discovering the cognitive heur-
istics that people use in dealing with such probabilistic cues, thus extending
Simon’s “bounded rationality” to an “adaptive rationality”. The whole
theoretical “adaptive toolbox” of special purpose heuristics proposed by
Gigerenzer and his colleagues is accessible in Gigerenzer (2000), Gigerenzer
and Todd (1999), and Gigerenzer and Selten (2001). For a brief treatment, see
Gigerenzer and Kurz (2001).

CONTEMPORARY EMPIRICAL RESEARCH IN THE

BRUNSWIKIAN TRADITION

Empirical investigations in this tradition range over many areas, including

applied contexts. We considered as “Brunswikian” studies that included the
following characteristics: (1) the subject is confronted with multiple samples
 The Brunswikian approach 29
from some population of uncertain environments; (2) the judgements made
by the subject have a many–one character, i.e., vicarious functioning is pos-
sible; (3) the investigator has attempted to model and represent the causal
texture of the environment.

Multiple Cue Probability Learning (MCPL)

In an MCPL study, a person learns a set of probabilistic cues (Xi in lens
model notation) to some criterion (Ye), then uses those cues in predicting that
criterion. Hence, it is an investigation of how people think when combining
cues. MCPL tasks typically involve a subject confronting an environment
with several possibly intercorrelated cues with varying relationships to the
criterion (i.e., the environment has a causal texture). The number of cues can
be manipulated, as can the forms of the functions relating cues and criterion,
whether the integration rule is additive or configural, and so forth. A subject
is presented several blocks of trials, perhaps 25 per block, with each trial
involving the presentation of the cues (Xi), a prediction (Ys) of the criterion
by the subject, then the presentation of the criterion value (Ye). The latter
operation is called outcome feedback. On the first trial of the first block, the
subject’s prediction is a guess. From then on, the subject takes account of
the feedback and can discover, by inductive reasoning, the causal texture of
the environment. The subject’s performance is described by the lens model
equation.
The question of whether subjects could even learn such probabilistic re-
lationships was answered early on. They can. In relatively simple tasks, sub-
jects’ regression models at asymptotic performance levels are highly similar to
those of the environment. We now know that people bring to the task an a
priori hierarchy of function forms relating the cues to the criterion, from
positive linear, to negative linear, inverse U, and finally U shaped (Brehmer,
1974). People have great difficulty learning if the function forms are nonlin-
ear, but can profit greatly from “cognitive feedback”, which might include a
description of the nonlinearities and nonaddivities in the environment
(Balzer, Doherty, & O’Connor, 1989; Doherty & Balzer, 1988). Empirical
research using the MCPL paradigm peaked in the mid 1970s (Holzworth,
2001), and performance in the paradigm is so well understood that it has been
used as a tool to study such diverse phenomena as the cognitive effects of
psychoactive drugs (Gillis, 2001; Hammond & Joyce, 1975) and cognitive
changes with ageing (Chasseigne & Mullet, 2001).

The single system case: Judgement analysis3

The MCPL paradigm directly incorporates two systems, the environment and
the organism, represented by the two sides of the lens model. Judgement
analysis directly involves just one, the organism, typically in the study of
expert judgement. A single system paradigm may seem inconsistent with
30 Doherty and Tweney
Brunswik’s fundamental approach, but the inconsistency is more apparent
than real, because the investigator analyzes and represents the environment
prior to the investigation. Judgement analysis entails having the subject judge
multiple instances of multiattribute profiles. As in MCPL research, the
default analysis is multiple regression. Judgement analysis is done when
actual criterion data are unavailable or not of relevance to the research issue
at hand (Cooksey, 1996).
For example, Ullman and Doherty (1984) used judgement analysis to
investigate the diagnosis of hyperactivity. In Study 1, 11 mental health profes-
sionals made diagnostic judgements on 52 seven-cue profiles of boys, some of
whom had been referred to the university clinic for assessment for hyperactiv-
ity. The individual differences in diagnoses of the same children were so great
that the chapter reporting the research was called Two determinants of the
diagnosis of hyperactivity: The child and the clinician.
Many of the subjects in Study 1 indicated that diagnoses should be based
on more information, and listed the additional information they would have
wanted. Consequently, Study 2 presented the diagnosticians with 95 hypo-
thetical boys described in terms of 19 attributes. There were 74 subjects: 22
clinical and school psychologists, 15 paediatricians and child psychiatrists, 22
teachers and special educators, and 15 other mental health workers. The
individual differences in cue usage and judgemental base rates were even
greater than in Study 1, and swamped whatever group differences there may
have been. The median interjudge correlation among diagnosticians was only
.51. In both studies, the judges tended to use far less information than
their subjective reports indicated, a commonly found mismatch between
subjectively and statistically assessed measures of cue usage.
The individual differences found in Ullman and Doherty (1984) are com-
monly observed in judgement research. One source of those differences is
that expert judgement is needed precisely in those domains where objective
criteria are absent, domains such as psychodiagnosis, social policy formation,
and employment interviewing. A second is that investigators tend to be inter-
ested in just those situations in which people will be expected to disagree. The
evidence suggests, however, that the overwhelming individual differences in
judgement research reflect real individual differences, which means that the
best way to investigate expert judgement is idiographically. Otherwise,
interesting individual differences are hidden in the error terms of group
comparisons.
Judgement analysis studies with experts have routinely found that additive
models account for virtually all of the predictable variance in the judgement.
Configural (or, in ANOVA terminology, interaction) terms have repeatedly
been found unnecessary. Part of the reason is statistical, in that the outputs of
different models of the same dataset tend to be highly correlated, and
parsimony dictates that we adopt the simpler model. A more powerful
reason lies again in the selection of cues. Suppose two investigators are inter-
ested in investigating the factors influencing the choice of marital partners.
 The Brunswikian approach 31
Investigator A uses a single set of multiple cue descriptions of potential mates
to be used for both males and females; B uses separate sets of male and
female profiles to be judged by female and male subjects, respectively. Serving
as a subject for Investigator A, a heterosexual male would give a low rating to
all male profiles without even looking at the other attributes. Conversely, this
same subject, noting that the profile described a female, would examine the
other cues and rate the profile accordingly. This is a configural policy, in that
the impacts of the nongender cues depend on the level of the gender cue. To
Investigator B, the same subject would appear additive. The difference shows
up nicely in the scatterplots of the ratings against the cue values. Again,
decisions made in good faith by experimenters who want their research to
produce knowledge may lead to knowledge that is true in the laboratory, but
leads to a serious overgeneralization beyond it. It is ironic that this occurs in
research with Brunswikian origins, given Brunswik’s insistence on representa-
tive design! Nevertheless, even when people describe their judgement polices
as configural, additive models often predict their judgements remarkably
well. It appears that people are capable of configural thinking, but claim to be
thinking configurally even when additive models suggest otherwise.
Judgement analysis and extensions of the basic logic of the single system
design to multiple systems have been used with a wide variety of issues,
including the judgement of what police handgun ammunition would be
acceptable to the diverse constituencies in Denver, Colorado (Hammond &
Adelman, 1976), social perception (Funder, 2001; Gillis & Bernieri, 2001),
highway design (Hammond, et al., 1997), medical ethics (Smith & Wigton,
1988), medical judgement (Wigton, 1988, 1996), fairness in faculty salar-
ies (Roose & Doherty, 1978) and in graduate school admissions (Maniscalco,
Doherty, & Ullman, 1980), scientists’ assessment of the health risk associated
with plutonium (Hammond & Marvin, 1981), the potential contribution of
mediation to conflict resolution (Holzworth, 1983), self insight (Reilly &
Doherty, 1989, 1992), and conflict resolution (Rohrbaugh, 1997). For sum-
maries see Brehmer and Brehmer (1988) and Hammond and Stewart (2001a).
Many applications of judgement analysis involve decision making, e.g.,
Hammond and Adelman’s (1976) “Denver bullet study”, which artfully sep-
arated facts and values, so it might be useful to point out some of the distinc-
tions between judgement analysis and a procedure known as decision analysis
(e.g., von Winterfeldt & Edwards, 1986).
A decision analysis typically calls for a stakeholder to consider at length a
single, complex problem, to decompose it into a set of prescribed categories
(options for action, states of the world, attributes, probabilities, utilities),
then aggregate the resulting values according to an algorithm that maximizes
subjective expected utility. The externalization and decomposition of the
problem are accomplished by analytical thought processes by the stakeholder,
often with the aid of consultants.
In judgement analysis, a stakeholder makes holistic judgements on a large
number of instances of the same sort of complex, multiattribute problem
32 Doherty and Tweney
dealt with by the decision analyst. The decomposition is made via multivari-
ate statistical analysis on a computer. Thus, judgement analysis is more
attuned to the study of the sort of quasirational thinking that is a mix of
intuitive and analytic processes, whereas decision analysis is supposed to be
just that, analytic.

Probabilistic Mental Models (PMMs)

Gigerenzer and the ABC group focus on the basic cognitive processes
involved in multiattribute judgement, We take as an example of their research
one application from the adaptive toolbox, the “Take-The-Best” algorithm,
which Gigerenzer, Hoffrage, and Kleinbölting (1991) proposed as a model of
how people answer general knowledge questions. The task they presented
German subjects was the choice of which of two randomly selected German
cities was larger. PMM theory assumes that subjects have information avail-
able in memory concerning each city, ranging from not even having heard of
the city to having a large number of cues of varying cue validities. These cue
validities, coded as conditional probabilities rather than correlation coef-
ficients, are analogous to cue utilization coefficients (ris) in the lens model. In
judgement analysis, the cues are presented to subjects in a profile format; in
Take-The-Best, only the city names are presented; the cues are in memory.
Gigerenzer et al. illustrated the Take-The-Best approach with the question,
“Which city has more inhabitants? (a) Heidelberg or (b) Bonn?”. The first
step in the algorithm is the construction of a local mental model (MM), i.e.,
whether the subject either knows the answer or has information from which
the answer can be deduced with certainty. If a local MM cannot be brought
to bear on the question, as is often the case, then a PMM is constructed that
uses all relevant information known about the two cities. The cue with the
highest cue validity might be as simple as whether the subject recognizes the
name of one of the cities but not the other, the so-called recognition cue.
Larger cities are mentioned more often in the environment, on the average,
than smaller ones. Therefore recognition is correlated with population size,
and is an ecologically valid cue. A person who is well adapted to the environ-
ment (in this regard) can therefore use recognition as a cue to size, but only if
only one city name is recognized. If so, information search in memory is
terminated and the decision is made based on this one best cue (hence the
name, Take-The-Best). If both names are recognized, then the recognition
cue cannot be used, and the cue of the next highest ecological validity is
activated.
The cue with the next highest validity, say whether the city has a team in the
German soccer league, is then activated and a comparison made. If this cue
discriminates between Heidelberg and Bonn, the choice is made and the
information search in memory is terminated. But neither of the cities has
such a team, so the subject spontaneously drops down the cue hierarchy to
the next best cue, and so forth. The postulated process proceeds until a cue
 The Brunswikian approach 33
discriminates or until a guess must be made. PMM also accounts for con-
fidence in judgement, with the confidence judgement being determined by the
cue validity of the discriminating cue.4
Social Judgement Theory and PMM may seem starkly different, but they
are similar in deep ways. Both are in the Brunswikian tradition of probabil-
istic functionalism that focuses on cognitive successes in real environments,
and for both, vicarious functioning is an essential feature of thinking. Both
heed Brunswik’s call for “the investigation of environmental or ecological
texture”.

A RECAPITULATION: BRUNSWIK IN THE 20TH CENTURY

Brunswik was concerned with everyday success in the world, a success medi-
ated by perception and thinking. In his eyes, thinking, as well as perception,
was distally focused; both thinking and perception were part of “psychology
as a science of objective relations” (Brunswik, 1937). A fundamental idea of
Brunswik’s was that to understand a cognitive process one must understand
the environment confronted by that process. That idea derives from
Brunswik’s functionalist orientation, and leads in turn to his insistence on
probabilism. These ideas had, for Brunswik, inescapable methodological
implications, including the principle of behaviour–research isomorphism,
representative design, situation sampling, and his idiographic-statistical
approach.

NEW DIRECTIONS: BRUNSWIK IN THE 21ST CENTURY

How are Brunswik’s influences on the investigation of thinking and reason-

ing likely to play out in the future? To approach this question we note, first of
all, the very wide scope of research now being carried out by those who
consider themselves Brunswikians. Thus, studies of team decision making,
interpersonal perception, the role of cognitive feedback in medical education,
the relation between physicians’ diagnostic judgements and treatment
choices, and methodological studies modifying and extending representative
design; these and many others appeared on the programme of a recent meet-
ing of the Brunswik Society. In addition, Hammond (2000) has recently
extended the concept of constancy, so fundamental to our understanding of
perception, to play a central role in a theory of stress.
Second, we note that our own research, over the years, has shown a per-
vasive underlying Brunswikian theme – even when we were not aware of such
a relation! Thus, together and separately, we have had a long-standing interest
in the nature of scientific thinking. With hindsight, we see now that scientific
thinking is a compelling case of distal focus. Scientists study a distal world
inaccessible to perception and everyday thinking; their task, in Brunswik’s
34 Doherty and Tweney
terms, is to determine the relevant proximal cues, to assess their ecological
validity, and to develop an explicit “lens model” (that is, a theory).
In science, vicarious functioning becomes deliberate and conscious rather
than intuitive, but the processes by which this is achieved rely as much on
intuitive thinking (analogies, metaphors, hunches) as on analytical thinking
(deductive thinking, explicit mathematical reasoning). Brunswik’s approach
makes perfect sense as a framework for the study of science. In contrast
to Popper (another student of Bühler’s), such an approach is not “mere”
psychology, of little or no interest to the true understanding of science!
Some of our own research has involved the kind of stripped-down labora-
tory simulations of scientific thinking introduced by Peter Wason in his 2–4–6
and selection tasks (1960, 1968; see, e.g., Tweney et al., 1980). In other
research (e.g., Mynatt, Doherty, & Tweney, 1978) we used a more complex
computer-based environment, a so-called artificial universe, asking a small
number of subjects to conduct experiments in order to understand its laws.
Since we designed the artificial universe, we understood the environmental
laws and could assess the strategies used by the subjects as they sought to
unravel its complexity by sampling different parts of the environment. Our
idiographic descriptions of each subject were based on the portions of the
environment that each subject had actually confronted. By matching draw-
ings of the relevant portions with appropriate sections of taped think-aloud
protocols, we were able to uncover many of the heuristics used by subjects
and to relate this to what Brunswik would call the “achievement” of each
subject. In other research, we extended the artificial universe paradigm by
using probabilistic scenarios, that is, scenarios that met Brunswik’s suppos-
ition about the actual universe, that its causal texture was proximally semi-
erratic (e.g., O’Connor, Doherty, & Tweney, 1989). And, using traditional
experimental designs, we explored the effect of varying base rates, varying
semantic context, and varying task instruction on subjects’ use of probabil-
istic information in inference tasks (e.g., Doherty, Chadwick, Garavan, Barr,
& Mynatt, 1996).
In recent years, others have used an “artificial universe” approach to bridge
the gap between the artificiality of the usual laboratory task and the complex-
ity of real-world scientific thinking. Thus, David Klahr and his students used
a programmable robot toy, “Big Trak”, in a series of studies in which subjects
had to determine experimentally the function of an obscure command within
the otherwise simple programming language of the device (Klahr, 2000).
Klahr showed that subjects searched separate “Experiment Spaces” and
“Hypothesis Spaces” and described characteristic developmental differences
in how such searches were carried out. Kevin Dunbar (1995) used a com-
puterized model of gene interaction to conduct similar studies, then extended
his inquiries by examining interactions among scientists in four molecular
biology laboratories.
The use of “real” science as the object of study also characterizes studies
of the work of Michael Faraday, whose extensive laboratory notebooks have
Exploring the Variety of Random
Documents with Different Content
1.C. The Project Gutenberg Literary Archive Foundation (“the
Foundation” or PGLAF), owns a compilation copyright in the
collection of Project Gutenberg™ electronic works. Nearly all the
individual works in the collection are in the public domain in the
United States. If an individual work is unprotected by copyright law
in the United States and you are located in the United States, we do
not claim a right to prevent you from copying, distributing,
performing, displaying or creating derivative works based on the
work as long as all references to Project Gutenberg are removed. Of
course, we hope that you will support the Project Gutenberg™
mission of promoting free access to electronic works by freely
sharing Project Gutenberg™ works in compliance with the terms of
this agreement for keeping the Project Gutenberg™ name associated
with the work. You can easily comply with the terms of this
agreement by keeping this work in the same format with its attached
full Project Gutenberg™ License when you share it without charge
with others.

1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside the
United States, check the laws of your country in addition to the
terms of this agreement before downloading, copying, displaying,
performing, distributing or creating derivative works based on this
work or any other Project Gutenberg™ work. The Foundation makes
no representations concerning the copyright status of any work in
any country other than the United States.

1.E. Unless you have removed all references to Project Gutenberg:

1.E.1. The following sentence, with active links to, or other

immediate access to, the full Project Gutenberg™ License must
appear prominently whenever any copy of a Project Gutenberg™
work (any work on which the phrase “Project Gutenberg” appears,
or with which the phrase “Project Gutenberg” is associated) is
accessed, displayed, performed, viewed, copied or distributed:
 This eBook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this eBook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.

1.E.2. If an individual Project Gutenberg™ electronic work is derived

from texts not protected by U.S. copyright law (does not contain a
notice indicating that it is posted with permission of the copyright
holder), the work can be copied and distributed to anyone in the
United States without paying any fees or charges. If you are
redistributing or providing access to a work with the phrase “Project
Gutenberg” associated with or appearing on the work, you must
comply either with the requirements of paragraphs 1.E.1 through
1.E.7 or obtain permission for the use of the work and the Project
Gutenberg™ trademark as set forth in paragraphs 1.E.8 or 1.E.9.

1.E.3. If an individual Project Gutenberg™ electronic work is posted

with the permission of the copyright holder, your use and distribution
must comply with both paragraphs 1.E.1 through 1.E.7 and any
additional terms imposed by the copyright holder. Additional terms
will be linked to the Project Gutenberg™ License for all works posted
with the permission of the copyright holder found at the beginning
of this work.

1.E.4. Do not unlink or detach or remove the full Project

Gutenberg™ License terms from this work, or any files containing a
part of this work or any other work associated with Project
Gutenberg™.

1.E.5. Do not copy, display, perform, distribute or redistribute this

electronic work, or any part of this electronic work, without
prominently displaying the sentence set forth in paragraph 1.E.1
with active links or immediate access to the full terms of the Project
Gutenberg™ License.

1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if you
provide access to or distribute copies of a Project Gutenberg™ work
in a format other than “Plain Vanilla ASCII” or other format used in
the official version posted on the official Project Gutenberg™ website
(www.gutenberg.org), you must, at no additional cost, fee or
expense to the user, provide a copy, a means of exporting a copy, or
a means of obtaining a copy upon request, of the work in its original
“Plain Vanilla ASCII” or other form. Any alternate format must
include the full Project Gutenberg™ License as specified in
paragraph 1.E.1.

1.E.7. Do not charge a fee for access to, viewing, displaying,

performing, copying or distributing any Project Gutenberg™ works
unless you comply with paragraph 1.E.8 or 1.E.9.

1.E.8. You may charge a reasonable fee for copies of or providing

access to or distributing Project Gutenberg™ electronic works
provided that:

• You pay a royalty fee of 20% of the gross profits you derive
from the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
 about donations to the Project Gutenberg Literary Archive
Foundation.”

• You provide a full refund of any money paid by a user who

notifies you in writing (or by e-mail) within 30 days of receipt
that s/he does not agree to the terms of the full Project
Gutenberg™ License. You must require such a user to return or
destroy all copies of the works possessed in a physical medium
and discontinue all use of and all access to other copies of
Project Gutenberg™ works.

• You provide, in accordance with paragraph 1.F.3, a full refund of

any money paid for a work or a replacement copy, if a defect in
the electronic work is discovered and reported to you within 90
days of receipt of the work.

• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.

1.E.9. If you wish to charge a fee or distribute a Project Gutenberg™

electronic work or group of works on different terms than are set
forth in this agreement, you must obtain permission in writing from
the Project Gutenberg Literary Archive Foundation, the manager of
the Project Gutenberg™ trademark. Contact the Foundation as set
forth in Section 3 below.

1.F.

1.F.1. Project Gutenberg volunteers and employees expend

considerable effort to identify, do copyright research on, transcribe
and proofread works not protected by U.S. copyright law in creating
the Project Gutenberg™ collection. Despite these efforts, Project
Gutenberg™ electronic works, and the medium on which they may
be stored, may contain “Defects,” such as, but not limited to,
incomplete, inaccurate or corrupt data, transcription errors, a
copyright or other intellectual property infringement, a defective or
damaged disk or other medium, a computer virus, or computer
codes that damage or cannot be read by your equipment.

1.F.2. LIMITED WARRANTY, DISCLAIMER OF DAMAGES - Except for

the “Right of Replacement or Refund” described in paragraph 1.F.3,
the Project Gutenberg Literary Archive Foundation, the owner of the
Project Gutenberg™ trademark, and any other party distributing a
Project Gutenberg™ electronic work under this agreement, disclaim
all liability to you for damages, costs and expenses, including legal
fees. YOU AGREE THAT YOU HAVE NO REMEDIES FOR
NEGLIGENCE, STRICT LIABILITY, BREACH OF WARRANTY OR
BREACH OF CONTRACT EXCEPT THOSE PROVIDED IN PARAGRAPH
1.F.3. YOU AGREE THAT THE FOUNDATION, THE TRADEMARK
OWNER, AND ANY DISTRIBUTOR UNDER THIS AGREEMENT WILL
NOT BE LIABLE TO YOU FOR ACTUAL, DIRECT, INDIRECT,
CONSEQUENTIAL, PUNITIVE OR INCIDENTAL DAMAGES EVEN IF
YOU GIVE NOTICE OF THE POSSIBILITY OF SUCH DAMAGE.

1.F.3. LIMITED RIGHT OF REPLACEMENT OR REFUND - If you

discover a defect in this electronic work within 90 days of receiving
it, you can receive a refund of the money (if any) you paid for it by
sending a written explanation to the person you received the work
from. If you received the work on a physical medium, you must
return the medium with your written explanation. The person or
entity that provided you with the defective work may elect to provide
a replacement copy in lieu of a refund. If you received the work
electronically, the person or entity providing it to you may choose to
give you a second opportunity to receive the work electronically in
lieu of a refund. If the second copy is also defective, you may
demand a refund in writing without further opportunities to fix the
problem.

1.F.4. Except for the limited right of replacement or refund set forth
in paragraph 1.F.3, this work is provided to you ‘AS-IS’, WITH NO
OTHER WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR ANY PURPOSE.

1.F.5. Some states do not allow disclaimers of certain implied

warranties or the exclusion or limitation of certain types of damages.
If any disclaimer or limitation set forth in this agreement violates the
law of the state applicable to this agreement, the agreement shall be
interpreted to make the maximum disclaimer or limitation permitted
by the applicable state law. The invalidity or unenforceability of any
provision of this agreement shall not void the remaining provisions.

1.F.6. INDEMNITY - You agree to indemnify and hold the Foundation,

the trademark owner, any agent or employee of the Foundation,
anyone providing copies of Project Gutenberg™ electronic works in
accordance with this agreement, and any volunteers associated with
the production, promotion and distribution of Project Gutenberg™
electronic works, harmless from all liability, costs and expenses,
including legal fees, that arise directly or indirectly from any of the
following which you do or cause to occur: (a) distribution of this or
any Project Gutenberg™ work, (b) alteration, modification, or
additions or deletions to any Project Gutenberg™ work, and (c) any
Defect you cause.

Section 2. Information about the Mission

of Project Gutenberg™
Project Gutenberg™ is synonymous with the free distribution of
electronic works in formats readable by the widest variety of
computers including obsolete, old, middle-aged and new computers.
It exists because of the efforts of hundreds of volunteers and
donations from people in all walks of life.

Volunteers and financial support to provide volunteers with the

assistance they need are critical to reaching Project Gutenberg™’s
goals and ensuring that the Project Gutenberg™ collection will
remain freely available for generations to come. In 2001, the Project
Gutenberg Literary Archive Foundation was created to provide a
secure and permanent future for Project Gutenberg™ and future
generations. To learn more about the Project Gutenberg Literary
Archive Foundation and how your efforts and donations can help,
see Sections 3 and 4 and the Foundation information page at
www.gutenberg.org.

Section 3. Information about the Project

Gutenberg Literary Archive Foundation
The Project Gutenberg Literary Archive Foundation is a non-profit
501(c)(3) educational corporation organized under the laws of the
state of Mississippi and granted tax exempt status by the Internal
Revenue Service. The Foundation’s EIN or federal tax identification
number is 64-6221541. Contributions to the Project Gutenberg
Literary Archive Foundation are tax deductible to the full extent
permitted by U.S. federal laws and your state’s laws.

The Foundation’s business office is located at 809 North 1500 West,

Salt Lake City, UT 84116, (801) 596-1887. Email contact links and up
to date contact information can be found at the Foundation’s website
and official page at www.gutenberg.org/contact

Section 4. Information about Donations to

the Project Gutenberg Literary Archive
Foundation
Project Gutenberg™ depends upon and cannot survive without
widespread public support and donations to carry out its mission of
increasing the number of public domain and licensed works that can
be freely distributed in machine-readable form accessible by the
widest array of equipment including outdated equipment. Many
small donations ($1 to $5,000) are particularly important to
maintaining tax exempt status with the IRS.

The Foundation is committed to complying with the laws regulating

charities and charitable donations in all 50 states of the United
States. Compliance requirements are not uniform and it takes a
considerable effort, much paperwork and many fees to meet and
keep up with these requirements. We do not solicit donations in
locations where we have not received written confirmation of
compliance. To SEND DONATIONS or determine the status of
compliance for any particular state visit www.gutenberg.org/donate.

While we cannot and do not solicit contributions from states where

we have not met the solicitation requirements, we know of no
prohibition against accepting unsolicited donations from donors in
such states who approach us with offers to donate.

International donations are gratefully accepted, but we cannot make

any statements concerning tax treatment of donations received from
outside the United States. U.S. laws alone swamp our small staff.

Please check the Project Gutenberg web pages for current donation
methods and addresses. Donations are accepted in a number of
other ways including checks, online payments and credit card
donations. To donate, please visit: www.gutenberg.org/donate.

Section 5. General Information About

Project Gutenberg™ electronic works
Professor Michael S. Hart was the originator of the Project
Gutenberg™ concept of a library of electronic works that could be
freely shared with anyone. For forty years, he produced and
distributed Project Gutenberg™ eBooks with only a loose network of
volunteer support.
Project Gutenberg™ eBooks are often created from several printed
editions, all of which are confirmed as not protected by copyright in
the U.S. unless a copyright notice is included. Thus, we do not
necessarily keep eBooks in compliance with any particular paper
edition.

Most people start at our website which has the main PG search
facility: www.gutenberg.org.

This website includes information about Project Gutenberg™,

including how to make donations to the Project Gutenberg Literary
Archive Foundation, how to help produce our new eBooks, and how
to subscribe to our email newsletter to hear about new eBooks.
back
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookgate.com