Thinking Like an Engineer :

Studies in the Ethics of a Pro-

title:
fession Practical and Profes-
sional Ethics Series
author: Davis, Michael.
 2/1183

publisher: Oxford University Press

isbn10 | asin: 0195120515
print isbn13: 9780195120516
ebook isbn13: 9780585245690
language: English
subject Engineering ethics.
publication date: 1998
lcc: TA157.D32 1998eb
ddc: 174/.962
subject: Engineering ethics.
 Page i

Thinking Like an Engineer

 Page ii

PRACTICAL AND PROFESSIONAL ETHICS

SERIES

Published in conjunction with the

Association for Practical
and Professional Ethics

SERIES EDITOR
Alan P. Wertheimer, University of Vermont
 5/1183

EDITORIAL BOARD
Sissela Bok, Harvard University
Daniel Callahan, The Hastings Center
Deni Elliott, University of Montana
Robert Fullenwider, University of Maryland
Amy Gutmann, Princeton University
Stephen E. Kalish, University of Nebraska-Lin-
coln
Thomas H. Murray, Case Western Reserve
University
Michael Pritchard, Western Michigan
University
Henry Shue, Cornell University
David H. Smith, Indiana University
Dennis F. Thompson, Harvard University
Vivian Weil, Illinois Institute of Technology

Brian Schrag, Executive Secretary of the Asso-

ciation
for Practical and Professional Ethics
 6/1183

Practical Ethics
A Collection of Addresses and Essays
Henry Sidgwick
With an Introduction by Sissela Bok

Thinking Like an Engineer

Studies in the Ethics of a Profession
Michael Davis
 Page iii

Association for Practical and Professional Ethics

Thinking Like an Engineer

Studies in the Ethics of a Profession

Michael Davis
Oxford University Press

Oxford New York

Athens Auckland Bangkok Bogota Bombay
Buenos Aires Calcutta Cape Town Dar es Salaam
Delhi Florence Hong Kong Istanbul Karachi
Kuala Lumpur Madras Madrid Melbourne
Mexico City Nairobi Paris Singapore
Taipei Tokyo Toronto Warsaw

and associated companies in

Berlin Ibadan

Copyright © 1998 by Michael Davis

Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016

Oxford is a registered trademark of Oxford Univers

 9/1183

All rights reserved. No part of this publication may

duced, stored in a retrieval system, or transmitted, i
form or by any means, electronic, mechanical, phot
recording or otherwise, without the prior permissio
ford University Press.

Library of Congress Cataloging-in-Publication Data

Davis, Michael, 1943-

Thinking like an engineer: studies in the ethics of a
Michael Davis.
p. cm.(Practical and professional ethics series)
Includes bibliographical references and index.
ISBN 0-19-512051-5
1. Engineering ethics. I. Title. II. Series.
TA157.D32 1998
174'.962dc21 97-24240

135798642

Printed in the United States of America on acid-free

 Page v

For Jeffrey, who was an engineer, sort of;

for all my former students who are now;
and for Alexander, who may be, some day.
 Page vii

PREFACE
This book is a contribution both to engineering
ethics and to the philosophy of a profession, en-
gineering. Teachers of courses in engineering
ethics, the philosophy of any profession, or even
philosophy of technology should find much in
the book useful, but its proper audience is any-
one, engineer or not, scholar or not, who has
ever wondered, "What is an engineer?"
 12/1183

What is engineering ethics? The word "ethics"

can be used in at least three senses. In one sense,
it is a mere synonym for ordinary morality. In
another, it names a field of philosophy (moral
theory, the attempt to understand morality as a
rational undertaking). In a third, it refers to
those special standards of conduct that apply to
members of a group just because of that mem-
bership. When I describe this book as a contri-
bution to engineering ethics, "ethics" has both
the second and third sense. The ethics in this
book is ethics in the second sense, philosophy;
but insofar as understanding standards makes
both following them and improving them easier,
what I do should contribute to engineering eth-
ics in the third sense of ethics as wellthat is, to
the interpretation, application, and revision of
engineering's special standards of conduct.
 13/1183

As a work in ethics, this book resembles such

philosophical textbooks as Harris, Pritchard, and
Rabins, Engineering Ethics: Cases and Con-
cepts. 1 It nonetheless differs from them in at
least two ways. First, it is not a survey but a
series of essays, a supplement to textbooks
rather than their competitor. The book concen-
trates on a few particularly important points,
corresponding to the book's four parts. Part I,
the first three chapters, puts engineering in his-
torical perspective, making clear both how new
engineering is and in what that newness con-
sists. Part II is an extended meditation on the
Challenger disaster. Each of its three chapters
considers one aspect of the complex relationship
between engineering ideals and engineering
practice
Page viii
 15/1183

today. Here we see in detail how social organiz-

ation and technical requirements combine to
define how engineers should (and presumably
do) think. Part III's two chapters clarify the im-
portance of protecting engineering judgment and
identify the chief means of doing it. These three
parts together give considerable content to the
notion of "thinking like an engineer." Part IV,
the last three chapters, is concerned with testing
this philosophical construction empirically.
Chapter 9 reports results of a study of how en-
gineers and managers work together in ten dif-
ferent companies. Chapter 10 attempts to clarify
the concept of professional autonomy in such a
way that social scientists should be able to tell
us how much professional autonomy engineers
have. The epilogue draws from the book's argu-
ment four questions concerning engineering that
the social sciences, including history, could an-
swer in a way helpful to engineering ethics. The
epilogue is an invitation to those working in the
 16/1183

social sciences to contribute to engineering

ethics.

That is one difference between this book and

textbooks in engineering ethics, a difference in
intensity. The other difference is one of exten-
sion. This book is as concerned with the "engin-
eering" in "engineering ethics" as with the "eth-
ics." It is a contribution to the philosophy of
engineering.
 17/1183

What is the philosophy of engineering? Like the

philosophy of science, of law, or of art, the
philosophy of engineering tries to understand its
subject as a rational undertaking. It does not of-
fer a philosophy of engineeringthat is, a (contro-
versial) conception of how engineering should
be done. It attempts to say what engineering
iswithout becoming a mere subtopic in the
philosophy (or sociology) of technology. Al-
though the philosophy of technology focuses on
what engineers (and others) help make, the
philosophy of engineering focuses on engineers
themselveson what they try to do and why.
 18/1183

I have learned from Walter Vincenti's What

Engineers Know and How They Know It, 2 but
that book and this one differ substantially. First,
Vincenti is both an engineer and a historian. I
am neither. He has a grasp of technical principle
and historical documents that I never will. Se-
cond, although Vincenti's work contributed to
mine, his has a narrower focus. He tries to un-
derstand engineering as a developing body of
technical knowledge, a discipline; I, on the other
hand, try to understand engineering as a profes-
sion. Knowledge, though of course a part of
what makes an engineer, is only a part. At least
as important is the way the knower moves (or, at
least, is supposed to move) from knowledge to
action. That movement from knowledge to ac-
tion is the "thinking" of my title. The thesis of
this book, if it has only one, is that this thinking
is fundamentally ethical (in both my first and
third senses).
 19/1183

The philosophy of engineering may seem too

technical a field for philosophers: Who could
know better than engineers how engineers
think? The question answers itself. Of course,
engineers know better than anyone else how
they think. That, however, does not decide who
should do philosophy of engineering. Generally,
scientists know science better than philosophers
of science, lawyers know the law better than
philosophers of law, and artists know art better
than most philosophers of art. Philosophers still
do philosophy of science, law, and art, doing
something that scientists, lawyers, and artists
cannot do for themselves. Although some of
these philosophers are amphibians, philosopher-
scientists, philosopher-lawyers, or philosopher-
artists, even some of the best are not. That is a
fact, but it raises the
 Page ix

question: How is it possible for those who know

less to teach those who know more? Answering
that question requires a little "philosophy of
philosophy."

Philosophy (at its best) puts our tacit knowledge

into words. It makes the obvious obvious. The
first philosopher, Socrates, distinguished him-
self from the "wise men" of ancient Greece by
asking rather than telling. He asked the pious
what piety is, politicians what politics is, and so
on. Those he asked had great trouble putting
what they knew into words; indeed, much they
said turned out, on Socrates' examination, to be
false.
 21/1183

How have engineers done compared to the ex-

perts Socrates questioned? Certainly, many en-
gineers feel that nonengineers generally do not
understand what engineers do, that the achieve-
ments of engineers are appreciated less than
they should be, and that engineering does not do
as well as it should in recruiting the next genera-
tion. Scientists, architects, lawyers, and even
MBAs generally seem to carry off the prizes.
Yet, when engineers try to make their own case,
what happens? Even Samuel Florman, as literate
a polemicist as any profession can claim, is sur-
prisingly unhelpful. His The Existential Pleas-
ures of Engineering 3 is a powerful defense of
technology, but one from which engineers are
largely absent. Change a few words and the
book could be a defense of scientists, industrial-
ists, or even mere inventors rather than engin-
eers. His The Civilized Engineer4 fails in anoth-
er way, pleasing engineers rather than informing
nonengineers.
 22/1183

The power of philosophers is not in their initial

knowledge of a field butas Socrates stressedin
their initial ignorance of it. That ignorance is not
ordinary ignorance, the unassuming or presum-
ing of the benighted; it is, instead, experienced,
open, systematic, cooperative, and dogged. Such
ignorance can help those who know a field to
put their knowledge into words even those who
do not know the field can understand. The result
is paradoxical. Even the expert seems to learn
from having what she said put into a
philosopher's wordsas one learns something
when seeing for the first time the pattern in a
mosaic known by heart. The expert may then
conclude that the philosopher "really" knows
more about the expert's field than the expert her-
self, forgetting that the philosopher could only
reveal what he revealed by drawing it out of her.
While philosophers often seem generators of
knowledge, they are, as Socrates put it, merely
its midwives.
 23/1183

This book is the product of more than a decade

working with engineers, trying to understand
what so absorbed them and about which they
could say so little. I began by thinking that en-
gineering was primarily about things, a complex
but fundamentally unimaginative application of
science, mere "problem solving" (as even engin-
eers will describe it, if you let them). I have
come to understand engineering quite differ-
ently: as the practical study of how to make
people and things work together betteran under-
taking as creative as art, as political as law, and
no more a mere application of science than art
or law is. That is the understanding I have tried
to put into words here. I will consider myself
successful if engineers reading this book say,
"Yes, of course, exactly" and nonengineers add,
"So, that's what they do: I had no idea!"
 24/1183

I publish this book without apology for the mis-

takes it must contain. The only way to write
without mistakes is to write nothingor, at least,
nothing interesting. I have done my best to be
interesting, taking controversial positions if I
believed
 Page x

them right and defending them as best I could,

hoping thereby to incite others to add their
views, explained and justified, whether or not
they agree with mine. Only through critical dis-
cussion that is rational and informed can either
engineering ethics or the philosophy of engin-
eering grow as a field of study. If, in the pro-
cess, I am shown to have erred, I will not
complain.
 26/1183

Though I publish this book without apology, I

do not publish it without trembling. For his ef-
forts, Socrates was put to death. Apparently,
some experts do not take well to philosophical
ignorance. If I fare better than that master of
philosophy, it will be because of those engineers
(practitioners, academics, and students) who
pulled me aside, explained what I got wrong,
and then patiently answered one question after
another until I got it right. My notes thank those
I remembered, but my memory for names is not
good. I hope those I forgot will forgive the
forgetting.
 27/1183

I owe special thanks to two colleagues: to Vivi-

an Weil, for helping me, more than a decade
ago, to see that engineers might be at least as
philosophically interesting as lawyers, and to
Robert Ladenson, for convincing me to join a
small band of philosophers following their call-
ing among the engineers at the Illinois Institute
of Technology (IIT). Though I accepted the in-
vitation more because I trusted him than because
I believed what he said, I now doubt that any
other course of action could have had as good an
outcome. I had taught at three other universities
with engineering schools; IIT was the first
where the philosophers and engineers had much
to say to each other.
 28/1183

Chapters 1, 3, and 5 through 10 have appeared

in print before much as they do here. Chapter 4
is a much enlarged version of an essay previ-
ously published. Chapter 2 and the epilogue see
print for the first time. Though acknowledge-
ment of prior publication is made at the appro-
priate place, I should like to thank the editors of
the journals in which those chapters initially ap-
peared, as well as Alan Wertheimer, the editor
of this Oxford series, and two of his reviewers,
Deborah Johnson and Michael Pritchard, for
suggesting many of those improvements now in-
corporated in the text.

M.D.
CHICAGO
DECEMBER 1997
CONTENTS

Part I
Introduction to Engineering

1. Science, Technology, and Values

2. A History of Engineering in the United States

3. Are ''Software Engineers" Engineers?

Part II
Engineers in Context

4. Codes of Ethics and the Challenger

5. Explaining Wrongdoing

6. Avoiding the Tragedy of Whistleblowing

 30/1183

Part III
Protecting Engineering Judgment

7. Conflicts of Interest in Engineering

8. Codes of Ethics, Professions, and Conflict of Inte

Part IV
Empirical Research

9. Ordinary Technical Decision-Making: An Empir

10. Professional Autonomy: A Framework for Emp

Epilogue: Four Questions for the Social Sciences

Appendix 1: Questionnaire for Engineers

Appendix 2: Questionnaire for Managers

Appendix 3: Interviewee Characteristics

Notes

Bibliography

Index
 Page 3

PART I
INTRODUCTION TO ENGINEERING
 33/1183

This work of philosophy begins with a long for-

ay into the history of engineering. Foraging in
another's field is always risky. One can easily
get lost, fall into traps the owners long ago
learned to skirt, or find oneself suddenly out-
numbered and outgunned. I am taking the risk
for four reasons. First, I believe that reading his-
tory can lead to philosophical insights. The past
gives the present context. Second, I believe that
some historians, those I have been reading,
sometimes miss the obviousor, at least, get the
emphasis wrongand therefore tend to mislead
those trying to understand engineering. I believe
I can do better. Third, although I am trespassing,
I have precedent on my side. Philosophers have
long made themselves useful by pointing out the
obvious in fields not their ownwhich is all I in-
tend to do. Fourth, and most important, I believe
that my trespass will pay off. Understanding the
history of engineering better, we shall under-
stand engineering better.
 34/1183

This foray has two important outcomes. First, it

works out a definition of engineering as an oc-
cupation, a way to distinguish engineers from
nonengineers. In other words, it defines the field
this book is to study. Second, it makes a case for
distinguishing between engineering as an occu-
pation and engineering as a profession. It makes
clear the importance of understanding engineer-
ing as a profession rather than as a mere intel-
lectual discipline or occupation of "knowledge
workers." To understand engineering as a pro-
fession is, I argue, to make ethics central to what
engineers do.
 Page 5

1
Science, Technology, and Values
Is engineering just applied science, a field as
free of values as science itself? Or is engineer-
ing just technology, a field already well studied
by those who study technology? Are the values
of engineering, if there are any, just the values
of technology, whatever those are? Or does en-
gineering contribute something more? What?
Why?
 36/1183

We must answer these important questions as

soon as possible. But before we can, we must
clarify the terms. "Science," "technology," and
"values," like ''engineering" and "ethics," are
used in enough different ways to be dangerous.
Clarifying these five terms and others related to
them requires a foray into history. History ex-
plains some of the confusion about these terms
and helps us choose meanings useful to the
work ahead.

Techne and Sophia:

Twins Ancient but Unequal
 37/1183

I begin with etymology. "Technology" is a com-

pound of two words from ancient Greek, techne
and logos. Techne means manual art. So, for ex-
ample, a tekton was a carpenter or builder; an
"architect" was a master builder. The suffix
form of logos, "ology," means a putting into
words, an explanation or study. So, when our
word "technology" still meant what Greek tells
us it means, technology was the explanation or
study of manual art, just as biology is the ex-
planation or study of bios, "life". It was a field
in which gentlemen entered the workshop to re-
cord the artisan's secrets for later publication. 1

That, of course, is not what technology means

now. Despite its Greek root, "technology" is
really a new word, recoined in the middle of the
last century for a new idea.2 What idea?
 Page 6

Ancient Greece was a slave-owning society and,

like other slave owners, Greeks tended to asso-
ciate manual labor with slaves. Because no free
man would want to be mistaken for a slave, the
ancient Greeks generally avoided doing what
slaves do. For example, because slaves tended to
rush about on their master's business, free men
were supposed to walk slowly. 3 Greeks had
such a low opinion of manual labor that they
even rated sculpture less noble than painting be-
cause the sculptor, unlike the painter, had to
sweat over his work like a slave.4

There were a few exceptions to this low opinion

of manual labor. One was athletics. Athletics,
however sweaty, was not something slaves did.
War was another exception. Hacking one anoth-
er with swords, though hard and dirty work, was
a job for free men.
 39/1183

The Greeks contrasted teche with sophia. Al-

though often translated as "intellectual know-
ledge" or even "science," sophia is probably bet-
ter translated as ''wisdom." From sophia comes
our word "philosophy" (the love, that is, the pur-
suit, of wisdom). For the Greeks, philosophy in-
cluded mathematics, physics, economics, and
similar sciences. Because philosophy was
primarily a matter of thought, not manual art,
philosophy was appropriate to free men.

The Greeks of Greece's Golden Age loved

sophia, and she rewarded them accordingly. The
Greeks of that period can claim credit for begin-
ning the tradition of philosophy now dominant
over most of the world, the one to which I be-
long. They can also claim credit for beginning a
number of the sciences, including geometry,
biology, and political science.
 40/1183

Their achievements in poetry, architecture, and

history are no less impressive. Not so their con-
tributions to techne. Of course, there were some
contributionsfor example, improved design of
war galleys. But you must hunt for them.
Europe's Dark Ages seem to have given us
many more useful devices than did Greece's
Golden Age.5
 41/1183

By now, perhaps, you can see two reasons to

distrust that ugly word "technology."5 First,
there is the implicit opposition between sophia
and techne. Today we think of science and tech-
nology as related, not opposed. So, for example,
one reason politicians give for funding scientific
research is that it will pay off in new technolo-
gies.6 Second, there is the word's meaning in
Greek. For us, technology is notas its Greek
parts suggesta study of manual art but, primar-
ily, our way of referring to all those inventions
that make manual labor easier, more productive,
or unnecessary. In this sense, technology began
with the first tool someone made; the new tech-
nologies we hear about are new technologies in
this sensenew tools someone has made.
 42/1183

Of course, there is yet another sense of techno-

logy, one derived from this second but referring
to a studyas in, for example, the title "institute
of technology" (or "technological university").
An institute of technology is not, as the Greek
suggests, a place to study manual arts
(carpentry, machining, and so on)a mere tech-
nical school. An institute of technology is, in-
stead, a place to study practical inventions: how
to make them and how to organize them (and
those who use them) to make other useful
things. The Greeks, who had a word for almost
everything, seem not to have had a word for
that.
 Page 7

What does this history have to do with us? Con-

sider, for example, how we dress for work:
Some of us dress in "white collars"that is, fine
shirts, ties, good slacks, dresses, sport coats, or
the like. Others wear "blue collars"that is, coarse
shirts, denim pants, coveralls. Generally, those
in white collars have higher status than those in
blue. Salary is secondary, as is social usefulness.
A carpenter has less status than an accountant
earning half as much. Why? Though carpentry
requires a trained mind, it requires as well, like
other blue-collar work, much sweaty labor sur-
rounded by dust and debris. Because such labor
would quickly ruin good clothing, the white col-
lar guarantees some distance between its wearer
and such "slavish labor." And, because it does
that, the white collar confers status.
 44/1183

No matter the origin of our parents, we are, in

this respect at least, all more or less descendants
of the ancient Greeks. Even if we ourselves like
manual labor, we do not respect it as much as
we do mental labor. 7 I doubt that this is good,
especially for engineers. But it does seem to be
a stubborn fact about us. We are prejudiced
against blue collars, not only those who work in
them but even those who associate with those
who work in them.8
 45/1183

That prejudice shows up even in a phrase seem-

ingly having nothing to do with it"science and
technology." Why does technology always come
second? The explanation cannot be historical. If
technology refers to inventions making manual
labor easier, technology is older than science by
thousands of years. And, even if the
"technology" in "science and technology" refers
instead to the systematic study of practical in-
vention, technology would be no younger than
science in the corresponding sensethe systematic
study of nature. Until quite recently, ''science"
included all systematic knowledge, whether of
nature or invention, including even jurispru-
dence and theology.
 46/1183

Nor can the explanation of the inevitable prior-

ity of science be alphabetical order. Substitute
"engineering" for "technology" and the order re-
mains the same: science and engineering (as in
the journal Science and Engineering Ethics), not
engineering and science. Nor can the explana-
tion be practical importance. Technology bakes
our bread; science only helps us to understand
how. Nor can the explanation be mere accident.
Accident would produce more variation. The or-
der seems fixed: science and technology. Why?

The answer, I think, is that the order indicates

relative status. Science has higher status than
technology; hence, it gets first mention.
 47/1183

Well, shouldn't science have higher status? After

all, isn't technology just applied science?
Doesn't science come first in the order of devel-
opment? Doesn't science lay down the law, like
a master whereas technology merely applies it,
like a slave? Even engineers may be tempted to
answer yes to these questions. But the answer is:
No, technology is not merely applied science.

Science, Technology, and Engineering

One can understand the words "science" and

"technology" to refer to comparable concepts.
Science is explicit, systematic knowledge of
how "nature" works; technology is explicit, sys-
tematic knowledge of how to make useful
things. Unfortu-
 Page 8

nately, usage today is not so neat. Although the

term "science" did once refer primarily to expli-
cit, systematic knowledge of nature, its meaning
has now shifted somewhat so that today it refers
as much, or instead, to a social undertaking: "a
voyage of discovery" (as scientists like to say)
rather than merely to what they discover. In this
sense, science consists of certain communities
engaged in trying to understand how nature
works. 9
 49/1183

Because "technology" refers only to our practic-

al inventions, or to the study of how to make
more, we lack a term comparable to this new
sense of science. What do we call communities
that invent useful things or, at least, add to our
knowledge of how to do it? "Technician" is
wrong: A technician is an assistant, one who
carries out routine work under direction of a sci-
entist, engineer, architect, physician, or the like.
"Technologist," though a natural choice, has not
caught on; "applied scientist,'' though once pop-
ular with sociologists, natural scientists, and
even engineers, is now fading.

Why? I think the reason is that the great major-

ity of people who would have to be called tech-
nologist or applied scientist already have a satis-
factory name: "engineer."
 50/1183

I said "great majority." I meant it. The United

States today has well over two million engin-
eers. That is more than all other technologists
together. Most other technologists are either ar-
chitects, chemists, physicists, biologists, physi-
cians, computer scientists, or mere inventors.
The United States has only about 135,000 archi-
tects, 388,000 natural scientists (including
chemists, physicists, and biologists), 450,000
computer scientists, and 600,000 physicians.10 I
have no figure for "mere inventors," but, since
most inventors seem to be engineers, there can't
be many "mere inventors." The number of phys-
icians contributing to technology also cannot be
large. Most physicians are not in research or de-
velopment but simply provide health care. So,
even assuming that most scientists are in techno-
logy, not pure research, engineers must outnum-
ber all other technologists combined by at least
two to one.
 51/1183

These numbers suggest an obvious solution to

the problem of what to call all those who make
technology: Call them engineers. But that
would, I think, be a terrible mistake. Chemists,
architects, physicians, biologists, and the like
are not engineers. Understanding why they are
not will help us understand both the values in-
herent in most technology, the technology en-
gineers develop, and the place of ethics in any
technology. It also brings us to the heart of our
subject. But it requires more history, though
mostly history less ancient than before.

The Beginnings of Engineering

 52/1183

Professions, aping aristocracy, like to trace their

origins back to ancient times. So, for example,
the American Medical Association's Principles
of Medical Ethics cites certain provisions of
Hammurabi's Code (about 2000 BC) as the
earliest known code of medical ethics.11 There
is, of course, some truth in such going back. The
healers of ancient Babylonia resemble today's
physicians in many ways. For example, like
modern physicians, they tried to cure the sick.
However, there are many differences as well,
and, for our purposes, the differences are more
important. For example,
 Page 9

Babylon's healers do not seem to have been or-

ganized as a profession or even as a guild. We
will understand professions better if we start
their history with the rise of modern markets
about two centuries ago, the accompanying dis-
solution of the old distinction between trades
and "liberal professions," and the slow emer-
gence of something new. Even an old occupa-
tion can be a new profession.
 54/1183

By 1850, especially in England, we begin to see

the modern pattern. The professions are connec-
ted with both a formal curriculum, ending with
an examination and a certification of some sort,
and explicit standards of practice, a code of eth-
ics. 12 Admittedly, those creating this new pat-
tern seem unaware of doing something new. But
there can be little doubt that they misunderstood
their own actions. Even some of the terms they
used were new. For example, the term "medical
ethics" was coined in 1803 by a physician, Tho-
mas Percival, for a book he thought was on an
old topic.13
 55/1183

What is true of most professions is true of en-

gineering. False pedigrees abound. Some histor-
ies of engineering begin with the Stone Age,
with the first tools. They confuse engineering
with mere technology.14 Other histories begin
more sensibly, recognizing that engineers gener-
ally do not do manual labor but prepare instruc-
tions for others to carry out. As the first tool al-
most certainly predates such a division of tasks,
these histories begin much later, with the first
projects large enough to have some people lay-
ing out a plan and others implementing it. They
begin with the building of Stonehenge, the Pyr-
amids, or some other wonder of ancient civiliza-
tion.15
 56/1183

Though better than the first, this second way of

beginning the history of engineering still has at
least two embarrassing consequences. One em-
barrassment is that it makes architects (or "mas-
ter builders") the first engineers. This is embar-
rassing because engineers generally agree that
architects today are definitely not engineers.
Another embarrassment is that this way of
telling the story makes a mystery of why our
word for engineer comes from French, rather
than Greek, like "architect," and why the French
have had the word for barely four hundred
years. Generally, we have a word for anything
important to us almost as soon as we have the
thing. There are no significant
"whatchamacallits."
 57/1183

So, when I tell the story of engineering, I start

four hundred years ago in France. Back then
there were things called "engines"but engine
then simply meant a complex device for some
useful purpose, a contraption showing intelli-
gence in designin short, a machine. The first
people to be called engineers were soldiers asso-
ciated with catapults, siege towers, artillery, and
other "engines of war." They were not yet en-
gineers in the sense that conerns us. They were,
rather, engineers in the sense that, even today,
the driver of a locomotive is an engineer. They
were engineers only in the sense that they oper-
ated (or otherwise worked with) engines.

Some soldiers are still engineers in something

like this sense: They belong to an engineering
corps. Though they do not know what engineers
know, they are directly involved in works of en-
gineering, though not precisely with engines of
war, a term no longer in common use.
 58/1183

Four hundred years ago the armies of France

were led by nobles, men on horseback who
learned war from their fathers or on the battle-
field or died in the attempt. The foot soldiers
came with the nobles. Most were peasants or ar-
tisans who
 Page 10

knew little of war until trained in camp. When

the war ended, the army dissolved, each noble
leading his own people home. In such an army,
an engineer was usually a carpenter, stone ma-
son, or other artisan bringing civilian skills to
war.

When Louis XIV ended the regency in 1661,

France still made war in this way. But, within
two decades, France had a standing army of
300,000, the largest, best trained, and best
equipped European fighting force since the Ro-
man legions. This achievement was widely
copied. To this day, most of our military
wordsfrom "army" itself to "reveille," from
"bayonet" to "maneuver," from "private'' to
"general"are French. "Engineer" is just one of
these military terms.
 60/1183

Until 1676 French engineers were part of the in-

fantry. But in that year the engineers were or-
ganized into special units, the corps du génie. 16
This reorganization had important con-
sequences. A permanent corps can keep much
better records than isolated individuals; can ac-
cumulate knowledge, skills, and routines more
efficiently; and can pass them on. A corps can
become a distinct institution with its own style
and reputation. More than a group of protoen-
gineers, the corps du génie were, potentially,
both a center of research in engineering and a
training ground for engineers (in something like
our sense)officieurs du génie.
 61/1183

The corps du génie did not take long to realize

this potential. Within two decades, it was known
all over Europe for unusual achievements in
military construction. When another country
borrowed the French word "engineering" for use
in its own army, it was for the sort of activity
the corps du génie engaged in.17 That was
something for which other European languages
lacked a word.
 62/1183

The corps du génie was not, as of 1700, a school

of engineering in our sense; it was more like an
organization of masters and apprentices. Indeed,
strange as this may seem now, at that time
neither France nor any other European state had
a permanent military academy (in anything like
our sense), much less a school of engineering.
There was no settled curriculum for training of-
ficers generally or engineers in particular, or
even a very clear idea that a curriculum was ne-
cessary. Only during the 1700s did the French
slowly come to understand what they wanted
from an engineering education and how to get it.
But, by the end of the 1700s, they had a cur-
riculum from which today's engineering cur-
riculum differs only in detail; they had also in-
vented engineering.18
 63/1183

An army needs fortifications for protection,

mines under enemy fortifications, roads to
march on, and bridges to cross. Civilians either
need the same things or need other things that
require similar skills to build. So, in 1716, the
French established another corps of engineers,
the corps des ponts et chaussés, to build and
maintain the nation's bridges, roads, and canals
(as important to the army as to commerce). This
corps set up a school for training its officers, the
first engineering school to survive long enough
to matter. Like the military engineers, these civil
engineers were admired all over Europe. Those
who copied their method copied their name as
well.19
 64/1183

What was their method? Engineers, military as

well as civil, resembled architects in being able
to make drawings for construction projects, de-
velop detailed instructions from those drawings,
and oversee the execution of those instructions.
They nonetheless seem to have differed from ar-
chitects in at least three ways.
 Page 11

First, engineers were much better trained in

what was then the new mathematics and physics
than the architects were. They had the ability to
consider systematically questions most archi-
tects could only deal with intuitively or ignore.
20

Second, because the strategies of engineering

had their roots in the necessities of war, engin-
eers paid more attention to reliability, speed, and
other practicalities. So, for example, the system-
atic testing of materials and procedures in ad-
vance of construction was early recognized as a
characteristic of engineers.21 At least in compar-
ison, the architect seemed an artist, one for
whom beauty claimed much of the attention an
engineer would devote to making things work.
 66/1183

Third, to be an engineer was to be trained as an

army officer, to be disciplined to bear signific-
ant responsibility within one of world's largest
organizations. Engineers were therefore likely to
be better at directing large civilian projects than
were architects, most of whom would have had
experience only of much smaller undertakings.

These three advantages tend to reinforce one an-

other. For example, not only do large projects
require more planning in advance and more dis-
cipline in execution, but they are also more
likely to require better mathematical analysis
and to justify extensive testing of materials and
procedures. For this, and perhaps other reasons,
civil engineers slowly took over much of the
work that once would have been the domain of
architects. They were a new power in the world.
 67/1183

Early experiments in engineering education cul-

minated in the École Polytechnique. Begun in
1794 as the École des Travaux Publics (the
school of public works), it changed its name the
following year, for the first time connecting en-
gineering and techne. I don't know why the
French changed the school's name. The school
never trained architects, much less artisans or
mechanics. It was a school of engineering, de-
serving the "poly" only for offering preparation
for many fields of engineering, military and
naval engineering, as well as civil.22
 68/1183

The École Polytechnique's curriculum had a

common core of three years. The first year's
courses were geometry, trigonometry, physics,
and the fundamentals of chemistry with practic-
al applications in structural and mechanical en-
gineering. There was a good deal of drawing,
some laboratory and workshop, and recitations
after each lecture. The second and third year
continued the same subjects, with increasingly
more application to the building of roads,
canals, and fortifications and the making of mu-
nitions. For their last year, students were sent to
one of the special schools: the school of artil-
lery, the school of military engineering, the
school of mines, the school of bridges and
roads, the school of geographical engineers (car-
tographers), or the school of ships.23
 69/1183

Engineers will immediately recognize this cur-

riculum, especially the four years, the progres-
sion from theory (or analysis) to application (or
design), and the heavy emphasis on mathemat-
ics, physics, and chemistry.

The École Polytechnique was the model for en-

gineering education for much of the nineteenth
century.24 The United States began using it very
early. Our first engineering school was the milit-
ary academy at West Point. By 1817, it had ad-
opted much of the École's curriculum, its meth-
ods of instruction, and even some textbooks.25 I
say more about West Point in the next chapter.
 Page 12

Values in Engineering

What values does engineering incorporate? A

decade ago, Eugene Ferguson, an engineer
turned historian, drew up a list of "imperatives
of engineering." 26 The list is neither complete
nor fundamentalnor, indeed, even entirely fair. It
will nevertheless help us understand
engineering.
 71/1183

Engineers, Ferguson claimed, (1) strive for effi-

ciency, (2) design labor-saving systems, (3)
design control into the system, (4) favor the very
large, the very powerful, orin electronicsthe
very small, and (5) tend to treat engineering as
an end in itself rather than as a means to satisfy-
ing human need. These imperatives are, accord-
ing to Ferguson, instincts engineers bring to
their work. Although engineers can resist them,
just as I can resist drinking water even if I am
thirsty, they are, in effect, the engineer's default
setting, what engineers will do unless they con-
sciously try to do something else.
 72/1183

Ferguson intended this list to be a criticism of

the way engineers work. It is, I think, both less
and more than that. The list is less than a criti-
cism because the first four imperatives seem, on
reflection, at least as much virtues as vices. The
list is also more than a criticism because it high-
lights certain enduring features of engineering,
permitting us to connect engineering's history
with today's practice. Let's take a closer look at
Ferguson's list.

"Efficiency" is the first imperative Ferguson

identifies. Ferguson points out, rightly, that "ef-
ficiency" is a slippery term, meaning "most
powerful'' here, "lowest cost" there, and
something else elsewhere. What he overlooks is
the concept's utility.
 73/1183

Engineers generally define efficiency so that

they can measure it (or its components), assign
numbers, and thereafter seek to control it. That
is not surprising. Like other professions, engin-
eering tends to analyze a situation so that its dis-
tinctive skills can be applied. One distinctive
skill of engineers is giving mathematical struc-
ture to practical problems. The concept of effi-
ciency allows them to exercise that skill.
 74/1183

Engineers have, no doubt, sometimes paid too

much attention to efficiency, especially forms of
efficiency that turned out not to matter. Indeed,
the history of engineering is in part the history
of measurable properties used for a time as
proxy for something that could not be measured
and then discarded when the proxy proved not
to have enough of a relation to what the engin-
eers actually cared about.27 Because engineer-
ing is a practical undertaking, it must learn from
practice. It cannot learn from practice without
making mistakes. Some of engineering's mis-
takes concern efficiency.
 75/1183

Engineers can be quite slow about giving up one

of these proxy measures. But, even this slow-
ness is understandable. Engineers are used to
working in large organizations, organizations in
which change is difficult and the consequences
are often hard to predict. They therefore have a
tendency to follow practices they would no
longer adopt. (Consider, for example, how
American engineers still specify non-metric
bolts or screws.) The world is a tough laborat-
ory. Many things better in theory are worse in
practice. How daring do we want engineers to
be with our lives?

The second imperative on Ferguson's list is a

preference for labor-saving devices. Engineers
will, Ferguson thinks, design to save labor even
when labor is cheap and the end result will be
higher production costs and more
unemployment.
 Page 13

The engineer's preference for labor saving is un-

derstandable as a product of engineering's milit-
ary origin. Since engineering began, the primary
labor pool of most armies has been their own
soldiers. Because no general wants his soldiers
doing construction when they could be fighting,
military engineers have always had an incentive
to look for means of saving labor even though
the labor saved was, in one sense, cheap
(indeed, free).
 77/1183

As military engineering became civil engineer-

ing, this tendency might have put engineers at a
disadvantage in their competition with other
technologists. Their designs might have proved
too costly. Those who hired engineers would,
however, soon have learned this. They would
then have compensated, either by calling in an
engineer less often or by making sure that the
engineer called in defined the desired outcome
taking cost into account.

If, as Ferguson's criticism suggests, such com-

pensation seldom occurs, the most likely reason
is that the engineer's preference for labor-saving
devices generally serves those who employ en-
gineers. The reason that preference might serve
their employers is not hard to see. Labor has a
tendency to become scarce, and so costly, when
it is not routinely saved.
 78/1183

Of course, that is only a tendency. Many of

those thrown out of work by a particular innova-
tion may live out their lives on the dole. Many
engineers would, no doubt, like to take such ef-
fects into account, and perhaps many of their
employers would let them. But, if engineers are
to take such considerations into account, they
will need both the relevant information and a
routine for using it.

Gathering such information belongs to the social

sciences, not to engineering as it is or as it is
likely to become. Any curriculum that could
give engineers the skills to develop significant
social statistics would probably be too long to
attract many students. Engineers should not be
blamed for failing to take into account social
consequences about which they can only guess.
 79/1183

However, when such information exists, devel-

oping ways to incorporate it into engineering
work is certainly something engineers can, and
should, do. Indeed, they have long done this
with the employer's share of the cost of produc-
tion. And, over the last two decades, thanks to
the Environmental Protection Agency (EPA),
engineers have become adept at incorporating
environmental costs into their designs (e.g., by
designing for disposal as well as for manufac-
ture and use). They could do the same for social
impact if they had numerical standards for as-
sessing impact and sources of information from
which the relevant numbers could be taken.
 80/1183

Engineers can help to develop such standards,

just as they helped to write EPA standards. But,
just as with environmental standards, standards
for permissible social impact are probably not
what most people would want engineers alone to
decideor even engineers with the help of law-
yers, accountants, corporate executives, and oth-
er specialists. Social impact raises political is-
suesthat is, issues everyone wants a part in de-
ciding. If engineers decline to develop such
standards unilaterally, should we blame them?
28

Ferguson's third imperative is designing controls

into the system. Engineers generally try to sep-
arate planning and execution. Intelligence is de-
signed into the system, requiring as little intelli-
gence as possible of the system's operators. The
assembly line is the typical example of this im-
perative. Engineers generally try to design an
 Page 14

assembly line so that the work is so simple that

only a few minutes training is necessary to learn
the job. The job is therefore likely to be repetit-
ive and boring; those doing the job are reduced
to little more than organic robots.

Engineering's military past certainly explains the

origin of this imperative. Soldiers sent over to
help on an engineering project, whether digging
trenches or putting a bridge over a river, don't
have much time to learn the job. The military
engineer must design the work so that anybody
can do it. (Architects, in contrast, seem, if any-
thing, to have a bias in favor of designs requir-
ing craftsmen.)
 82/1183

But its military past alone does not explain why

this imperative persists in civilian engineering
(or, at least, why engineers who do such things
should be so much in demand). The explanation
of that, like the persistence of engineering's
second imperative, must be that this tendency is
useful in civilian engineering as well. One re-
cent example suggests why that might be.
 83/1183

McDonald's restaurants now have cash-register

buttons with pictures of the various items on the
menu. The cashier need not know the price of
anything or even be able to read; the cashier
only has to recognize the pictures and push but-
tons accordingly. In a business where employee
turnover is high and education is low, where
prices change frequently and training is expens-
ive, this dumbing down of the cashier's job both
saves money for McDonald's and opens employ-
ment to many who might not otherwise qualify.
Whoever thought of that device, engineer or not,
was undoubtedly a hero to McDonald's. 29
 84/1183

The fourth imperative of engineering that Fer-

guson lists is a tendency to disregard human
scale, preferring the very large or the very small.
The reason for this imperative is that engineer-
ing was, and remains, a creature of large organ-
izations. Louis XIV's army, one of the largest
organizations of its day, created engineering to
do what civilian artisans could not do (or could
not do well enough). Even today, most engin-
eers work in large organizations. You do not
need an engineer to construct a single-family
house. A carpenter or architect will do, as they
always have. If, however, you want to construct
a thirty-story building, you need engineers.
 85/1183

The problem, I think, is not so much that engin-

eers disregard human scale as that they are sel-
dom needed for things on a human scale. Gener-
ally, asking engineers to work on a human scale
is like asking lawyers to prepare a partnership
agreement for two children opening a lemonade
stand. They can do it, but either they will do
what anyone else could do or they will do
something out of all proportion to the job.

In this respect, the very small can be like the

very large. For example, to make today's tiny
electronic circuits requires productive forces and
controls of which a single human being is incap-
able. Hence, there is work for engineers.
 86/1183

Ferguson's last imperative, putting technical

brilliance ahead of human need, is unlike the
others. It is a failing common to all professions.
(We all know the joke about the surgeon who
says, "The operation was a success, though the
patient died.") But this last "imperative of engin-
eering" is worse than a failing common to all
professions; it is a failing inconsistent with one
of engineering's fundamental values.

I have stressed the military origins of engineer-

ing.30 I have not pointed out that most of the
period we have been talking about, roughly the
1700s, is known as the
Page 15
 88/1183

Age of Enlightenment. This was the time when

many Europeans first came to believe that en-
lightenment, that is, scientific learning, would
bring peace, prosperity, and continuous im-
provement. For countless ages, the best hope of
the wise was that the world would not get much
worse. With the Age of Enlightenment, people
began to act on the belief that the world could
be made much better. Engineering has this be-
lief built into it. So, for example, early graduates
of the École Polytechnique were noted for "sci-
entific and democratic idealism and a desire to
work for human progress." 31 The same attitude
was evident in England at about the same time.
When, in 1828, the British Institution of Civil
Engineers, then nine years old, asked one of its
members, Thomas Tredgold, to define the term
"civil engineering," he gave an answer engineers
still quote: "Civil Engineering is the art of dir-
ecting the great sources of power in Nature for
the use and convenience of man. . . . The most
 89/1183

important object of Civil Engineering is to im-

prove the means of production and of traffic in
states, both for external and internal trade.''32

For Tredgold, engineering was committed to

making things "for the use and convenience of
man." But, for Tredgold, this was not simply a
matter of maintaining things as they are. Engin-
eering was supposed to "improve means of pro-
duction and traffic." Engineering was, by defini-
tion, an instrument of material progress.
 90/1183

But what about engineering today? Most engin-

eers would, no doubt, want to tinker with
Tredgold's definition, for example, by substitut-
ing "people" for "man." But few, if any, would
want to change its core. Engineering remains an
undertaking committed to human progress. So,
for example, the most widely adopted code of
engineering ethics in the United States begins:
"[Engineers uphold and advance the integrity,
honor, and dignity of the engineering profession
by] using their knowledge and skill for the en-
hancement of human welfare."33

Why Engineers are Not Scientists

That is enough about engineering. We are ready

to see how engineers differ from other technolo-
gists. I have already pointed out some of the
ways engineers differ from architects. I shall
now explain how they differ from applied
scientists.
 91/1183

I once did a workshop at the research lab of a

large petroleum company. The audience was
about half chemists and half chemical engineers.
I first asked the chemists, "If you had a choice
between inventing something useful and discov-
ering new knowledge, which would you
prefer?" The chemists thought this a hard ques-
tion: "After all," they reasoned, "it's hard to ima-
gine an interesting discovery in chemistry that
would not have a practical payoff." Eventually, I
asked for a show of hands. About half the chem-
ists voted for "something useful" and about half
for "new knowledge." The engineers, on the oth-
er hand, all voted for usefulness. For them, new
knowledge was just a means to improving hu-
man life.34
 92/1183

Unlike the chemists, the engineers had no com-

mitment to science as such. They used science,
much as they used other sources of insight. They
also contributed to science much as they con-
tributed to lawyering and other social practices,
for example, by helping to reduce the cost of
chemicals used in the manufacture of com-
puters. But they did this as nonscientists, as par-
ticipants in a voyage of invention rather than
discovery.
 Page 16

This difference between engineers and chemists

came as a surprise to these researchers. Many
had worked side by side for decades. They
thought that they shared the same values. But
we do not wear our values on our clothing like
an identity badgeexcept, of course, by declaring
ourselves to be of this profession rather than of
that. These researchers had not taken the differ-
ence in profession seriously. That is why they
were surprised.
 94/1183

This difference between scientists and engineers

is not a mere idiosycracy of one industrial labor-
atory. I have asked the same question at other
industrial laboratories since, with much the
same result. Nor is this difference between sci-
entists and engineers restricted to industrial
laboratories. I have asked my question at uni-
versity workshops attended by both engineering
and science faculty. The results were even
sharper. The only thing rarer than a university
scientist who voted for something useful was an
engineer who voted for new knowledge. 35 In
this respect at least, engineers are not scientists,
not even applied scientists. The primary com-
mitment of engineers is not to knowledge, theor-
etical or applied, as one would expect of scient-
ists, but to human welfare.36

Ethics and Engineering

 95/1183

Earlier, I described engineering as a "new power

in the world." Power, though neither good nor
bad in itself, is always dangerous. Because of
the scale on which engineers generally work,
engineering is particularly dangerous. Engineers
long ago realized this and set about to ensure, as
much as possible, that engineering would be
used for good rather than evil. They organized
as a profession, adopted codes of ethics, and
tried to put the codes into practice.
 96/1183

"Ethics" (as noted in the Preface) has at least

three common uses. It can refer not only to or-
dinary morality or the systematic study of ordin-
ary morality but also to those special morally
permissible standards of conduct every member
of a group wants every other member to follow
even if that would mean having to follow the
standards too. It is in this third special-standard
sense, I think, that members of a profession talk
of their "profession's ethics."37 In this sense, en-
gineers did not have a code of ethics until they
adopted one. In the United States, that was not
until early in this century.38

In this sense, too, their code of ethics is what

they make itas long as the standards they lay
down are consistent with ordinary morality.39
That means that engineering ethics can change
over time and even differ from country to coun-
try or field to field.
 97/1183

There is, then, an important difference between

moral values, such as the engineer's commit-
ment to improving human life, and ethics (spe-
cial standards of conduct). While engineers gen-
erally seem to have valued human welfare since
early in the history of engineering, failing to
treat the public welfare as paramount in their
work could not be unethical (in my third sense
of "ethical") until engineers adopted a standard
of conduct requiring them to treat the public
welfare as paramount. Nor could failing to treat
the public welfare as paramount as such be im-
moral until then. Morality imposes no such duty,
though it does require us to avoid doing certain
kinds of harm and perhaps to render certain
kinds of aid.40 The history of engi-
 Page 17

neering ethics reminds us that ethical standards,

like other engineering standards, are not discov-
eries but useful inventions.

Values such as honesty, safety, or efficiency are

reasons for acting, and so reasons for adopting
standards of conduct, but values as such, even
moral or professional values, do not tell us how
we should act. A value as such can only demand
consideration, that is, a certain weight in our de-
liberations. Values such as efficiency, safety,
and even honesty are considerations to be taken
into account in deciding what to do. They can-
not, as such, be obeyed or disobeyed. In con-
trast, standards of practice, including a code of
ethics, do tell us how we should act. They do
not lend themselves to weighing. They are im-
peratives we can only obey or disobey. And, as
we see later, they deserve special attention in
any discussion of engineering ethics.
99/1183
 Page 18

2
A History of Engineering in the United
States
 101/1183

This chapter continues our foray into the history

of engineering, showing (1) that engineering (in
the United States at least) is, in part, a fusion of
several older activities (managing, craft work,
science, and invention); (2) that conceptions of
engineering changed substantially over the last
two hundred years and also varied somewhat
from industry to industry (with one activity,
then another, seemingly central); and (3) that
there are nonetheless real limits to what can be
engineering, limits demonstrated especially by
attempts to train engineers that apparently failed
because they overemphasized one or another
activity. Though engineering is undoubtedly a
"social construct" in some sense or other, an en-
gineer is not an engineer simply because society
confers the title. The work engineers do has a
discipline of its own. Any adequate understand-
ing of engineering must acknowledge that dis-
cipline. Central to that discipline is a certain
way of educating engineers (a certain
 102/1183

curriculum) and a certain way of using what en-

gineers know (a code of ethics).

In the Beginning

The first engineers in the United States, or at

least the first to bear the title, were officers in
the Revolutionary War; the first school of engin-
eering here was a military academy, West Point.
1 This connection between engineering and the
military was no accident. As explained in
chapter 1, engineering began with the great
army built by Louis XIV after 1661. Though en-
gineers were soon called on for civilian project-
sto build roads, bridges, and canals; to construct
mines and oversee their operation; or to con-
struct shipsmost of the training of these "civil
engineers" was identical to that of military en-
gineers. So, for example, when the French reor-
ganized engineering education in 1794, creating
the École Polytechnique, they put students
103/1183
 Page 19

of military and civilian engineering side by side

for three years, separating them only in their
fourth (and final) year of training, when they
were sent to one or another school of "applica-
tion" (the school of military engineering, the
school of bridges and roads, and so on). After
1797, all students at the École Polytechnique
wore uniforms and lived under military discip-
line. 2
 105/1183

Establishing an engineering school in the United

States in the first years of the republic was not
easy. The first attempt occurred when George
Washington was still only a general. Other at-
tempts followed. Even with an Act of Congress
in 1802, more than a decade passed before West
Point had examinations, grades, or even a settled
curriculum. The curriculum settled on, four
years in length, was derived from the École
Polytechnique. Along with the curriculum came
a small library, recitations, examinations, one
French officer, several textbooks, and even the
use of blackboards.3
 106/1183

Though another two decades would pass before

anyone successfully copied West Point, the first
attempt came soon. Alden Partridge graduated
from West Point in 1805, taught mathematics
there for the next fourteen years, and briefly
served as superintendent, leaving under a cloud.
In 1820, he opened his own schoolthe American
Literary, Scientific, and Military Academyin his
home town, Norwich, Vermont (just across the
Connecticut River from Dartmouth College), to
train officers for the army and engineers for
public works.4 In 1824 he moved the academy
to Connecticut; in 1829 he moved it back to
Norwich. In 1834 the academy became Norwich
University, apparently without any change of
purpose, and so remains to this day, an experi-
ment complete and forgotten (though it moved
once more, in 1865, to Northfield, Vermont).
 107/1183

Though Captain Partridge's school enrolled al-

most as many students as West Point between
1820 and 1840, it did not do nearly as well as an
engineering school. Of West Point graduates
through 1837, 231 became civil engineers; of
Norwich graduates during the same period, only
about 30 did (and they generally held less re-
sponsible positions).5

The 1830s were more hospitable to copies of

West Point than the 1820s; the next decade,
even more so. The Virginia Military Institute
was founded in 1839; the Citadel, South
Carolina's military college, in 1842; and the
Naval Academy at Annapolis in 1845.6 What
was true of engineering education in general
was certainly true of civil engineering. The late
1830s mark the real beginning of civil engineer-
ing education in the United States.
 108/1183

The age of Rensselaer Polytechnic Institute, our

oldest school of civil engineering, may seem to
refute this claim. But Rensselaer, founded in
1823, is in fact evidence for the claim, not
against it.

Rensselaer was founded without either "poly-

technic" or "institute" in its name. Like Nor-
wich, it went through several changes, though it
never moved. Stephen Van Rensselaer, a gentle-
man farmer with a Harvard degree, gave the
school both his name and money to train teach-
ers of agriculture and mechanical arts for the
grammar schools of his locale. The original cur-
riculum was a single year (as one would expect
of a normal school of the day).

But by the 1830s, Rensselaer had become a kind

of scientific finishing school for graduates of
colleges of liberal arts like Harvard or Dart-
mouth. It may, in fact,
109/1183
 Page 20

rightfully claim to be the first American gradu-

ate school. Many of the graduates of this period
became important in American geology, botany,
and geography. 7
 111/1183

But Rensselaer was not yet an engineering

school. It did not award a degree in civil engin-
eering until 1835 and did not have a distinct en-
gineering curriculum until the late 1840s.8 That
curriculum, three years in length, along with the
school's present name, seems to owe much to an
1847 trip to Europe by the school's director,
young Benjamin Franklin Greene (who himself
graduated from Rensselaer in 1842 with one of
its first degrees in engineering).9 Yet, the addi-
tion of "polytechnic" to Rensselear's name may
not signal any direct connection with the École
Polytechnique. By then, Europe had other poly-
technics (all modeled, more or less, on the
French original).10 What the new name certainly
did signal was that thereafter Rensselaer would
focus on training engineers rather than scientists
and that French schools, rather than American
or English, directly or indirectly, provided the
model.11
 112/1183

Why did the first engineering schools in the Un-

ited States use French models? The answer is
simple: The French then provided the only prac-
tical models. The English, although already
leading Europe in manufacture in 1800, would
not have a respected school of engineering until
well after midcentury,12 and whether we even
say the English had civilian engineers in 1800
depends on how close we judge the analogy
between the skills of the mostly self-taught
mechanics, industrialists, and builders of Eng-
land and the French engineers whom they ad-
mired and studied.13 The English did well with
what was, in effect, training through apprentice-
ship in a craft. In 1800, the United States was al-
most without engineers (or protoengineers) to
whom apprentices could be sent.14 So, like most
of Europe, the United States copied France.
 113/1183

All our early engineering schools focused on

mathematics, physics, chemistry, and drawing.
There was also a good deal of bookkeeping, sur-
veying, measurement, and other practical sub-
jects. There was little of the Latin, Greek, or
Hebrew; classical literature; or rhetoric charac-
teristic of the liberal arts college of the day,
though there might be enough French (or Ger-
man) to read untranslated texts.
 114/1183

The difference between these early engineering

schools and the liberal arts colleges of the day
was not, however, that the engineering schools
taught science while the liberal arts colleges did
not. By 1800 Harvard, Brown, William and
Mary, North Carolina, and the other important
colleges already had professors of mathematics
and natural science.15 The early engineering
schools differed from the liberal arts colleges
primarily in offering an education that was ex-
plicitly practical in a way that the college educa-
tion of the day was not. But practical for what?
The historian Charles O'Connell tells a story that
suggests an answer.
 115/1183

In 1825, James Shiver led a team of civilians to

survey the route for an extension of the National
Road in Ohio. Because the road was a project of
the Army Corps of Engineers, Shiver reported to
Colonel Macomb, the Army's chief engineer, in
Washington. Shiver was soon reporting that his
team found it impossible to use the Army's
standard forms. Macomb wrote back that the
forms "were conceived to be more full and dis-
tinct, and consequently better adapted to the ful-
fillment of the purposes for which they were in-
tended," than what Shiver proposed instead.
But, because Macomb had dealt with civilians
before, he made allowances. The "civilian
 Page 21

brigade" could use Shiver's simpler forms for

now but should switch to the Army's forms "as
soon as they shall be understood." 16

Shiver was a competent civilian used to working

the way civilians then did. Macomb spoke for an
organization more complex than any other in the
United States. In truth, the Army's ways made
sense only in the Army. The United States was
then largely rural, with most citizens living in
towns with populations under 2,500. Its in-
dustry, although already inventive, still con-
sisted almost entirely of small companies. Such
companies did not need, or even understand, the
standardization the Army took for granted.17
 117/1183

Even a major project like a canal could still be

undertaken without engineers. Indeed, the
greatest of them all, the Erie Canal, was begun
about the time West Point settled on a cur-
riculum (1817) and completed about the time
Rensselaer was founded (1825). Though often
called "America's first engineering school," the
Erie was mostly a school of hard knocks. Those
in responsible charge were surveyors, lawyers,
or gentleman farmers. They learned as they
went, sometimes from visits to other canals or
from books and sometimes from experience.18
Whether these "canal engineers" are properly
engineers at all is, like the analogous question
about the British "engineers'' of the same period,
one that can be answered either way, depending
on whether one chooses to emphasize the analo-
gies with today's engineers (what they built) or
the disanalogies (their training and methods).
They are marginal cases. Treating them as clear
 118/1183

cases of engineers brings into engineering many

who clearly do not belong.

What was true of the early canals was not true

of the early railroads. Even the Baltimore &
Ohio Railroad (B & O), often compared to the
Erie Canal and called "America's first school of
railroad engineering," employed school-trained,
especially West Pointtrained, engineers from the
time work began in 1824.19

What explains this difference between the

canals and the railroads? At least four factors
seem relevant:
1. While canals were an old technology, railroads
were not. Insofar as railroads were a new techno-
logy, experience counted for less and a know-
ledge of fundamental principles for more.
 119/1183

2. Second, railroads required more centralized

planning than did canals. The chief economic ad-
vantage railroads had over canals was speed.
Speed was possible only if lines were clear, wa-
ter and wood were available at set distances, re-
pair crews could be sent out quickly, and so on.

3. By 1824 West Point had been in existence

long enough for its graduates to prove them-
selves likely to be useful to railroads.

4. West Point graduates brought with them styles

of organization that suited engineers. So, for ex-
ample, in 1829 Lieutenant Colonel Long, having
worked on the B&O for five years, published the
first Rail Road Manual, a book on which later
railroad engineers, schooled or not, would rely.20
There are many striking similarities between this
manual and the Army's.21
 120/1183

Even so, the railroads of the 1820s or 1830s

were not the domain of engineers they would
become. The true achievements of American en-
gineers of this period are of a different order.
For example, between 1825 and 1840, the
Army's arsenal in Springfield, Massachusetts,
developed procedures eventually much admired
in
 Page 22

Europe as "the American System." This system

made weapons parts interchangeable to a degree
never before achieved; it also subjected skilled
workers to a new discipline, including the sub-
stitution of an hourly wage for the traditional
piece rate. The arsenal was a model for later
mass production. 22

In 1850, the first year the census counted engin-

eers, only about two thousand Americans identi-
fied themselves as nonmilitary engineers, two
thousand in a population of about twenty-three
million (that is, about one in ten thousand).23
Today, in a population barely ten times larger,
we have a thousand times that number of engin-
eers (that is, about one in one hundred).
 122/1183

Engineering is sometimes described as a "cap-

tive profession" because most engineers work in
large organizations (General Motors, Westing-
house, Dow Chemical, IBM, and so on).24
Engineering is contrasted with such "free pro-
fessions" as law and medicine, where most
members practice as individuals or in small
groups (or, at least, did so until recently). Unfor-
tunately, the term "captive" gives the wrong em-
phasis to an important insight. Although we do
need engineers for the vast undertakings typical
of large organizations, engineering is no more a
captive of those organizations than the heart is a
captive of the body. The relationship between
engineering and certain large organizations, like
that between the heart and the body, is symbiot-
ic. Work in large organizations is not a night-
mare from which engineers will someday wake;
it is their natural habitat. We don't need the
skills of engineers to do what machinists, drafts-
men, architects, carpenters, millwrights, and the
 123/1183

like can do alone or in small groups. Engineers

are numerous only when there are large organiz-
ations to employ them in large undertakings. In
1850 the United States still had few such
organizations.

Middle Period:
The "Fragmenting" of Engineering

In the United States of 1850, civil engineers still

thought of engineering as a single occupation. In
1867, when a few hundred of them established
the first national engineering society, the Amer-
ican Society of Civil Engineers, any civilian en-
gineer could join.25 But, even then, engineering
had begun the branching into specialties that
would, by 1920, produce five major societies
(for civil, mining, mechanical, electrical, and
chemical engineering) and many smaller organ-
izations, each with membership requirements
excluding most other engineers.26
 124/1183

The history of the half century from 1870 to

1920 can be read as tragedy: the loss of the
primal unity of engineering under the impact of
industrialization. One history of mechanical en-
gineering even titles its chapter on this period
"Engineering: The Fragmented Profession."27
There are at least four reasons not to read his-
tory this way:
First, much of the history of engineering, not just
of this period, is a history of such branching. The
first branching separated French civil from
French military engineering in the middle of the
1700s.28

Second, the primal unity of engineering is itself

dubious. The period could equally well be por-
trayed as the one in which engineering became a
single identifiable occupation of which civil en-
gineering was but a part. In the United States of
1870, it was, I think,
 Page 23

still not clear what relation civil engineering had

to the "mechanic arts," bridge building, mining,
or metallurgy. Were they all engineering? Even
in 1896, when Columbia finally admitted that its
school of mining, founded in 1863, had long
since become what we would call a school of en-
gineering, the name became "the School of Min-
ing, Engineering, and Chemistry." Apparently,
even in the 1890s, "engineering" still did not in-
clude everything we now mean by that term. 29

Third, the half century from 1870 to 1920 was a

period of great success for engineering. In 1880,
the United States, with a population of forty mil-
lion, counted seven thousand civil engineersmore
than triple the number in 1850 (while the general
population barely doubled). Yet this impressive
increase gave no indication of what would hap-
pen during the next four decades. The 1920
census reported 136,000 engineers, twenty times
the number in 1880 (in a population that had
again barely doubled).30
 126/1183

Fourth, the enormous branching of engineering is

inevitable given the enormous growth of indus-
tries that rely on engineers.

Engineering has an important connection with

mathematics and natural science, as the similar-
ity between early engineering curricula and
today's suggests. But engineering is more than
mathematics and natural science: Engineers
know how to organize work, give instructions,
and check outcomes. This knowledge varies
from industry to industry. So, for example, a
civil engineer designing pipes that ordinary
plumbers are to install should not use tolerances
an aerospace engineer could use without a
second's thought.31
 127/1183

This field-specific knowledge is largely the res-

ult of experience, originally the experience of
individual engineers, "field experience" as well
as the results of tests in a laboratory or pilot
plant. Because engineers routinely record and
report their experience in the same way, this in-
dividual knowledge gets passed on to other en-
gineers with whom they work. Eventually, much
of it ends up in the tables and formulas that fill
the manuals written for those in the field. From
there, it works its way into customer specifica-
tions, government regulations, and courses
taught those entering the field. Though this
knowledge generally takes the form of graphs,
equations, mathematical formulas, and drawings
of things, it has little to do with natural science.
It is congealed experience of how humans and
things work together.32
 128/1183

Engineers often complain that when new tech-

nology worksthink, for example, of the space
shuttlescientists get the credit, but, when it fails,
engineers get the blame.33 Although engineers
are, I think, right about how praise and blame
are often distributed, I don't think they should
complain. That distribution is a compliment to
engineersthough one given with the back of the
hand. It implies that scientists only experiment
and experiments generally fail, whereas engin-
eers engineer and engineering generally suc-
ceeds. An engineer's failure is noteworthy for
the same reason a scientist's success isit is unex-
pected.34
 129/1183

What makes engineers so likely to succeed is

not their knowledge of mathematics and natural
science. That they share with scientists. What
makes them so likely to succeed is their know-
ledge of particular industries, what works and
what does not work there, what engineers call
"engineering science." Consider, for example,
the safety factor for steel struts supporting a
bridge. Setting a reasonable margin of safety for
such a structural component requires taking into
account, among other things, past failures to
catch flaws in materials, likely errors in main-
tenance, and unpre-
 Page 24

dicted changes in use of a structure that, prop-

erly maintained, can last for centuries. Such
knowledge is not the domain of any natural sci-
ence. It is sociological knowledge, a knowledge
of how people and tools work together, but it is
nonetheless engineering knowledge. Only en-
gineers know much about such things.
 131/1183

Here we reach another insight into engineering.

Though engineers often describe themselves as
applying natural science to practical problems,
they could just as easily, and more accurately,
describe themselves as applying knowledge of
how people work in a certain industry. Engin-
eering is at least as much management as it is
natural science. All engineers share the ability to
give mathematical structure to the problems
they encounter, the ability to draw on the natural
sciences for help in developing solutions, and
the ability to state each solution as "a design" or
set of useful specifications or directions. But
these designs, specifications, or directions are,
in effect, rules governing someone's work. 35
Engineering is, and always has been, technical
management.36
 132/1183

Technical management requires detailed know-

ledge of particular techniques. When such
knowledge becomes so great that no one can
learn it all, knowledge of techniques in one in-
dustry will exclude similar knowledge of tech-
niques in other industries. Engineers will have
to specialize and that specialization will tend to
break along industry lines.
 133/1183

But other occupationslaw and medicine, for ex-

amplespecialized without fragmenting in the
way engineering has. Lawyers have the Americ-
an Bar Association; physicians, the American
Medical Association. Why then should engin-
eers not have an American Engineering Associ-
ation rather than so many interlinking societies,
boards, councils, joint committees, and institutes
that no engineer knows more than a part? The
branching of engineering probably was inevit-
able; not so this fragmentation. Although I agree
that the fragmentation of engineering was not
inevitable, I think comparison with law and
medicine will help explain why it was nonethe-
less likely.
 134/1183

Until recently, a majority of lawyers and physi-

cians worked alone. Their employers, the client
or patient, might come in with any sort of prob-
lem. An unspecialized practice maintained a
common body of experience in law and medi-
cine for which engineering had no counterpart
since well before 1900.
 135/1183

Today, of course, that common experience has

largely disappeared. Both lawyers and physi-
cians now frequently specialize and their profes-
sional societies, once relatively homogeneous
and unified, are now divided into hundreds of
"sections" as diverse and almost as independent
as engineering's separate societies. Still, few
lawyers or physicians work the way engineers
long have. Though both lawyers and physicians
now commonly work in groups just as engineers
do, they seldom work in the same kind of group.
Few lawyers, and even fewer physicians, work
on projects requiring coordination among even a
hundred other lawyers or physicians. Few work
on projects in which everyone else has the same
specialty. Engineers, in contrast, generally work
with engineers in their own field: civils, with
civils; mechanicals, with mechanicals; and so
on.37 Often engineers must work with hundreds
or thousands of other engineers. (For example, a
single nuclear power plant needs several
 136/1183

hundred engineers on site just to operate.) The

names of specialties in law and medicine derive
from problems any client or patient can have.
The client or
 Page 25

patient still provides a common experience for

lawyers or physicians. That is not true of engin-
eering. In engineering, the specialties generally
take their name from a kind of employer or cli-
ent, the industry in which engineers of that kind
predominate. Engineering could remain a single
occupation only when engineers had so little to
do that they had little reason to specialize.

Who is an Engineer?
 138/1183

Almost from the beginning of engineering, en-

gineers have disagreed about the relative im-
portance of the scientific (especially, mathemat-
ical) knowledge engineers share and the specific
practical knowledge that tends to divide them.
Those emphasizing practice tended to take an
interest in professional ethics; those emphasiz-
ing science did not. 38 We can learn a good deal
about what engineering isor at least what it has
becomeby taking a look at how this disagree-
ment affected the education of engineers.

The practical emphasis in engineering education

has long appealed to practitioners, especially
those who began as apprentices rather than stu-
dents: Teach engineers what they need to know
to do the job they are going to do, the extremist
would say. Forget theory. Get the engineer into
the shop as soon as possible.39
 139/1183

At this extreme, the practical approach would

exclude not only courses in the humanities, so-
cial sciences, and other typical elements of a lib-
eral education but also much engineering sci-
ence. It would, in effect, substitute vocational
training for the university education that has
long been the norm for training engineers.40
 140/1183

The early history of engineering in the United

States includes many experiments with practical
education in a college or other academic institu-
tion, all more or less short-lived. For example,
Amos Eaton, who taught civil engineering at
Rensselaer in the 1830s, described its program
in this way: "The cloister begins to give way to
the field, where things, not words, are studied."
Eaton claimed that no mathematics more ad-
vanced than arithmetic was necessary to teach
engineering, that the most important part of en-
gineering could not be learned from any book,
and that the civil engineering text used at West
Point was good only for "closet reading."41 Yet,
during Amos's tenure, Rennselaer was no more
successful at training engineers than was Nor-
wich.42 And, when Greene replaced Amos,
Rensselaer moved much closer to the scientific
extreme which, by the standard of the times,
West Point represented.43
 141/1183

Beginning with the Erie Canal, many large un-

dertakings in the United States tried the practical
approach as a way of supplying technical skill
not obtainable in any other way. Whether these
count as attempts to train engineers in the shop
is an open question. I will give just one
example.

During the 1890s, General Electric offered a

course in "practical engineering" for $100. To
be eligible, one had to be a "young man" at least
twenty-one years old and have either a degree in
civil, mechanical, or electrical engineering or
two years' experience in practical electrical
work or two years in a machine shop. The
course of study, a year long, consisted of rotat-
ing through various departments of GE's Lynn
Works: four weeks in the shop plant doing wir-
ing, two weeks in the arc department assembling
arc lamps, and so on. There was no formal in-
struction.44
142/1183
 Page 26

What are we to make of this shop training?

Notice that for this course in practical engineer-
ing, two years of work experience were con-
sidered equal to a college degree in engineering.
By the 1890s, a first degree in engineering re-
quired four years, just as it does today. So, at
GE, practice was not only a substitute for formal
education, it was, it seems, considered, year for
year, twice as good. This is a striking attitude,
especially in a company that, like GE, was then
among the technologically most advanced. What
explains GE's attitude?

We must, I think, recognize that the meaning of

engineer (and engineering) has changed over
time. The term "engineer" was vague in 1890
and, though less vague than it used to be, is still
vague today. But it is not confused.
 144/1183

A term is confused when any case to which it is

thought to apply is disputable. A confused term,
such as "round square," has inconsistent criteria
of application. "Engineer" is not like that. There
are clear cases. On the one hand, someone with
a degree in civil or mechanical engineering and
several years of successful practice is certainly
an engineer. On the other hand, train operators
and boiler tenders, though usage allows them to
be called engineers, clearly are not engineers in
the sense relevant here. Such "technicians" are
engineers only in a sense belonging to an earlier
age.
 145/1183

Though not confused, "engineer," like other

terms, is still vague. In addition to the clear
cases, there are disputed cases. One contempor-
ary dispute concerns whether an individual can,
by getting the right experience, become an en-
gineer without a degree in engineering (for ex-
ample, with only a degree in physics or chem-
istry). Complicating this dispute is a subsidiary
dispute concerning which experiences are of the
right kind. Is supervising engineering work for a
decade or so the right kind? Or must an indi-
vidual actually do some engineering himself?
And what constitutes "doing engineering"? Why
isn't supervising engineers doing engineering?
 146/1183

Back in the 1890s, the boundaries were vaguer.

Then mechanical engineers were still at pains to
distinguish themselves from "mere mechanics"
who were, in turn, something more than today's
mechanics. Mechanics then were still expected
not only to repair machines but to make im-
provements as necessary. They were still regu-
larly allowed to invent. 45 Electrical engineers
had a similar problem distinguishing themselves
from "mere electricians" who were, in turn,
something more than today's electricians.46 And
so on. Perhaps what GE then meant by practical
engineering might today be identified by a two-
or four-year degree in technology rather than en-
gineering (or even by an advanced degree in
technological management). But, back in the
1890s, such distinguishing degrees were not an
option. Engineers had to find other ways to ex-
plain how they differed from mechanics, electri-
cians, and other craftsmen with whom they
shared some tasks and much technical
 147/1183

knowledge. Engineers found only two ways to

explain the difference.

One way was to understand engineering as a

kind of management.47 Engineers issue orders;
those with technical skills merely carry them
out. Engineers are officers in the army of pro-
duction. Though it has strong roots in the history
of engineering, this way of distinguishing engin-
eers from craftsmen is plainly inadequate. It
fails to explain why engineers should be in
charge. The explanation cannot simply be that
the employer so ordains. If being put in charge
of engineering work is all that
 Page 27

distinguishes engineers from other employees,

anyone put in charge of engineering work would
be an engineer. Engineers have generally sup-
posed that engineering requires more than that
(as, indeed, their employers have as well). 48
The other way, then, is to see that engineering
requires knowledge craftsmen do not have:
Engineers can give orders to craftsmen because
engineers know things that mere craftsmen do
not. This claim, though plausible, is plausible
only if the knowledge in question depends, at
least in part, on training outside the shop.
Knowledge of natural science and advanced
mathematics certainly is such knowledge.
Hardly anyone would suppose much of those
subjects could be learned in the shop.
 149/1183

That is one advantage of understanding engin-

eering as fundamentally "scientific" rather than
"practical". There are at least three others. First,
if engineering was to be a profession, like law or
medicine, not just a job title such as "manager,"
engineers could not let being an engineer de-
pend on how an employer happened to define
the engineer's job. Credentials, not employment,
had to define the engineer. Second, a common
academic training is generally considered one
crucial mark of a profession. Insofar as engin-
eers considered engineering a profession, or
wanted engineering to be one, they tended to
emphasize academic training. Third,
engineering's unity, insofar as it exists, depends
heavily on all engineers having an education
that they share with each other (a basis for the
"engineering method" engineers believe all en-
gineers share). Emphasis on what goes on in the
shop stresses just those features of engineering
that threaten to divide engineering into many
 150/1183

mutually incomprehensible occupations. In con-

trast, emphasis on ''engineering as science"
seems to confirm the sense most engineers have
that, for all the immense differences between
fields, virtually all engineers share something
that distinguishes them both from ordinary
workers and from ordinary managers.49

The question, "Who is an engineer?," sounds

like a philosopher's questionand it is. But it is
also a practical question: Every engineering so-
ciety that decides, as most do, to limit member-
ship (or a certain category of membership) to
engineers has to define engineer with more or
less precision. The historian Edwin Layton
taught us much about the consequences of ad-
opting various definitions. Definitions close to
the practical pole tend to turn engineering soci-
eties into trade associations; definitions close to
the scientific, to exclude many who shape the
projects engineers carry out and do much to
maintain discipline among engineers.50
 151/1183

Layton, however, taught us that while failing to

make clear how hard it is to say what an engin-
eer is. In particular, he failed to notice that, at its
extreme, engineering as science can be as dis-
astrous for engineering as "engineering as prac-
tice." Training engineers as scientists, if only as
"applied scientists," tends to turn out scientists
rather than engineers.51 Consider, for example,
the Lawrence Scientific School, founded in
1847 as part of Harvard, to teach: "1st, Engin-
eering; 2d, Mining, in its extended sense, in-
cluding metallurgy; 3d, the invention and manu-
facture of machinery."52 Plainly, Lawrence was
supposed to be an engineering school. By 1866,
Lawrence had graduated 147 students: 94 of
these became professors or teachers; 5 became
college presidents; but only 41 actually became
engineers (as against 126 from Rensselaer dur-
ing the same twenty years).53 The Massachu-
setts Institute of Technology opened in Boston
 152/1183

in 1865 in large part because Lawrence had

failed as a school of engineering.54
 Page 28

Nonetheless, during much of this century, espe-

cially after World War II, engineering education
moved ever closer to the scientific extreme. Pro-
grams in specialized fields of engin-
eeringeverything from agricultural engineering
to telephone engineeringdisappeared from the
undergraduate curriculum, leaving only the lar-
ger divisions: civil, chemical, electrical, and the
like. And even courses in these fields tended
more and more to emphasize general principles,
calculations, and laboratory work. Students were
left to learn the art of engineering after gradu-
ation, if at all. 55
 154/1183

Only recently did engineering schools begin to

move back toward practice. They did so largely
under pressure from industry and the board that
accredits engineering schools. But, this counter-
movement did not mean a return to the shop.
Engineering schools, instead, began to think of
engineering in a new wayas fundamentally con-
cerned with design.56 Some results of this new
thinking are already in placefor example, senior
courses in engineering design. Other results are
only now showing upfor example, as design ele-
ments in junior or even sophomore courses in
engineering science. And some results are only
at the stage of talk or experimentfor example, as
attempts to include in design courses everything
from the ethical issues a design might raise to
the practical problems of getting colleagues and
superiors to adopt one's design.57
 155/1183

In retrospect, these recent developments seem

both sound and overdue. The stereotype of en-
gineering as the logical or, rather, mechanical,
solution of practical problems by deduction
from scientific principles misses the creative
side of much engineering, something that should
have been obvious from the striking newness of
so many works of engineering, whether the
bridges of the early railroad engineers or the be-
wildering variety of today's computers.
 156/1183

Of course, engineering is not only inventive-

ness, just as it is not only science or only man-
agement. We want engineering rather than mere
invention in many departments of life in part be-
cause engineers work within constraints other
inventorswhether architects, applied scientists,
industrial designers, or mere handymendo not.
Engineers have distinctive routines for ensuring
safety, economy, reliability, durability, manu-
facturability, and so on. These routines, and the
engineering science behind them, are subordin-
ate to engineering design. But, though subordin-
ate, they are fundamental to engineering, much
as a certain pattern of rhyme and meter is to
making a sonnet.
 157/1183

Who then is an engineer? Today we must an-

swer: anyone who can design as engineers do.58
Unfortunately, we have only the roughest idea
of what engineering design is. Today, the philo-
sophy of engineering is where the philosophy of
science was a hundred years ago. We have
barely begun to understand that there is a ques-
tion.59

Ethics and the Profession of Engineering

So far in this chapter, I have discussed engineer-

ing primarily as an "occupation," not a "profes-
sion." I had a reason. While the old expression
''liberal profession" referred to any occupation
suitable for gentlemen, the modern use of "pro-
fession" requires moreorganization, with stand-
ards of admission, including both training
 Page 29

and character, and standards of conduct beyond

the merely technical. 60 In 1850 engineering
was still not a profession in this sense; nor was
it so in the United States even in 1900. Today it
is. What explains the change?

Until this century, engineering societies in the

United States were primarily scientific or tech-
nical associations. So, for example, the Americ-
an Society of Civil Engineers (ASCE) was es-
tablished with the purpose of "advancing know-
ledge, science and practical experience among
its members, by an exchange of thoughts, stud-
ies, and experience."61 There was no suggestion
either of improving the formal education of en-
gineers or of setting standards of conduct.
 159/1183

Indeed, the first efforts to set minimum stand-

ards for engineering education came from the
engineering schools, not from practicing engin-
eers. In 1893, at the Columbian Exposition in
Chicago, seventy engineering teachers organ-
ized the Society for the Promotion of Engineer-
ing Education (SPEE), later to become the
American Society of Engineering Education
(ASEE).62 While SPEE undertook a number of
valuable studies of engineering education, mak-
ing many influential recommendations, not until
1932 did the major engineering societies estab-
lish the Engineers' Council for Professional
Development (ECPD) to accredit engineering
curricula.63
 160/1183

The adoption of standards of conduct began

earlier. Indeed, in one sense, it began when en-
gineers first distinguished themselves from
those unable to work the way engineers do.
Engineering can be defined, in part, by stand-
ards of competenceand standards of competence
are, in a sense, standards of conduct. But every
skilled occupation has standards of conduct in
this sense, and some, like trade associations or
scientific societies, may be organized to main-
tain them. Ethical standards, not standards of
competence or organizations, seem to distin-
guish professions from other skilled occupa-
tions.64
 161/1183

Engineers in the United States lacked distinctive

ethical standards until the second decade of this
century. Why did they not adopt such standards
earlier? Why did they adopt them then? My
guess is that engineers did not adopt ethical
standards earlier for the same reason that most
of today's professions, including law and teach-
ing, did not. They did not see the need.65
 162/1183

Until this century, engineering was a clubby af-

fair. There were relatively few engineers and
those few worked in a small world in which gos-
sip maintained what discipline was necessary.
But by 1900 that time had passed. Cities grew
up where small towns had been. The big cities
of 1850 or 1870 had tripled, quadrupled, or
quintupled in size. The same thing happened to
the companies for which most engineers
worked.66 And engineering itself grew enorm-
ously. The few thousand engineers of 1870 had
become more than a hundred thousand and
seemed likely to continue to increase rapidly.
By 1900 most engineers were young. Old sys-
tems of apprenticeship were being swamped.
College or technical school was, or at least soon
would be, the primary route to a career in
engineering.
 163/1183

The old men of the profession naturally sought

new means to do what they could no longer do
by the old. A formal code of ethics must have
seemed one way to help the young understand
what was expected of them. So, early in this
century, each of the major engineering societies
set up a committee to draft a code of engineer-
ing ethics, but the drafting proved harder than
expected. The committees found that they
agreed on less than they had supposed; even de-
termining what that little was
 Page 30

took much effort. 67 The societies were not only

writing down what they agreed on, they were
also hammering out new agreements. What
began as an attempt to preserve the past ended
in a new professionin two senses. First, engin-
eers began professing something new; they
committed themselves to a specific code of eth-
ics. Second, their organizations were no longer
mere technical societies; they constituted an oc-
cupation organized to carry on certain work in
accordance with standards beyond what law,
market, and morality demandedthey were a
profession.
 165/1183

After World War I, there was a smaller round of

code writing; after World War II, another; then,
starting in the 1970s, the largest round yet. All
this code writing produced much coordination
among major engineering societies and substan-
tial agreement on what a code of ethics should
contain. Today, engineers have relatively clear
standards of conduct they can look to for guid-
ance and can cite when offering advice to one
another, when criticizing one another's work, or
when seeking to prevent certain conduct.
Chapters 4 and 8 provide examples of those
standards. What engineers still lack is a system-
atic way to protect members of their profession
who act ethically when an employer or client
wants something else. As with other profes-
sions, so with engineering: Ethics is unfinished
business.

My Method
 166/1183

We all have a tendency to see institutions, pro-

fessions, and even people as more or less com-
plete, as Platonic ideas dropped into history.
This is plainly a mistake when trying to under-
stand people. We all know that however smooth
the surface we show the world, we are all beings
ever changing or, at least, ever capable of
change.68
 167/1183

Because I believe this to be true of professions

as well, I try to describe engineering as an
evolving institution, one that people much like
us have made, not always intending what they
achieved, imperfect, as all human works are,
and therefore capable of improvement. I believe
that thinking of engineering in this way will
help engineers both to understand and to resolve
the ethical problems they face. I also believe
that thinking of engineering in this way helps
the rest of us understand engineering. In the
chapters to follow, we shall see whether that is
so.
 Page 31

3
Are "Software Engineers" Engineers?
Today, the field has emerged as a true engineering
discipline."
John J. Marciniak, "Preface," Encyclopedia of Soft-
ware Engineering (1994)

If you are a "engineer," you could be breaking the

law. It is illegal in 45 states to use that title, warns
Computerworld newspaper. People who aren't edu-
cated and licensed in 36 recognized engineering dis-
ciplines can't call themselves "engineers," and com-
puter professionals often don't qualify.
Wall Street Journal, June 7, 1994, p. 1.
 169/1183

For those interested in professions, the emer-

gence of what may be a new profession should
generate the same excitement that the discovery
of a new class of objects in the sky generates
among astronomers. Not only is it inspiring to
watch, it is a chance to put theory to work in un-
expected ways, a chance to separate the charm-
ing from the true. This chapter begins with the
emergence of software engineering as a distinct
discipline, occupation, and, perhaps, profession.
 170/1183

The term "software engineering" came into cur-

rency after a 1967 North Atlantic Treaty Organ-
ization conference on software design and test-
ing used that term in its title. 1 Today, thousands
of people are called software engineers, do
something called software engineering, and
have sophisticated employers willing to pay
them to do it.2 Yet, software engineering is no
ordinary engineering discipline. Few software
engineers have a degree in engineering. Some
are graduates of a program in computer science
who had a single course in "software engineer-
ing." (Typically, that course is taught by
someone with a degree in computer science
rather than engineering.) Most software engin-
eers are programmers with no formal training in
engineering.3 Are software engineers nonethe-
less engineers? What, if anything, makes this
question worth answering?
 171/1183

Let me answer the last question first: Defining a

field is more than semantics. How we define a
field can affect how it develops. Software engin-
eering may be a field whose progress is
threatened by the analogy with engineering, a
field pushed toward an unnecessarily rigid cur-
riculum.4 That is the first reason our questions
 Page 32

about software engineering are worth answer-

ing. Second is that trying to answer them will
help us understand engineering. What are its
boundaries? What is at stake when we draw
such boundaries? A third is that trying to answer
such questions tests the utility of our history of
engineering. What insight can this history give
us?
 173/1183

The insight may be disappointing. What I show

is that we can't tell whether software engineer-
ing is engineering. Only the future can tell. The
best we can get from engineering's history is in-
sight into why that is so. Getting that insight
leads me to defend two theses: First, that soft-
ware engineers are not engineers merely be-
cause they do much that engineers do or know
much that engineers know; second, that whether
software engineering can or should be a field of
engineering depends on whether software engin-
eers can or should be educated in the way engin-
eers are. These two theses rest on a third: Engin-
eering is (or, at least, should be) defined primar-
ily by its curriculum rather than, as we might
expect, by what engineers in fact do or know.
Because the defense of the other two theses rests
on the third, it is with the third that we must
begin.

The Standard Definition of Engineer

 174/1183

The standard definition of engineer is something

like this: An engineer is a person who has at
least one of the following qualifications: (1) a
college or university B.S. from an accredited en-
gineering program or an advanced degree from
such a program, (2) membership in a recognized
engineering society at a professional level, (3)
registration or licensure as an engineer by a gov-
ernment agency, or (4) current or recent em-
ployment in a job classification requiring engin-
eering work at a professional level. 5 The strik-
ing feature of this definition is that it presup-
poses an understanding of engineering. Three of
the four alternatives actually use the term "en-
gineering" to define engineer; and the other, al-
ternative (3), avoids doing the same only by us-
ing "as an engineer" instead of "to practice en-
gineering."6
 175/1183

This definition and others like it are important.

They determine who is eligible for admission to
engineering's professional societies, who may be
licensed to practice engineering, and who may
hold certain jobs. Such definitions are also em-
inently useful. For example, they help the
Census Bureau exclude from the category of en-
gineer drivers of railway engines, janitors who
tend boilers in apartment buildings, and soldiers
wielding shovels in the Army's Corps of Engin-
eers. These, though still called engineer, clearly
are not engineers in the relevant sense. They are
engineers only in a sense now obsolete.
 176/1183

However, the standard definitions do not suit

our purpose. They do not tell us whether a soft-
ware engineer is an engineeror even how to go
about finding out. A software engineer may, for
example, work at a job classified as requiring
software engineering (at a professional level).
That will not settle whether a software engineer
is an engineer: What an employer classifies as
"engineering" (for lack of a better word) may or
may not be engineering in the relevant sense.7

What will settle the question? In practice, the

decision of engineers. An organization of engin-
eers accredits baccalaureate and advanced pro-
grams in engineering. Other organizations of en-
gineers determine which societies with "engin-
eer" in the title are engineering societies and
whichlike the Brotherhood of Railway Engi-
 Page 33

neersare not. Engineers also determine which

members of their societies practice engineering
"at a professional level" and which do not.
Government agencies overseeing registration or
licensure of engineers, though technically arms
of the state rather than of engineering, generally
consist entirely of engineers. And, even when
they do not, they generally apply standards
(education, experience, proficiency, and so on)
developed by engineers. Engineers even determ-
ine which job classifications require
professional-level engineering work and which
do not.
 178/1183

The standard definition settles many practical

questions, but not ours. Because engineers di-
vide concerning whether software engineering is
really engineering, to say that software engin-
eers are engineers if they engage in professional
engineering isfor engineers and those who rely
on their judgment in such mattersmerely to re-
state the question. 8
 179/1183

That is a practical objection. There is a related

theoretical objection. A definition of engineer
that amounts to "an engineer is anyone who
does what engineers count as engineering" viol-
ates the first rule of definition: `Never use in a
definition the term being defined.' That rule
rests on an important insight. Though a circular
definition can be useful for some purposes, it
generally carries much less information than a
noncircular definition. So, for example, a dic-
tionary that defines ethics as "morality" and then
defines morality as "ethics" helps only those
who understand one of the terms but not the oth-
er. The smaller the circle, the less helpful a cir-
cular definition is.
 180/1183

How can we avoid the standard definition's cir-

cularity? The obvious way may be to define en-
gineering without reference to engineer and then
define engineer in terms of engineering. The
National Research Council (NRC) in fact tried
that approach, coupling its definition of engineer
with this definition of engineering:
Business, government, academic, or individual
efforts in which knowledge of mathematics and/
or natural science is employed in research, devel-
opment, design, manufacturing, systems engin-
eering, or technical operations with the objective
of creating and/or delivering systems, products,
processes, and/or services of a technical nature
and content intended for use.9
 181/1183

This definition is certainly informative insofar

as it suggests the wide range of activities which
today constitute engineering. It is nonetheless a
dangerous jumble. Like the standard definition
of engineer, it is circular: "Systems engineering"
should not appear in a definition of engineering.
The same is true of "technical" if used as a syn-
onym for engineering. (If not a synonym, tech-
nical is even more in need of definition than en-
gineering is and should be avoided for that reas-
on.) The NRC's definition also substitutes un-
certain listsnote the "and/or"where there should
be analysis. Worst of all, the definition is fatally
overinclusive. Not only are software engineers
engineers according to the definition, but so,
too, are many whom no one supposes to be en-
gineers, not only applied chemists, applied
mathematicians, architects, and patent attorneys
but, thanks to the and/or between mathematics
and natural science, even actuaries, accountants,
financial analysts, and others who use
 182/1183

mathematics to create financial instruments,

tracking systems, investment reports, and other
technical objects for use.
 Page 34

Though much too inclusive, this definition of

engineering shares with most others three char-
acteristic elements. First, it makes mathematics
and natural science central to what engineers do.
10 Second, it emphasizes physical objects or
physical systems. Whatever engineering is, its
principal concern is the physical world rather
than rules (as in law), money (as in accounting),
or even people (as in management). Third, the
definition makes it clear that, unlike science, en-
gineering does not seek to understand the world
but to remake it. Engineers do, of course, pro-
duce knowledge (for example, tables of toler-
ances or equations describing complex physical
processes), but such knowledge is merely (or, at
least, primarily) a means to making something
useful.11
 184/1183

Those three elements, though characteristic of

engineering, do not define it. If they did, decid-
ing whether software engineers are engineers
would be far easier than it has proved to be. We
could, for example, show that software engin-
eers are not engineers simply by showing that
they generally do not use the natural sciences in
their work. That many people, including some
engineers, believe software engineers to be en-
gineers is comprehensible only on the assump-
tion that these three characteristics do not define
engineering (except in some rough way). But if
they do not define engineering, what does?

Before answering, I describe three common mis-

takes about engineering to be avoided in any an-
swer. While these mistakes may seem far from
software engineering, they bring us to the best
point for understanding the relation between
software engineering and engineering proper.
 185/1183

Three Mistakes About Engineering

The NRC's definition of engineering uses "tech-

nical" twice, once as a catch-all ("or technical
operations") and once to limit the domain of en-
gineering ("of a technical nature and content").
It is the second use of technical that concerns us
now. It seems to be a common mistake in usage,
one even engineers make. We might summarize
the mistake this way: Engineering equals
technology.
 186/1183

There are at least three objections to this way of

understanding engineering. First, engineering
can equal technology only if we so dilute what
we mean by engineering that any tinkerer would
be an engineer (or, at least, be someone engaged
in engineering).12 Once we so dilute engineer-
ing, we are left to wonder why anyone might
want an engineer rather than some other techno-
logist who could do the same job.13 Why de-
mand a software engineer rather than a program-
mer, software designer, or the like to do soft-
ware design or development? What was the
point of inventing the term "software engineer-
ing"?14
 187/1183

Second, the proposition "engineering equals

technology" makes writing a history of engin-
eering (as distinct from a history of technology)
impossible. The history of engineering, accord-
ing to this proposition, is the history of techno-
logy. Every successful inventor is an engineer;
every successful manager of industry is an en-
gineer; and so on. We are left to wonder why
our term for engineerunlike our term for archi-
tect, mathematician, or artisanis so recent. Why
does engineering have a history distinct from
technology when engineering is technology?15
Why do engi-
 Page 35

neering organizations devote any effort to defin-

ing engineering? Why don't they just define
technology and technologist and then say, "Ditto
engineer"?

Third, "engineering equals technology" trans-

forms talk of engineering ethics into talk of the
ethics of technology. It turns professional ethics
into public policy. Whatever engineering ethics
is, it is, in part at least, the ethics of a profes-
sionnot merely standards governing the develop-
ment, use, and disposal of technology but stand-
ards governing a certain group of technologists.
 189/1183

That reference to profession suggests a second

mistake commonly made about engineering, one
we might summarize this way: Engineering is,
by nature, a profession. What makes this mis-
take attractive is the idea that a professional is a
"knowledge-worker," that special knowledge
defines each profession (as well as the underly-
ing occupation). Any occupation that requires a
lot of training is a profession. 16 Engineering re-
quires a lot of training; hence, it must be a pro-
fession. Connecting profession with knowledge
helps exclude from the profession of engineer-
ing those who, though they may function as en-
gineers (or, rather, as "mere technicians"), lack
the requisite knowledge to be engineers strictly
so called ("engineers at the professional level").
Claiming that engineering is, by nature, a pro-
fession provides an antidote to the first mistake,
but only by making another.
 190/1183

What is this second mistake? Thinking of engin-

eering as, by nature, a profession suggests that
organization has nothing in particular to do with
profession. As soon as we have enough know-
ledge, we have a profession. There could be a
profession of one.
 191/1183

Thinking this way makes much of the history of

engineering mysterious. Why, for example, did
engineers devote so much time to setting minim-
um standards of competence for anyone to claim
to be an engineer? Why did they set these stand-
ards rather than others? Why did they suppose
setting such standards relevant to being a profes-
sion? Like other professions, engineering has a
corporate history that such nonprofessions as
shoe repair, inventing, and politics lack. Any
definition of engineering must leave room for
that history. What is striking about the history of
engineeringindeed, of all professionsis the close
connection between organization, special stand-
ards, and claims of profession.
 192/1183

A third mistake may help explain the appeal of

the second. We might summarize it this way:
The engineering profession has always recog-
nized the same high standards. There are at least
two ways that this mistake is defended. One ap-
peals to the "nature" (or "essence") of engineer-
ing. Any occupational group that did not recog-
nize certain standards would not be engineersor,
at least, would not be engaged in engineering.
Engineers have organized to set standards to
avoid being confused with those who are not
"really" engineers. The standards simply record
what every good engineer knows; they codify
rather than legislate.
 193/1183

The other argument for this mistake appeals to

the moral nature of the engineer. It is said that
engineers are always generally conscientious.
To be conscientious is to be careful, to pay at-
tention to detail, to seek to do the best one can.
To do this is to be ethical. Professional ethics is
being conscientious in one's work. To be a con-
scientious engineer is, then, to be by nature an
ethical engineer.17 Engineering societies adopt
standards to help society know what it should
expect of engineers,
 Page 36

not to tell a conscientious and technically adept

engineer what to do. Informing society is, ac-
cording to this view, enough to explain the ef-
fort engineers put into codes of ethics.

What is wrong with the proposition that engin-

eering is, by nature, ethical? Like the other two
mistakes, this third makes understanding the his-
tory of engineering harder. Why have engineers
changed the text of their codes of ethics so of-
ten? Why do experienced engineers sometimes
disagree about what should be in the code of
ethics (as well as about what should be in their
technical standards)? Why do these disagree-
ments seem to be about how engineers should
act, not about what to tell society?
 195/1183

If we examine a typical code of engineering eth-

ics, we find many provisions that demand more
than mere conscientiousnessprovisions requir-
ing, for example, engineers to help engineers in
their employ to continue their education or to
make public statements only in a truthful and
objective manner. 18 Such codes are less than a
hundred years old.19 Before they were adopted,
an engineer only had to be morally upright and
technically proficient to do all that could reason-
ably be expected. In those days, engineers had
no responsibilities beyond what law, market,
and ordinary morality demanded (and, so, had
no need to inform society what to expect). The
claim that engineering has always accepted the
same high standardsthat, for example, failing to
inform a client of a conflict of interest was al-
ways unprofessionalis contrary to what we have
learned about engineering.

Membership in the Profession of Engineering

 196/1183

As we saw in chapter 2, engineering education

in the United States, almost from its beginning,
had two strands: One was a series of unsuccess-
ful experiments with various alternatives to the
West Point curriculum; the other was the evolu-
tion of the West Point curriculum into the stand-
ard for engineering education in the United
States. The details of that story do not matter
now.20 What does matter is that the education of
engineers became more and more the province
of engineering schools, and these in turn became
more and more alike. For engineers, an engineer
became someone with the appropriate degree
from an engineering school or, absent that, with
training or experience that was more or less
equivalent; hence, the standard definition of
"engineer" with which this chapter began.
 197/1183

The point of this story is not that engineering

will always have the same curriculum it does
today. The engineering curriculum has changed
a great deal since West Point was founded in
1802; for example, there is now more calculus
and less drafting. No doubt, the curriculum will
continue to change. Perhaps the second year of
calculus will disappear, with ecology or indus-
trial psychology taking its place. The point of
the story of engineering as I told it is, rather,
that just as today's curriculum grew out of
yesterday's, so tomorrow's will grow out of
today's. Any new field of engineering has to find
a place in that curriculum. Finding a place may
mean changing the curriculum; what it cannot
mean is starting fresh. Finding a place in a cur-
riculum is a complex negotiation of social ar-
rangements. It is like joining a family. You can
change your name to Davis if you like, make
yourself look like a
198/1183
 Page 37

member of my family (perhaps even genetic-

ally), and declare yourself a member of my fam-
ily, but that won't make you one. To be a mem-
ber of my family, you must come in by birth,
marriage, or adoption.

Some fields of engineering (for example, nucle-

ar) seem to be born engineering, but others
(mining, for example) seem to come in by the
occupational equivalent of marriage or adoption.
For any field not born engineering, the only way
to become a field of engineering is by
"marriage" (or "adoption"). Failing that, it can-
not be a field of engineering. It can only don
quotation marks to show irony, start another
family of the same nameas railway engineering
has, but without its historical justificationor
choose a more suitable name.
 200/1183

The history of a profession tells how a certain

occupation organized itself to hold its members
to standards beyond what law, market, and mor-
ality would otherwise demand. The history of a
profession is the history of organizations, stand-
ards of competence, and standards of conduct.
For engineering in the United States, that history
began after the Civil War. It is a confused story
because the profession was taking shape along
with the occupation. Many early members of its
professional societies would not qualify for
membership today.
 201/1183

Nonetheless, I think we can see that as engineers

became clearer about what engineers were (or,
at least, should be), they tended to shift from
granting membership in their associations ("at a
professional level") based on connection with
technical projects, practical invention, or other
technical achievements to granting it based on
two more demanding requirements: (1) specific
knowledge and (2) commitment to use that
knowledge in certain ways (that is, according to
engineering's code of ethics). The first is occu-
pational. This requirement is now typically iden-
tified with a degree in engineering. The second
is professional. Although many professions
(law, especially) make a commitment to the
profession's code of ethics a formal requirement
for admission, engineering has not (except for
licensed professional engineers, or P.E.s). In-
stead, the expectation of commitment reveals it-
self when an engineer is found to have violated
the code of ethics. The defense, "I'm an engineer
 202/1183

but I didn't promise to follow the code and

therefore did nothing wrong," is never accepted.
The profession answers, "You committed your-
self to the code when you claimed to be an en-
gineer." 20
 203/1183

Attempts to understand software engineering as

engineering have, I think, generally missed this
complexity in the concept of the profession of
engineering. Consider, for example, Mary
Shaw's observation: "Where, then, does current
software practice lie on the path to engineering?
It is still in some cases craft and in some cases
commercial practice. A science is beginning to
contribute results, and, for isolated examples,
you can argue that professional engineering is
taking place.21 Substitute "applied science" for
"engineering" in the first sentence in this pas-
sage and for "professional engineering" in the
second, and there is little to argue with. But, as
it stands, its final sentence is simply false. There
is nothing in what Shaw describes to suggest
that "professional engineering is taking place.''
 Page 38

The Fundamental Problem in Software

Engineering

The term "software engineering" was coined in

the mid-1960s to describe "the need for software
manufacture to be [based] on the types of theor-
etical foundations and practical disciplines that
are traditional in the established branches of en-
gineering." 22 Thinking about software engin-
eering thus began with the assumption that the
established branches of engineering share cer-
tain theoretical foundations and practical discip-
lines. This is an assumption that engineers gen-
erally share, calling the theoretical foundation
"science'' or "engineering science" and the prac-
tical discipline "engineering method." Yet, even
the history of software engineering puts that as-
sumption in doubt.
 205/1183

The early proponents of software engineering

disagreed concerning what engineering's theor-
etical foundations and practical disciplines are.
Some understood engineering as essentially ap-
plied science, with a theoretical foundation in
physics, chemistry, and mathematics. Others un-
derstood engineering as primarily a body of
techniques for design. For them, engineering
was primarily a way of moving from concep-
tion, through specification, to prototype, testing,
and final fine-tuning. For most, however, engin-
eering was primarily a way of organizing and
managing a process of design, development, and
manufacture, of ensuring that work would be
completed on time, within budget, and to the
customer's satisfaction.23
 206/1183

In fact, what the established branches of engin-

eering share, perhaps all they share, is a com-
mon core of courses (physics, chemistry, math-
ematics, and so on), which may or may not
provide a theoretical foundation for engineering.
Beyond that, there are important overlaps
between this and that field, many family re-
semblances and analogies, but nothing more (or,
at least, nothing more of importance). For a long
time, perhaps from its very beginning, engineer-
ing was a protean mix of activities held together
by a common education. The common education
clearly had connections with what engineers did,
but the connections were not always clear, even
to engineers.24
 207/1183

So, if software engineering is to be, strictly

speaking, a field of engineering, it has to require
of its practitioners a degree in engineering (or its
equivalent).25 Right now, the software engineer-
ing curriculum is more flexible than
engineering's. It is, I think, an empirical ques-
tion, one that remains open, whether students of
software engineering would be better software
engineers if they followed engineering's more ri-
gid curriculum rather than, say, taking more
computer science, psychology, and management
courses than engineering's curriculum allows.
How much physics, calculus, thermodynamics,
and the like does one need to design, develop,
and maintain software?
 208/1183

The answer to this question is not obvious.

Indeed, in its present form, the question is prob-
ably unanswerable. How much physics, calcu-
lus, thermodynamics, and the like a software en-
gineer needs may well depend on the kind of
software in question (not whether or not it is
"life critical" but what sort of knowledge its de-
signer should have to do it right). Although we
might worry about someone developing soft-
ware for engineering applications who didn't
know what engineers know, would we feel the
same about such a person developing a com-
puter game for children or a diagnostic program
for physicians?
 Page 39

Software engineering was not born engineering.

If it is ever to be part of engineering ("an engin-
eering discipline"), it must come in by
"marriage" or "adoption". That will require sub-
stantial changes in software engineering, engin-
eering proper, or both. Software engineering
may have to bring its curriculum up to standards
for engineering accreditation, or engineering
may have to change its curriculum to make
room for software engineering (for example, by
dropping the required chemistry course), or both
engineering and software engineering may have
to change. Software engineering cannot become
engineering simply by adopting the name, by
copying engineering methods, or even by having
some authoritative body like the IEEE declare it
engineering. Indeed, software engineers will not
necessarily be members of the engineering pro-
fession even if they receive an engineering
education.
 210/1183

Education only satisfies the occupational re-

quirement. There is also the professional re-
quirement, commitment to the engineers' code
of ethics. 26 So far, software engineers seem to
believe they can have a code of their own.27

Like the occupational requirement, the profes-

sional requirement leaves some room for man-
euver. Software engineers can have their own
code in addition to the engineers' code (that is, a
code with obligations beyond those all engineers
share). Software engineers can also try to work
out a common code with engineers, changing
what engineers require of themselves. What they
cannot do is be engineers "at the professional
level" yet refuse to share engineering's profes-
sional commitments.
 211/1183

Will software engineering ever join

engineering's family? That is a question for
prophets. What I tried to do here is to use soft-
ware engineering to reveal the complexity in the
concept of a profession of engineering.
However, I must add that the benefits of making
software engineering a "true engineering discip-
line" strike me as less certain than the discussion
so far makes them seem. Training in engineer-
ing as such will not ensure that projects come in
on time, within budget, or to the customer's sat-
isfaction. Although engineering education has
always had elements of managementmore in the
first half of this century than nowengineers al-
ways have problems delivering on time, within
budget, and to the customer's satisfaction, espe-
cially in fields such as computer development,
where experience is thin. The obvious ability of
engineers in many fields to keep their promises
seems more an indication of the maturity of the
field than of any special knowledge of engineers
 212/1183

as such. Aren't physicians and auditors just as

able to deliver on their promises?
 213/1183

Nothing said here is meant to raise questions

about the status of software engineering as a dis-
cipline, an occupation, or even a profession. My
concern is how to conceptualize this new but
already respectable occupation. Perhaps we
would understand it better if we stopped trying
to borrow concepts from engineering and in-
stead borrowed them from architecture or indus-
trial design, areas in which chemistry, physics,
and mathematics are less important, pure inven-
tion more so, and codes of ethics less detailed.
Or, perhaps we should borrow concepts from
construction management. Software engineering
may be more like overseeing the building of a
great public work (a bridge, skyscraper, or
power plant) than like doing the engineering for
it. Construction managers are at least as good as
engineers at delivering on time, within budget,
and to the customer's satisfaction.28 Or, perhaps
software engineering is more like what lawyers
 214/1183

do when they create new negotiable instruments

or complex land-use agreements.
 Page 40

The question to be asked, then, is not whether

software engineers are engineers. Clearly, while
some are, most are not. The question is, rather,
whether (or when) they should be.

My conclusion is that there is no fact of the mat-

ter here, only a complex of social decisions
about standards of training and conduct in need
of attention. Like engineering, software engin-
eering is a social project, not a natural species.
 Page 41

PART II
ENGINEERS IN CONTEXT
 217/1183

Having defined engineering, we are ready to

identify the place of ethics in the practice of en-
gineering today. Though this, too, requires some
history, I begin with a recent event, the Chal-
lenger explosion. Of course, that event is itself
important to engineering. For the public, that
event was the most traumatic engineering dis-
aster in recent memory, more traumatic even
than the two nuclear disasters, Three Mile Island
and Chernobyl. Unlike these, it produced an en-
gineer hero, Roger Boisjoly. Engineers found in
the Challenger explosion confirmation of much
they feared was wrong with corporate decision
making. However, I have two other reasons for
beginning with the Challenger, both more
mundane. First, the enormous documentation
the disaster produced allows us to get closer to
decisive events than is possible in most engin-
eering. Here is a drama from which we can learn
much; here, too, are details to provoke thought
about engineering's "mission" and the place of a
 218/1183

code of ethics in accomplishing it. Here we may

see the profession in action. Second, however
dramatic, the events leading up to the disaster
have many characteristics of ordinary engin-
eeringespecially, a large organization, coopera-
tion and conflict between engineer-managers
and ordinary engineers, a mix of technical and
business considerations, the problem of defining
what is and what is not a question of engineer-
ing, and even ethical considerations in what may
at first seem mere technical decisions. The Chal-
lenger disaster is, in many respects, no more
than ordinary engineering writ large. Both here
and in part IV, it will help us understand what
engineers do, what can go wrong ethically, and
what can be done to prevent ethical wrongdoing.
 Page 43

4
Codes of Ethics and the Challenger
The Public knows that doctors and lawyers are bound
to abide by certain recognized rules of conduct. Not
finding the same character of obligations imposed
upon engineers, people have failed to recognize them
as members of a profession.
A.G. Christie (1922), engineer

With respect to each separate profession we must be-

gin by analyzing the functions it performs in society. A
code of ethics must contain a sense of mission, some
feeling for the peculiar role of the profession it seeks
to regulate.
Lon Fuller (1955), lawyer
 220/1183

On the evening of January 27, 1986, Robert

Lund, vice president for engineering at Morton
Thiokol, had a problem. The Space Center was
counting down for a shuttle launch the next day.
Earlier that day, Lund presided at a meeting of
engineers who unanimously recommended
against the launch. He concurred and informed
his boss, Jerald Mason. Mason informed the
Space Center. Lund expected the flight to be
postponed. The Space Center had a good safety
record. It had achieved it by not allowing a
launch unless the technical people approved.
 221/1183

Lund did not approve because the temperature at

the launch site would be close to freezing at lift-
off. The Space Center was worried about the ice
already forming on the boosters, but Lund was
worried about the O-rings that sealed the boost-
ers' segments. They were a good idea, permit-
ting Thiokol to build the huge rocket in Utah
and ship it in pieces to the Space Center two
thousand miles away. Building in Utah was so
much more efficient than building on-site that
Thiokol was able to underbid the competition.
The shuttle contract had earned Thiokol $150
million in profits. 1 But the O-rings were not
perfect. If one failed in flight, the shuttle could
explode. Data from previous flights indicated
that the rings tended to erode in flight,
 Page 44

with the worst erosion occurring on the coldest

temperature preceding lift-off. Experimental
evidence was sketchy but ominous. Erosion
seemed to increase as the rings lost resiliency
and resiliency decreased with temperature. Un-
fortunately, almost no testing had been done be-
low 40° F. The engineers had had to extrapolate.
But, with the lives of seven astronauts at stake,
the decision seemed clear enough: safety first.
 223/1183

Well, it had seemed clear earlier that day. Now

Lund was not so sure. The Space Center was
"surprised" and "appalled" by the evidence on
which the no-launch recommendation was
based. The Space Center's senior managers
wanted to launch, but they could not launch
without Thiokol's approval. They urged Mason
to reconsider. He reexamined the evidence and
decided the rings should hold at the expected
temperature. Joseph Kilminster, Thiokol's vice
president for shuttle programs, was ready to sign
a launch approval, but only if Lund approved.
Lund's first response was to repeat his objec-
tions. But then Mason said something that made
him think again. Mason asked him to think like
a manager rather than an engineer. (The exact
words seem to have been, "Take off your engin-
eering hat and put on your management hat.")
Lund did so and changed his mind. On the next
day the shuttle exploded during lift-off, killing
all aboard. An O-ring had failed. 2
 224/1183

Should Lund have reversed himself and ap-

proved the launch? In retrospect, of course, it
seems obvious that he should not have. But we
would hardly have any problems at all if we
could foresee all the consequences of what we
do. Fairness to Lund requires us to ask whether
he should have approved the launch given only
the information actually available. We need to
consider whether Lund, an engineer, should
have thought like a manager rather than an en-
gineer. But first we need to know the difference
between thinking like a manager and thinking
like an engineer.
 225/1183

One explanation of the difference would stress

technical knowledge. Managers are trained to
handle people. Engineers are trained to handle
things. To think like a manager rather than an
engineer is to focus on people rather than on
things. According to this explanation, Lund was
asked to concern himself primarily with how
best to handle his boss, the Space Center, and
his own engineers. He was to draw on his know-
ledge of engineering only as he might draw on
his knowledge of a foreign language, for ex-
ample, to help him understand what his engin-
eers were saying. He was to act much as he
would if he had never earned a degree in
engineering.
 226/1183

That explanation of what Mason was asking of

Lund seems implausible. But if it seems im-
plausible, what is the alternative? If Mason did
not want Lund to treat his knowledge of engin-
eering as peripheral (as it seems Mason, also an
engineer, did not do when he earlier reexamined
the evidence himself), what was he asking Lund
to do? What is it to think like an engineer if not
simply to use one's technical knowledge of en-
gineering? As we saw in chapter 2, that is a
question engineers have been asking for at least
a century. Answers are often expressed as a
code of ethics. So, it seems, one way to begin to
answer our question is to learn more about those
codes.
 Page 45

History of Engineering Codes

The first civilian engineering organization in the

United States, the Boston Society of Civil
Engineers, was founded in 1848. Others fol-
lowed, with the first truly national organization
appearing two decades later. Though leaders of
these early organizations sometimes spoke of
the "high character and integrity" engineers
needed to serve the interests others committed to
them, the history of engineering codes of ethics
in the United States begins much later.
 228/1183

In 1906, the American Institute of Electrical

Engineers (AIEE) voted to embody in a code the
ideas expressed in an address by its president,
Schuyler S. Wheeler. After much debate and
many revisions, the AIEE's board of directors
adopted a code in March 1912. The AIEE code
was adopted with minor amendments by the
American Society of Mechanical Engineers
(ASME) in 1914. Meanwhile, the American In-
stitute of Consulting Engineering, the American
Institute of Chemical Engineers (AICHE), and
the American Society of Civil Engineers each
adopted a code of its own. By 1915, every major
engineering organization in the United States
had a code of ethics. 3
 229/1183

These first codes were criticized almost as soon

as they were adopted.4 One common criticism
was that they were too concerned with duties to
employers and fellow engineers. Among the
provisions so criticized were the following: Sec-
tion B.3 of the AIEE code required engineers to
"consider the protection of a client's or
employer's interest [their] first professional ob-
ligation, and [to] . . . avoid every act contrary to
this duty." An engineer's duties to the public
were merely "to assist the public to a fair and
correct general understanding of engineering
matters, to extend generally knowledge of en-
gineering . . . to discourage the appearance of
untrue, unfair or exaggerated statements on en-
gineering subjects," and otherwise to be careful
what one said in public (section D.16-19).
Another common criticism was that, though
they often speak of "employers" as well as of
''clients," the early codes seemed designed
primarily for engineers who contracted with
 230/1183

many clients and were not dependent on any one

of them. "Bench engineers," employeed engin-
eers without significant management responsib-
ilities, always the majority, seemed almost for-
gotten.5 But perhaps the most serious criticism
was that occasionally one code permitted con-
duct that others forbade. For example, the ASCE
code (section 1) forbade an engineer to "accept
any remuneration other than his stated charges
for services rendered his client," whereas the
AIEE code (section B.4) permitted payments to
the engineer from suppliers or other third parties
if the client consented. (These inconsistencies
were important, of course, only if all engineers
belonged to one professiononly if, for example,
civil engineers and electrical engineers would be
held to the same standard.)
 231/1183

Attempts to respond to such criticism began al-

most immediately. Among the first was the code
that the American Association of Engineers
(AAE) adopted in 1927. However, the AAE,
though intended to include all engineers and
briefly influential, was almost dead by then.
None of the other early attempts to respond to
these criticism amounted to much more. But, on
the eve of World War II, the American Engin-
eering Council (AEC) organized a committee to
develop a code for all engineers. Each major en-
gineering society was represented. When the
AEC dis-
 Page 46

solved, the Engineers Council for Professional

Development (ECPD) took over sponsorship.
The resulting code was a conscious effort to
synthesize the major provisions of earlier codes.

The ECPD code was enormously successful at

creating at least the appearance of unity among
engineers. All eight major engineering organiza-
tions either "adopted or assented" to it in 1947.
By 1955, it was accepted, at least in large part,
by eighty-two national, state, or local engineer-
ing organizations. That was, as one commentat-
or put it, "probably the greatest progress to be
made ever before or since toward the realization
of a single set of ethical standards for all engin-
eers." 6
 233/1183

But the ECPD code was not as successful as it at

first seemed. Some organizations, while "assent-
ing" to the code, retained their own codes as
well to preserve certain detailed provisions that
seemed to suit their circumstances better than
the corresponding provisions of the ECPD code.
As time went on, these organizations tended to
rely more and more on their own code. The
ECPD code slowly lost influence.
 234/1183

The ECPD revised its code in 1963, 1974, and

1977 in an attempt to reverse this trend. Though
many of the revisions were substantive, perhaps
the most important were structural. Four "funda-
mental principles" replaced the "foreword,"
twenty-eight ''canons" were reduced to seven
"fundamental canons"; and a set of "guidelines"
was added. These structural changes were inten-
ded to allow an organization to adopt the prin-
ciples and canons without the guidelines, if it
did not want to accept the whole package.
Though the Guidelines are supposed to be read
in the light of the principles and canons, they
are, in fact, an independent code.
 235/1183

The ABET replaced the ECPD soon after these

revisions were made. The revisions nevertheless
gave the ECPD code new life (though under the
new name). The revised code (that is, the funda-
mental principles and canons) was adopted, at
least in part, by most major engineering organiz-
ations in place of their own code. There are two
important exceptions, however.

The NSPE initially adopted the 1947 version of

the ECPD code but substituted its own code in
1964 and has since revised it several times.
Though it still has much in common with the
original ECPD code, the NSPE code differs
somewhat both in structure and content. The
NSPE Code is important for two reasons.
 236/1183

First, the NSPE has a "Board of Ethical Review"

(BER) which answers ethics questions members
of the society submit. Some other engineering
societies have similar advisory committees, but
the NSPE's is by far the most active in publish-
ing the advice. BER "opinions" are printed sev-
eral times a year in the NSPE's magazine, Pro-
fessional Engineer. About 250 opinions were
collected and published in six volumes, the last
covering the period 1981-1989. These opinions
are a valuable resource for questions of engin-
eering ethics.
 237/1183

Second, because professional engineers are li-

censed by states, the NSPEthrough its state soci-
etieshas a role in the regulation of professional
engineers much like the role of state medical so-
cieties in the regulation of physicians. The
NSPE code is at least potentially enforceable
(though only against registered engineers) in a
way that other codes of engineering ethics are
not.

The other independent code, that of the Institute

of Electrical and Electronic Engineers (IEEE), is
important for different reason. The IEEE, with
more than 300,000 members, is the largest en-
gineering organization in the United States. Its
Page 47
 239/1183

1979 code represented an alternative to the oth-

ers. Much briefer than the NSPE's (though signi-
ficantly longer than the ABET code without the
guidelines), it applied only to "members" of
IEEE. Some of its provisions were unusual as
well. For example, Article II enjoined engineers
"to treat fairly all colleagues and co-workers re-
gardless of race, religion, sex, age, or national
origin"engineering codes generally protect other
engineers from unfair treatment but not all
coworkerswhile Article III expressly limited
what engineers owe employer and client to what
was consistent with "other parts of this Code"
but did not (as other engineering codes now do)
declare the public health, safety, and welfare
"paramount." In 1991, the IEEE abandoned that
code for a much shorter one hardly distinguish-
able in content from the ABET code (without
the guidelines). Unfortunately, historians have
yet to tell us how the 1979 code failed or why
the IEEE continues to insist on a code of its own
 240/1183

rather than just adopting the virtually equivalent

ABET code.

These four codesNSPE code, ABET code, IEEE

code (1991), and ABET guidelinestoday serve
as ethical benchmarks for engineers generally.
No doubt, others will follow.

Codes of Ethics Today

Most professions regularly amend their codes of

ethics. Many undertake drastic revisions more
than once. But engineering seems to be unique
in the number of competing codes proposed and
adopted over the years. Why is the history of
codes different for engineering? Is engineering,
or engineering ethics, itself unique?
 241/1183

Chief among the explanations often advanced

for the number of codes is that engineering is
simply too diverse for one code of ethics to ap-
ply to all. Some engineers are independent prac-
titioners. Some are employees of large organiza-
tions. Some are managers. Many are closely su-
pervised. Some, whether in large organizations
or on their own, are more or less their own boss.
Engineers do too many different things for the
same standards to apply to all. In sum, engineer-
ing is not a single profession but a family of his-
torically related professions.
 242/1183

There is a false note in this explanation for the

number of codes of ethics. If the divisions in en-
gineering were similar to that, say, between
medicine and dentistry, why would engineers
establish "umbrella" organizations and devote so
much time to trying to achieve one code for all
engineers? Physicians and dentists have not
made similar efforts to write a single code of
ethics for their two professions. The three-quar-
ters of a century during which engineers tried to
write a code for all engineers islike the existence
of schools of engineeringitself evidence that en-
gineers all belong to one profession, however
divided and diverse its membership. Indeed, the
effort to write a single code might be an attempt
to preserve the unity of the profession. On this
view, the number of codes proposed and adop-
ted is an instance of what engineers call the
"NIH" (not invented here) phenomenon. The
number of independent professional organiza-
tions, not the existence of several engineering
 243/1183

professions, explains the number of competing

codes. 7

The NIH phenomenon occurs when each side

has good reasons for its view. Perhaps this is
such a case. One side is certainly right to point
out that a short code is easy to remember or con-
sult. It can be conspicuously posted to remind
engineers
 Page 48

of their obligations. A short code is also easier

to get approved because its necessary generality
automatically obscures disagreement over de-
tails of conduct. But the other side can also
point out that a long code provides much more
information. It can take into account special cir-
cumstances, make exceptions explicit, and oth-
erwise provide more guidance, at least for those
willing to take the time to read it through. A
long code makes it less likely that engineers
who think they agree on standards will suddenly
discover that they do notat a moment that the
discovery is costly. Some professionals (for ex-
ample, lawyers and accountants) long ago opted
for a long code like the NSPE's or ABET's
guidelines. Others, for example, dental hygien-
ists and social workers, opted for a short code
like the IEEE's or ABET's fundamental prin-
ciples and canons.
 245/1183

Though the various engineering codes differ in

length, they differ in content as well. Because
these latter differences involve more than pride
of authorship, the NIH phenomenon only partly
explains why engineers cannot agree on a single
code. History doubtless has a part. The NSPE
cannot give up its code without substantially re-
ducing the value of its BER opinions. Other
considerations may also be relevant. For ex-
ample, the NSPE's code is designed for use in
state disciplinary committees; neither ABET's
code nor the IEEE's is. They are designed
primarily for self-discipline.
 246/1183

Whatever the explanation of the number of

codes, there is no doubt that their variety could
make it hard for an engineer to know what to do.
An engineerfor example, an IEEE member li-
censed as a professional engineer (P.E.)might be
subject to three codes (the IEEE's, NSPE's, and
ABET's). Which should she consult? If the
codes differ on some point, which, if any,
should she consider binding? What should other
engineers think of her if she chooses to do what
one code allows even though another forbids it?
What should they do?
 247/1183

These difficulties are not as serious as they may

seem. In general, the various codes are not en-
forced by the organizations adopting them.
Though the language often resembles that of
statute, codes of ethics are in fact more like
guides to conscience or public judgmentthat is,
moral rules. An engineer who violates the code
of one of the organizations to which she belongs
is not likely to be expelled (or even formally
censured). She is even less likely to have her "li-
cense to practice" revoked (since most engineers
are not licensed at all). Apart from pangs of con-
science, the only repercussion she is likely to
suffer is the poor opinion of those who know her
well enough to know what she did. Her primary
concern should be one of justifying her conduct
to those concerned, including herself. (We re-
turn to the problem of resolving conflicts
between codes in chapter 8.)
 248/1183

But thinking of codes of ethics as moral rules

rather than legal rules suggests new difficulties.
If codes of ethics are merely moral rules, why
worry about them at all? Why shouldn't each en-
gineer let his private conscience be his guide?
Why should he have to consider what some or-
ganization of engineers has to say about what he
should do? What expertise can engineering soci-
eties have in morals? Aren't the experts in mor-
als, if there are any, philosophers or clergy
rather than engineers? To answer these ques-
tions, we must look deeper into the relationship
between professions and codes of ethics.
 Page 49

Professions and Codes

A code of ethics generally appears when an oc-

cupation organizes itself into a profession. Why
this connection between codes of ethics and pro-
fessions? There are three common explanations.

One explanation, "definition by paradigm," has

would-be professions imitating the forms of
widely recognized professions. To be a profes-
sion is to be like the most respected professions,
the paradigms. Because the paradigmsespecially
law and medicinerequire long training, special
skills, licensing, and so on, so should any other
group that wants to be considered a profession.
Because both law and medicine have a code of
ethics, engineering would naturally suppose it
needs one too if it is to be a "true profession."
 250/1183

Much may be said for this first explanation of

why engineering has a code of ethics. For ex-
ample, the American Bar Association (ABA)
adopted its first code of ethics in 1908four years
before the first American engineering society
did. Engineers certainly did not ignore the
ABA's action. 8
 251/1183

Nonetheless, this explanation is inadequate for

our purposes. The emphasis on imitation does
not explain why the "paradigm professions" ad-
opted codes or why engineers copied the ABA
in adopting a code of ethics but not the code it-
self, enforcement procedures, or licensing re-
quirement. The emphasis on imitation also
makes it hard to understand why engineers think
what the code says important. After all, if a pro-
fession only needs a code so it can be like other
professions, why should it matter much what the
code says? Does what the code says matter only
because the paradigm profession thinks what the
code says matters? Why should the paradigm
profession think that? But perhaps most signific-
ant, the emphasis on other professions does not
explain why some early American codes of en-
gineering ethics were modeled on the code of
the British Institute of Civil Engineers rather
than on some American paradigm like the
ABA'sor why, in England, the first professions
 252/1183

to adopt codes were the relatively low-status

apothecaries and solicitors (rather than the high-
status physicians or barristers).9

One attempt to make up for these inadequacies

of definition by paradigm yields "the contract
with society" approach to understanding the re-
lation between professions and codes of ethics.
According to this approach, a code of ethics is
one of those things a group must have before so-
ciety recognizes it as a profession. The code's
content is settled by considering what society
would accept in exchange for such benefits of
professionalism as high income, prestige, and
trust. A code is a way to win the advantages so-
ciety grants only to those imposing certain re-
straints on themselves. A profession has no oth-
er interest in having a code of ethics.
 253/1183

Although this second explanation is a significant

advance over the first, it is still far from ad-
equate. In particular, it gives us little help in an-
swering such questions as the following: Why
should engineers be so concerned about the de-
tails of their code when, it seems, society recog-
nizes engineering as a profession and does not
care which of the various codes engineers ad-
opt? Why did the original engineering codes
take so much space laying down rules about
how engineers should treat one another when it
seems society is likely not to care about such
things or to be posi-
 Page 50

tively adverse? The inability of the second ex-

planation to help us answer such questions sug-
gests that we should look for a better one.
 255/1183

A third explanation of the relation of profession

and codes of ethics is better than the other two.
This explanation views a code as primarily a
"contract between professionals." According to
this explanation, a profession is a group of per-
sons who want to cooperate in serving the same
ideal better than they could if they did not co-
operate. Engineers, for example, might be
thought to serve the ideal of efficient design,
construction, and maintenance of safe and useful
physical systems. 10 A code of ethics would
then prescribe how professionals are to pursue
their common ideal so that each may do the best
he can at minimum cost to himself (and to the
publicif looking after the public is part of the
ideal). The code is to protect each from certain
pressures (for example, the pressure to cut
corners to save money) by making it reasonably
likely that most other members of the profession
will not take advantage of his good conduct. A
 256/1183

code protects members of a profession from cer-

tain consequences of competition.

According to this explanation, an occupation

does not need society's recognition to be a pro-
fession. It needs only a practice among its mem-
bers of cooperating to serve a certain ideal.
Once an occupation becomes a profession, soci-
ety has a reason to give the occupation special
privileges (for example, the sole right to do cer-
tain kinds of work) if society wants to support
serving the ideal in question in the way the pro-
fession serves it. Otherwise, society may leave
the profession unrecognized. So, according to
this third explanation, what is wrong with the
first two is that they confuse the trappings of
profession with the thing itself.11
 257/1183

If we understand a code of ethics as the way a

profession defines relations between those who
want to serve a common ideal, we may construe
the number of different codes of ethics as show-
ing that engineers are not yet fully agreed on
how they want to pursue their common ideal.
Engineering would, in this respect, still be a pro-
fession in the making. Thinking of engineering
in this way is, under the circumstances, nonethe-
less consistent with thinking of engineering as a
profession. The substantive differences between
codes is not great. The differences in structure
and language are more obvious than important
in the choice of conduct. Engineers agree on all
essential terms of their "contract."
 258/1183

Understanding a code of ethics as a contract

between professionals, we can explain why en-
gineers should not depend on mere private con-
science when choosing how to practice their
profession, and why they should take into ac-
count what an organization of engineers has to
say about what engineers should do. What oth-
ers expect of us is part of what we should take
into account in choosing what to do, especially
if the expectation is reasonable. A code provides
a guide to what engineers may reasonably ex-
pect of one another, what "the rules of the
game" are. Just as we must know the rules of
baseball to know what to do with the ball, so we
must know engineering ethics to know, for ex-
ample, whetheras engineerswe should merely
weigh safety against the wishes of our employer
or instead give safety preference over those
wishes. A code of ethics should also provide a
guide to what we may expect other members of
our profession to help us do. If, for example,
 259/1183

part of being an engineer is putting safety first,

then Lund's engineers had a right to expect his
support. When Lund's boss asked him to think
like a manager rather than an en-
 Page 51

gineer, he should, as an engineer, have respon-

ded, "Sorry, if you wanted a vice president who
would think like a manager rather than an en-
gineer, you should not have hired an engineer."
12
 261/1183

If Lund had responded in this way, he would

have responded as the "rules of the engineering
game" require. But would he have done the right
thing, not simply according to those rules but
really (that is, all things considered)? This is not
an empty question. Even games can be irrational
or immoral. (Imagine, for example, a game in
which you score points by cutting off your fin-
gers or by shooting people who pass in the street
below.) People are not merely members of this
or that profession. They are also individualsmor-
al agentswith responsibilities beyond their pro-
fession, individuals who cannot escape con-
science, criticism, blame, or punishment just by
showing that they did what they did because
their profession told them to.
 262/1183

We now have an explanation of why engineers

should, as engineers, take into account their
profession's code of ethics, but we have no ex-
planation of why anyone should be an engineer
in the relevant sense.
 263/1183

We may put this point more dramatically. Sup-

pose Lund's boss had responded in this way to
what we just imagined that Lund told him: "Yes,
we hired an engineer, butwe supposedan engin-
eer with common sense, one who understood
just how much weight a rational person gives a
code of ethics in decisions of this kind. Be reas-
onable. Your job and mine are on the line. The
future of Thiokol is also on the line. Safety
counts a lot. But other things do too. If we block
this launch, the Space Center will start looking
for someone more agreeable to supply boosters.
We all could be out of a job." If doing as one's
professional code says is really justified (that is,
justified all things considered), we should be
able to explain to Lund (and his boss) why, as a
rational person, Lund should support his
profession's code as a guide for all engineers
and why, under the circumstances, he could not
rationally expect others to treat him as an
exception.
 264/1183

Why Engineers Should Obey Their Profession's

Code
 265/1183

We can begin our explanation of obedience to

the code by outlining two alternatives some
people find plausible. One is that Lund should
do as his profession requires because he "prom-
ised," for example, by joining an engineering
society that has a code of ethics. We must dis-
miss this explanation because Lund may never
have done anything we could plausibly call
promising to follow a code. Lund may, for ex-
ample, have failed to join any professional asso-
ciation that has a code (as perhaps half of all
American engineers do). Would we excuse him
from conducting himself as an engineer should?
No. The obligations of an engineer do not seem
to rest on anything so contingent as a promise,
oath, or vow. The "contract" between profes-
sionals of which we spoke cannot literally be a
contract. It seems more like a "quasi-contract"
or "contract implied in law"that is, an obligation
resting on what is fair to require of someone
given that she benefited by some action of her
 266/1183

own (for example, by claiming to be an

engineer).

That Lund should do as his profession requires

because "society" says he should is another
plausible answer. We may dismiss it in part be-
cause it is not clear that
 Page 52

society does say that. One way society has of

saying things is through law. No law binds all
engineers to abide by their profession's code of
ethics (as the law does bind all lawyers). Of
course, society has another way of saying things
than by law, that is, by public opinion. But it
seems doubtful that the public knows enough
about engineering ethics to have a distinct opin-
ion on the questions we are considering. More
important, it is not clear why public opinion or
law should decide what it is rational or moral to
do. Certainly there are both irrational laws (for
example, those requiring use of outmoded tech-
niques) and immoral laws (for example, those
enforcing slavery). The public opinion support-
ing those laws could not have been much less ir-
rational or immoral than the laws themselves.
 268/1183

The two answers we dismissed share one not-

able feature. Either would, if defensible, provide
a reason to do as one's profession says inde-
pendent of what in particular the profession hap-
pens to say. The answers do not take into ac-
count the contents of the code of ethics. They
are "formal." The answer I now give is not
formal. I show that supporting a code of ethics
with a certain content is rational by showing that
supporting codes with a content of that sort is
rational.
 269/1183

Consider the ABET code. It is divided into fun-

damental principles and fundamental canons.
The fundamental principles simply describe in
general terms an ideal of service. Engineers "up-
hold and advance the integrity, honor and dig-
nity of the engineering profession by: I. using
their knowledge and skill for the enhancement
of human welfare, II. being honest and impar-
tial, and serving with fidelity the public, their
employers and clients [and so on]." What ration-
al person could object to other people with her
skills trying to achieve that ideal? (Or at least,
what rational person could object as long as do-
ing so did not interfere with what she was do-
ing?) Surely every engineer, indeed, every
member of society, is likely to be better off
overall if engineers uphold and advance the in-
tegrity, honor, and dignity of engineering in this
way.
 270/1183

If the fundamental principles lay down goals,

the fundamental canons lay down general duties.
For example, engineers are required to "hold
paramount the safety, health and welfare of the
public," to "issue public statements only in an
objective and truthful manner," to "act in profes-
sional matters for each employer or client as
faithful agents and trustees," and to "avoid all
conflicts of interest." Each engineer stands to
benefit from these requirements both as ordinary
person and as engineer. As ordinary person, an
engineer is likely to be safer, healthier, and oth-
erwise better off if engineers only make truthful
public statements, and so on. But how can en-
gineers benefit as engineers from such require-
ments? To explain that, we have to try a thought
experiment.
 271/1183

Imagine what engineering would be like if en-

gineers did not generally act as the canons re-
quire (while satisfying the requirements of law,
market, and ordinary morality). If, for example,
engineers did not generally hold paramount the
safety, health, and welfare of the public, what
would it be like to be an engineer? The day-to-
day work would, of course, be much the same.
But every now and then an engineer might be
asked to do something which, though profitable
to the employer or client and legal, would put
other people at risk, some perhaps about whom
the engineer cared a great deal. Without a pro-
fessional code, an engineer could not object as
an engineer. An engineer could, of course, still
object "personally" and refuse to do the
 Page 53

job. But, if he did, he would risk being replaced

by an engineer who would not object. An em-
ployer or client might rightly treat an engineer's
personal qualms as a disability much like a tend-
ency to make errors. The engineer would be un-
der tremendous pressure to keep "personal opin-
ions" to himself and get on with the job. His in-
terests as an engineer would conflict with his in-
terests as a person; his conscience, with his self-
interest.
 273/1183

That, then, is why each engineer can generally

expect to benefit from other engineers' acting as
their common code requires. The benefits are
clearly substantial enough to explain how indi-
viduals could rationally enter into a convention
that would equally limit what each can do. I
have not, however, shown that every engineer
must benefit overall from such a convention, or
even that any engineer will consider these bene-
fits sufficient to justify the burdens required to
achieve them. Professions, like governments,
are not always worth the trouble of maintaining.
Whether a particular profession is worth the
trouble is an empirical question. Professions
nonetheless differ from governments in at least
one way relevant here. Professions are voluntary
in a way that government is not. No one is born
into a profession. One must claim professional
status to have it (by taking a degree, for ex-
ample, or by accepting a job for which profes-
sional status is required). We therefore have
 274/1183

good reason to suppose that people are engin-

eers because, on balance, they prefer to have the
benefits of being an engineer, even given what
is required of them in exchange. 13

If, as we shall now assume, the only way to ob-

tain the benefits in question is to make it part of
being an engineer that the public safety, health,
and welfare come first, every engineer, includ-
ing Lund, has good reason to want engineers
generally to adhere to (something like) the
ABET code. No one wants to be forced to
choose between conscience and self-interest.
But why should an engineer adhere to the code
himself when, as in Lund's case, he may seem
likely to benefit by departing from it? The an-
swer should be obvious.
 275/1183

Lund would have to justify his departure from

the code by appeal to such considerations as the
welfare of Thiokol and his own self-interest.
Appeal to such considerations is just what Lund
could not incorporate into a code of ethics for
engineers or generally allow other engineers to
use in defense of what they did. Lund could not
let such an exception be incorporated into the
code because its incorporation would defeat the
purpose of the code. A code is necessary in
large part because, without it, the self-interest of
individual engineers would lead them to do what
would harm everyone overall. Lund could not
allow other engineers to defend what they did
by appeal to their own interests or that of their
employer for much the same reason. To allow
such appeals would contribute to the breakdown
of a practice Lund has good reason to support.
 276/1183

I believe this argument explains why, all things

considered, Lund should have done as his
profession's code requires, not why he should
have done so in some premoral sense. I am an-
swering the question "Why be ethical?" not
"Why be moral?" I therefore have the luxury of
falling back on ordinary morality to determine
what is right, all things considered, that is, tak-
ing into account the fact of profession. The mor-
al rule on which this argument primarily relies is
"the principle of fairness" ("Don't cheat"). Be-
cause Lund voluntarily accepts the benefits of
being an engineer
 Page 54

(by claiming to be an engineer), he is morally

obliged to follow the morally permissible con-
vention that helps to make those benefits pos-
sible. 14 What I have taken pains to show is how
that convention helps to make those benefits
possible and why, even now, Lund has good
reason to endorse the convention generally.
 278/1183

Of course, I am assuming that engineers do in

fact generally act in accordance with the ABET
code (whether or not they know it exists). If that
assumption were mistaken, Lund would have no
professional reason to do as the code says. The
code would be a dead letter, not a living prac-
tice. It would have much the same status as a
"model statute" no government ever adopted, or
the rules of a cooperative game no one was
playing. Lund would have to rely on private
judgment. But relying on private judgment is
not necessary here. Lund's engineers made their
recommendation because they thought the safety
of the public, including the astronauts, para-
mount. They did what (according to the ABET
code) engineers are supposed to do. Their re-
commendation is itself evidence that the code
corresponds to a living practice.15
 279/1183

So, when Lund's boss asked him to think like a

manager rather than an engineer, he was in ef-
fect asking Lund to think in a way that Lund
must consider unjustified for engineers gener-
ally and for which Lund can provide no ration-
ally defensible principle for making himself an
exception. When Lund did as his boss asked, he
in effect let down all those engineers who
helped to build the practice that today allows en-
gineers to say no in such circumstances with
reasonable hope that their client or employer
will defer to "professional judgment" and that
other members of their profession will aid them
if the client or employer does not defer.
 280/1183

Lund could, of course, explain how his action

served his own interests and those of Thiokol
(or, rather, how they seemed to at the time).16
He could also thumb his nose at all talk of en-
gineering ethics (though that might lead to the
government's barring him from work on any
project it funds, to fellow engineers refusing to
have anything to do with him, and to his em-
ployer coming to view him as an embarrass-
ment). What he cannot doassuming we have
identified all relevant considerationsis show that
what he did was right, all things considered.

But have we identified all relevant considera-

tions? I certainly think so. But, for our purposes,
it does not matter. I did not examine Lund's de-
cision to condemn him but to understand the
place of a code of ethics in engineering. There is
more to understand.

Using a Code of Ethics

 281/1183

So far, we have assumed that Lund did as his

boss asked: he thought like a manager rather
than like an engineer. Assuming that allowed us
to provide a relatively clear explanation of what
was wrong with what Lund did. What was
wrong was that Lund acted like a manager when
he was an engineer and should have acted like
one.

We must now put that assumption aside and

consider whether engineering ethics actually
forbids Lund to do what it seems he did, that is,
weigh his own interests, his employer's, and his
client's against the safety of the seven astro-
nauts. Ordinary morality seems to allow such
weighing. For example, no one would think I
did something morally wrong if I drove my
child to school rather than let him take the
 Page 55

bus, even though being on the road increases the

risk that someone will be killed in a traffic acci-
dent. Morality allows us to give special weight
to the interests of those close to us. If engineer-
ing ethics allows it, too, thenwhatever he may
have thought he was doingLund would not actu-
ally have acted unprofessionally. Let us imagine
Lund reading our four "benchmark" codes.
What would they tell him? What could he infer?
 283/1183

Of the seven fundamental canons of the current

ABET code, only two seem relevant: (1) "[hold-
ing] paramount the safety, health and welfare of
the public" and 4) "[acting] in professional mat-
ters for each employer or client as faithful
agents or trustees." What do these provisions tell
Lund to do? The answer is not obvious. Does
"public" include the seven astronauts? They are,
after all, employees of Thiokol's client, the
Space Center, not part of the public as, say,
those ordinary citizens are who watch launches
from the beach opposite the Space Center. And
what is it to be a "faithful agent or trustee" of
one's client or employer? Is it to do as instructed
or to do what is in the client's or employer's in-
terests? And how exactly is one to determine
those interests? After all, the actual result of
Lund's decision was a disaster for both employer
and clientbut a disaster Lund, his employer, and
his client (or, at least, their representatives)
thought themselves justified in risking. And
 284/1183

what is Lund to do if the public welfare requires

what no faithful agent could do? What is it to
''hold paramount" the public welfare?
 285/1183

The IEEE code of 1979-1990, for all its innova-

tions, would not have helped Lund much.
Article III.1 more or less repeats the faithful-
agent requirement of ABET canon 4. Article
IV.1 more or less repeats the requirement of the
ABET canon 1 (though without formally declar-
ing the public interest "paramount"). Members
of the IEEE are supposed to "protect the safety,
health and welfare of the public and speak out
against abuses in these areas affecting the public
interest." The duties of a faithful agent are,
however, limited by other provisions of the code
whereas the duty to protect the public is not.
The public welfare takes precedence whenever it
conflicts with the duties of a faithful agent. The
old IEEE code thus provides a plausible inter-
pretation of "hold paramount." This would be
helpful if we knew what was included in the
public safety, health, and welfare. Unfortu-
nately, the IEEE code (like ABET's) tells us
nothing about that. The only relevant provision
 286/1183

of the new IEEE code is the first: "to accept re-

sponsibility in making engineering decisions
consistent with the safety, health, and welfare of
the public, and to disclose promptly facts that
might endanger the public or the environment."
Although the new code does not declare the
public safety, health, and welfare paramount in
just those words, it achieves the the same effect
by combining omission of all references to em-
ployer or client with a requirement that IEEE
members take responsibility for the public
health, safety, and welfare.
 287/1183

Though the NSPE code is much more detailed

than the other two, its details are only somewhat
more helpful here. The first "rule of practice"
simply repeats the language of ABET canon 1;
the fourth rule does the same for canon 4. Rule
1a follows the IEEE code in giving priority to
the public safety, health, and welfare over all
other considerations but gives more detail to
how one should "disclose" any danger. If over-
ruling Lund's judgment were to endanger the
public "safety, health, property, or welfare,"
then, according to NSPE rule 1a, Lund would
have a positive
 Page 56

duty to bring the matter to the attention of "the

appropriate authority." The appropriate author-
ity might, it seems, be someone other than the
client or employer. Rule 1b partially defines
"safe for public health, property, and welfare" in
terms of conformity to "accepted standards."
That would be helpful if the problem that con-
cerned Lund were conventional enough for cer-
tain standards to win acceptance. Unfortunately,
the use of O-rings in question here was so new
that the engineers had no manual of "safety
specs'' to which to turn. That was part of Lund's
problem.
 289/1183

The NSPE code illustrates the advantage of de-

tailed provisions. The more detailed a code, the
more guidance it is likely to provide on ques-
tions an engineer is worried about. The current
NSPE code could, for example, have contained
a provision like canon 11 of the NSPE code of
1954: "[The engineer] will guard against condi-
tions that are dangerous or threatening to life,
limb or property on work for which he is re-
sponsible . . ." That would have made Lund's
duty clear. Unfortunately, the NSPE code no
longer contains that provision. Why? One pos-
sibility is that the drafters of the current code
thought the provision redundant given the duty
to hold the public safety paramount. Another
possibility is that the NSPE codeand ABET
guidelinesnow require engineers to be con-
cerned only for the public safety, health, and
welfare rather than, as canon 11 seems to be,
everyone's. Perhaps, after due consideration, the
drafters of the various codes decided it was too
 290/1183

much to ask engineers to worry about the safety

of their client's or employer's employees as well
as the safety of the public. How is an engineer to
understand a code of ethics if, as often happens,
it does not clearly address a problem?
 291/1183

That question is surprisingly easy to answer if

we keep in mind the connection between profes-
sions and codes of ethics. The language of any
document must be interpreted in light of what it
is reasonable to suppose its authors intended.
For example, if the word "bachelor" appears un-
defined in a marriage statute, we interpret it as
referring to single males, but if the same word
appears undefined in directions for a college's
graduation ceremony, we instead interpret it as
referring to all students getting their baccalaur-
eate, whether male or female, single or married.
That is the reasonable interpretation because we
know that marriages usually involve single
males (as well as single females) rather than
people with baccalaureates while just the re-
verse is true of graduation ceremonies. So, once
we figure out what it is reasonable to suppose
engineers to intend by declaring the public
safety, health, and welfare paramount, we
should be able to decide whether interpreting
 292/1183

"public" so that it includes employees is what

they intend (or, at least, what, as rational per-
sons, they should intend).

The "authors" of a code of engineering ethics

(both those who originally drafted or approved it
and those who now give it their support) are all
more or less rational agents. They differ from
most other rational agents only in knowing what
engineers must know to be engineers and in per-
forming duties they could not perform (or could
not perform as well) but for that knowledge. It is
therefore reasonable to suppose that their code
of ethics would not require them to risk their
own safety, health, or welfare, or that of anyone
for whom they care, except for some substantial
good (for example, high pay, easy application of
the code, or service to some ideal to which they
are committed). It also seems reasonable to sup-
pose no code they
 Page 57

"authored" would include anything people gen-

erally consider immoral. Engineers being gener-
ally much like the rest of us, we have, all else
equal, no reason to suppose engineers as a group
to be bent on immoral conduct.

But what if that were not true? What if most en-

gineers were moral monsters or self-serving op-
portunists? What then? Interpretation of their
code would certainly be different, and probably
harder. We could not understand it as a profes-
sional code (a system of morally permissible
rules). We would have to switch to principles of
interpretation we reserve for mere folkways,
Nazi statutes, or the like. We would have to
leave the presuppositions of ethics behind.
 294/1183

But, given those presuppositions, we can easily

explain why a code of engineering ethics would
make holding the public safety paramount a
duty taking precedence over all others, including
the duty to act as a faithful agent or trustee of
one's employer: Rational engineers would want
to avoid situations in which only their private
qualms stood between them and a use of profes-
sional knowledge they considered wrong or oth-
erwise undesirable. Each would want to be reas-
onably sure the others' knowledge would serve
the public even when the interests of the public
conflicted with those of employer or client.
Given this purpose, what must "public" mean?
 295/1183

We might interpret "public" as equivalent to

"everyone" (in the society, locale, or whatever).
On this interpretation, the public safety would
mean the safety of everyone more or less
equally. A danger that struck only children, or
only those with bad lungs, or the like, would not
endanger the public. This interpretation must be
rejected. Because few dangers are likely to be-
fall everyone more or less equally, interpreting
public to mean everyone would yield a duty to
the public too weak to protect most engineers
from having to do things that would make life
for them (and those for whom they care) worse
than it would otherwise be.
 296/1183

We might also interpret public as referring to

"anyone" (in the society, locale, or whatever).
On this interpretation, the public safety would
be equivalent to the safety of some or all. Hold-
ing the public safety paramount would mean
never putting anyone in danger. If our first inter-
pretation of public made provisions protecting
the public too weak, this second would make
such provisions too strong. For example, it is
hard to imagine how we could have airplanes,
mountain tunnels, or chemical plants without
some risk to someone. No rational engineer
could endorse a code of ethics that virtually
made engineering impossible.
 297/1183

We seem, then, to need an interpretation of pub-

lic invoking some relevant feature of people
(rather than, as we have so far, just their num-
ber). We might, for example, think that what
makes people members of a public is their relat-
ive innocence, helplessness, or passivity. On this
interpretation, "public" would refer to those per-
sons whose lack of information, technical know-
ledge, or time for deliberation renders them
more or less vulnerable to the powers an engin-
eer wields on behalf of his client or employer.
An engineer should hold paramount the public
safety, health, and welfare to ensure that engin-
eers will not be forced to give less regard to the
welfare of these "innocents" than they would
like.
 298/1183

On this interpretation, someone might be part of

the public in one respect but not in another. For
example, the astronauts would be part of the
public in respect of the O-rings because, not
knowing of that danger, they could not abort the
launch
 Page 58

because of it. The astronauts would, in contrast,

not be part of the public in respect of the ice
forming on the boosters because, having been
fully informed of that danger, they could abort
the launch if they were unwilling to risk the ef-
fects of the ice. On this third interpretation, pub-
lic seems to create none of the difficulties it did
on the two preceding interpretations. We now
have a sense of "holding the public safety para-
mount" we may reasonably suppose rational en-
gineers to endorse.
 300/1183

On this interpretation, all four codes of ethics

would require Lund either to refuse to authorize
the launch or to insist instead that the astronauts
be briefed to get their informed consent to the
risk. Refusing authorization would protect the
public by holding the safety of the astronauts
paramount. Insisting that the astronauts be
briefed and decide for themselves would hold
the safety of the public paramount by transfer-
ring the astronauts from the category of member
of the public to that of informed-participant in
the decision. Either way, Lund would not, under
the circumstances, have treated his own in-
terests, those of his employer Thiokol, or those
of his client, the Space Center, as comparable to
those of the public.
 301/1183

Is this the right answer? It is if we took every

relevant consideration into account. Have we?
How are we to know whether we have? We can,
of course, go through a checklist. But how are
we to know that the list is complete? Past exper-
ience is an indication, but now and then
something unprecedented occurs. So, what are
we to do? In engineering ethics, as in the rest of
engineering, it is often easier to demonstrate the
fault of alternatives than to demonstrate that this
or that answer must be right. This is such a case.
Although we cannot demonstrate that our third
interpretation is the right one, we can demon-
strate that the only obvious alternative is wrong.
 302/1183

That alternative is that public refers to all "inno-

cents" except employees of the client or employ-
er in question. Employees might be excluded
because they are paid to take the risks associated
with their work. On this interpretation, Lund
would not have to hold the safety of the astro-
nauts paramount. They would not be part of the
public.
 303/1183

What is wrong with this fourth alternative? Con-

sider how we understood innocents. These are
persons whose lack of information, training, or
time for deliberations renders them vulnerable
to the powers an engineer wields on behalf of
his client or employer. An employee who takes
a job knowing the risks (and being able to avoid
them) might be able to insist on being paid
enough to compensate for them. She might then
be said to be paid to take those risks. But she
would, on our third interpretation, also not be
part of the public to which an engineer owed a
paramount duty. She would have given in-
formed consent to the risk in question. On the
other hand, if the employee lacked information
to evaluate the risk, she would be in no position
to insist on compensation. She would, in other
words, be as innocent of, as vulnerable to, and
as unpaid for the risks in question as anyone in
the public. Nothing prevents an engineer, or
someone for whom an engineer cares, from
 304/1183

being the employee unknowingly at risk. So, ra-

tional engineers have as much reason to want to
protect such employees as to protect the public
in general. "Public" should be interpreted ac-
cordingly. 17
 Page 59

Lesson

One notable feature of engineering disasters is

that they seldom have just one cause, whether
bad design, natural catastrophe, operator error,
or even conscious wrong-doing. Engineering
creates systems relatively immune to disaster,
systems that require many failures before any-
thing big can happen.
 306/1183

Like most engineering disasters, the Challenger

explosion could not have occurred without
many things going wrong. Disputes about "the"
cause are therefore "academic" (in the bad sense
of that otherwise honorable word). The
Challenger's design was certainly imperfect (but
what engineering design is not?). The O-rings
were just one of many troubles. Thiokol's (and
NASA's) decision procedure was imperfect, too.
The redesign of the O-rings should have had top
priority at least a year before the explosion.
There were also budget problems, as Congress
came to take the Space Program for granted.
Morton's purchase of Thiokol may have
changed Thiokol's decision making in subtle
ways, making it less engineering oriented. And
so on. 18 Had any of these factors been other-
wise, Lund might have had an easy decision to
make on January 27, 1986.
 307/1183

Any engineer who has had an important part in

some large project will find the Challenger story
familiar. The complex causal chain that led from
the early 1960s to the Challenger's explosion is
not alien to engineering. Quite the contrary, en-
gineering is inseparable from budgets, problems
of coordinating work, choice of design on prac-
tical (as well as purely technical) grounds, and
so on. Such "political" considerations help to ex-
plain why one night Lund became the last barri-
er to disaster, but they do not explain away his
decision or render it irrelevant. Among the
many lessons the Challenger's story has to teach
is that, in practice, the ethics of engineers is as
important to the success of engineering as good
design or testing is.

Professional Responsibility
 308/1183

Given the argument developed so far, engineers

clearly are responsible for acting as their
profession's code of ethics requires. Do their
professional responsibilities go beyond the
code? That depends on what we mean by "re-
sponsibilities" here. If we mean (special) "oblig-
ations" or ''duties," that is, acts required, then the
answer is no; it takes a convention, the code of
ethics itself, to create those. If, however, we
mean less, (something like) "tasks they have
good reason to take on or assign," then engin-
eers certainly do have professional responsibilit-
ies beyond the code.19 Engineers should not
only do as their profession's code requires but
also support it less directly by encouraging oth-
ers to do as the code requires and by criticizing,
ostracizing, or otherwise calling to account
those who do not. Engineers should support
their profession's code in these auxiliary ways
for at least four reasons.
 309/1183

First, supporting the code helps protect engin-

eers and those they care about from being in-
jured by what other engineers do. Second, sup-
porting the code helps ensure each engineer a
working environment in which resisting pres-
sure to do what the engineer would rather not do
is easier than it would otherwise be. Third,
engineers
 Page 60

should support their profession's code because

supporting it helps make their profession a prac-
tice about which they need feel no morally justi-
fied embarrassment, shame, or guilt. And fourth,
considerations of fairness call on an engineer to
take on his share of these additional responsibil-
ities insofar as other engineers do the same and
he (by claiming to be an engineer) benefits from
their doing so.
 Page 61

5
Explaining Wrongdoing
How often is a man, looking back at his past actions,
astonished at finding himself dishonest!
Cesare Beccaria, On Crimes and Punishments, chap.
39
 312/1183

What first interested me about professional eth-

ics were the social questions: the problems faced
by those trying to act as members of a profes-
sion, the options available, the reasons relevant
to deciding between those options, and the
methods of assessing those reasons. I thought of
myself as advising decision makers within a
complex institution. I could, it seemed, contrib-
ute to right action in the professions simply by
applying skills developed in political and legal
philosophy to this new domain. But, like many
others who began to do applied ethics in this
way, I soon learned that matters are not that
simple.
 313/1183

Part of studying professional ethics is reading

the newspaper accounts, congressional testi-
mony, and court cases that wrongdoing in the
professions generates. I read such documents to
identify new problems. But, in the course of
reading so much about wrongdoing, I began to
wonder how much use my advice could be.
Though the wrongdoers were usually well-edu-
cated and otherwise decent (like the characters
in chapter 4), much of what they did seemed
clearly wrong. Surely they did not need a philo-
sopher to tell them so. I also began to wonder at
how little the wrongdoers themselves had to say
about why they did what they did. They seemed
far less articulate about that than many illiterate
criminals. 1
 314/1183

Having begun to wonder about the motivation of

the wrongdoers I was studying, I turned to the
philosophical literature explaining wrongdoing.
I was surprised at how little there wasand at how
unhelpful it was. The wrongdoers I studied did
not seem to do wrong simply because they were
weak-willed, self-deceiving, evil-willed, ignor-
ant, or morally immatureor even because they
combined several of those failings. At most,
those failings seemed to play a subsidiary part in
what they did. Yet the philosophical literature
offered no sustained discussion of anything else.
Only when I turned to the more practical literat-
ure of organization analysis did I find more.
And, even there I did not find enough. I still did
not have a satisfactory
 Page 62

explanation of the wrongdoing I was studying. I

concluded that we lack an adequate psychology
of wrongdoing.

This lack is unfortunate for both a practical and

a theoretical reason. The practical reason is that
the less well we understand wrongdoing, the
less able we are to devise strategies likely to re-
duce wrongdoing. The theoretical reason is that
insofar as we cannot explain wrongdoing, there
is a substantial hole in our understanding of en-
gineering ethics.
 316/1183

This chapter has three objectives: (1) to provide

some evidence for the claim that evil will, weak-
ness of will, self-deception, ignorance, and mor-
al immaturity, even together, do not explain
much wrongdoing of concern to students of en-
gineering ethics; (2) to add one interesting al-
ternative to the explanations now available; and
(3) to suggest the practical importance of that al-
ternative. Ultimately, though, this chapter has
only one objective: to invite others to pick up
where I leave off. Those interested in engineer-
ing ethics need to think more about the psycho-
logy of wrongdoing.

Three Examples of Wrongdoing

 317/1183

Let's begin with the testimony of a minor figure

in the General Electric price-fixing scandal of
the 1950s. No longer facing criminal or civil
charges, he described how he got into trouble as
follows: "I got into it . . . when I was young. I
probably was impressed by the manager of mar-
keting asking me to go to a meeting with him
[where price-fixing discussions took place]. I
probably was naive." 2
 318/1183

This explanation of our price fixer's wrongdoing

is more interesting for what is missing than for
what is actually there. We hear nothing about
greed, temptation, fooling oneself, or anything
else we tend to associate with those destined to
do wrong. What we do hear about is ordinary
socialization. After working nine years at GE as
an engineer, our price fixer (at age 32) was pro-
moted to "trainee in sales." His superior then
showed him how things were done. Yet,
something is wrong. The witness twice indicates
that this is only "probably" what happened.
Though the acts in question are his, he talks
about them like a scholar analyzing someone
else's acts. A screen has come down between
him and the person he was only a few years
before.
 319/1183

An inability to understand their own past wrong-

doing is, I think, not uncommon in wrongdoers
such as this engineer. One more example shows
that he is not unique. In 1987, the Wall Street
Journal carried a follow-up on the fifty people
by then convicted of inside trading during the
1980s. Here is part of what we learn from one of
them, a civil engineer: "When it started, I didn't
even know what inside information was." But
around 1979, he says, he began reading about
people being arrested for inside tradingyet he
continued trading even though he knew it was
illegal. It's a decision he wouldn't repeat. "In the
long run," he says, "you are going to get
caught."3
 320/1183

This testimony comes from a small investor, not

a broker, analyst, arbitrageur, or the like. Unlike
the price fixer, this inside trader was never
really inside the relevant organization. He
simply received information from inside that he
had no right to receive. Yet, for our purposes, it
doesn't matter. What matters is that he
 Page 63

knew early on that he was doing something il-

legal and stood a fair chance of being arrested.
He went on trading nonetheless. Why? He
doesn't say and, more important, he doesn't
seem to know. He does not object to the law (as
many economists do). He does not claim that
what he was doing was "really" all right. Nor
does he report himself overcome by greed,
temptation, or evil will. All that he says is that
he would not have done it if he had known then
what he knows now. What changed? What does
he know now that he did not know then? It can-
not be what he says, that is, "In the long run,
you are going to get caught." He has no way to
know that every inside trader will get caught in
the long run. Statistics on the occurrence of il-
legal inside trading do not exist; but, if inside
trading is like other crimes, a substantial per-
centage of those engaged in it will never be
caught. 4
 322/1183

Our witness's overstatement of the risks of being

caught is, I think, better understood as a way of
calling attention to what he risked and would
not risk again. He now regrets doing what he did
because he now appreciates what was at stake in
a way he did not at the time. He does not so
much have new information as a new perspect-
ive on the information he had all along. It is this
new perspective that makes it hard for him to
understand how he could have done what he did.
The person we are listening to differs in an im-
portant way from the person who engaged in in-
sider trading a few years before, even if he does
not know anything he did not know before. It is
as if he sees the world with new eyes.

Explaining Lund's Decision

 323/1183

These two wrongdoers are minor figures in ma-

jor scandals. We may therefore wonder whether
the major figures differ in some significant way.
I don't think they do. Recall the events the night
before the Challenger exploded (described in
chapter 4).
 324/1183

The story of the Challenger resembles many

cases discussed in engineering ethics. Although
no one broke the law, as many did in the Gener-
al Electric price-fixing scandal, there was
wrongdoing. And, in retrospect, everyone recog-
nized thator at least sensed it. Mason quickly
took early retirement. Kilminster and Lund were
moved to new offices, told they would be "reas-
signed," and left to read the handwriting on the
wall. Morton Thiokol didn't treat them as if their
errors were merely technical, nor did it defend
their decision the way it would a decision it be-
lieved in. Thiokol's defense consisted largely of
lame excuses, attempts to suppress embarrassing
information, and similar self-convicting
maneuvers.
 325/1183

What went wrong? From the perspective of en-

gineering ethics, it should now be obvious.
Lund, an engineer who held his position in part
because he was an engineer, had a professional
duty to act like an engineer. He was not free to
take off his engineering hat (though he could
wear other hats in addition). For an engineer,
public safety is the paramount consideration.
The engineers could not say the launch would
be safe, so, Lund should have delayed the
launch. Seven people died, in part at least be-
cause Lund did not do what, as an engineer, he
was supposed to do.5
 326/1183

One of the features that made the Challenger

disaster an instant classic is a clear clash of le-
gitimate perspectives. Lund was not just an en-
gineer. He was also a manager. Managers are
not, by definition, evil doers. Government
openly supports institutions to train managers.
Few people would want to forbid managers to
practice
 Page 64

their trade. As a vice president of Morton

Thiokol, part of Lund's job was to wear a man-
agement hat. What then was wrong with his giv-
ing approval after putting on that hat?

The answer must be that, in the decision proced-

ure of which he was part, his job was to stand up
for engineering. He was supposed to represent
engineering judgment in management decision.
He was vice president for engineering. When he
took off his engineering hat, he simply became
another manager. He ceased to perform the job
for which he was needed. That, in retrospect, is
why, even from management's perspective, he
did something wrong.
 328/1183

Why then did he take off his engineering hat

that night? Lund's explanation for deciding as he
did was, and remains, that he had "no choice"
given the Space Center's demand. Self-interest
does not explain what Lund means by "no
choice." To approve the launch was in effect to
bet his career that the Challenger would not ex-
plode, to bet it against the best technical advice
he could get. If he refused to approve the
launch, he would at worst be eased out of his
position to make way for someone less risk
averse. He would have no disaster on his record
and a good chance for another good job either
within Thiokol or outside. Self-interest seems to
support a decision not to launch.
 329/1183

What about moral immaturity? 6 This is a pos-

sibility, but only that. The records tell nothing
about anyone's moral development. Participants
said nothing about social pressure, law, ordinary
morality, or professional ethics. They spoke en-
tirely in the bland technical language engineers
and managers use to communicate with one an-
other. Whatever we say about Lund's moral de-
velopment as of that night would be mere specu-
lation. Something similar is true of the hypo-
thesis that Lund acted with evil intent. By all re-
ports, Lund was too decent a person for that.
 330/1183

What about carelessness, ignorance, or incom-

petence?7 I think none of these explanations will
do. Too much time went into the decision to dis-
miss it as simply careless. Because Lund had the
same training as his engineers and all the in-
formation they had, we can hardly suppose him
to be ignorant in any obvious sense. Nor can we
declare him to be incompetent. Too many exper-
ienced peopleboth at Thiokol and at NASAcon-
curred in Lund's decision for it to be incompet-
ent. Lund may well have been operating at the
limit of his ability or beyond. But that is not ne-
cessarily incompetence. We generally speak of
incompetence only when we have competent al-
ternatives. Here we have no reason to believe
that anyone who could have occupied Lund's
place that night would have done better.
 331/1183

Can we then explain Lund's decision by weak-

ness of will?8 Did he know better but yield to
temptation, give in to pressure, or otherwise
knowingly do what he considered wrong be-
cause he lacked the will to do better? The evid-
ence does not support this explanation either.
Mason's advice, "Take off [your] engineering
hat and put on [your] management hat," does
not sound like tempting Lund to act against his
better judgment. It sounds much more like an
appeal to Lund's better judgment, an appeal
from engineering instinct to management
rationality.

Lund did, of course, give in to pressure. But

"pressure" does not so much explain his de-
cision as describe it on the model of a physical
process (for example, the collapse of a beer can
when we stamp on it). We still need to explain
why the appeal
 Page 65

to management rationality was so convincing

when nothing else was. (We need something
like the physical theory that allows us to under-
stand why the beer can collapses under our
weight but not under the weight of, say, a cat.)
Well, we might say, that's easy enough. The ap-
peal to management rationality allowed Lund to
fool himself into thinking he was doing the right
thing.

Explaining Lund's decision in this way, as a res-

ult of self-deception, is, I think, much closer to
the mark. Mike Martin, coauthor of a text in en-
gineering ethics, recently published a book on
self-deception. Among his many examples of
self-deceivers are participants in the GE price-
fixing scandal. I have no doubt that he could
find some self-deception in Lund as well. 9
 333/1183

Yet, if we can, we should avoid explaining

Lund's act by self-deception. Self-deception,
though common, is an abnormal process. It is
something we do, not simply something that
happens to us. We must knowingly fail to think
about a question in the way we believe most
likely to give the right answer. We must then
not think about the unreliability of the answer
we get even, or especially, when we must act on
it. In its extreme form, self-deception may in-
volve believing something while being aware
that the evidence decisively supports the oppos-
ite belief. Self-deception is, as such, a conscious
flight from reality.10
 334/1183

We should not explain the conduct of respons-

ible people in this way unless the evidence re-
quires it. The evidence hardly requires it in
Lund's case. We can explain why Lund did what
he did by a process which, though similar to
self-deception, is normal, familiar, and at least
as probable on the evidence as any of the ex-
planations we have considered so far. Let's call
the process "microscopic vision."11

Microscopic Vision Examined

 335/1183

What is "microscopic vision"? Perhaps the first

thing to say about it is that it is not "tunnel vis-
ion." Tunnel vision is a narrowing of one's field
of vision without any compensating advantage.
Tunnel vision is literally a defect in vision and
figuratively a defect in our ability to use the in-
formation we have available, a radical single-
mindedness, a monomania. Tunnel vision is of-
ten associated with self-deception. Any advant-
age it might yield would be accidental.
 336/1183

Microscopic vision resembles tunnel vision only

insofar as both involve a narrowing of our field
of vision. But, whereas tunnel vision reduces the
information we have available below what we
could effectively use, microscopic vision does
not. Microscopic vision narrows our field of vis-
ion only because that is necessary to increase
what we can see in what remains. Microscopic
vision is enhanced vision, a giving up of inform-
ation not likely to be useful under the circum-
stances for information more likely to be useful.
If tunnel vision is like looking through a long
tunnel to a point of light at the other end, micro-
scopic vision is like looking into a microscope
at things otherwise too small to see. Hence, my
name for this mental process.
 337/1183

Microscopic vision is also not nearsightedness

or myopia. A nearsighted person has lost the
ability to see things far off. His acuity close up
is what it always was. But when he looks into
the distance, he sees only a blur. Like tunnel vis-
ion, myopia is partial blindness; microscopic
vision is, in contrast, a kind of insight. The near-
 Page 66

sighted person needs glasses or some other aid

to regain normal vision. A person with micro-
scopic vision need only cease using his special
powers to see what others see. He need only
look up from the microscope.

Every skill involves microscopic vision of some

sort. A shoemaker, for example, can tell more
about a shoe in a few seconds than I could tell if
I had a week to examine it. He can see that the
shoe is well or poorly made, that the materials
are good or bad, that the wearer walks in a par-
ticular way, and so on. But the shoemaker's in-
sight has its price. While he is paying attention
to people's shoes, he may be missing what the
people in them are saying or doing. Microscopic
vision is a power, not a handicap, but even
power has its price. We cannot look into the mi-
croscope and see what we would see if we did
not.
 339/1183

Though every skill involves microscopic vision,

the professions provide the most dramatic ex-
amples. In part, the professions provide these
examples because the insight they give is relat-
ively general. The microscopic vision of a law-
yer, engineer, doctor, minister, or accountant
concerns central features of social life, which
the microscopic vision of a shoemaker does not.
In part, though, professions provide the most
dramatic examples of microscopic vision be-
cause both the long training required to become
a professional and the long hours characteristic
of professional work make the professional's mi-
croscopic vision more central to his life. A pro-
fession is a way of life in a way shoemaking is
not (or, at least, is not anymore).
 340/1183

Consider, for example, the stereotypes of pro-

fessionalsthe pushy lawyer, the comforting doc-
tor, the quiet accountant, and so on. There are
no similar stereotypes of the shoemaker, the car-
penter, or the personnel director. These skills
don't seem to shape character as much. We joke
about professional myopiafor example, the en-
gineer who, about to be excused from a death
sentence because the guillotine's blade jammed
during the preceding execution, volunteers to
"fix the problem." Behind the joke is an appreci-
ation of the power a profession has to shape, and
therefore, to misshape, the consciousness of its
members. Real professional myopia is probably
rare. Few professionals seem to lose altogether
the ability to see the world as ordinary people
do. Common, however, is a tendency not to look
up from the microscope, a tendency unthink-
ingly to extend the profession's perspective to
every aspect of life.
 341/1183

Managers are not professionals in the strict

sense. Though managing now has schools like
those of the professions, managers lack two fea-
tures essential to professionals strictly so called:
first, a formal commitment to a moral ideal, and
second, a common code of ethics. Indeed, man-
agers seem to me to lack even a clear sense of
themselves as managersthat is, as custodians of
other people's wealth, organization, and reputa-
tion. I am surprised by the number of managers
who think of themselves as entrepreneurs or
capitalistsas businesspeople who risk their own
money, not someone else's.
 342/1183

Nonetheless, managing today does have many

of the characteristics of a profession, including
distinctive skills and a corresponding perspect-
ive, butmost importanta way of life that can
make their microscopic vision seem all that mat-
ters. We have the stereotype of the manager who
can't see the toxic waste beyond the end of his
budget. Behind that stereotype is a certain real-
ity. Managers, especially senior managers, work
almost entirely with other managers. Their days
are spent
 Page 67

"in the office" doing the things that managers

do. Often, those days are quite long, not 9 to 5
but 8 to 7 or even 7 to 8. They read management
magazines and go to management meetings.
They come to see the world from the perspect-
ive of a manager.
 344/1183

There is a natural process by which people are

made into managers. But most companies are
not satisfied with "normal acculturation." They
have special programs to train managers in a
certain style of management. Roger Boisjoly,
one of the engineers who tried to get Lund to
stick to his no-launch recommendation, was
himself briefly a manager. He went back to be-
ing an engineer because he wanted to be closer
to work on the shuttle. Though he mocked
Thiokol's management programs as "charm
schools," he has also pointed out that they
helped to make the managers at Thiokol a co-
hesive team. They helped to give the engineer
turned manager a clear sense of the priority of
the manager's way of looking at things. 12
 345/1183

What is the difference between the way an en-

gineer might look at a decision and the way a
manager might? For our purposes now, what is
important is the way engineers and managers
approach risk.13 I think engineers and managers
differ in at least two ways. First, engineers
would not normally include in their calculations
certain risksfor example, the risk of losing the
shuttle contract if the launch schedule were not
kept. Such risks are not their professional con-
cern; such risks are properly a manager's con-
cern. Second, engineers are trained to be conser-
vative in their assessment of permissible risk.
Often they work from tables approved by the
appropriate professional association or other
standard-setting agency. When they do not have
such tables, they try not to go substantially bey-
ond what experience has shown to be safe.
Engineers do not, in general, balance risk
against benefit. They reduce risk to permissible
levels and only then proceed. Managers, on the
 346/1183

other hand, generally balance risk against bene-

fit. That is one of the things they are trained to
do.

So, we have two perspectives on the same prob-

lem: the engineer's and the manager's. Which is
better? The answer is neither. The engineer's
perspective is generally better for making engin-
eering decisions; the manager's perspective is
generally better for making management de-
cisions. Either perspective has a tendency to
yield a bad result if applied to the wrong kind of
decision. Indeed, that is nearly a tautology. If,
for example, we thought a certain decision bet-
ter made by managers than engineers, we would
describe it as (properly) a management, rather
than an engineering, decision.14
 347/1183

If that is so, it's not too hard to understand why

Lund changed his mind the night before the
Challenger exploded (and why he might still
claim that he had "no choice"). Once he began
thinking about the launch as an ordinary man-
agement decision, he could rationally conclude
that the risk of explosion was small enough to
tolerate given the demands of NASA and how
much was at stake for Morton Thiokol. But why
would Lund think about the launch like a man-
ager rather than an engineer?

As I described the difference between the

engineer's perspective and the manager's, the
two approaches to risk are inconsistent. Lund
had to choose. In a way, Mason's plea to Lund
to take off his engineering hat and put on his
management hat accurately stated the choice
Lund faced. In another way, however, it did not.
 Page 68

Mason's plea assumed that Lund's decision to

launch was an ordinary management decision.
This would be just what any manager would
normally assume (unless trained to think other-
wise), especially a manager who was himself an
engineer. For an engineer to be made a manager
is generally considered a promotion, an opening
of new horizons. Managers are in charge of en-
gineers. They regularly receive engineering re-
commendations and then act on them, taking in-
to account (supposedly) more than the engineers
did. Engineers generally defer to managers. 15
 349/1183

Anyone who thought of relations between en-

gineers and managers in this way would, I think,
succumb to Mason's pleaunless he had a clear
understanding of what made him different from
other managers. Lund, it seems, had no such un-
derstanding. Indeed, Mason's plea probably
shows that no one in the senior management at
Morton Thiokol did. Mason could hardly urge
Lund to take off his engineering hat in front of
so many managers if it was common knowledge
that Lund had a duty to keep his engineering hat
on. Perhaps those who originally organized the
decision procedure at Thiokol understood things
better. If so, they failed to institutionalize their
understanding. Without some way to preserve
that understanding, ordinary management under-
standing would eventually take over. The Chal-
lenger explosion was the natural outcome of or-
dinary management.
 350/1183

My purpose here is not to defend Lund's de-

cision but simply to understand how he might
have made it without being careless, ignorant,
incompetent, evil-willed, weak-willed, morally
immature, or self-deceiving. I explained his de-
cision as rational from the perspective of an or-
dinary manager and then explained why that
perspective might seem the right one at the de-
cisive moment. Earlier, I pointed out that, in ret-
rospect, everyone seemed to see not only that
Lund made an unfortunate decision but that the
decision he made was wrong. I have now ex-
plained why. Lund was not an ordinary man-
ager; he was supposed to be an engineer among
managers.
 351/1183

Next, I would like to generalize what we learned

from Lund: We have a tendency to suppose that
doing the right thing is normal, doing the wrong
thing is abnormal, and when something goes
wrong, the cause must be something abnormal,
usually a moral failing in the wrongdoer. What
the analysis so far suggests is that sometimes at
least the wrong may be the result of normal pro-
cesses.16

The analysis also suggests something more:

Managers sometimes say of obeying the law, of
doing what's morally right, or of maintaining
professional standards, "That should go without
saying"or, in other words, that the importance of
such things is so obvious that pointing the im-
portance out is unnecessary. The truth, I think, is
almost the reverse. I return to this point later.

Price Fixing and Inside Trading

 352/1183

Lund represents one category of wrongdoer-

those whose conduct is merely unprofessional.
There is anotherthose whose wrongdoing is il-
legal. Could what I said of Lund apply to this
other category as well? Let's try to answer that
question by briefly examining the two lawbreak-
ers with whom we began. We are now ready to
understand why they might have done what they
did.

First, the price fixer at General Electric. Arriv-

ing at his new job eager to learn, he found that
much he had learned as an engineer did not
quite fit. Everyday was
 Page 69

a struggle to ''get up to speed." One day his su-

perior invited him to go to a meeting. The meet-
ing consisted of sales managers from the other
two major turbine manufacturers, clearly re-
spectable people and clearly engaged in fixing
market prices. There was no question that the
meeting was secret. But what conclusion should
he draw from that? A company like GE has
many secrets. The meeting did not take long and
soon the future price fixer was doing other
things. After a few more meetings like this, he
was allowed to go without his superior. Soon the
meetings were routine.
 354/1183

He may initially have had qualms about the

meetings. We often have qualms about a prac-
tice with which we are unfamiliar. But we learn
to suspend judgment for a decent interval. Often
the qualms disappear as understanding in-
creases. Of course, the price fixer may have had
more than the usual reasons for qualms. He may
have received in the mail a copy of a GE policy
that forbade what he was doing ("policy 20.5").
But the policy would have come from the law
department, not from anyone in his chain of
command. Nothing would have made that mail-
ing seem more important than other mailings
from law or other non-line departments that
managers routinely ignore. What occupied the
price fixer's time, his field of vision, as we
might say, was learning to be a manager. He had
little time to think about matters that seemed to
matter to no one with whom he dealt. Eventu-
ally, he would stop thinking about the price fix-
ing as price fixing. 17
 355/1183

I have now told the price fixer's story in terms

appropriate to microscopic vision. I described
him as developing in a normal way a sense for
what matters and what does not matter in a cer-
tain environment. The process is similar to the
"desensitization" that a surgeon must undergo
before she can calmly cut off a human limb or
put a knife into a still-beating heart. Though the
process I described does not require learning to
block anything out, only failing to use some in-
formation because one is busy using other in-
formation, it does share at least one important
feature with self-deception (something ordinary
desensitizing does not): The price fixer was
misled. The process nonetheless differs from
self-deception in at least two ways:
 356/1183

1. Although self-deception presupposes in the

self-deceiver some sense of the unreliability of
the procedure he is using to learn about the
world, the process I described presupposes no
such thing. The price fixer might well have be-
lieved his procedure would yield an accurate, al-
beit incomplete, picture of the world in which he
worked.

2. Although self-deception presupposes that the

procedure used is in fact generally unreliable, we
need not presuppose that here. The procedure I
described might well be generally reliable. The
problem is that the price fixer's procedure was
not "designed" to distinguish legal from illegal
management. That procedure may well have
made sense in the 1950s. Who then would have
thought that the managers of a company like GE
would have engaged in extensive illegal con-
duct? The discovery must have astonished many
people, including our price fixer.
 357/1183

Next consider the inside trader. He got warnings

of a sort the price fixer did not. He actually read
about indictments of people for inside trading.
Yet, he continued to trade on inside information.
Why? Perhaps he deceived himself about the
chances of being caught. A more interesting
possibility, though, is that he never thought
about being caught. He began inside trading
with a clear conscience. He read of its
 Page 70

illegality only after he had grown used to inside

trading. Perhaps those with whom he cooperated
showed no fear. Busy with many things beside
inside trading, normal prudence would have told
him that, if he feared everything the newspapers
invited him to fear, he would live with number-
less terrors. We must use the judgment we de-
velop in daily life to put newspaper stories in
perspective. His immediate environment seemed
as safe as ever. So, why should he worry about
what he read in the newspaper? Again, we have
a normal process leading to an abnormal result.
 359/1183

I do not, of course, claim that this is how it was.

I am suggesting a hypothesis, not demonstrating
it. However, I should, I think, point out one
piece of evidence that suggess that my explana-
tion is at least partly right. Recall that neither of
these wrongdoers seems to understand how he
came to do what he did. That is what we should
expect if my version of their stories is more or
less right. As I told their stories, each did wrong
in part at least because he did not use certain
factspolicy 20.5 in one case, arrests of inside
traders in the otherin a way that would have led
him to a true understanding of what he was do-
ing or to the conduct such an understanding
would normally lead to. The facts did not trigger
the fear of punishment or concern about doing
wrong that seems normal outside the environ-
ment in which they were working. The facts
seemed to have been pushed from consciousness
by other facts, as most of the world is pushed
aside when we look into a microscope.
 360/1183

Once the two wrongdoers were pulled from the

microscope, they would cease to see the world
as they had. They would not, however, have that
sense of having "known it all along" so charac-
teristic of coming out of self-deception (because
they could not have known it until they looked
at the world somewhat differently). They would
instead be aware of seeing something thatin the
light of the evidence now before themmust have
been there all along. Microscopic vision is a
metaphor for a mental process, a "mind-set" or
''cognitive map." Because the mechanics of such
mental processes are no more visible to the per-
son whose mind it is than it is to an outside ob-
server, these two wrongdoers need not have
been aware of what made them attend to other
things until now. They could honestly be per-
plexed about how they could have missed for so
long what is now as plain as day.

Some Practical Lessons

 361/1183

In this chapter, I tried to describe wrongdoing as

the outcome of a social process the literature on
wrongdoing overlooked. I tried to avoid assum-
ing such serious moral failings as weakness of
will or self-deception. Of course, I did not de-
scribe my wrongdoers as paragons of rationality
or virtue. The price fixer and inside trader were
certainly naive; they lacked enough insight into
the way the world works to recognize obvious
signs of trouble. Lund, however, seems no more
naive than the rest of us. I can imagine myself
doing what he did.

If the process I described in fact explains much

wrongdoing in large organizations, we may
draw some interesting conclusions about how to
prevent that wrongdoing. The most obvious,
perhaps, is that screening out potential wrongdo-
ers is probably impractical. Who would be let in
by a procedure that screened out the
 Page 71

wrongdoers we discussed here? We must in-

stead consider how to prevent wrongdoing by
the relatively decent people an organization
must employ.
 363/1183

The problem as I described it is that normal pro-

cesses can lead to important information going
unused at a decisive moment. Lund's training as
a manager would not prepare him to see how
special his role was. The future price fixer's way
of learning his job would not alert him to the
risks of illegality, much less to any moral objec-
tions to fixing prices. The inside trader's experi-
ence would make him discount the warning
signs in the newspaper. Though microscopic
vision is not a flight from reality, it does involve
a sacrifice of one part of reality to another. Usu-
ally, the sacrifice is worth it. Sometimes it is
not. When it is not, we need to change the mi-
croscopic vision of those working in the envir-
onment in question or change the environment.
Sometimes we need to change both. Often,
changing one changes the other too.
 364/1183

How might we change the environment? One

way is simply to talk openly and often about
what we want people to notice. For example,
Lund would probably have refused to do as
Mason suggested if the people back at Morton
Thiokol's headquarters in Chicago regularly re-
minded him that he was no ordinary manager:
"We are counting on you to stand up for engin-
eering considerations whatever anyone else
does." Indeed, had Mason heard headquarters
say that to Lund a few times, he could hardly
have said what he did say. He might well have
deferred to Lund's judgment even though NASA
was pressuring him. "Sorry," he could have said,
"my hands are tied."
 365/1183

Business professors especially, but ordinary

managers as well, often decline to talk about
what they call ethics because they do not want
to "sermonize." Sermons, they say, cannot lead
people to do the right thing. If adults haven't
learned to be ethical by now, or don't want to,
what can a sermon do?

These professors and managers seem to use the

word "ethics" as a catch-all for whatever "value"
considerations seem so obvious that they would
be embarrassed to raise them. I must admit to
some doubts about their consistency here. These
same people regularly sermonize about profit.
They do not find mention of profit embarrassing
though we might suppose that, for them, profit
would be the most obvious value consideration
of all.
 366/1183

Still, whatever doubts I have about their consist-

ency, I can easily respond to their concern about
the ineffectiveness of sermonizing. However ob-
vious the sermon's content, the sermon itself can
help to keep legal, moral, and professional con-
siderations in an organization's collective field
of vision. That, I think, is why both business
professors and ordinary managers talk so much
about profit. That is how they keep profit a
primary concern. So, doing the same for ethical
considerations should, by itself, be a significant
contribution to getting decent people to do the
right thing.
 367/1183

Sermons are, of course, hardly the best way to

do that. Better than sermons are such familiar
devices as a code of ethics, ethics audit, ethics
seminar for managers, discussion of ethics in the
course of ordinary decision making, and rewards
for those who go out of their way to do the right
thing ("reward" including not only praise but
also the other valuables that normally go to
those who serve their employer well,
 Page 72

especially money and promotion). But, whatever

the merits of these particular devices in them-
selves, they all have this important characteristic
in common. They help to keep employees alert
to wrongdoing. They help to maintain a certain
way of seeing the world.
 369/1183

What about teaching? Part of teaching is getting

people used to thinking in a certain way. What
academics call disciplines are in fact forms of
microscopic vision. We should, therefore, pay as
much attention to what we don't teach as to what
we do teach. If we limit ourselves to teaching
technical aspects of a discipline, those we teach
tend to develop a perspective including only
those technical aspects. They do not automatic-
ally include what we don't teach. Indeed, they
would be quite unusual students even to see how
to include such extras. If, then, we teach engin-
eering without teaching engineering ethics, our
graduates will begin work thinking about the
technical aspects of engineering without think-
ing about the ethical aspects. They will not dis-
miss the ethical aspects. They will not even see
them. 18
 370/1183

Of course, we can teach what we should and

still do little good. Good conduct in business or
a profession presupposes a suitable social con-
text. If a morally sensitive graduate goes to
work in a company where ethics is ignored, he
will, if he stays, slowly lose his sense of the eth-
ical dimension of what he does. His field of vis-
ion will narrow. Eventually, he may be as blind
as if we had taught him nothing. This claimI
should stressis not about moral development (as
that term is now commonly understood). Our
graduate may well score no worse on a Kohl-
berg test than he did before.19 He will simply
cease to think of a certain range of decisions as
raising questions to which moral categories are
important. The questions will seem "merely
technical," "an ordinary business decision," or in
some other respect "merely routine."
 371/1183

To develop such "moral blindness" is, of course,

a misfortune. Insofar as it results from circum-
stances over which one has little or no con-
trolfor example, ignorance of what a company is
really likeit is a misfortune for which one cannot
be held responsible. Even so, the resulting moral
blindness does not provide an excuse for wrong-
doing. Moral blindness is itself a character flaw.
That an act arose from a flaw in character
(rather than from good character) adds to the
grounds for condemnation rather than reduces
them.
 372/1183

That an organization can make the teaching of

ethics ineffective by blinding its employees to
moral concerns is no reason not to teach ethics.
But it is good reason to conceive ethics teaching
as part of a larger process, and good reason too,
to try to transmit that conception to our students.
Caroline Whitbeck provides a good example of
what can be done. As part of a course in engin-
eering design, she had her students contact local
companies to find out how an engineer in the
company could raise an issue of professional
ethics related to design. Her students thus
learned to think of their future employers as in
part "ethics environments." But that is not all
her students did. Their inquiries made it more
likely that their future employers will think
about how an engineer could raise an ethics is-
sue. So, Whitbeck was also helping to improve
the ethics environment in which her students
would someday work.20
373/1183
 Page 73

6
Avoiding the Tragedy of Whistleblowing
The strength of the pack is the Wolf, and the strength
of the Wolf is the pack.
Rudyard Kipling, "The Law of the Jungle"
 375/1183

I focused on the decision to launch the Chal-

lenger to help us understand some of the com-
plex interrelations between technical and organ-
izational factors so characteristic of much that
engineers do. Yet what the Challenger disaster
is most remembered for today is an engineer,
Roger Boisjoly. Boisjoly stepped forward after
the event, against the wishes of his employer
and at the risk of his career, to tell the truth
about what happened. He was not the first en-
gineer to "blow the whistle" on an employer, nor
is he likely to be the last. Not without reason
does the literature of engineering ethics and the
teaching of it emphasize the subject of whis-
tleblowing. Whistleblowing is one way engin-
eers have to show that the public health, safety,
and welfare means more to them than employer,
career, and even their own material welfare.
Whistleblowing also reminds us of the political
side of engineering, the importance of what en-
gineers say and how they say it.
 376/1183

Most discussions of whistleblowing seek to jus-

tify whistleblowing or to distinguish justified
from unjustified whistleblowing, or they report
who blows the whistle, how, and why; or they
advise on how to blow the whistle or how to re-
spond to an employee about to blow the whistle
or what to do once she has; or they make recom-
mendations for new laws to protect whis-
tleblowers. In one way or another, such discus-
sions treat whistleblowing as inevitable. I do not
do that here. Instead, I consider ways to avoid
whistleblowing.

It is not that I oppose whistleblowing. I think

whistleblowing is, on balance, a necessary evil
(and sometimes even a positive good). I cer-
tainly think justified whistleblowers should be
protected legally. 1 They should not be fired for
their good deed or punished for it in any other
way. But I doubt that much can be done to pro-
tect them. This chapter explains why.
377/1183
 Page 74

The explanation stresses the destructive side of

whistleblowing, making it easier to understand
those who mistreat whistleblowers. Insofar as it
does that, it both gives the organization's case
for mistreatment and shows the importance of
avoiding whistleblowing. We must get the bene-
fits of whistleblowing without making people
and organizations pay the enormous price that
whistleblowing typically exacts. 2
 379/1183

I address this chapter both to those who have a

substantial say in how some organization runs
and to those who may someday have to blow the
whistle on their own organization. These groups
overlap more than most discussions of whis-
tleblowing suggest.3 However, that is not why I
chose to address both here. My reason runs
deeper. I believe that even if those two groups
did not overlap, they would still share an interest
in making whistleblowing unnecessary; that
both groups can do much to make whistleblow-
ing unnecessary; and that each will be better
able to do its part if it understands what the oth-
er group can do.

The Informal Organization Within the Formal

 380/1183

No matter how large or small, every formal or-

ganization includes one or more informal
groups. An academic department, for example,
is a network of poker buddies, movie buffs,
cooks, and so on. Departmental conversation is
not limited to what must be said to carry on de-
partmental business. Ordinary life, ordinary atti-
tudes, permeate the formal structure. Much of
what makes the formal organization succeed or
fail goes on within and between these informal
groups. Who likes us is at least as important in
most organizations as what we do. Success is
not simply a matter of technical skill or accom-
plishment. We must also have enough friends
properly placedand not too many enemies. Per-
haps only at hiring time do academics talk much
about personality, but every academic knows of
a department that fell apart because certain
members did not get along and others that sur-
vived financial troubles, campus disorders, and
tempting offers to individual members in part at
 381/1183

least because the faculty got along so well with

each other.4

Though my example is an academic department,

nonacademics will confirm that much the same
is true in industrial plants, research laboratories,
and even government bureaus. Most of what
makes such organizations work, or fail to work,
can't be learned from the table of organization,
formal job descriptions, or even personnel eval-
uations. Thinking realistically about whis-
tleblowing means thinking about the informal
aspects of formal organization as well as the
formal. I focus on those informal aspects here.

Blaming the Messenger

 382/1183

"Whistleblower" is a capacious term. Whis-

tleblowers can be anonymous or open, internal
or external, well-intentioned or not so well-in-
tentioned, accurate or inaccurate, justified or un-
justified. Perhaps strictly speaking, some of
these are not whistleblowers at all.5 But I have
no reason to speak strictly here. For my pur-
poses, "whistleblower" may refer to any mem-
ber of a formal organization who takes informa-
tion out of channels to try to stop the organiza-
tion from doing something he believes morally
wrong (or to force it to do something he believes
morally required).6
 Page

Almost any organization will fire a whistleblower i

it can, whether she was right or not; will ruin her jo
prospects if it can; and if it can do neither will still
what it can to make her life miserable. An otherwis
humane organization can treat a whistleblower sav-
agely. 7 Why?

The most frequent answer is that those who mistrea

whistleblowers do so because they expect to benefi
from having fewer whistleblowers. The self-interes
of individuals or their organization explains the
mistreatment.

Though no doubt part of the truth, this explanation

only a small part. In general, we are far from perfec
judges of self-interest. Our judgment does not im-
prove simply because we assume an organizational
role. We can still be quite irrational. Recall how
Shakespeare's Cleopatra responds to her messenger
report that Antony married Octavia:
 384/1183

. . . Hence
Horrible villain! or I'll spurn thine eyes
Like balls before me; I'll unhair thy head;
Thou shalt be whipp'd with wire and stew'd in brine . . .
. . . let ill tidings tell
Themselves when they be felt.8

Though Cleopatra ordered him to spy on Antony, th

messenger hears more harsh words, receives severa
hard blows, and has a knife angrily put to his throat
before he is allowed to leave with a small reward.
 385/1183

Today's formal organizations can treat the bearer of

bad news much as Shakespeare's love-sick Cleopatr
did. So, for example, in a recent book on corporate
life, Robert Jackall grimly recounts what happened
several executives with bad news to tell their respec
ive organizations. Though each discovered wrongd
ing it was his duty to discover, reported it through
channels, and saw the wrongdoer punished, though
none of them was responsible for the wrong reporte
and though the organization was better off for the r
port, the lucky among Jackall's executives had their
part in the affair forgotten. Some paid with their ca-
reer.9

We usually think of information as powerand it is.

But thinking of information that way is no small
achievement when the information wrecks our plan
Even experienced managers can find themselves
telling subordinates, "I don't want to hear any more
bad news."
 386/1183

The rationality of a formal organization is an ideal

never more than partially achieved. We must keep
that in mind if we are to understand what happens t
so many whistleblowers. An organization that woul
"whip with wire and stew in brine" the simple beare
of bad news is not likely to respond well to the whi
tleblowereven if, as often happens, the whistleblow
serves the organization's longterm interests. The
whistleblower is, after all, not only a bearer of bad
news; he is bad news.

Whistleblowing as Bad News All Around

Discussions of whistleblowing tend to emphasize th

undeniable good the accurate whistleblower does
(both directly, by revealing wrongdoing, and indir-
ectly, by de-
 Page 76

terring further wrongdoing). The incidental

harm tends to be overshadowed, perhaps be-
cause so much of it seems deserved. The harm
done by inaccurate whistleblowing receives
much less attention. 10

Whatever the reasons for ignoring the bad news

about whistleblowing, the fact remains that
much bad news is ignored and, for our purposes,
the bad news is crucial. So, let us recall how
much bad news there is:

Whistleblowing is always proof of organization-

al trouble. Employees do not go out of channels
unless the channels seem inadequate.
 388/1183

Whistleblowing is also proof of management

failure. Usually several managers directly above
the whistleblower have heard his complaint,
tried to deal with it in some way, and failed to
satisfy him. However managers view the
whistleblower's complaint, they are bound to
view their own failure to "keep control" as a blot
on their record.

Whistleblowing is also bad news for those on

whom the whistle is blown. What they were do-
ing in obscurity is suddenly in the spotlight.
They have to participate in "damage control"
meetings, investigations, and the like that would
not otherwise demand their scarce time. They
have to write unusual reports, worry about the
effect of publicity on their own career, and face
the pointed questions of spouse, children, and
friends. And they may have to go on doing such
things for monthsor even years.
 389/1183

Insofar as whistleblowing has such effects, no

one within the organization will be able to hear
the whistleblower's name without thinking un-
pleasant thoughts. No manager will be able to
make a decision about the whistleblower
without having bad associations color her judg-
ment. The whistleblower not only makes con-
scious enemies within his organization but also
creates enormous biases against himself, biases
hard to cancel by any formal procedure.

And there is more. What must the whistleblower

suffer to blow the whistle? At the very least, he
must lose faith in the formal organization. If he
kept faith, he would accept whatever decision
came through formal channelsat least once he
exhausted all formal means of appeal.
 390/1183

For anyone who was a loyal employee for many

years, losing faith in the organization is likely to
be quite painfulrather like the disintegration of a
marriage. My impression is that few whis-
tleblowers accept a job thinking that they might
someday have to blow the whistle. They seem to
start out as loyal employeesperhaps more loyal
than most. One day something happens to shake
their loyalty. Further shocks follow until loyalty
collapses, leaving behind a great emptiness.
Managers tend to think of whistleblowers as
traitors to the organization, but most whis-
tleblowers seem to feel that, on the contrary, it is
the organization that betrayed them.11
 391/1183

Moreover, before the whistleblower was forced

to blow the whistle, she trusted the formal or-
ganization. She took its good sense for granted.
Now that is no longer possible. Faith becomes
suspicion. Because what we call organizational
authority is precisely the ability of the organiza-
tion to have its commands taken more or less on
faith, the "powers that be" now have as much
reason to distrust the whistleblower as she has to
distrust them.12 She no longer recognizes their
authority. She is much more likely to blow the
whistle than before. She is now an enemy
within.
 Page 77

Something equally bad happens to relations

between the whistleblower and her coworkers.
Whistleblowing tends to bring out the worst in
people. Some friends will have become implac-
able enemies. Others hide, fearing "guilt by as-
sociation." Most simply lose interest, looking on
the whistleblower as they would someone dying
of cancer. These desertions can leave deep scars.
And even when they do not, they leave the
whistleblower an outsider, a loner in an organiz-
ation in which isolation for any reason makes
people vulnerable.
 393/1183

All this suggests some hard questions: How can

a whistleblower work as before with people
whose loyalty he no longer shares? How can
coworkers treat him as they did before when he
is no longer quite one of them? How can he
hope for promotion, or even retention, in an or-
ganization in which he can put no trust, in which
he has no friends, and for which he is likely to
make further trouble? These are questions a law
cannot answer.

Helping the Whistleblower and the Organization

 394/1183

What then can be done for whistleblowers? One

option is to find them other jobs. That is not
easy. Potential employers generally shun known
whistleblowers. Then, too, the whistleblower
may not be as good an interviewee as before.
Many whistleblowers seem to signal the bad
news even when they do their best to conceal it.
They may, for example, sound emotionally ex-
hausted, ask questions that suggest distrust, or
just seem prickly. They are like people going
through a bad divorce.
 395/1183

Because few potential employers want someone

else's troubles, we must draw this paradoxical
conclusion: The whistleblower's best hope for
continuing her career may be her old employer.
That the old employer may be her best hope is
the chief reason to support laws protecting whis-
tleblowers. Though a law can offer the whis-
tleblower little direct protection, it can prod the
organization to think about making peace with
the whistleblower. This, however, is still a small
hope. The organization can make peace with the
whistleblower only if it can reestablish her loy-
alty to the organization and her trust in those
with whom she must work. That is not easy.
 396/1183

Clearly, the formal organization itself must

change enough for the whistleblower to have
good reason to believe that she will not have to
go out of channels again. The changes probably
have to be substantial. Most organizations auto-
matically resist substantial change. But formal
changes alone are not enough to reestablish the
whistleblower's informal relations with superi-
ors, subordinates, and coworkers. What is
needed in addition is something like marriage
counseling, group therapy to expose and resolve
all the feelings of betrayal, distrust, and rejec-
tion whistleblowing inevitably generates. The
whistleblower will not be safe until she is rein-
tegrated into the informal organization.
 397/1183

Some government agencies require employees

involved in a whistleblowing case to participate
in such group therapy. The results so far have
not been good. Managers, especially, seem to
view such therapy as just one more hoop to
jump through on the way to the inevitable. 13 To
work, the therapy probably needs to be volun-
tary for all participants, something not easily
legislated.
Page 78
 399/1183

That is why even this best hope for the whis-

tleblower, reconciliation with the organization,
is so small. We need to find better ways to pro-
tect whistleblowers. In the long run, peace
between the whistleblower and the organization
is as good for the organization as for the whis-
tleblower. The whistleblower is not really an en-
emy. An organization that has whistleblowers
needs them. The whistleblower is like the knock
at the door that wakes someone in a house on
fireunwelcome, but better than sleeping until the
fire reaches the bed. An organization that pun-
ishes its whistleblowers blinds itself to troubles
better faced. So, for example, when the Bay
Area Transit Authority (BART) fired three elec-
trical engineers for reporting trouble with the
program that was to run its new operatorless
trains, it merely turned a technical problem into
scandal waiting to happen. When a train jumped
the tracks after rushing through a station at
which it was supposed to stop, BART had to
 400/1183

face the technical problem the engineers had

identified, to deal with public concern over the
operatorless trains, and to explain why it had ig-
nored the problem until then.

To say that whistleblowers generally tell an or-

ganization what it needs to know is not to deny
the disadvantages of whistleblowing described
earlier but to explain why we should try to make
whistleblowing unnecessary rather than try to
prevent whistleblowing in other ways. I now
turn to the chief means of making whistleblow-
ing unnecessary.

How Organizations can Avoid Whistleblowing

If whistleblowing means that an organization

has trouble using bad news, one way for an or-
ganization to avoid whistleblowing is to im-
prove its ability to use bad news. We may dis-
tinguish three approaches: procedural, educa-
tional, and structural.
 401/1183

The "procedural" approach builds invitations to

report bad news into the ordinary ways of doing
business. These procedures can be quite simple,
for example, a space on a form for "disadvant-
ages" or "risks." Such a blank almost forces the
person filling out the form to say something
negative. Those above him are also more likely
to treat bad news reported in this way as part of
"doing the job'' than they would the same bad
news reported without that specific invitation. 14

The first approach also includes more complic-

ated procedures, for example, "review meetings"
the purpose of which is to identify problems.
The review meeting works like a blank space.
When the emphasis is on revealing bad news,
more bad news is likely to come out. Revealing
bad news is more likely to seem part of the job.
 402/1183

Of course, how things seem is in part a matter of

the mental set that the people involved bring to
the procedure. That set is determined in large
part by what happened in the organization be-
fore. Organizational atmosphere can turn any
procedure into a mere formality. If, for example,
people who fill in the disadvantage blank or
speak up at a review meeting are commonly
treated like Cleopatra's messenger, the proced-
ures will bring in little bad news. Part of making
procedures work is making sure those involved
think about them in the right way. That is espe-
cially important when the procedures are new
and patterns of response have not yet de-
veloped.15
 Page 79

In a way, then, my first approach, the procedur-

al, presupposes the two others. Those participat-
ing in various procedures need to understand
how important bad news can be. They also need
regular reminders because everyday experience
tends to teach them how much bad news hurts.
Education can provide one reminder; a structure
of formal incentives can provide another.
 404/1183

I intend "education" to be understood broadly

(so broadly in fact that the line between educa-
tion and formal incentives all but disappears).
Training sessions in which superiors or special
trainers stress the importance of hearing the
worst are only part of what I have in mind.
Everyday experience is also part of education.
Subordinates are more likely to take the formal
training to heart if they are regularly thanked for
giving superiors bad news, if they see that bring-
ing bad news is treated much as bringing good
news is, and so on.
 405/1183

Superiors are, of course, more likely to treat

well those subordinates who bring bad news if
the organization makes it rational to do so. But
treating such subordinates well is generally ra-
tional only if the organization routinely uses bad
news in ways that encourage reporting itor, at
least, do not discourage reporting it. An
organization's ability to do this routinely de-
pends on its structure.
 406/1183

For example, suppose an organization holds a

manager responsible only for what gets reported
"on her watch." Suppose, too, that her subordin-
ate informs her that her predecessor improved
the division's profits by skipping routine main-
tenance and now much of the machinery is in
poor condition. The manager will not want to re-
port this to her superiors. She would be bringing
news that threatens everyone who must pass it
on. She therefore does not want to hear the bad
news herself. She has good reason to tell her
subordinate, "Let sleeping dogs lie." Perhaps the
dogs will not howl until her successor takes
over.
 407/1183

Now, suppose instead that the organization has

routine ways of assigning responsibility to a
manager for what she does while in a position
even if the bad consequences only become ap-
parent later. In such an organization, a manager
has good reason to want subordinates to report
the bad news about her predecessor's work as
soon as they learn of it. She need not fear such
"sleeping dogs." They will not wake to howl for
her blood. And, if she lets them lie, she may
later have to explain how she could have missed
them.
 408/1183

Most organizations tend to treat the person in

charge as responsible for whatever bad news he
must report. Few have any routine for assigning
responsibility to anyone else (perhaps because
such a routine would be quite expensive). 16
Hence, in that respect at least, most organiza-
tions have structures tending to discourage bad
news. Leaving managers in charge for long
terms, say, ten or twenty years, would probably
compensate for this tendency. Few problems lie
dormant that long. Today, however, managers
seldom stay in one position for even five years.
If they do not rise quickly within an organiza-
tion, they are likely to move to another. This
mobility means that most organizations must
rely on other means of giving managers reason
to welcome bad news.
 409/1183

The most common approach these days is to

create alternative channels for bad news so that
no one in an organization is in position to block
its flow upward. The oldest of these alternative
channels is probably the regular outside audit.
Another is
 Page 80

an "open door" policy that allows subordinates

to go directly to a senior official, bypassing sev-
eral layers of management. Another is changing
the traditional chain of command into something
much more like a lattice, so that subordinates
have less to fear from any particular superior
and have routine access to more than one. (I say
more about such arrangements in chapter 9.)
Such arrangements give a manager reason to be
thankful that he heard the bad news from a sub-
ordinate rather than from a superior and reason
to try to respond in a way likely to satisfy the
subordinate. The subordinate saved the manager
from being "blindsided." Such arrangements
tend to make whistleblowing unnecessary.

We are now ready to consider how individuals

can avoid becoming whistleblowers.

How to Avoid Having to Blow the Whistle

 411/1183

The simplest way to avoid having to blow the

whistle may seem to be to join an organization
in which whistleblowing will never be neces-
sary. Unfortunately, things are not that simple.
Organizations are human contrivances; none is
perfect.
 412/1183

Still, organizations differ quite a bit. By choos-

ing the right organization, an individual can sub-
stantially reduce the chance that he will have to
blow the whistle (much as, it is said, one can re-
duce substantially the chance of repenting "at
leisure" by not "marrying in haste"). The ques-
tion is how the organization handles bad news.
The answer is found in the organization's pro-
cedures, educational programs, and structure,
not the ones "on paper,'' of course, but the ones
actually in effect. The difference can be crucial.
For example, if the organization has an open-
door policy, is the door ever used? Because or-
ganizations always work imperfectly, an open
door that is never used is probably not an unne-
cessary channel but one no one dare use. Using
such a channel will probably be treated as
whistleblowing.
 413/1183

Any organization described as "one happy fam-

ily" should be examined with special care. Or-
ganizations, like families, generally have argu-
ments, tensions, and the like. That is how they
grow. The organization that recalls only good
times is not the one that had no bad times but
the one that has no use for bad news. It is ex-
actly the kind of organization in which whis-
tleblowing is most likely to be necessary. I
prefer an organization in which old battles are
recalled blow by blow and the general happiness
must be inferred from the fact that all parti-
cipants survived to work together again. 17

Having chosen the right organization, can an in-

dividual do anything more to reduce the chance
that he will someday have to blow the whistle?
Certainly. But he will have to think in strikingly
political terms.
 414/1183

First, he will want to develop his own informal

channels to augment formal channels. So, for
example, suppose a new engineer W officially
reports to A, but B carries more weight with
their common superior. W might want to get to
know B. Perhaps they share an interest in chess.
Once W makes friends with B, W is in position
to pass information around A should A try to
suppress it. A can hardly object to W playing
chess with B. Yet, once A knows W and B are
chess buddies, A will be less likely to suppress
information W wants passed up. A knows W has
a channel around him.
 Page 81

Second, employees should form alliances with

colleagues and subordinates, people who share
their responsibilities. No one should have to
stand alone against a superior. Whenever pos-
sible, the superior should have to respond to a
common recommendation. Managers are likely
to treat a group concern much more seriously
than a single individual's. Employees should try
to work through groups as much as possible.
 416/1183

Third, however, not any group will do. The

group should be sensitive to the moral concerns
likely to force an individual to blow the whistle.
The organizations most in need of whis-
tleblowers are also most likely to be so organ-
ized that employees become morally less sensit-
ive the longer they work for the organization. 18
So, an individual hoping to avoid whistleblow-
ing probably should cultivate the moral sensitiv-
ity of potential allies. There are many ways to
do this. The simplest is to bring in items from
the newspaper raising problems similar to those
the organization could face and pass them
around at lunch, asking how "we" could handle
them. If potential allies share the same profes-
sion, they might try getting the local profession-
al society to host discussions dealing with the
ethical problems that come up in work they
do.19
 417/1183

Fourth, but not least, an individual hoping to

avoid whistleblowing needs to cultivate his own
ability to present bad news in a way most likely
to get a favorable response. Part of doing this, of
course, is presenting the information clearly,
with enough technical detail and supporting
evidence. But there is more to it than that. Some
people become whistleblowers for lack of a pun-
gent phrase.20 A master of words is less likely
to have to blow the whistle than someone who,
though understanding a peril, has trouble com-
municating it.
 418/1183

That is not all. Presenting bad news in a way

likely to get a favorable response also includes
what used to be called rhetoric. A little sugar
helps the medicine go down. Is there a good side
to the bad news? If so, why not present that
first? If there is no good side, how about
presenting the bad news in a way likely to bring
out the personal stake the decision maker has in
responding favorably? Such tactics are usually
not mentioned in a discussion of whistleblow-
ing. Yet, it seems to me, many people end up as
whistleblowers because they did not pay enough
attention to the feelings of their audience.
 419/1183

Those who have substantial say in how an or-

ganization runs might, then, want to consider
some educational programs that our earlier dis-
cussion did not. In particular, they might want to
consider training employees in such political
skills as how to present bad news effectively
and how to maneuver it through channels. They
also might want to review their hiring practices.
For example, will the personnel office reject an
applicant who asks whether the company has an
open-door policy, treat such a question with in-
difference, or consider it as a plus? Any organiz-
ation that does not treat such questions as a plus
will not select for people with skills needed to
make whistleblowing unnecessary.

Concluding Remarks
 420/1183

The world can be a hard place. Individuals can

do everything in their power and still end up
having to choose between blowing the whistle
and sitting by while
 Page 82

innocent people suffer an injustice that can be

prevented. Whistleblowers are tragic characters.
Their decency pushes them to bring great suffer-
ing on themselves and those about whom they
care most. Their only alternative, sitting by,
would save from harm those about whom they
care mostbut at an incalculable cost (failing to
do what they have a duty to do). Their organiza-
tion will probably be better off in the long runif
it survives. But, in the short run, it too will
suffer.
 422/1183

When events leave only this choice, most of usat

least when we are not directly involvedwould
hope the individual on whom that choice is
forced will find the strength to blow the whistle.
Heroism is the best we can hope for then. But,
looking up from this chain of unhappy events,
we can see how much better off everyone would
have been had heroism been unnecessary. That
is why I focus on making whistleblowing
unnecessary.
 Page 83

PART III
PROTECTING ENGINEERING
JUDGMENT
 424/1183

Until recently, no texts concerned with the en-

gineering profession devoted significant space
to conflict of interest. Perhaps that was because
the topic struck most text writers as foreign to
engineering, as a subject belonging to business
ethics rather than to the ethics of a profession
concerned with shaping the material world. Yet,
conflicts of interest turn up in engineering quite
often. All engineers must exercise professional
judgment on behalf of a client or employer; in-
deed, there is little to engineering beside the ex-
ercise of engineering judgment. Insofar as con-
flicts of interest (those tugs, whether of desire or
duty, that tend to bias professional judgment)
undermine the reliability of engineering judg-
ment, conflicts of interest threaten the utility of
engineering. Unnoticed, conflicts of interest can
interfere in ways both subtle and gross with the
ability of engineers to do what they should.
 425/1183

The two chapters in this part approach the sub-

ject of conflict of interest in quite different
ways. Chapter 7 examines an important case of
conflict of interest, bringing out the relevant
principles and discussing their importance for
engineers as professionals and as moral agents.
The case is an absorbing cautionary tale. It is,
however, given pride of place for two other
reasons. First, we know more about it than we
know, or are likely to know, about most other
instances of conflict of interest in engineering.
Second, it reveals something of the role profes-
sional societies have in both the practice of en-
gineering and the control of technology. Along
the way, it offers occasions to think more about
the relation between professional ethics and or-
dinary morality.
 426/1183

Chapter 8 reaches many of the same issues from

another perspective, that of an occupation
"born" engineering (as we would say after
chapter 3). We are invited to consider what its
code of ethics should say about conflict of in-
terest, and, indeed
 Page 84

about whether it should remain a part of engin-

eering. We have an opportunity to think further
about the rationale for controlling conflict of in-
terest in engineering, about how much freedom
a profession has when writing a code of ethics,
about what considerations are relevant to writ-
ing a code, and about what the relation is
between the code and a profession.
 Page 85

7
Conflicts of Interest in Engineering
 429/1183

On May 17, 1982, the United States Supreme

Court upheld a civil judgment against the Amer-
ican Society of Mechanical Engineers (ASME)
for violating the Sherman Anti-Trust Act. 1
ASME v. Hydrolevel (Hydrolevel as the case is
commonly called) may be to engineering ethics
what Watergate was to legal ethics. Most of the
individuals involved were engineers, persons
who held high office in industry and in ASME.
Some may in fact have engaged in conduct they
knew to be unlawful. Certainly it was widely be-
lieved that they did. But the special interest of
Hydrolevel here is that there was something ser-
iously wrong with the way the principals con-
ducted themselves, supposing (as all claimed)
that each acted with the best motives and
without realizing that what he was doing was
wrong. Whatever else Hydrolevel is, it is a case
of conflict of interest in engineering. To under-
stand what the principals did wrong (even if
they acted with the best motives) is to
 430/1183

understand much about conflict of interest and

engineering.

Hydrolevel:
the Facts

On April 12, 1971, ASME received an inquiry

concerning a forty-three-word paragraph in its
eighteen-thousand-page "Boiler and Pressure
Vessel code." The code is one of about four
hundred model standards that ASME maintains.
While only advisory, these standards have a
powerful influence. Federal regulations have in-
corporated many of them by reference, as have
many cities, states, and Canadian provinces. Be-
cause of the influence and complexity of the
codes, it is often necessary to have them inter-
preted. ASME responds to at least ten thousand
requests for interpretation each year. Like the
codes themselves, these interpretations are only
advisory.2
431/1183
 Page 86

The inquiry concerned paragraph HG-605a,

which provides in part: "Each automatically
fired steam or vapor system boiler shall have an
automatic low-water fuel cutoff, so located as to
automatically cut off the fuel supply when the
surface of the water falls to the lowest visible
part of the water-gauge glass." 3 The purpose of
the paragraph is to prevent the "dry firing" that
can damage (or even cause an explosion of) a
boiler with too little water in it. The inquiry
came from McDonnell and Miller, Inc. of Ch-
icago (M&M), which had for decades domin-
ated the market for low-water fuel cutoffs. The
inquiry simply asked, "Is it permissible to incor-
porate a time-delay feature in the cutoff so that
it will operate after the boiler water level
reaches some point below the visible range of
the gauge glass?''4
 433/1183

The inquiry was signed by Eugene Mitchell,

M&M vice president for sales. Mitchell made
the inquiry because a competing firm, Hydro-
level Corporation of Farmington, New York
(Hydrolevel), had entered the low-water cutoff
market a few years before with a cutoff that in-
cluded a time delay and early in 1971 had won a
contract from the Brooklyn Gas Company, an
important M&M customer. If ordinary use of
Hydrolevel's time-delay cutoff was consistent
with ASME safety standards (and it was com-
monly believed to be consistent), M&M might
well lose its predominance in the market. If,
however, there was even a little doubt about the
safety of Hydrolevel's cutoff, M&M sales staff
could easily protect M&M's share of the market.
Mitchell knew that Hydrolevel's cutoff could be
installed safely. But he also thought the cutoff
could not be installed to cut off before the water
level fell below the visible range of the gauge
glass without being positioned much higher than
 434/1183

other cutoffs.5 The unusual position would itself

introduce unattractive complexity into
Hydrolevel's marketing (because installers
would put a new gauge exactly where the old
gauge was unless they were carefully instructed
to put it someplace else). If Mitchell could get
ASME to say that HG-605a meant that the water
level in the gauge could not drop from sight
without immediately triggering a fuel cutoff,
M&M sales staff could argue that the Hydro-
level cutoff would violate ASME standards if
positioned in the ordinary way. They might also
argue that it would violate ASME standards
wherever positioned. The same sixty-second
delay that could prevent unnecessary cutoffs
could, it seemed to Mitchell, also allow a hot
and suddenly almost waterless boiler to crack or
explode.
 435/1183

Mitchell discussed this sales strategy several

times with John W. James, M&M's vice presid-
ent for research. James had been a member of
the ASME subcommittee responsible for heating
boilers (the "Heating Boiler Subcommittee of
the Boiler and Pressure Vessel Committee")
since 1950 and also had a leading part a few
years before in rewriting the code of which
HG-605a was part. James suggested a meeting
with T.R. Hardin, chair of the Heating Boiler
Subcommittee. The meeting occurred in late
March 1971. Hardin (in town for other business)
came by the M&M office and the three (along
with M&M's president) went to dinner. During
dinner, Mitchell asked Hardin about HG-605a.
Hardin answered that he believed it meant what
it said: The water level should not drop from
sight without triggering the cutoff immediately.
Soon after that meeting James drafted a letter of
inquiry to ASME, sending a copy to Hardin,
 436/1183

who made some suggestions that were incorpor-

ated into the final draft.
Page 87
 438/1183

The inquiry was addressed to W. Bradford Hoyt,

secretary of the Boiler and Pressure Vessel
Committee. Hoyt treated it as a routine inquiry,
directing it to the appropriate subcommittee's
chair, T.R. Hardin. Hardin then prepared a re-
sponse without referring his action to the whole
subcommittee for approval. He was entitled to
do this provided the response was treated as an
"unofficial communication." Hoyt signed the
unofficial communication that Hardin drafted
and sent it out on ASME stationery. That letter,
dated April 29, 1971, advised that a low water
cutoff must "operate immediately" when the wa-
ter level falls below the lowest visible point of
the gauge glass and that a cutoff with a time
delay gave "no positive assurance that the boiler
water level would not fall to a dangerous point
during a time delay period.'' 6 Although the re-
sponse did not say that Hydrolevel's time delay
was dangerous, that was a plausible inference.
M&M used the ASME letter to discourage
 439/1183

potential customers from buying Hydrolevel's

cutoff. The strategy seemed to work.

Hydrolevel learned of the ASME letter early in

1972 through a former customer and immedi-
ately requested a copy from ASME. This was
duly sent on February 8, 1972, the name of the
inquirer (Mitchell) being omitted as ASME
policy required (to preserve confidentiality).
 440/1183

Hydrolevel was, of course, not happy with the

interpretation. On March 23, Hydrolevel wrote
Hoyt a nine-page letter explaining why ASME
should change its ruling. Hoyt sent Hydrolevel's
request to the Heating Boiler Subcommittee. On
May 4, the subcommittee voted to confirm the
intent of the original response. James, who by
then had replaced Hardin as chair of the sub-
committee, abstained from participation in the
subcommittee deliberations on that question but
reported the vote to the Boiler and Pressure Ves-
sel Committee. The full committee voted to
send Hydrolevel an "official communication."
Dated June 9, 1972, it "confirmed the intent" of
the letter of April 29, 1971, but advised Hydro-
level that, although cutoffs with time delay were
not expressly forbidden, they had to be posi-
tioned to cut off before the water level fell from
sight.7 While James did not participate when his
committee decided how to respond to Hydro-
level, he did (at the drafting committee's
 441/1183

request) help to draft a critical sentence of that

response.8

Hydrolevel found the response insufficient to

permit it to compete successfully with M&M.
There still seemed to be doubt about the safety
of Hydrolevel's low-water cutoff.

That is where events stood for two years. Then,

on July 9, 1974, the Wall Street Journal pub-
lished an article describing Hydrolevel's diffi-
culty trying to sell a fuel cutoff many in the in-
dustry thought to be in violation of ASME's
code. The article suggested "close ties between a
dominant company in an industry and the pro-
fessional society that serves as its watchdog."
The only close tie the article noted was that
James, an M&M vice president, was vice chair
of the appropriate ASME subcommittee when
M&M made its original inquiry and chief drafter
of the code involved.9
 442/1183

The article produced an uproar within ASME.

For example, the vice president of ASME's Re-
gion 11 wrote: "If the facts are as stated in the
article, it would seem that Mr. James should not
only be relieved of his duties on the Board of
Codes Com-
mittee but he should also be kicked out of ASME fo
ASME's Professional Practices Committee then inv
proper or unethical in James' conduct, and commen
in a forth-right manner as chair of his subcommittee
Committee did not have all the facts. James did not
meeting in Chicago with Hardin, of his (or Hardin's
quiry, or of his part in drafting the June 9 response
came out until March 1975 during hearings before t
Antitrust and Monopoly.11 Hydrolevel filed suit a f
charging M&M, ASME, and Hardin's employer, Ha
surance Company, with unlawful restraint of trade.
Names to Remember
Hardin Chair of Heating Boiler Subcommittee (of ASME
and vice president of Hartford Boiler Inspection an
Hoyt Secretary of ASME Boiler and Pressure Vessel Co
for that committee and its subcommittees.
HydrolevelHydrolevel Corporation of Farmington, New York
out of business, the plaintiff in Hydrolevel, the leg
James M&M vice president for research, a drafter of rele
sure Vessel code, vice chair of Heating Boiler Sub
and chair of that subcommittee after Hardin retired
 444/1183

M&M McDonnell and Miller, Inc. of Chicago, makers of

market before the entry of Hydrolevel's time-delay
Mitchell M&M vice president for sales.

What Did they do Wrong?

Assuming that Hardin and James acted from honest

was morally wrong with what they did? There are a
to answer that question.

One way would point to the consequences of what

ample, they might have driven Hydrolevel out of bu
provement in boiler safety. Let us call this way of e
wrong consequentialist.

A second way to explain what makes an act wrong

of some social rule, for example, to the violation of
law. We might call this way of explaining what ma
ism (because it makes the act's moral rightness or w
how social roles happen to be defined).
 445/1183

A third way to explain what makes an act wrong is

act itself (given the context in question) that makes
its actual or probable consequences and whether an
 Page 89

example, an act might be morally wrong simply

because it is an instance of lying or a betrayal of
trust. This way of answering our question is
sometimes called "moral absolutism" because
the answer is not relative to this or that social
rule. But it is probably less misleading to call it
duty-based (or "deontological") because it relies
on considerations of duty directly (even though
the duties may themselves be defended in part at
least by appeal to the overall consequences of
having such duties). These duties are sometimes
called natural (or "absolute") to distinguish them
from the "conventional'' (or "relative") duties
imposed by law or other merely social rules. 13

Which of these three ways of explaining why an

act is wrong should we employ to determine
what, if anything, Hardin and James did wrong?
Let us consider them one at a time.

Consequences
 447/1183

What Hardin and James did certainly had con-

sequences. For example, M&M printed ASME's
April 29 response in a booklet entitled "The Op-
positionWho They Are, How to Beat Them."
The booklet, distributed to sales staff late in
1971, included a message from Mitchell de-
scribing Hydrolevel's time-delay cutoff and stat-
ing that such a device "would defeat the intent
of the ASME code and this should definitely be
brought to the attention of anyone considering a
device which included a time delay in the low
water cutoff circuitry."14 The ASME letter gave
legitimacy to Mitchell's opinion and seems to
have had much to do with driving Hydrolevel
out of business.15
 448/1183

So, on the one hand, what Hardin and James did

had some bad consequences. Their acts helped
to drive Hydrolevel out of business, and that
was bad for Hydrolevel. On the other hand,
what they did helped M&M keep its share of the
cutoff market and that was certainly good for
M&M. But the evaluation of consequences can-
not end with that. The consequences of what
Hardin and James did went on. Hydrolevel sued.
M&M settled out of court for $750,000. Hart-
ford also settled. It paid Hydrolevel $75,000.
ASME went to trial and lost. The judgment
against ASME$7,500,000was equal to three-
fourths of its annual budget. That litigation was
bad for M&M, Hartford, and ASME but good
for Hydrolevel. ASME appealed, lost on the de-
cision but won rehearing on the damages. The
case was settled when ASME agreed to
$4,750,000 in damages.16 In the end, Hydro-
level (or rather its owners) may have gained
more than it lost (ignoring substantial attorney's
 449/1183

fees). Its winnings in court amounted to far

more than its profits over a decade. But how
should we balance all these good and bad con-
sequences to decide whether what Hardin and
James did was right? Do they merely balance
out (because one side's loss seems to be
another's gain)? Or do they total up to a bad out-
come overall? Or to a good one? Consequential-
ism requires some method of balancing con-
sequences one against another to reach an over-
all evaluation. What method should we use?

Might other consequences let us get around this

problem of method? Certainly. For example,
driving Hydrolevel out of business might have
suppressed a new boiler cutoff which, if widely
used, would have reduced boiler explosions and
otherwise improved boiler operation (for ex-
ample, by reducing unnecessary boiler shutoffs).
 Page 90

Driving Hydrolevel out of business would then

have had consequences so bad that few people
would think they could be outweighed by any
advantage to M&M or Hydrolevel.

Was Hydrolevel's cutoff that much better than

M&M's? If that is the sort of question we must
answer before we can say what was wrong with
what Hardin and James did, we cannot say what
was wrong with what they did. We do not have
the answer and we are not likely to get it. Labor-
atory tests are only suggestive and we are not
likely to have a good "field test" now.
 451/1183

So, any decision we make based on the con-

sequences of Hydrolevel's demise must rely
either on educated opinion or on something less
reliable. Educated opinion is the judgment of
those whose experience and learning make them
relatively reliable guides in answering questions
of the sort posed. Educated opinion seems to be
divided. Both Hydrolevel and ASME had out-
side experts at trial, some testifying to the su-
periority of Hydrolevel's cutoff, others testifying
to the possible dangers of its use. 17 When out-
side experts disagree, we naturally turn to some
body of experts capable of sorting out the opin-
ions of individuals and arriving in that way at an
authoritative consensus. Because our concern
here is boiler safety, the natural place for us to
turn for such an authoritative statement on boiler
safety would, of course, be ASME's boiler code
and the committee with authority to interpret it.
Unfortunately, that code and committee are part
of our problem.
 452/1183

But there is one more alternative to consider.

The market itself is a possible source of the in-
formation we need. Victory in a free market is
good evidence that the victorious product is bet-
ter than its competitors (and victory in a perfect
market, decisive evidence). M&M's victory over
Hydrolevel is, however, not good evidence that
M&M's cutoff was safer than Hydrolevel's.
There are two reasons for this. First, the market
measures overall value, not safety, as a distinct
factor. One product might be less safe than an-
other but sell better for other reasons, for ex-
ample, because its low cost more than pays for
the higher insurance premiums its use entails.
Second, Hydrolevel in effect charged that M&M
used unfair means to win its victory. If M&M
did rig the market, the market cannot tell us
even whether M&M's cutoff was better than
Hydrolevel's overall. The verdict of this market
may not have been the verdict of a free market.
Did M&M rig the market? We don't know. If
 453/1183

ASME's letter of April 29 was a sensible read-

ing of the code and the code itself was correct,
then M&M's use of that letter to discredit Hy-
drolevel would not have been unfair (unless
there was something wrong with the way M&M
obtained the letter). All else equal, information
should not distort the market.18

So, it seems, we cannot make a reliable judg-

ment that Hydrolevel's demise served or dis-
served the public good. If we cannot do that, we
cannot make a reliable judgment that overall,
the consequences of ASME's response of April
29, 1971, and June 9, 1972, were bad (or good).
Without such a judgment, we cannot provide an
appealing consequentialist explanation of what
was wrong with what Hardin and James did. So,
if we are to explain what was wrong with what
Hardin and James did, we must, it seems, do it
by showing that they violated a social rule or
that they failed to act in accordance with some
natural duty.
454/1183
 Page 91

Social Rules and Individual Conscience

"Rule" (as used here) includes standards of con-

duct evident from practice (that is, so-called un-
written rules) as well as those standards ex-
pressly adopted. A "social rule" is a rule that
may vary from one society to another just be-
cause one society chooses one rule while anoth-
er society chooses another rule. (I reserve the
term "moral rule'' for rules that do not vary in
this way.) We can distinguish at least three sorts
of social rules Hardin or James might have viol-
ated: (1) ASME rules, (2) rules governing all en-
gineers (in the United States), or (3) federal law.
Let us consider each possibility one at a time.

ASME Rules
 456/1183

Did Hardin or James violate any ASME rule?

Hardin did meet with M&M executives to dis-
cuss a question likely to come before him as
chair of the Heating Boiler Sub-committee.
Indeed, he expressed his opinion on the question
and helped draft an inquiry to obtain that opin-
ion from ASME. The Senate subcommittee in-
vestigating the Hydrolevel case found that ob-
jectionable. But ASME officials did not. For ex-
ample, Melvin R. Green, managing director of
the Research Codes and Standards Section of
ASME, defended what Hardin did in this way:
 457/1183

I think you must recognize that you are trying to

get words in a letter, so that you clarify a provi-
sion in the code. And to get the proper words, I
do not really see that there was anything wrong
with that, because I, when I was secretary of the
Boiler and Pressure Vessel Committee, had
people who would telephone inquiries to me and
they would sayI would give an answer on the
telephone and then they would say, "Well, how
can I get this in writing?" I would suggest the
wording for the inquiry, so that they could get
the response to clarify that particular part of the
code. 19

So, Hardin's meeting with M&M executives and

his help in drafting the inquiry to his own com-
mittee may not have violated an express or tacit
ASME rule. Indeed, they seem to fit in with
ASME's ordinary procedures.
 458/1183

Hardin's other acts also seem to be consistent

with ASME rules. As noted earlier, Hardin was
entitled to answer M&M's inquiry if he thought
it sufficiently routine (provided his response
was treated as an "unofficial communication").
He apparently thought the inquiry was suffi-
ciently routine, a case of the code meaning just
what it said.20 And it is not clear even now that
he was wrong to think so. The Heating Boiler
Subcommittee and the entire Boiler and Pres-
sure Vessel Committee later "confirmed the in-
tent" of his original response. Hardin treated the
response as an unofficial communication just as
ASME rules required.
 459/1183

Hardin did not, it is true, reveal to the Profes-

sional Practices Committee his meeting with
M&M executives or his part in drafting the in-
quiry. We do not know why he did not. The
most favorable explanation is that he was not
asked directly and did not see why he should
raise the matter himself. If ordinary ASME pro-
cedure was as Green indicates, the Professional
Practices Committee would not have cared what
part Hardin had in drafting the original inquiry
(and so, they probably would not have found
that revelation worth the trouble of a hearing).
Indeed, Green told
 Page 92

the Senate subcommittee that he considered

Hardin's conduct perfectly ethical, even taking
into account what the subcommittee uncovered.
21

ASME rules seem to be equally kind to what

James did. Green defended James's self-effacing
part in drafting the April 12 inquiry in this way:
Well, here again, I think you must understand the
voluntary standard system. Many people, who
serve a great deal of their time in a code activity,
try to identify themselves with the code activity.
And there is another part [of their life] where
they would be in the Government or employed
by industry. If they have [to make] an inquiry
from the Government agency or that company,
they will have an associate within the [agency or]
company, who will sign the inquiry and send it to
us. . . . [It] is just a matter of trying to keep their
house in order.22
 461/1183

In other words, having someone else sign the

letter was James's way of keeping his work on
the Heating Boiler Subcommittee distinct from
his work at M&M. That he wrote the inquiry
should be irrelevant to the response it received.
If signed by Mitchell, the inquiry would not
have the authority that might come from having
the name of the subcommittee's vice chair sub-
scribed to it. Should James have signed the in-
quiry? Nothing in ASME rules required him to.
Indeed, Green makes it sound as if ASME prac-
tice would have condemned James for signing
the inquiry had he done so. James would not
have "kept his house in order." So, even accord-
ing to ASME practice, there was no reason for
James to inform ASME's Professional Practices
Committee of his part in drafting the original
letter.
 462/1183

What about James's part in drafting the official

ASME communication of June 9, 1972? Again,
James seems to have done as ASME rules al-
lowed. A subcommittee chair would not nor-
mally step down even though he helped to draft
the inquiry the response to which was under re-
view (and, it seems, even though he worked for
a company that had an interest in the outcome).
According to Hoyt, James stepped down only
because Hydrolevel's letter to ASME com-
plained that ASME seemed to be out to destroy
a new product. That James should help draft the
June 9 response to Hydrolevel was, Hoyt said,
"perfectly normal because the chairman is in the
best position, on the basis of experience, to
know what the intent of his subcommittee is."23
James was "merely trying to be helpful in select-
ing words that would be appropriate to clarify
the subject."24 Green concurred. As far as he
was concerned, there was nothing in what James
did contrary to ASME practice.25
 463/1183

NSPE Code of Ethics

That Hardin and James did not violate any

ASME rule does not mean that they did what
was professionally proper. As engineers, their
conduct is also subject to evaluation under the
code of Ethics of the National Society of Profes-
sional Engineersat least insofar as the code itself
expresses a well-articulated standard of conduct
for engineers. (ASME, like many other engin-
eering societies, also has a code of ethics of its
own, basically, the ABET code, but at the time
that code also lacked a general provision con-
cerning conflict of interest and its interpretation
of the provisions it had did not differ from the
NSPE's.) A cursory reading of the NSPE code
may indicate that Hardin or James violated sev-
eral provisions.26 But a closer reading
 Page 93

makes everything much less clear. Let us con-

sider those potentially relevant provisions one at
a time beginning with the most specific.

"Faithful Agent."
 465/1183

The Senate subcommittee investigating the Hy-

drolevel complaint suggested that Hardin and
James each did something wrong because each
had a conflict of interest that should have
stopped him from doing what he did. A federal
court of appeals made the same point. 27 Section
III.5 of the NSPE code of ethics specifically dis-
cusses "conflicting interests." "Engineers shall
not," it says, "be influenced in their professional
duties by conflicting interests." This seems clear
enough, but the only examples the code gives of
such conflicting interests are (1) accepting "fin-
ancial or other considerations, including free en-
gineering designs, from material or equipment
suppliers for specifying their products," and (2)
accepting ''commissions or allowances, directly
or indirectly, from contractors or other parties
dealing with clients or employers of the Engin-
eer in connection with work for which the
Engineer is responsible." So, if this is all NSPE
means by conflicting interests, neither Hardin
 466/1183

nor James had a conflict of interest. They accep-

ted no consideration from material or equipment
suppliers for specifying their product. They also
accepted no commission in connection with
work for which they were responsible. Is there
any reason to limit the term "conflicting in-
terest" (or "conflict of interest") to cases like
those expressly listed in section III.5? Well, that
depends on considerations beyond the mere let-
ter of this section, doesn't it? We must look
further.
 467/1183

Section II.4 also deals with matters most people

would think involve conflicts of interest.
"Engineers shall," it says, "act in professional
matters for each employer or client as faithful
agents or trustees." Would a "faithful" agent or
trustee allow himself to act as Hardin or James
did? There are really three questions here. First,
can one be a faithful agent or trustee and yet
have a conflict of interest of the sort Hardin or
James had? (Of course, we haven't yet con-
cluded that Hardin or James had any conflict of
interest as the code of ethics defines that term.)
Second, how would a faithful agent or trustee
act if he had such a conflict? And, third, did
Hardin or James act differently?
 468/1183

Section II.4(a) provides at least a partial answer

to the first two of these three questions. "Engin-
eers shall," it says, "disclose all known or poten-
tial conflicts of interest to their employers or cli-
ents by promptly informing them of any busi-
ness association, interest, or other circumstance
which would influence or appear to influence
their judgment or the quality of their services."
Section II.4(a) thus understands the term "con-
flict of interest" to include more than the two
examples of conflicting interests mentioned un-
der section III.5. A conflict of interest can, it
seems, be any business association, interest, or
other circumstance that could influence or even
just appear to influence an engineer's judgment
or the quality of his service. Both Hardin and
James had conflicts of interest in this sense.
That Hardin gave his opinion on the cutoff in-
quiry informally (and perhaps without due con-
sideration) could reasonably be supposed to
have influenced his judgment when he later
 469/1183

actually undertook to respond more formally

(and, because that can reasonably be supposed,
there is at least an appearance of such influ-
ence). Similarly, James could not be certain that
his contribution to the ASME letter of June 9,
1972, was
 Page 94

not influenced in part by how the exact wording

seemed to affect his company's prosperity. 28

Section II.4(a) does not, however, rule out con-

flicts of interest. All it requires is that a faithful
agent or trustee disclose any conflict he has to
his employer or client. A faithful agent or trust-
ee may have a conflict of interest and still be
faithful, but only if he discloses the conflict to
his employer or client.
 471/1183

What then can we conclude from the rules?

Neither Hardin nor James seems to have con-
cealed any conflict of interest from their respect-
ive employer (M&M and Hartford). Both, it is
true, concealed conflicts from ASME (or, at
least, failed to disclose them to anyone at
ASME). So, section II.4(a)and section II.4 it-
selfwould condemn both Hardin and James if,
but only if, serving as a volunteer on an ASME
committee constituted "acting in professional
matters" and ASME was their "employer" or
''client." Is it proper to interpret "professional
matters" to include working as an unpaid volun-
teer for an engineering society? (Must not a pro-
fessional be paid if he is to act as a profession-
al?) Is it proper to interpret "employer" or (more
likely) "client" to include ASME?
 472/1183

The code provides little help with these ques-

tions. Among the other examples of being a
"faithful agent and trustee" listed under section
II.4 only two seem worth noting. Section II.4(d)
provides that "engineers in public service as
members, advisors, or employees of a govern-
mental body or department shall not participate
in decisions with respect to services solicited or
provided by them or their organizations in
private or public engineering practice." Section
II.4(e) adds that "engineers shall not solicit or
accept a professional contract from a govern-
mental body on which a principal officer of their
organization serves as member." There is noth-
ing under section II.4 about professional societ-
ies or other nongovernmental bodies.
 473/1183

Only one conclusion of interest to us seems to

follow from these two examples of being a
faithful agent or trustee. There may be enough
of a distinction between "employer" and "client"
so that ASME could reasonably be thought
Hardin's or James's "client" (even though ASME
could not be the employer of either in any but
the most strained sense). Though section III.4 it-
self only refers to "employer or client," (d)
refers to "service" as a "member" of a govern-
mental body or as an "advisor" (as well as of
"service" as a governmental "employee"). So, if
an engineer is a member of a governmental
body (or even an unpaid advisor of one), the
body might be his "client" (in the appropriate
sense) even though he is not being employedthat
is, paidas an engineer. On the other hand, the
section can also be read so that "client" is just
another word for "employer" or at most for
someone whom an engineer is paid to serve in
 474/1183

some professional capacity (for example, when

serving the customer of his employer).

Section III.4(e) does not clarify which interpret-

ation is intended. According to (e), an engineer
does something wrong if he solicits or accepts a
professional contract from a governmental body
on which he serves. But the section does not tell
us why that would be wrong. There are at least
two reasons why soliciting or accepting such
contracts might be wrong. The reasons point to
different interpretations of "client." One reason
soliciting or accepting such contracts might be
wrong is that the engineer would fail to be a
faithful agent or trustee of the government in
question. He would have taken advantage of
someone, the government, that is already his cli-
ent (because
 Page 95

he is serving on one of its bodies). The other

reason soliciting or accepting such contracts on
behalf of a client might be wrong is that an en-
gineer cannot be a faithful agent or trustee of a
private client he is working for if he risks get-
ting that client into trouble by obtaining govern-
mental contracts for the client through misuse of
his public trust. Section III.4(e) does not help us
to understand whether just any service on a gov-
ernmental bodyand so, by analogy, on an ASME
committeeis acting in a "professional matter" (so
that the duty to act as a faithful agent or trustee
applies at all). For example, is helping one's
committee draft a letter a case of acting in a pro-
fessional matter?

Miscellaneous Provisions.
 476/1183

Three other seemingly promising provisions of

the code turn out to be even less helpful. Section
II.3(c) provides that "engineers shall issue no
statements, criticisms or arguments on technical
matters which are inspired or paid for by inter-
ested parties, unless they have prefaced their
comments by explicitly identifying the inter-
ested parties on whose behalf they are speaking,
and by revealing the existence of any interest
the engineers may have in the matters." This
section would certainly condemn what Hardin
and James did if, for example, Hardin responded
as he did in part because he hoped to benefit his
employer (or because he hoped to postpone his
own overdue retirement from M&M by proving
himself especially useful). The statements
Hardin and James made would then have been
"inspired" (if not exactly "paid for") by inter-
ested parties they had not explicitly identified.
We are, however, assuming that both Hardin and
James acted from the best motives, that is, that
 477/1183

the acts in question were not paid for by an in-

terested party or inspired by anything but con-
cern for the public safety and welfare. So, sec-
tion II.3(c) cannot help us decide what was
wrong with what Hardin or James did.
 478/1183

Section III.1 may, in contrast, seem likely to be

more helpful just because it is more general.
"Engineers shall," it says, "be guided in all their
professional relations by the highest standards
of integrity." Section III.1(f) gives as an ex-
ample of being so guided that "engineers shall
avoid any act tending to promote their own in-
terests at the expense of the dignity and integrity
of the profession." These two sections seem
promising because, if Hardin and James improp-
erly misled others about their intentions, their
interest in M&M's inquiry, or their part in
ASME's response, they could not have been
guided by the "highest standards of integrity.'' If,
in addition, they did all that to endear them-
selves to their respective employers, they would
also have promoted their own interests at the ex-
pense of ASME and so at the expense of the
profession as a whole. If, however, what Hardin
and James did was not "improper," their acts
could still be consistent with the "highest
 479/1183

standards of integrity" and so not something the

code of ethics condemns. So, were their acts
proper or improper? That depends on what the
"highest standards of integrity" are. We must
look elsewhere in the code for guidance con-
cerning that.

Section III.3(a) may seem to provide such guid-

ance. This section provides that "engineers shall
avoid the use of statements containing a material
misrepresentation of fact or omitting a material
fact necessary to keep statements from being
misleading." Hardin and James both omitted
statements of fact necessary to keep others (for
example, the Professional Practices Committee)
from drawing false conclusions
Page 96
 481/1183

(for example, that Hardin had no part in drafting

the original inquiry or that James had no part in
drafting the response of June 9, 1972). But were
these facts "material"; that is, were they facts
that should have been revealed to keep others
from drawing conclusions they had a right to be
protected against (for example, the conclusion
that Hardin or James had acted properly when in
fact they had not)? Well, that depends on what
the ultimate conclusion should have been,
doesn't it? If, for example, Hardin and James
would have been judged to have acted just as
properly had all the facts they failed to reveal
been revealed, would we consider those unre-
vealed facts "material"? It seems not. So, it
seems we cannot know that their conduct fell
below ''the highest standards of integrity" until
we know whether what they did was proper. Ap-
peal to the NSPE code thus seems to lead us to a
dead end just as appeal to actual or probable
consequences did.
 482/1183

But that is not quite true. We have one turn yet

to take. Written rules are seldom self-interpret-
ing. We must bring to the "letter" of a rule an
understanding of the "spirit"that is, the underly-
ing purposes, policies, and principles that
provide a context we can use to understand what
the rule is supposed to do. For example, to un-
derstand what is ruled out by a general prohibi-
tion of "conflicting interests," we need to know
what the community that prohibited such in-
terests means by the term. We also need to
know what reasons it had for such a prohibition
and how the rule would have to be interpreted to
do what the community wants done.

BER as Authority.
 483/1183

Where then should we go for help with inter-

preting the NSPE code of ethics? One place is
NSPE's Board of Ethical Review (BER). That
brings us to the most difficult problem generated
by trying to provide a relativist explanation of
what makes an act wrong. What if we do not
find the BER's interpretation convincing? What
if we think engineers should not do as the BER
says they should? Are we necessarily wrong
either about what the code says or about what
engineers should do? The relativist answer is
plain. If what makes an act right is that it is per-
mitted or required by the appropriate social rule,
and if the appropriate social rule means what
those with authority to interpret it say it means,
then of course, we must be wrong if we disagree
with the BER's interpretation of the code. 29 The
BER has authority to interpret the code (because
the NSPE gave the BER that authority). We are
disagreeing with those who speak for the society
in question.
 484/1183

So, here is the problem with moral relativism.

Certainly it seems that the BER (or even the
NSPE as a whole) can be wrong. The BER
might, for example, issue an opinion interpret-
ing the code in a certain way. Everyone might
agree that this is in fact how the code should be
interpreted if, say, "read strictly." Yet a majority
of the society might think that the code should
not be read strictly in cases of this sort. The
BER itself might eventually change its mind
about how the code should be interpreted (or un-
dergo a change of membership leading to a
changed interpretation). And even if the BER
did not change its interpretation, the NSPE
might itself change the code to prevent such
"strict" interpretation. Now, if the BER (or
NSPE) can be wrong about what engineers
should do, there must be a standard of what en-
gineers should do beyond what the BER (or
NSPE) says, some standard of right action for
engineers beyond what this or that engineering
 485/1183

society happens to say. What might that stand-

ard be?
 Page 97

Conscience as Authority.

One answer often given is "individual con-

science." We have, it is said, an inborn sense of
right and wrong. We need only be "true to
ourselves" to do right. We must do what we
''feel" to be right whatever anyone else thinks.
That is all we can require of ourselves and all
we should require of one another. The right act
is, it is said, simply that act the individual feels
to be right.

Though this appeal to individual conscience

may appear the very opposite of appeal to social
rules, it really is similar. If the appeal to social
rules for a standard of rightness can accurately
be described as group-centered moral relativism,
this appeal to individual conscience might, with
equal accuracy, be called individual-centered
moral relativism.
 487/1183

Individual-centered moral relativism (or "sub-

jectivism") is not without attractions. We all re-
cognize that individuals are beings not to be op-
erated entirely from the outside. Each must do
what she chooses, and each should choose by
her own standards. What right have we to ask a
person to do other than she thinks right? We of-
ten go out of our way to respect each other's
moral integrity. We sometimes let others do
what we think wrong because each "has a right"
to act on her own conception of the good. We
sometimes even excuse wrong acts because the
person who did them "meant well." Neverthe-
less, there are at least two reasons to reject indi-
vidual conscience as the ultimate standard of
moral right and wrong.
 488/1183

One reason for rejecting individual-centered

moral relativism is that such relativism makes it
impossible for an individual to do wrong as long
as he feels that what he is doing is right. The
distinction between an act appearing right (to
the actor) and its being right dissolves if the ulti-
mate standard of right and wrong becomes how
the act appears to the actor, how he "feels" about
it. That someone feels no horror at the prospect
of committing murder, no remorse or regret af-
terward, would (according to individual-
centered relativism) be enough to show that he
did not do wrong. A person's moral insensitivity
would be a guarantee of the propriety of what he
did. That certainly seems inconsistent with our
understanding of moral right and wrong.
 489/1183

The other reason for rejecting individual-

centered relativism is related to this first one.
We began this chapter by assuming that all the
engineers involved in Hydrolevel acted from
honest motives, that all of them felt that what
they were doing was right. There is no evidence
that any of them had a pang of conscience be-
forehand or experienced any remorse afterward.
If we accepted individual-centered moral re-
lativism, we would have to agree that Melvin
Green's concluding remarks to the Senate sub-
committee constituted the last word on the pro-
fessional propriety of what Hardin and James
did. "Every professional works by a canon of
ethics," he explained, "and I think it is up to the
professional who is serving in that position at
that time to make this kind of judgment." 30
Hardin and James made their judgment and (ac-
cording to Green) that is all we can require of
them.
 490/1183

Individual-centered relativism thus cuts off eth-

ical discussion as soon as it begins. As long as
Hardin and James acted in a way they judged
best, there is nothing to criticize in what they
did. Indeed, even if they asked in advance what
to do, the best advice anyone could give them
would be to do what they felt proper, whatever
that might be. Telling them any more would be
telling them what we should do were we in their
place, not what they should do. Individual-
centered moral relativism makes
 Page 98

most reasoning about moral right and wrong a

lonely and pointless activity. The work of the
BERindeed, the work of all those who advise
others what to docould be helpful only insofar
as it helps the individual to reach some judg-
ment, whatever it might be. One might as well
throw dice as ask the BER. That also seems in-
consistent with our understanding of right and
wrong.

Laws
 492/1183

It may seem that these problems of group-

centered moral relativism could be resolved
simply by appealing from the rules of the NSPE
to those of some more inclusive society, for ex-
ample, the laws of the United States. But that is
not so. All the problems simply follow along.
The law in question here, the Sherman Anti-
Trust Act, prohibits "[unreasonable] restraint of
trade." "Unreasonable" is a word that leaves
plenty of room for the interpretive problems we
already encountered in the NSPE code of ethics.
Of course, courts do have authority to interpret
laws (just as the BER has authority to interpret
the code). But, though they have such authority,
their interpretation is not necessarily right (no
more so than the BER's is). Not only do courts
sometimes change their mind and "overturn"
precedent, they may also find the rules they laid
down repudiated by the legislature. There is
nothing unreasonable about telling a court that it
made a mistake and should decide differently
 493/1183

next time. Nor is there anything unreasonable

about telling Congress that it was wrong to pass
a certain law. There appear to be standards of
right and wrong independent of the particular
rules of this or that society, even if the society is
a whole nation. If we are to explain what if any-
thing Hardin and James did wrong, we must
eventually appeal to such an independent
standard.

Natural Standards
 494/1183

What standard of right and wrong could there be

beside social rules? The traditional answer is
"rules of reason" (or "natural laws"). What is a
rule of reason? For our purposes, the following
rough definition will do: A rule of reason is a
statement of how one should act that all rational
persons support, advocate, endorse, or recognize
as somehow binding (or, at least, would recog-
nize as binding if they were to consider the
statement in a certain way, for example, impar-
tially, in a "cool hour," or at their rational best).
There are many such rules. The rules of arith-
metic, for example, are rules of reason (as
defined here). They state standards every ration-
al person recognizes (or, at least, would recog-
nize if she gave them much thought) as the way
to add, subtract, multiply, and divide if she
wants to get answers other rational persons can
accept as accurate. Rules of prudence, though
different from rules of arithmetic, are also rules
of reason. Prudence is choosing actions most
 495/1183

likely to serve our overall long-term interests.

All rational persons recognize their own in-
terests as relevant to determining what to do
(relevant but, of course, not necessarily decis-
ive). 31

Rational persons support, advocate, endorse, or

recognize rules of reason only because, and only
insofar as, there is good reason for so doing.
(Acting for good reason is a large part of what it
means to be rational.) Thus, another way to un-
derstand what a rule of reason is is to understand
it as a rule that, all things considered,
 Page 99

is better supported by good reasons than any al-

ternative. Rational persons support, advocate,
endorse, or recognize certain rules as rules of
reason at least in part because the weight of
evidence and argument support treating them
(rather than any alternative) as binding.
 497/1183

Among rules of reason, the most important for

our purposes are moral rules. What is a "moral
rule?" A moral rule, let us say, is any rule in-
structing rational persons how to act, which
each rational person would want all others to
follow even if their following it meant that he
would have to follow it too. 32 Moral rules do
not necessarily state what people in fact do (ex-
cept insofar as they are good people). Moral
rules tell us only what rational persons have
good reason to want each other to do, what it
would be in a rational person's overall interest to
have others do whether or not he followed the
rules himself. Unlike the rules of arithmetic or
prudence, moral rules presuppose that rational
persons are able to help or harm one another if
they choose. Moral rules lay down requirements
for the treatment of others, acts due others as
persons, our "natural" duties.33
 498/1183

We must, however, be careful to distinguish

between the reasons for supporting, advocating,
endorsing, or recognizing moral rules in general
(their justification) and what may lead us as in-
dividuals to follow or ignore any particular rule
(our reason or motive for acting as we do).
What justifies moral rules is that having them is
in everyone's interest. But people may in fact do
what morality requires (when they do) for any
number of reasons. Some may act as morality
requires because they were brought up to do so
and doing wrong has no appeal. They act mor-
ally because they are of morally good character.
Some may act morally because they wish others
well. Such persons act morally because they
possess the special virtue of altruism or benevol-
ence. Others may do what morality requires be-
cause, though tempted to do wrong, they try to
do what they believe right (and succeed). Such
persons act morally to preserve their moral in-
tegrity. Others may do what morality requires
 499/1183

because they fear criticism, prison, or divine

wrath. Such persons act morally because they
are prudent.

Most people probably act morally from a com-

bination of these or other motives. As long as
they do what is required (with the appropriate
intention), what they do is right and their motive
will be relevant only in assessing their character
or moral worth. If, however, they do something
wrong, their motive may be relevant in another
way. "He meant well" cannot justify an act (that
is, show it to have been right), but it may
provide a reason for not blaming someone as
much as would otherwise be appropriate. For
example, the man who steals bread to feed his
family is still a thief, but he does not deserve as
much blame or punishment as the man who
steals the same amount to gamble or because he
enjoys the thrill of crime. The man who steals
bread to feed his family clearly means well in a
way the gambler or thrill seeker does not.
 500/1183

Moral rules are, in one sense, absolutethat is,

they take precedence over any consideration
conflicting with them. But they take precedence
only in a sense. They do not take precedence in
the sense that they in fact always win out in the
deliberations of a rational person. They may not.
Winning out in the deliberations of even the
most rational person involves considerations
other than those that justify moral rules. (I may,
for example, benefit from breaking my promise
to you even
 Page 100

though I would suffer were there no general

practice of keeping promises. It would then be
in my interests both to support promise-keeping
in general and to break my promise in this case.)
Moral rules also do not take precedence in the
sense that reason requires me to do what moral-
ity says, whatever other interests are at stake.
Reason does not require that. (For example,
breaking a morally-binding promiseat great cost
to othersis not necessarily irrational even if my
only reason for breaking it is that I would suffer
somewhat more if I kept it than if I broke it.)
Moral rules take precedence over other consid-
erations only in the sense that we want them to
win out in general, that we want everyone else
to be taught that they should win out all the
time, that we would help make them win out by
condemning those who do not give them preced-
ence, and so on. Moral rules are, in this sense,
absolute almost by definition.
 502/1183

There are, however, also two senses in which

moral rules are not absolute. They are, first, not
absolute in the sense that would be captured by
saying that "reason requires" them to take pre-
cedence. Reason does not require that much.
Moral rules are those rules everyone wants
everyone else to follow. They are not necessar-
ily the rules any rational person, merely because
he is rational, wants to follow himself. His fol-
lowing them is not itself necessarily desirable
(unless he is a person of good character). For
example, when I keep an expensive promise I
wish I had not made, I ordinarily do it not be-
cause it is rational to "want to" but because I
"must." Reason requires moral rules to take pre-
cedence over other considerations (and so, to be
"absolute") only from "the moral point of view.''
(The moral point of view is the way rational per-
sons must look at things when laying down rules
to guide all rational persons in their relations
with one another.) The merely prudent person
 503/1183

(the person not moved by benevolence or his

own moral dignity) may break her promise
without being irrationalif she can get away with
itthough a person of good character or high pur-
pose may not. Good character or high purpose
can change what it is rational to do. 34
 504/1183

That brings us to the second sense in which

moral rules are not absolute. They are not excep-
tionless.35 The rule "Don't kill" might, for ex-
ample, better be written "Don't kill except. . . ."
Although all rational persons would agree that
killing should in general be prohibited, few, if
any, would agree that all killing should be pro-
hibited. Perhaps the easiest exception to justify
is killing in self-defense. If the reason we sup-
port a general prohibition against killing is that
we fear involuntary death at the hands of others,
the exception for self-defense might be justified
by reasons much like those justifying the gener-
al prohibition: An exception for self-defense
would in general be invoked only against those
breaking the rule against killing, would tend to
discourage rule breaking by making rule break-
ing more risky than it would be if people of
good character or moral integrity could not in
good conscience kill in self-defense, and would
otherwise serve our rational interest in a safe
 505/1183

life. Exceptions to moral rules help make it easi-

er to do what is right, heading off possible con-
flicts between morality and prudence.

Our quest for a standard of right and wrong by

which to evaluate what Hardin and James did
leads us to ask whether Hardin or James did
anything morally wrong. The answer we must
now consider is that they did do something mor-
ally wrong
 Page 101

because each had a conflict of interest making it

morally wrong to do what he did. Because the
NSPE code includes a general prohibition of
conflicts of interest, justifying that answer in ef-
fect provides a moral justification for a certain
reading of the code. 36 And because moral con-
siderations take precedence over all others (in
the sense explained earlier), that answer would
take precedence over those already considered
even if they had not proved inconclusive.

What is Morally Wrong with a Conflict of

Interest?
 507/1183

Section II.4(a) of the NSPE code of ethics as-

sumes a certain understanding of conflict of in-
terest. Let us begin by trying to make that un-
derstanding explicit. The section assumes, on
the one hand, that an engineer will be acting for
an "employer" or "client" and, on the other, that
he will be exercising ''judgment" (or providing a
"service") of a certain sort the quality of which
might be influenced (for the worse) by certain
associations, interests, or circumstances. The
sort of "judgment" of concern to the code is that
judgment (or service) an engineer provides
when "acting in a professional matter"that is,
when exercising the special skills, powers, or
authority he has because he is an engineer rather
than, say, a mere citizen, business-person, or
employee. Though competent to provide the
judgment, his ability to do so is nevertheless
compromised because he has a conflict of in-
terest. His judgment is subject to "influences" by
improper considerations or, at least, appears to
 508/1183

be. There is reason to believe he may not do

what a "faithful agent or trustee" with his skills,
powers, and authority (as an engineer) would or-
dinarily do for the person in whose interests he
is supposed to be acting.

The NSPE code limits its concern to "profes-

sional matters." That very limitation suggests
that an engineer might have a conflict of interest
even when not acting as an engineer. The code
seems to apply an analysis of conflict of interest
more general than engineering ethics. The no-
tion of conflict of interest the code assumes is, it
seems, one any rational person should be able to
understand, engineer or not. So, let us try to
state that general analysis of conflict of interest
first, see how it works in a case with which we
are all familiar (and about which we have relat-
ively settled opinions), assure ourselves that the
analysis implicit in the code is one we can ac-
cept (if indeed it is), and only then try to under-
stand what it tells us about Hardin and James.
 509/1183

General Analysis of Conflict of Interest

We might generalize the code's analysis of con-

flict of interest in this way:
A conflict of interest is any situation in which (1)
a person (for example, an engineer) is in a rela-
tionship with another person (for example, a cli-
ent or employer) requiring him to exercise judg-
ment on behalf of that other person and (2) there
is good reason to believe that, though competent
to provide that judgment,37 he may not do it as
he should (for example, as an equally competent
agent or trustee of that client would) because of
some special interest, obligation, or other con-
cern of his.38
 Page 102

Does this analysis fit our settled opinions about

conflict of interest in general? Can we provide a
moral justification for those opinions? Let us
consider a relatively clear case of conflict of in-
terest that has nothing to do with engineering.
 511/1183

Suppose that a judge is to hear a case between

two large corporations, that she is known to be a
good judge in general and an expert in the law
affecting this case. But suppose, too, that she
has substantial holdings in one of the two cor-
porations. Such a judge certainly has a conflict
of interest. Does she have a conflict of interest
according to the analysis we derived from the
code? The answer seems to be yes. She is in a
relationship with another requiring her to exer-
cise judgment. Her role as judge puts her in the
position of having to decide the case before her
according to her judgment of what the law re-
quires. She is supposed to provide impartial
judgment to both parties to the case. And that is
exactly what there is reason to believe she may
not do. The circumstances are such that, al-
though she is exceptionally competent to judge
cases of this kind, she may nevertheless be un-
able to judge this one as she should. Her interest
in one of the corporations may bias her
 512/1183

judgment in favor of that corporation. Money

talks.
 513/1183

Of course, there is no guarantee that she will

listen. This judge might, for example, be able to
allow for her natural bias when deciding the
case. She may be able to "bend over backward"
to cancel its effect. But, even if she can in fact
cancel the effect of her bias, how can she or
anyone else know that she has succeeded? This
is not the sort of bias judges routinely cancel
out. Canceling the effect of pecuniary interest is
not part of ordinary judicial training or skill. We
cannot then rely on this judge's judgment that
she has canceled the effect of that interest be-
cause her judgment of that may itself be affected
by the same influences. It is also unlikely that
she will be able to show in some other way that
she succeeded in canceling the effect of that in-
terest. Judging is in part a matter of forming an
informed opinion about controversial questions.
There is no mechanical way to check such judg-
ments for the effect of interest. (If there were,
we could replace judges with clerks.) We can, of
 514/1183

course, bring in other judges to examine the

same evidence the first judge examined and
form opinions of their own. But beside being
impractical (why not just replace the first judge
instead?), such double-checking would simply
produce other opinions. We would learn that
other judges would agree or disagree with the
first judge, but not whether she succeeded in
canceling the effect her interest had on her judg-
ment. There would remain the question whether
she would have decided differently had she not
had that conflict of interest. So her ownership of
the stock cannot be shown not to affect her judg-
ment. Because the inability to prove absence of
actual bias is itself a good reason to doubt her
judgment, she will have a conflict of interest
even if she decides the case "correctly" and for
all the right reasons.
 515/1183

This conflict of interest would not be "an appar-

ent conflict of interest," a mere appearance. A
conflict of interest is "merely apparent" if the
relevant parties have available information cap-
able of showing that the interest in question has,
under the circumstances, no tendency to bias
judgment. So, in our example, we could dispel
the appearance of conflict by showing that the
judge no longer owns the stock or, while owning
it, cannot know that she does because her hold-
ings are in a blind trust.
 Page 103

Responding to Conflict of Interest

A conflict of interest is like dirt in a sensitive

gauge. For the same reason rational persons
want reliable gauges, they want those on whose
judgment they rely to avoid conflict of interest
(insofar as practical). We would, for example,
ordinarily want the judge to decline to hear the
case (or to sell her stock before hearing it). We
do not want her "bending over backward" to
compensate for possible bias because we have
no way to know how such bending will turn out.
Will she bend over far enough? Will she bend
over too far?
 517/1183

If that is what conflict of interest is, what can we

do about it? Most conflicts of interest can be
avoided. We can take care not to put ourselves
in a position in which contrary influences or di-
vided interests might undermine our ability to
do what we are supposed to do. But, however
much care we take, we will not always succeed
in that. Our relations with one another are too
many and too varied for us to keep track of them
all. We cannot always foresee how they will af-
fect one another and so cannot take the precau-
tions necessary to prevent all conflicts of in-
terest. Still, though conflicts of interest cannot
always be avoided, they can always be escaped.
We can end the association, divest ourselves of
the interest, or otherwise get beyond the influ-
ence that might otherwise compromise our
judgment.
 518/1183

But is it always practical to do that? Do we

really want people never to act for us just be-
cause they have an interest that makes their
judgment somewhat less reliable than it would
otherwise be? Should there be an absolute pro-
hibition on acting with a conflict of interest?
These are not hard questions. Consider the judge
again. Suppose she retires. Some time later the
two corporations have a similar dispute but this
time agree to arbitrate rather than endure the ex-
pense of another trial. They come to the judge
because of her reputation and the integrity she
displayed in withdrawing during their previous
dispute. She has not sold the stock. Would we
want her either to refuse to arbitrate or to sell off
the potentially biasing stock?
 519/1183

One might suppose that the answer is clearly

yes, she should refuse or sell. After all, the own-
ership of the stock is still a consideration that
could influence her judgment and an arbitrator-
like a judgeis expected to provide unbiased
judgment. On the other hand, the two corpora-
tions may be willing to run the risk of that influ-
ence to benefit from the judge's special insight
into their problem (just as we might prefer to
use a sensitive but slightly unreliable gauge
rather than one which, though fully reliable, is
too crude for the measurements we want to
make). The general rule against conflict of in-
terest protects the person who properly relies on
the judgment of another. If such protection were
sometimes to make people worse off and there
were some other way to provide much the same
protection without making the people involved
worse off, would it not be reasonable to make an
exception to the general prohibition? Would this
 520/1183

not be like treating self-defense as an exception

to the general prohibition of killing?

Consider the retired judge once more. Suppose

she reasons in this way: "I could not have
agreed to such an arrangement when I was a
judge because the public as well as these two
parties were relying on my judgment. My de-
cision in the case would have been a precedent
for others. Here there is no question of preced-
ent; no
 Page 104

one will rely on my judgment but these two cor-

porations. They have come to me because they
trust me and because they want to save money.
They have not asked me whether I still own the
stock. Obviously, they don't care. If I were to
sell off the stock now, I would lose a lot of
money, much more than they are willing to pay
for this job. So, I must keep the stock. I can,
however, do the job fairly even if I own the
stock. I'm quite sure of that. So, there's no reas-
on why I should not accept the arbitration
without further ado."
 522/1183

Is there anything troubling about the judge reas-

oning in this way? Certainly there is. The judge
seems to be taking too much on herself. She de-
cided that the reason the corporations did not
ask about the stock is that they did not care
about it rather than, say, that they forgot about it
or expected the judge to inform them if she still
owned it. She also decided that she can arbitrate
the case fairly even if she has a conflict of in-
terest rather than leaving that decision to those
whose agent or trustee she is to be. She decided
what they will risk (and, however "sure" she is,
there remains a reasonable chance that she is
wrong). Her reasoning is, in a word, "paternal-
istic." She assumes that it is morally permissible
for one rational person (without the other's in-
formed consent) to decide significant aspects of
the other's life because she believes herself at
least as able to judge such things as the other is.
 523/1183

It is easy to see what is wrong with our judge's

reasoning. Each rational person wants to live ac-
cording to his own conception of the good (his
own judgment of what his interests are and how
they should be balanced), not according to
someone else's. We do not want people deciding
what is better for us simply because they believe
they know better. That is true even when they
may in fact know better and their decision does
not impose any significant risk of harm. How
much more true when, as usually happens, they
lack the information about us that we ourselves
have and the decision would impose significant
risks on us! Because it is something all rational
persons would generally oppose, imposing risks
on another rational person for that other's good
but without the other's informed consent must in
general be morally wrong.
 524/1183

It seems, then, that before our retired judge

agrees to arbitrate the case, she should disclose
her conflict of interest. Indeed, she should dis-
close any information that might cast doubt on
her ability to perform as the two parties would
otherwise reasonably expect. She may advise
them that she believes she can overcome the
conflict (because she does believe that). But she
must be sure that they are fully informed of
what the conflict is and fully appreciate the risks
of putting their case to an arbitrator laboring un-
der such a disability. Only then can she be reas-
onably sure that if they go ahead with the arbit-
ration, the decision to go ahead will be "theirs,
not hers," that is, the result of their informed
judgment, not in part the result of her not reveal-
ing information they would have found relevant.
Disclosure has another benefit as well. It allows
our judge to discuss with the two corporations
ways to compensate for any tendency toward bi-
as she might have.
 525/1183

To sum up: A conflict of interest exists if (1) an

individual is in a relationship with another justi-
fying that other's reliance on the proper exercise
of her judgment in that other's interest and (2)
the individual has an interest tending to interfere
with the proper exercise of that judgment. In
general, conflicts of interest should be avoided
or, if unavoidable, ended as soon as possible. In
special cases, however, a
 Page 105

conflict may be tolerated if tolerating it will be-

nefit the person who is relying on the judgment
in question, but then only if there is full disclos-
ure to that person and that person intelligently
consents to the relationship nonetheless. Dis-
closure does not end a conflict of interest. What
it ends is the passive deception of allowing an
individual's judgment to appear more reliable
than it in fact is. 39

Judges, Hardin, and James

 527/1183

If all this makes sense, it should not be hard now

to see what was wrong with what Hardin and
James did. Let us begin with Hardin. Hardin ini-
tially gave his opinion on the interpretation of
HG-605a in the friendly atmosphere of dinner
with M&M executives. Such an atmosphere
does not invite hard thought. We cannot know
whether Hardin would have given a different
opinion under other circumstances. Indeed, even
he cannot know that. We can reasonably con-
clude, however, that his opinion might well have
been different if, say, Hydrolevel executives had
taken him to dinner first or had been present at
the dinner with M&M executives. Having "gone
on record" as accepting a certain interpretation
of the code, Hardin would have found it embar-
rassing to change his mind once the inquiry was
officially submitted in writing. His first response
thus tended to undermine his ability to consider
the written inquiry with the open mind he might
otherwise have had. He had, in other words, a
 528/1183

conflict of interest from the moment he first

gave his opinion at dinner on a question likely to
come before his committee. (Because giving
one's opinion on a question tends to prejudice
one's judgment thereafter, judges generally re-
fuse to discuss any case that might come before
them.) Hardin's helping to draft the inquiry may
have strengthened further his feeling of owing
M&M the opinion he gave at dinner. But, had he
not given his opinion in advance, his part in
drafting the inquiry would hardly have seemed
important.
 529/1183

What should Hardin have done about the con-

flict of interest once it developed? He could
have declined to respond to the inquiry when
Hoyt referred it to him, passing it on to his sub-
committee (minus James) and leaving it to them
to decide what to do with it without his particip-
ation. Or he could have informed Hoyt that he
had already committed himself on the question
informally (and helped to draft the inquiry),
leaving to Hoyt the decision whether Hardin
should participate. Had Hardin done either, no
one would have had reason to doubt his integrity
(and his employer might have been saved
$75,000).
 530/1183

Of these two alternatives, however, declining to

participate seems much the better. Declining to
participate resolves the problem altogether;
while disclosing the problem to Hoyt simply
makes it Hoyt's problem rather than Hardin's.
Whenever there is a conflict of interest, there is
someone ("the client") entitled to rely on the
judgment in question. Conflict-of-interest prob-
lems cannot be resolved by disclosure unless the
disclosure is made to the client. Sometimes it
takes some thought to determine who the client
is (or, more often, who all the clients are). This
is such a case. Who is Hardin's client here? The
answer is not ASMEor, at least, not only ASME.
ASME holds itself out as an authority on boiler
safety. It invites the general public to rely on its
safety codes and on the interpretations its com-
mittees make of them. And the public does rely
on them. ASME, though not a governmental
body,
531/1183
 Page 106

is still a "public agency," that is, an agency that

purports to serve the public interest. So, Hardin's
client (or at least one of them) is ultimately the
general public. Had Hardin made full disclosure
to Hoyt and Hoyt told him to go ahead, Hardin
would still not have made full disclosure to the
public. He would have allowed Hoyt to act for
his (and Hoyt's) client. He would have treated
Hoyt as trustee or guardian of the public in-
terest. That may sometimes be necessary, for ex-
ample, when revealing information to one client
would do serious harm to another and withdraw-
ing would do similar harm. (Not all paternalism
is morally wrong.) But, given the ease with
which Hardin could have escaped the conflict
altogether (without any risk of harm to the pub-
lic interest), it does not seem necessary or even
desirable for him to have, in effect, allowed
Hoyt to act for the public without the public's
informed consent.
 533/1183

Identifying Hardin's ultimate client as the gener-

al public, not ASME (or M&M), also helps to
explain why Hardin should have revealed more
to the Professional Practices Committee than he
did. The Professional Practices Committee, like
Hardin's own Heating Boiler Subcommittee,
was acting as trustee of the public, not simply as
an agent of ASME. (That is so because ASME
implicitly guarantees the integrity of its proced-
ures when it invites the public to rely on its
codes and committees.) The standard of disclos-
ure was, then, not what was customary within
ASME but what the public might reasonably
think relevant (or what it was in the interests of
the public to know) should it wish to evaluate
the reliability of the ASME interpretation in
question. Hardin should have revealed his meet-
ing with M&M executives because the meeting
might reasonably have looked suspicious to
members of the public. He should not have kept
that information to himself just because
 534/1183

hecorrectlybelieved ASME officers would agree

there was nothing inappropriate about it. The
decision whether to trust his judgment was the
public's, not his, because he invited their trust by
answering the M&M inquiry in his capacity as
chair of the Heating Boiler Subcommittee. For
the same reason, he should have revealed his
part in helping to prepare the original inquiry.
 535/1183

I leave evaluation of James's conduct as an exer-

cise for the reader. Consider in particular the
following questions: What if anything was
wrong with not signing the original inquiry?
What if anything was wrong with reporting to
the full Heating and Pressure Vessel Committee
the recommendation of his subcommittee con-
cerning the Hydrolevel objection to Hardin's ori-
ginal response? What if anything was wrong
with helping to draft the letter of June 9, 1972?
What if anything was wrong with failing to re-
veal those acts to the Professional Practices
Committee? What part does mere appearance
play in answering these questions? If there was
anything wrong with any of these acts, what
should James have done instead (while remain-
ing a faithful employee of M&M)? Why?
 Page 107

8
Codes of Ethics, Professions, and Con-
flict of Interest
Chapter 7 used moral arguments to interpret
provisions of the NSPE code of ethics. This may
suggest that morality somehow determines what
a profession should require of its members. This
chapter uses an emerging field of engineering,
clinical engineering, to show thatat least for
conflict of interestmorality leaves professions a
substantial range of choice concerning what
should or should not be required of members.
Morality, while limiting what professional eth-
ics can be, does not determine what it is. This
chapter also provides an opportunity to respond
to some criticism of the book's general approach
to engineering ethics.
 537/1183

What is Clinical Engineering?

Clinical engineering is part of another relatively

new field, biomechanical engineering (barely
twenty-five years old). Besides clinical engin-
eering, biomechanical engineering includes re-
habilitation engineering (the engineering that
goes into, for example, the choice, attachment,
and maintenance of artificial limbs) and bio-
mechanical research (the engineering that goes
into, for example, the design, building, and test-
ing of those limbs). As in "bioethics," the "bio"
in "biomechanical" signals a relation with medi-
cine (rather than with "life" as such); the "mech-
anical,'' an origin in mechanical engineering.
(Electrical engineers, especially, seem to prefer
to call the field by the more informative "medic-
al [or biomedical] engineering.")
 538/1183

Clinical engineers share with other biomechan-

ical engineers a working relation with medicine.
They differ from other biomechanical engineers
in the relation they have. Working in hospitals
(or other medical enterprises), clinical engineers
oversee the vast technological structure that
makes modern medicine possible. Because that
technological structure is a complex mix of con-
ventional mechanical systems and
 Page 108

the newest in electronics, the typical clinical en-

gineer may have a degree in either electrical or
mechanical engineering. 1
 540/1183

Clinical engineers are engineers. They have de-

grees in engineering; employ much the same
method, skills, and knowledge other engineers
employ; and, like other engineers, are concerned
with designing, developing, and operating safe
and useful physical systems. Clinical engineers
are, nonetheless, not ordinary engineers. Most
engineers work in an organization where engin-
eering is a central concern. Even in a finance-
oriented company like General Motors, engin-
eering is the mother tongue, the language of
most of those with whom most engineers must
deal. That is not true of a hospital. Medicine is
the mother tongue there. A clinical engineer
may be the only engineer the hospital employs.
And even when (as often happens) he has a few
colleagues, together they form only a small part
of the organization. Most of their dealings are
with physicians, nurses, medical administrators,
and others to whom engineering is alien.
 541/1183

This alone suggests that the hospital may be an

environment in which ordinary engineering eth-
ics is not appropriate. There are other reasons to
think so. For instance, engineers generally agree
that the safety, health, and welfare of the pub-
licrather than that of the client (or employ-
er)comes first. Yet, for physicians, nurses, and
other health care professionals, the safety,
health, and welfare of the patient comes first.
The public interest, like the interests of col-
leagues or other third parties, is secondary.

Such differences in environment suggest ques-

tions such as these: How, if at all, should the ob-
ligations of clinical engineers differ from those
of other engineers? What should be the para-
mount obligation of clinical engineers?
 542/1183

Such questions are not easy to answer. That is

reason enough not to try to answer them here.
But there is another reason. I am not a clinical
engineer. Deciding what clinical engineers
should profess is part of the profession in which
they claim membership. Others can only advise.
As a philosopher, I can best advise by clarifying
the questions I have put so that clinical engin-
eers, the members of the profession in question,
find them easier to answer. I can, of course, give
all sorts of other advice, too, but why do it? I
lack the experience on which such advice should
rest. And my theory of engineering will pass an
important test even if it only helps clarify the
questions I put.
 543/1183

But how can clinical engineers answer those

questions, however clarified? Are not the ques-
tions ultimately philosophical? They are not
(though they certainly can generate philosophic-
al puzzles). They are, rather, much like other
questions engineers routinely face, for example,
questions of safety or reliability. They can be re-
solved in much the same way. Engineers will
have to make educated guesses, test them, share
results with peers, reassess their guesses based
on the tests and peer response, make such modi-
fications as seem appropriate, test again, and so
on.2
 544/1183

At this point, some readers may be formulating

objections. After all, I have suggested both that
there is no Archimedean point from which to
deduce a code of ethics for a profession and that
the members of a profession have a privileged
position with respect to determining what their
code of ethics is (and should be). I even sugges-
ted that writing a code of ethics for engineers is
much like other engineering tasks. Clearly, I
went well beyond the claims of chapter 4. I have
much to
 Page 109

explain. But, first, I want to examine an ethical

problem typical in clinical engineering. That ex-
amination provides evidence crucial to my
explanation.

A Problem of Professional Ethics

 546/1183

Consider this relatively simple problem: You

are an engineer in charge of clinical engineering
at Big Bill Hospital. Your work there introduced
you to the products of Hi-Tec, mostly very ex-
pensive diagnostic equipment. Hi-Tec is a relat-
ively large company, with good service as well
as good equipment. You have been impressed
by everything of theirs you have seen. Indeed,
after some bad experiences with Hi-Tec's com-
petitors, you have recommended some pur-
chases from Hi-Tec even when the competitor's
price was significantly lower. When your stock-
broker lists Hi-Tec's stock as a good buy, you
consider buying a few hundred shares at $14
each. Should you?
 547/1183

Some things are obvious. Big Bill's purchases

are not large enough to affect the overall profit-
ability of Hi-Tec. You will not be able to make
money by giving business to Hi-Tec rather than
to its competitors. You will not have what most
people would think of as a clear conflict of in-
terest. On the other hand, you will have a con-
nection with Hi-Tec that could affect your pro-
fessional judgment. 3 Hi-Tec will, as it were, be
a member of your financial family. Although
you may be sure that the connection will not in-
fluence you, you recognize that others cannot be
so sure.4 If they knew you owned Hi-Tec stock,
they might wonder about your impartiality when
you recommended a Hi-Tec product over some
other. Your recommendation might carry less
weight than it would otherwise.
 548/1183

Let us suppose that, like many employers, Big

Bill does not require you to reveal ownership of
publicly traded stock. So, you are not, as an em-
ployee, required to tell Big Bill if you buy the
stock. Your employer leaves you free to choose
between at least these three options: (1) passing
up the stock, (2) buying the stock and saying
nothing, and (3) buying the stock and informing
your employer. Which should you choose?
 549/1183

A physician faced with such a question might

profitably turn to section 8.06(1) of the Prin-
ciples of Medical Ethics of the American Med-
ical Association (AMA): "A physician should
not be influenced in the prescribing of drugs,
devices or appliances by a direct or indirect fin-
ancial interest in a pharmaceutical firm or other
supplier." Physicians are not forbidden to have
financial interests that could influence their
judgment in ways not in the best interest of their
patient. They are only forbidden to be influ-
enced by such interests. So, because you believe
that owning the stock will not influence your
judgment, you could if you were a physician
rather than a clinical engineer, properly buy the
stock.5
 550/1183

But you are a clinical engineer, not a physician.

Where then can you go for guidance when your
field has no code of ethics of its own? If you
were a member of the Institute of Electrical and
Electronic Engineers (IEEE), you might turn to
its code of ethics. Clause 2 would tell you to
"avoid real or perceived conflicts of interest
whenever possible, and to disclose them to af-
fected parties when they do exist."6 For an elec-
trical engineer, the crucial question is not
whether the interest will influence her but
whether she has such an interest ("real" or
"perceived").
 Page 110

If, instead, you were trained as a mechanical en-

gineer, you might turn to the Ethical Guidelines
(4.a) of the American Society of Mechanical
Engineers. You would get much the same an-
swer: "Engineers . . . shall promptly inform their
employers or clients of any business association,
interests, or circumstances which could influ-
ence their judgment or the quality of their ser-
vices." (ABET's code would answer in the same
way.)
 552/1183

The American Association of Engineering Soci-

eties (AAES) sets an even higher standard. Ac-
cording to its Model Guide for Professional
Conduct, "Engineers disclose to affected parties
all known or potential conflicts of interest or cir-
cumstances which might influenceor appear to
influencejudgment or impair fairness or quality
of performance." Even if you were sure owning
the stock could not affect your judgment, you
would, according to the AAES, be obliged to in-
form the hospital because owning the stock
might appear to influence your judgment (if ever
anyone there came to know of the stock).
 553/1183

The similarity between the IEEE, ASME, and

AAES codes suggests that engineers generally
agree that they must be held to a higher standard
than physicians. 7 No such agreement exists. If
you were a licensed professional engineer, you
might instead have turned to the Code of Ethics
of the National Society of Professional Engin-
eers. You would then find that article III, section
5 reads much like the AMA's code: "Engineers
should not be influenced in their professional
duties by conflicting interests." There is nothing
about disclosure of what merely "could" influ-
ence your judgment.8

So, what are you, an ordinary clinical engineer,

to do? You could, of course, hold yourself to the
highest standard possible. But why do that if
other engineers would not do the same and your
doing so burdens you without benefit to your
employer? Why should you not make a little ex-
tra money if you can do so properly?
 554/1183

Analyzing the Problem

Of course, the question is what exactly is proper

here. Ordinarily, no matter what stock you
bought, you would not want to notify your em-
ployer. Your investments are your own busi-
ness. You have even more reason to keep any
purchase of Hi-Tec from your employer. You
want to avoid unjustified undermining of your
professional authority. According to the NSPE,
as an engineer you can properly keep that in-
formation confidential, as long as you don't al-
low ownership of Hi-Tec stock to influence your
professional judgment. But, according to the
IEEE, ASME, and AAES, if you own any Hi-
Tec stock, you have a professional obligation to
tell your employer. So what should you do?
 555/1183

You might call up other clinical engineers in the

area and ask them what they would do. Engin-
eers in specialized fields sometimes develop a
consensus about certain ethical questions just as
they do about the reliability of certain instru-
ments. If the question comes up enough, you
may get a relatively clear answer. If, however,
the question does not come up enough, you are
likely to get a range of half-thought-through
opinions leaving you more or less where you
began. Let us suppose you find no consensus.
What then?
 Page 111

The way I set up the problem, the best you can

do as a clinical engineer is muddle through.
Without a standard governing clinical engineers
as such, there are several morally permissible
options. The choice among these is a personal,
rather than a professional, matter (at least until
you decide whether you are a professional en-
gineer subject to the NSPE code, a mechanical
engineer subject to the ASME code, or a mem-
ber of IEEE subject to its code). You can't know
what you should do as a clinical engineer.
 557/1183

This problem, though real enough for a practi-

cing engineer, is, I admit, not "philosophically
interesting" in the way those problems are that
hold the attention of philosophers for genera-
tions. The problem is philosophically interesting
in a more humble way however. Not being a
hard problem, it is just the sort that any plaus-
ible theory of professional ethics should be able
to handle easily. It is philosophically interesting
as a test case, revealing something important
about codes, something most theories of profes-
sional ethics miss.
 558/1183

If you, as clinical engineer, find "muddling

through" an unsatisfying way to resolve the
problem, you have at least one reason to want
clinical engineering to have its own code of eth-
ics. A code can turn a morally indeterminate
question like this into a question of professional
ethics with a relatively determinate answer. A
code of ethics does that by creating a convention
for all members of the profession to follow. If
generally following the convention gives clinic-
al engineers something they all wantwhether
freedom to make money on the side, the greater
trust of employers, or some combination of
these or other goodseach clinical engineer has
reason to want all others to follow the conven-
tion even if their following it means doing the
same. If the others generally do follow the con-
vention, realizing it in practice, each clinical en-
gineer is morally obliged to do the same. Claim-
ing to be a clinical engineer is a voluntary act.
No engineer can fairly take the benefits flowing
 559/1183

from that claim without doing her share to main-

tain the practice creating the benefits. A morally
permissible code of professional ethics itself
defines each member's fair share in helping to
make possible the benefits the code generates. 9

So, a code of ethics is, as such, not merely good

advice or a statement of aspiration. It is a stand-
ard of conduct which, if generally realized in the
practice of a profession, imposes a moral oblig-
ation on each member of the profession to act
accordingly. A profession's code of ethics neces-
sarily sets a standard below which no member
of the profession can properly fall. Any docu-
ment not intended to have that effect is not,
strictly speaking, a code of ethics, whatever it
may be called.10
 560/1183

A code of ethics sets a minimum standard, no

matter how high the standard if, but only if, the
standard set is generally realized in the practice
of the profession. For that reason, a profession
should not set its standard very high. For ex-
ample, if the code set its standard so high that
only a few could hope to survive in the profes-
sion, either all but a few saints would avoid the
profession or most of those in the profession
would ignore the standard. The code would
either define a dying (or dead) profession or
serve as a mere statement of aspiration in a pro-
fession defined in other terms. A living code of
ethics is always a compromise between ideal
and reality.

So, one reason clinical engineers might want to

have a code of ethics of their own is to make it
possible for them to tailor their professional ob-
ligations to the
 Page 112

special realities of hospitals. This may mean set-

ting their standards higher or lower than those of
other engineers (or just setting different stand-
ards). But, what it must mean is setting their
standards higher than their employer's. A code is
pointless when moralityincluding the morally
permissible promise contained in the employ-
ment contractleaves only one option. In this re-
spect, living by a professional code is necessar-
ily "public service"that is, benefiting the public,
those served, by giving them more than they
would be entitled to but for the code.

Codes at Work

Adopting a realistic code is part of making an

occupation a distinct profession. But it is only
part. Let me now briefly describe some other as-
pects of professionalism, making clear how
central a code is to them.
 562/1183

A code of ethics cannot actually guide conduct

unless those to be guided know the code. Be-
cause a code necessarily sets a standard higher
than ordinary morality, even a morally decent
person is not likely to do what the code requires
unless he knows what in particular it requires.
The code must be learned in just the way other
engineering standards are. The code can be
taught as part of the profession's basic cur-
riculum or its continuing education. 11 It can
also be taught in less formal waysfor example,
by publishing articles on particular questions of
ethics in the profession's journals.
 563/1183

Education is probably the primary means by

which a profession puts its code into practice.
But every profession needs something more,
some means of enforcement. The minimum is
the informal enforcement that comes from one
member of a profession saying to another, "But
that would be unethical." Such a rebuke is
barely distinguishable from education. Beyond
this minimum are group pressure; peer review;
reputation in the profession; formal certification
of various sorts; disciplinary committees with
the power to censure, suspend, or expel from the
profession; and state licensing with the power to
bar from employment.
 564/1183

Educational and enforcement activities almost

define profession in the public mind. Without
them a learned occupation is only a field of
study and endeavor, a discipline, not a profes-
sion (strictly so called). These educational and
enforcement activities all presuppose a code of
ethics of some sort, that is, a minimum standard
common to all members of the profession mak-
ing improper what would otherwise be proper.
The code need not be written, but the more that
is in writing the easier it will be to teach the
code, especially in a young profession. So, clin-
ical engineering, though no longer a new discip-
line, is still only potentially a profession distinct
from engineering. What it lacks is its own code
of ethics.

Objections Answered
 565/1183

We come, then, to the objections. John Ladd

represents the chief objection to my approach.
For him, talk of "a code of ethics" rests on a
confusion of morality and law:
 Page 113

Ethics, sometimes called "critical morality," is

logically prior to all these institutions and mech-
anisms of social control [like law or the "value
system" of some group]. . . . The principles of
ethics (or morals) are not the kind of thing that
can be arbitrarily created, changed, or rescin-
ded. . . . [T]hey are "discovered" rather than cre-
ated by fiat. They are established through argu-
ment and persuasion, not by imposition of ex-
ternal social authority. 12
 567/1183

On my view, however, "ethics" as used in "code

of professional ethics" does refer to a "value
system of a group," the profession itself. It is a
mechanism of social control, a way of coordin-
ating the conduct of people engaged in a com-
mon occupation. Unlike morality, a code of pro-
fessional ethics cannot be logically prior to all
institutions of social control. A code of profes-
sional ethics is an institution of social control.
However, a code of professional ethics can still
be distinguished from law (and other forms of
merely external authority). On my view, a code
of professional ethics cannot, as such, be the
work of external authority. To be a code of pro-
fessional ethics, a code must be a morally per-
missible standard of conduct each member of
the profession wants all others to follow even if
their following it would mean he must do the
same. Professional ethics thus resembles moral-
ity in having an ''internal" aspect (being
something each wants). Argument and
 568/1183

persuasion are essential to developing and main-

taining a code as a living practice.13 Profession-
al ethics is "social" because it involves coordin-
ating the conduct of a group, the profession. But
it is "control" not by external authority but in
part at least by the conviction of those subject to
it that claiming membership in the profession is
voluntary; that the profession prescribes certain
ways of acting for its members; that acting in
the prescribed ways will, on balance, serve the
interests of all members of the profession; and
that therefore none can fairly claim membership
while failing to act as prescribed.
 569/1183

Though ethics resembles morality in this intern-

al aspect, it is not just ordinary morality,
something that is or should be common to all ra-
tional persons, to all rational persons in this so-
ciety, or even to all rational persons in a particu-
lar occupation. Groups like clinical engineers
actually "create" their ethics; they do not simply
"discover" them in the larger society. Or, at
least, they do not discover them in the larger so-
ciety in a sense different from the way a legis-
lature "discovers" the law.
 570/1183

We must make a distinction Ladd does not.

There is a sense in which law, like ethics, must
be discovered. We do not make law arbitrarily.
That would ordinarily be irrational. We look for
reasons. We try to choose wisely. Still, what the
law is is determined in part by what we actually
decide. It cannot be deduced from any (interest-
ing) general principles (even combined with a
description of circumstances). So, for example,
no principles settle which side of the road
Japanese law will require vehicles to drive on in
the year 2101, or how much I should pay in
taxes next year, or even how much sulfur diox-
ide steel plants along Lake Michigan's Indiana
shore should put into the air tomorrow. Legislat-
ors must make law to decide such questions.
And the laws they make will be the law even if
they do not in fact choose wisely (as long as
they satisfy procedural requirements, substant-
ive constitutional restraints, and the minimum
requirements of ordinary morality).
571/1183
Page 114
 573/1183

Members of a profession make professional eth-

ics in the same way. Clinical engineers need
have no special rules about conflict of interest.
They can continue to allow those practicing in
their field to follow ordinary moral standards,
the standards set by their employers, or the
standards of their respective professions. They
are free to "legislate" or not, as they choose.
They do, however, have reason to "legislate." If
they want clinical engineers to be known as
people who handle conflict of interest in a spe-
cific way, they have to determine what way that
is. If they want clinical engineers to be respected
for the way they handle conflict of interest, they
have to set a standard higher than would other-
wise prevail. Such reasons certainly do not leave
clinical engineers free to set any standard
whatever. They do, however, leave them free to
choose among a fair number. Each clinical en-
gineer has to balance individual convenience in
complying with this or that standard against the
 574/1183

benefit to everyone of a common standard.

Reasonable people may disagree about where
the balance is to be struck. For example, the
choice between the AAES and the ASME treat-
ment of conflict of interest is neither a choice
between right and wrong nor a choice without a
difference. Clinical engineers cannot "discover''
what their profession's ethics are; they have to
decide what they will be.
 575/1183

By now it should be obvious that writing a code

of ethics is like other engineering tasks in one
way: Writing a code of ethics has (as Caroline
Whitbeck might say) the structure of a design
problem. 14 Ordinary morality, licensing laws,
the interests of clinical engineers, and the like
correspond to the specifications with which en-
gineers usually begin a project. Specifications
constrain outcome but seldom determine it.
Engineers are free to invent new options or cull
old ones from current or past practice. Having
developed a list of options, they try to choose
the best one. There is often no standard decision
procedure yielding a unique answer. Considera-
tion may weigh against consideration. Different
engineers may initially choose differently. The
engineers involved then discuss the problem un-
til a consensus emerges (perhaps doing addition-
al tests or drawing on other sources for relevant
information).15 Perhaps no one gets his first
 576/1183

choice, but each gets a choice he considers bet-

ter than stalemate.

That is one way in which writing a code of eth-

ics for clinical engineering is like other engin-
eering work. There is another. The history of en-
gineering is, in part, a history of standardiza-
tionthat is, a history of constructing tables, for-
mulas, or procedures defining safety, reliability,
convenience, or other elements of good practice.
These standards cover everything from strength
of beams to be used in highrise buildings to the
distance between threads on a screw. The boiler
codes described in chapter 7 constitute just one
example. Every such standard lays down a rule
of conduct for engineers. Engineers develop
these standards because all engineers doing
things the same good way is better than each en-
gineer choosing a way she considers best.
Coordinating conduct pays in such cases.
 577/1183

The standards so defined are not, however,

likely to freeze in the way pure conventions tend
to. Most change from time to time as experience
generates new options or shifts the weight of
evidence favoring this or that old one. But
change cannot be justified simply by the fact
that a new standard would beat the old one in a
"fair fight." "Ideally best," "best in the original
position," or best in some other timeless way is
not good enough. The proposed standard must
be so much better than the
 Page 115

present one that the benefits of changing to the

new one at least pay the costs of changing over.
Engineers must give history its due.

Indeed, history has much to do with maintaining

consensus on engineering standards, ethical as
well as technical. Individual engineers may have
strong views about which standard would be be-
stfor example, the best layout for keys on a
typewriter or computer (almost anything but the
present one). Yet, for most standards, there is
also agreement that the present standard,
whatever its faults, is better than none at all.
And, for most standards, there is no consensus
on which alternative is better. When standardiz-
ation is important, that is enough to justify fol-
lowing the present standard until a new con-
sensus emerges. Like the rest of the world's
work, the work of engineers cannot wait for
perfection.
 579/1183

Engineering standards are, of course, not neces-

sarily what any engineer or set of them intends
them to be. They are public facts, usually words,
numbers, or symbols on paper. They are always
vague or incomplete to some degree. 16 They re-
quire "interpretation" (or, as engineers might
say, "interpolation"). A code of ethics is no dif-
ferent from other engineering standards in this
respect, either. Just as ASME has a committee
to interpret its boiler code, so the NSPE has a
committee to interpret its code of ethics. Each
profession is a continuing discussion, funda-
mentally political in the good sense the ancient
Athenians used to distinguish government by
persuasion from tyranny. To join a profession is,
in part, to enter that discussion, gaining some
control over a common enterprise by giving up
the right to act as a mere individual (what the
Athenians called an "idiot"). To claim to be an
engineer is not simply to claim to know what
engineers know; it is to claim to act as engineers
 580/1183

act. To claim to be a mechanical engineer, for

example, is to claim a share in a specific histor-
ical enterprise carried on according to certain
standards of conduct from technical codes like
ASME's boiler code to ASME's ethical
guidelines. Anyone who wants to be a mechan-
ical engineer but not to act according to those
standards has some explaining to do, especially
to his employer and those who might otherwise
rely on him for engineering judgment.
 581/1183

Indeed, insofar as the distinction between "eth-

ical standard" and "technical standard" suggests
that technical standards have nothing to do with
ethics, the distinction is misleading. For any
profession, part of acting ethically is satisfying
technical standards. What we call a code of eth-
ics can equally well be thought of as the most
general of technical standards, the framework
into which the more detailed ones may be sor-
ted. So, clinical engineering is a mere field of
engineering insofar as clinical engineers are
"standardized" in the way mechanical or elec-
trical engineers are. They are an emerging pro-
fession in part because they have already de-
veloped some distinctive technical standards.
But they will not achieve full status as a distinct
profession until they adopt their own code of
ethics. And nothing requires them to do that.
They may instead (properly) choose to remain a
part of the engineering profession.
582/1183
 Page 117

PART IV
EMPIRICAL RESEARCH
The first three parts of this book drew irregu-
larly but often on history for insight into engin-
eering. Though this part again begins with his-
tory, including an already familiar example, the
Challenger disaster, its focus is on the social sci-
ences. Chapter 9, a report of empirical research,
reveals something of the day-to-day work of en-
gineers, their place in business, and their rela-
tionship to management. It is also, like chapter
5, an attempt to use what we learn about engin-
eers to protect them from some hard ethical
choices.
 584/1183

Chapter 10 tries to convert the abstract claim

that most engineers cannot be professionals be-
cause, as employees, they lack autonomy into a
set of concrete claims capable of empirical test-
ing, but it leaves the testing to social scientists.
Chapter 10 thus suggests the useful cooperation
possible between philosophers and social scient-
ists interested in professions.

The book's epilogue is an invitation to the social

sciences (and history) to contribute more to en-
gineering ethics. It formulates four questions
about engineering, the answers to which would
be useful to those of us who work in engineering
ethicsfour questions the social sciences seem ad-
mirably suited to answer. The epilogue under-
scores the preliminary character of the book as a
whole: its attempt to open up a field of study, to
suggest hypotheses and lines of research, and,
perhaps, to offer a target for those who think
differently.
585/1183
 Page 119

9
Ordinary Technical Decision Making:
An Empirical Investigation
 587/1183

For Canada's engineers, part of belonging to the

profession is wearing a finger ring, originally a
plain band of iron, now generally steel, in
memory of the collapse of a great "iron" bridge
across the St. Lawrence at Quebec in 1907. 1
Remembering that disaster, in which more than
seventy workers died because of an engineer's
error, is supposed to help today's engineers
avoid similar errors. No other engineering soci-
ety I know of has anything quite like this phys-
ical momento, but there is in it something char-
acteristic of engineering. Engineers do not bury
their mistakes. They record them, study them,
and put into practice what they learn from them.
Engineering handbooks, with their tables of tol-
erances, safety factors, standard methods, and so
on, are in part the intellectual equivalents of
Canada's iron ring, an attempt to use failure.
 588/1183

This chapter is also an attempt to use engineer-

ing failure; it differs from most such attempts in
being concerned with ethical rather than tech-
nical failure. I begin with the assumption that
whenever an engineer faces an ethical problem,
something went wrong. There are at least three
possible ways to explain what went wrong: (1)
the individualsomeone (the engineer or someone
else) acted inappropriately; (2) the organization-
althe organization lacks a satisfactory policy or
procedure to prevent the problem or at least to
make it a "no brainer"; or (3) the technicalthe
absence of some device that would prevent the
problem from arising (for example, a testing
device eliminating the uncertainties that leave a
decision to "engineering judgment").
 589/1183

These three ways of explaining an ethical prob-

lem are not mutually exclusive. Indeed, often,
each sheds some light. What we identify as "the
explanation" of a problem probably has more to
do with the solution we think best than with how
the problem arose. When we think the problem
best handled by an individual, we
 Page 120

emphasize individual conduct. When we think

the problem best handled by changing the or-
ganization, we emphasize the role of practice or
policy. When we think the problem best handled
by bringing in a new machine, changing the
physical layout of a building, or otherwise re-
arranging things, we emphasize the technical
(usually without acknowledging the ethical di-
mension in any technical choice). And, when no
one approach seems adequate, we are likely to
describe the problem as "complex."

This chapter approaches certain problems of en-

gineering ethics organizationally. We can identi-
fy policies or practices that, by improving com-
munication between engineers and managers,
will avoid some of the ethical problems engin-
eers would otherwise have to resolve as indi-
viduals. This is a work in "preventive ethics."
 591/1183

The Problem

What engineers do is important. A defect in the

design of an airplane, a failure to maintain qual-
ity in the manufacture of a chemical, or even a
mistake in operating a power plant can ruin a
company, undermine trust in government, or kill
hundreds of innocent people. Our comfort,
prosperity, and safety depend on feats of engin-
eering which, because of their scale and com-
plexity, are necessarily feats of management,
too. Anything in the relationship between engin-
eers and managers that can threaten the integrity
of their work also threatens our common well-
being. The tendency for technical communica-
tion between managers and engineers to break
down is certainly such a threat. The Challenger
disaster provides two stories that illustrate how
serious a threatand suggest the potential signi-
ficance of this chapter.
 592/1183

Acting as a member of the Presidential Commis-

sion investigating the Challenger disaster,
Richard Feynman, a Nobel Prize-winning physi-
cist, interviewed managers and engineers in the
shuttle program. He soon found that managers
could differ substantially from engineers even
about what seem readily determinable facts. 2
For example, Feynman asked both a middle-
level manager and three engineers working for
him on the shuttle's engines "what the probabil-
ity of failure for a flight is, due to failure in the
engines." The engineers all said about one in
two hundred. Feynman's description of the
manager's answer is too good to paraphrase:
[The manager] says, "100 percent." The engin-
eers' jaws drop. My jaw drops. I look at him,
everybody looks at him and he says, "uh . . . uh,
minus epsilon?"

"OK. Now the only problem left is what is

epsilon?"
 593/1183

He says, "1 in 100,000." So I showed [him] the

other answers and said, "I see there is a differ-
ence between engineers and management in their
information and knowledge here."3

The disagreement Feynman thus uncovered had

nothing directly to do with the Challenger dis-
aster. Failure of a booster O-ring, not the
shuttle's engines, caused the disaster. The com-
pany responsible for the boosters, Morton
Thiokol, had nothing to do with the shuttle from
which the boosters would have detached after
the first few minutes of flight. Neither these
shuttle engineers nor their manager was an em-
ployee of Morton Thiokol.
 Page 121

Yet, the disagreement is relevant. Feynman

asked his questions of these engineers and their
managers only after finding that the managers
and engineers working on the booster differed
substantially in their assessment of the probabil-
ity of the booster's failure. The differences there
had come as a surprise. The probabilities were
easily calculated (or, at least, everyone agreed
on how to do the calculations). Once Feynman
realized such differences existed, he wondered
how widespread they were.

He began his interview with the shuttle engin-

eers by asking about any disagreement between
the engineers and their manager. The manager
assured him there were none, explaining why by
pointing out that he, too, had been trained as an
engineer. 4
 595/1183

Feynman did not ask the manager why training

as an engineer should guarantee agreement with
the engineers he managed. No doubt the man-
ager assumed that being able to understand tech-
nical information is enough to ensure that he
would understand it in the way others with the
same technical training did. This assumption
certainly seems plausible. What, then, explains
the disagreement? Feynman suggests that the
manager's misunderstanding was produced by a
work environment, "a game, just as in the case
of the solid rocket boosters, of reducing criteria
and accepting more and more errors that weren't
designed into the device."5 Feynman does not
explain how this process could lead managers to
get simple facts wrong or why ordinary engin-
eers were not affected in the same way. In truth,
Feynman's suggestion does not so much answer
a difficult question as identify the difficulty. We
may get a better sense of the difficulty by con-
sidering an event crucial to the disastrous
 596/1183

decision to launch, one already described in

chapter 4.6

The night before the Challenger blew up, one

manager advised another to "take off [his] en-
gineering hat and put on [his] management hat."
This advice apparently led the manager, a vice
president at Morton Thiokol, to change his eval-
uation of the risk of O-ring failure and approve
the launch (knowing that the launch would not
occur without his approval). The manager was
himself an engineer who, earlier that day, had
decided against the launch after receiving the
unanimous recommendation of his engineering
staff. The night-time reversal occurred under
pressure from NASA but without any new in-
formation about the risks involved. "Putting on
[his] management hat" seems to have changed
the way he thought about the data before him.
Here the gap between engineers and managers
seems to have existed within one individual, an
engineer-manager.
 597/1183

Feynman did not find this gap between engin-

eers and managers everywhere in the shuttle
program. For example, in avionics, "everything
was good: the engineers and managers commu-
nicated well with each other."7 So, the gap is not
inherent in relations between engineers and
managers. It must open as a result of specific
practices. Once it opens, management may (like
Feynman's NASA manager) make decisions on
the basis of something less than all information
readily available. Insofar as full information
tends to make decisions better (better by almost
any reasonable criterion), management has re-
duced its chances of making a good decision.

Good managers want to avoid making decisions

on less than the best information available of
course. The research reported here is intended to
contribute to that end. Our research group began
with four related questions:
 Page 122

1. Can the communications gap Feynman dis-

covered occur in other organizations that employ
engineers and managers?

2. Is there a readily usable procedure for identi-

fying the communications gap before disaster
strikes?

3. What can be done by engineers or managers in

an organization to help prevent the communica-
tions gap from opening or help close it once it
has opened?

4. What can be done by colleges and universities

training future engineers or managers to help
prevent the gap from opening or to help close it
once it has opened?

Relevant Literature
 599/1183

Descriptions of the manager's life in large or-

ganizations are common. Many touch on ethical
problems a manager may face. Few do more
than touch on them. An important exception is
Robert Jackall's Moral Mazes, a particularly
grim evocation of corporate life. His managers
work in a largely amoral environment in which
technical knowledge seems largely irrelevant
and satisfying the boss is the only criterion of
success. Jackall discusses engineers only in the
context of whistleblowing and without giving
any indication that engineers might differ in any
significant way from managers or other employ-
ees. 8 Still, if the managers he describes are
even a rough approximation of the managers
with whom engineers work, the communications
gap between managers and engineers would be
both common and difficult to eliminate.
 600/1183

Engineering is a profession. What does the liter-

ature explicitly discussing relations between
managers and professionals have to offer? That
literature is surprisingly small. Most of it seems
designed for the personnel department (or, per-
haps, for a generic MBA program). Albert
Shapero's Managing Professional People is typ-
ical. Much is said about how to recruit creative
professionals, how to keep them creative, and
how to evaluate them. Shapero is especially
good on such personnel questions as whether to
keep salaries confidential and how to break in a
new hire. He even provides some useful advice
about encouraging communications between
professionals. But he says virtually nothing
about what happens to the information, designs,
and recommendations professionals generate.
Shapero gives no hint that professionals and
managers might disagree in the way those work-
ing on the Challenger did.9
 601/1183

The one significant exception we found to this

personnel-department orientation in the literat-
ure concerned with managing professionals is
the work of Joseph Raelin, especially The Clash
of Cultures: Managers and Professionals. The
title itself suggests the important difference
between Raelin's work and that of others writing
about relations between professionals and man-
agers. For Raelin, there can indeed be a "clash"
between managers and professionals. Raelin ex-
plains this clash by the difference in culture
between professionals and managers. Profes-
sionals have a code of ethics setting standards
they must satisfy whatever their employer may
think. Professionals, as such, always have loyal-
ties beyond their employer. Managers, on the
other hand, have no such divided loyalty. They
are therefore much more susceptible to organiz-
ational pressures. Raelin therefore urges man-
agers to rely
602/1183
 Page 123

on their professionals for guidance in decisions

with an important ethical component. 10

Yet, even Raelin's work does not help us under-

stand the gap between managers and engineers.
Raelin's own discussion of the Challenger dis-
aster ignores the fact that virtually all the man-
agers involved in the Challenger disaster were
engineers.11 His emphasis on the ethical also
seems misplaced. The disagreement between
managers and engineers on the night before the
shuttle exploded was not explicitly ethical. And
the disagreement Feynman reports is over an
easily calculable fact, the probability of failure.
 604/1183

Chapter 5 (in this volume) takes a different tack,

one closer to Feynman's.12 It stresses the close
relation between the work we do and the way
we think. Because the work engineers do is dif-
ferent from that of managers, engineers may be
expected to think somewhat differently.
However, the exact differences depend on the
specific working environment. Chapter 5 sug-
gests that the working environment at Morton
Thiokol (Feynman's "game") would have made
a certain "tunnel vision" part of how managers
normally thought about risk. Thinking like a
manager rather than an engineer there would,
then, mean giving less weight to engineering
considerations than an outsider might think jus-
tified. The managers, in effect, went blind.
 605/1183

Chapter 5 (like Raelin's work) is a contribution

to the literature on the relation between organiz-
ational structure and ethics. James Waters
makes another suggestive contribution to this
literature. Though Waters's chief example, GE's
price fixing in the 1950s, does not involve a
breakdown of technical communications, there
was a communications breakdown. Waters ar-
gues that seemingly unproblematic aspects of
the organization blocked the normal tendency of
people to oppose conduct they judged illegal,
unethical, or unwise.13

Technical communications seem far from ethics.

Why, then, are we drawn to the ethics literature?
The answer is that technical communications are
often the vehicle for making ethics practical.
 606/1183

Consider that the cases commonly used to teach

business ethics include a surprising number that
seem to involve a breakdown in communica-
tions between managers and engineers of just
the sort that concerns us. Among these cases are
the Ford Pinto's exploding gas tank, the DC-10's
cargo door, Three Mile Island, and BART.
Though Waters's 1978 article focuses on GE's
price fixing, he briefly discusses another work-
horse of business ethicsthe scandal over brakes
Goodrich developed for the Air Force's A-7D
project. Waters notes what is a common feature
of the other scandals as well: the great differ-
ence between the way the senior managers and
the engineers directly involved interpreted cru-
cial events, an apparent failure of middle man-
agers to pass along important information, a
failure in principle avoidable because it arose
predictably from procedure or organizational
structure rather than by accident.14 Engineers
saw serious problems where, apparently,
 607/1183

management, especially upper management,

saw nothing significant.

Our problem is connected with engineering eth-

ics in the same way. Most of the scandals cited
here can (and frequently do) appear in a course
in engineering ethics. Our problem is more than
a problem for business and engineering,
however. There is evidence of a similar gap
between government managers and their engin-
eers.15 There also seem to be analogous com-
munications breakdowns where no engineers
 Page 124

are involvedfor example, between army generals

and their technical staff and between airline
mechanics and their managers. 16

Our problem's connection with business ethics is

nonetheless important. Unlike the literature on
professional ethics, the literature on business
ethics is relatively rich in suggestions for pre-
venting or eliminating the sort of communica-
tions gap that concerns us. For example, Waters
makes five suggestions:17
1. Remove ambiguity concerning organizational
priorities (e.g. by a corporate code of ethics)

2. Include concrete examples in directives con-

cerning what is permitted or forbidden

3. Provide concrete steps for internal whis-

tleblowers (e.g., ombudsman)
 609/1183

4. Develop an appropriate organizational vocabu-

lary (e.g., by organizationwide ethics training
that includes discussion of specific cases likely
to arise in the organization)

5. Launch regular ethical investigations similar

to the annual audit.

Waters also remarks that the problems that con-

cern him seem to arise in relatively hierarchical
organizationsthat is, highly compartmentalized
organizations with a strict chain of command
that makes it difficult for information to flow
"horizontally" (from department to department)
or "vertically" (around a particular manager).

Raelin's recommendations are similar to

Waters's. The only important additions are:
6. Mentorship to help socialize new engineers
professionally
 610/1183

7. Rewards for those who bring in bad news the

organization is better off having (as well as the
usual rewards that go to those who bring in good
news)18

Similar recommendations abound.19 But virtu-

ally missing from the literature is the suggestion
that the organization should explicitly encour-
age professionals to adhere to their profession's
code of ethics, provide in-house training in that
code, or otherwise encourage loyalty to the pro-
fession.20 Even studies of professional diversity
in the workplace are rare.21

We have so far omitted mention of two more

categories of relevant literature. One is the liter-
ature on managing innovation, especially the
classic study by Burns and Stalker.22 That liter-
ature seems to confirm the connection already
suggested, between good ethics and good man-
agement, while offering another perspective on
it.
 611/1183

The other category of relevant literature so far

omitted should have been obvious to us from the
beginning. The Challenger explosion was a
man-made disaster. It was, however, very late
when we discovered Barry Turner's classic work
on manmade distasters. Much of his analysis fo-
cuses on breakdowns of communication, some
quite subtle. Unfortunately, he says little about
prevention.23

Hypotheses

We began this study with the assumption that

business and government tend to treat engineer-
ing as a "staff function" and management as a
"line function." That
 Page 125

seemed safe. The staff-line distinction has been

a relatively stable feature of American business
ever since the middle of the last century when
America's first big businesses, the railroads, or-
ganized on the model of the U.S. Army,
America's first big organization. 24
 613/1183

In its pure form, the division between staff and

line works like this: Engineers (and other pro-
fessionals) are regarded as having special know-
ledge of how to do certain work (drafting,
designing, checking, evaluating safety, and so
on). They answer to a manager, but no matter
how high they stand in the organization, no one
(except perhaps a few assistants) answers dir-
ectly to them. The engineers are not "in the
chain of command." Managers, on the other
hand, whether or not they have technical know-
ledge, are regarded as having special responsib-
ility for deciding what to do and how to do it.
Managers answer to those "above" and com-
mand those "below." Engineers on the staff of a
particular manager provide information, advice,
and technical assistance.25 Engineers are con-
cerned with facts; manager, with decisions. A
historian of technology recently summed up this
"military model" of engineer-manager relations
rather nicely (while assuming it to be an
 614/1183

accurate description of engineer-manager rela-

tions today): ''The organizational structure of en-
gineering today does not encourage practitioners
to ask questions beyond narrowly technical
onesmuch less to raise objections."26
 615/1183

While recognizing that practice is seldom pure,

we assumed that the staff-line distinction would
nonetheless produce a division of labor in which
engineers tended to think about questions one
way while managers tended to think about them
another. In particular, we expected engineers
generally to defer to managers, to present op-
tions and let the managers decide. We also as-
sumed that engineers and managers would bring
somewhat different standards of evaluation to
their work. For example, engineers, adhering to
professional standards of success, would want to
"do things right," even if the added expense or
time required was substantial. The managers
would instead adhere to company standards of
success; they would want to "get things done"in
time and within budgeteven if that meant cutting
corners or taking substantial risks. We expected
this difference in perspective to make the per-
spective of managers at least partially opaque to
 616/1183

engineers and the perspective of engineers par-

tially opaque to managers.
 617/1183

Finally, we began with the assumption that the

current literature on improving communications
between managers and engineers was probably
inadequate. The shuttle program had a complex
system of consultation to ensure engineering
"input" at every step in making any important
decision. That system included much of what
the literature recommended. Information (or, at
least, the paper it was printed on) moved upward
relatively freely, with no one in a position to
block it. Communications between engineers
and managers still broke down on a grand scale,
however, and the result was a disaster no one
wanted. Because the shuttle program did not
seem to differ in any fundamental way from oth-
er undertakings that employ large numbers of
engineers, we assumed that the same thing could
happen in any other undertaking of that sort.
Clearly, then, something more than NASA's
complex system of consultation was needed.
 618/1183

These assumptions lead naturally to the follow-

ing hypotheses:
 Page 126

1. That the boundary between engineer and man-

ager would be relatively clear in most organiza-
tionsso that, for example, an engineer would
know whether or not she had become a manager

2. That engineers would be primarily concerned

with safety and quality while managers would be
primarily concerned with costs and customer
satisfaction

3. That engineers would tend to defer to manage-

ment judgment, because management had ulti-
mate responsibility for decisions (so that, for ex-
ample, one way to improve communications
between managers and engineers would be to
find ways to encourage engineers to be more as-
sertive in their dealings with managers)

4. That the more hierarchical an organization, the

more difficult communications between man-
agers and engineers would be and the more likely
that a communications gap would open
 620/1183

5. That we could develop a procedure for identi-

fying a gap between engineers and managers if
one existed

6. That we could add to the stock of procedures

to prevent a gap from appearing or to help close
it once it appeared

Method

We early recognized that the empirical literature

was inadequate for our purposes in three
respects.

First, little of the literature specifically discussed

engineers. Most of what did discuss engineers
was too abstract to give any feel for how man-
agers and engineers deal with each other day to
day.
 621/1183

Second, the only works that did give such a feel

were the congressional hearings, court cases,
and investigative reporting that scandals gener-
ate. Engineers were, we assumed, likely to err
on the side of safety and quality. Such errors
may hurt corporate profits or even ruin a com-
pany, but they do not produce a public scandal.
Managers, on the other hand, seemed likely to
err on the side of profit or consumer satisfaction.
Because such errors tend to threaten safety or
quality, they are likely to create just the sort of
disaster the public would be interested in. Thus,
the scandal literature, standing alone, seemed
likely to be skewed against managers.
 622/1183

Third, engineers are seldom in a position to pro-

duce an interesting disaster by themselves. Man-
agers have to be involved. When managers are
involved, they have to take the blame, whether
or not they relied on their engineers. It would be
their decision, however poorly they were ad-
vised. Engineering advice thus tends to be invis-
iblewith one exception. When the disaster hap-
pens because the manager did not take the en-
gineers' advice, the engineers' advice suddenly
becomes visible. Why, everyone wants to know,
did the manager not take that advice? It is, then,
not surprising that the scandals getting the most
attention are those where communications
between managers and engineers broke down.
When a manager correctly overrules an engin-
eer, nothing newsworthy happens.
 623/1183

So, we could not rely solely on the scandals lit-

erature for an understanding of how managers
and engineers normally work together. We
needed to investigate directly how engineers and
managers work together under more or less nor-
mal conditions (that is, without the selective
hindsight disaster gives).
 Page 127

We developed one questionnaire for engineers

and another for managers (see Appendixes 1 and
2). We then tested the questionnaires at one
company and made minor revisions, mostly cla-
rifications in wording so that, for example, it
was clear that we were interested in disagree-
ments on "technical" rather than "personnel"
matters. We then interviewed at three more
companies. Only then did we add the starred
questions, preserving the original numbering to
make reference easier.
 625/1183

The questionnaire had four functions: (1) to tell

us what the engineer or manager did, his daily
routine and place in the organization's work; (2)
to tell us what his relations were with manage-
ment (if he was an engineer) or with engineers
(if he was a manager); (3) to help us identify
those practices that contributed to good commu-
nications and those that did not; and (4) to see
whether we could identify a breakdown in com-
munications of the sort Feynman found in the
shuttle program. The questionnaire was de-
signed to structure an open-ended interview last-
ing about ninety minutes.
 626/1183

After developing the questionnaire, we contac-

ted companies that employed engineers. The
smallest employed four engineers (two without
degrees); the largest, more than ten thousand.
Except for one construction company, all were
engaged in manufacturing. They ranged from
companies with relatively benign technologies
such as electronics to a company with a relat-
ively dangerous technology (manufacture of
petroleum-based chemicals); from companies
that are primarily parts suppliers to companies
that produce primarily for end markets; from a
company with one location to several large mul-
tinationals (one of which was closely held).
 627/1183

These companies were not chosen by chance.

Our original budget kept interviewing within the
Chicago metropolitan area. Even after the
budget was revised to allow interviews at two
locations beyond an hour's drive of Chicago, we
were selective. We assumed that few companies
would be willing to let just anyone interview
their employees on company time. We therefore
limited our contacts to companies at which one
of the research group had an "in." 27 The result
of this mode of selection may be a bias in favor
of "good companies."
 628/1183

Perhaps for this reason, one sort of bias we ex-

pected did not occur. We expected some self-se-
lection (even though we promised that our re-
port would name the company only to acknow-
ledge its help or to recommend one of its pro-
cedures). Agreeing to participate meant that the
company had to think what we were doing was
important enough to be worth the time we
would take out of the working day of its man-
agers and engineers. The company also had to
feel comfortable having outsiders probe into
day-to-day operations. Every company we asked
to participate wanted to know what we were go-
ing to do before they agreed. Management saw
our project proposal and both versions of the
questionnaire. We made no effort to conceal our
interest in ethics. Any company without a sense
of social responsibility or without a clear con-
science would, we thought, refuse. To our sur-
prise, not one of the ten companies we contacted
refused, insisted on control over what we
 629/1183

published, or even suggested that the company

be allowed to comment before we published. All
did, however, ask for a copy of the final report.
(We gave each a chance to comment on a draftto
see whether our sense of their technical decision
making fit theirs.)
Page 128
 631/1183

Once a company agreed to cooperate, we (the

research group) indicated that we were not inter-
ested in interviewing just any manager or engin-
eer. We were interested in the "interface
between management and engineering func-
tions." We wanted engineers who dealt with
managers and managers who dealt with engin-
eers. We left it to the company to choose the
managers and engineers to be interviewed. Their
choice seemed determined primarily by who,
among those who would be appropriate, could
be available on the day we were to interview.
Generally, we got to interview a manager and
one or more engineers who worked together
rather than two unconnected individuals. In a
significant number of cases, there were last-
minute substitutions because "something came
up" (for example, an emergency at a distant
plant or a meeting date had changed). Often, it
seems, a company simply asked for volunteers
from among those in the appropriate category.
 632/1183

We never had a sense that we were interviewing

from a "stacked deck."

Small companies had no trouble understanding

what we meant by "manager" and "engineer."
But, to our initial surprise, companies with large
numbers of engineers did. In these companies,
there was no single interface between the engin-
eering and management functions. Two, three,
or even four levels of organization might stand
between employees regarded as "just engineers"
and others regarded as ''just managers." In such
companies, we said we wanted to interview
some from each level, beginning with "bench
engineers" and ending with the first level of
"just managers." For this reason, we conducted
more interviews in large companies than in
small.
 633/1183

All interviews were conducted at the company

on company time and usually within a few feet
of where the engineer or manager worked, either
in a conference room or in a private office. The
only people present during an interview were
the interviewers and the interviewee. We did not
use a tape recorder. Generally, we had two inter-
viewersone to ask questions and one to take
notes. 28 Occasionally, the note taker would ask
a clarifying question. The interviews began with
introductions, an explanation of the interview's
purpose, an assurance of anonymity for the in-
terviewee, and a promise to identify the com-
pany only to thank it for its cooperation or to
point out a procedure others might want to copy.
 634/1183

The interviewer then asked, "Manager or engin-

eer?" This often occasioned a brief discussion
useful in understanding how the organization
thought about engineering. We abided by the
individual's decision. This method had one
troubling consequence. Some "group leaders"
(those who look after the work of four to six
bench engineers) are treated as engineers, while
others with the same responsibilities in the same
company are treated as managers. This is less
troubling than it may seem. We were, after all,
concerned with understanding our interviewees'
work from their perspective. However, we have
taken one precaution against any bias this meth-
od might introduce. Whenever we quote a group
leader while contrasting the perspective of man-
ager with that of engineer, we indicate that the
person quoted is not only an engineer or man-
ager but also a group leader.
 635/1183

Once the interviewee decided that he was a

manager or an engineer, the interviewers
worked from the appropriate questionnaire
(adding a spontaneous question now and then).
Though we tried to get a copy of the question-
naire to each interviewee at least a week before
the interview, about half the interviewees did
not see
 Page 129

the questionnaire in advance. Those who re-

ceived a copy in advance indicated that they had
read it and given it some thought. A few had
even made notes. Our impression is that those
who had the questionnaire in advance tended to
give fuller answers. Otherwise, the answers giv-
en by those who had the questionnaire in ad-
vance did not seem to differ from the answers of
those who did not. No interviewee gave any in-
dication that he had discussed his answers with
a superior.
 637/1183

We interviewed a total of sixty engineers and

managers. All but one were male (indicating, we
think, how few women these companies employ
in engineering work). These sixty represented
all the major fields of engineering: mechanical,
electrical, chemical, civil, and metallurgical.
They included engineers in design, testing, and
operations (both manufacturing and construc-
tion). Not all were trained in the United States.
At least one was trained in each of the following
countries: Canada, Netherlands, what was then
West Germany, what was then East Germany,
Poland, India, and Japan. Most of the foreign-
trained respondents had worked as engineers be-
fore coming to the United States. In the large
companies, the most senior managers inter-
viewed were middle level; in the smaller com-
panies, they were close to the top of the
company.
 638/1183

Initially we expected that engineers would have

engineering degrees and managers would have
management degrees. Although most engineers
in most companies were in fact "degreed," we
occasionally came across an older "engineer"
who had been "promoted from the shop floor."
In one company, however, promotion from the
floor was still common. That was also the one
company in which we interviewed three man-
agers who had neither been trained as engineers
nor worked as engineers. We sought out that
company when we realized our initial sample of
managers consisted entirely of former engineers
(most with a baccalaureate in engineering,
whether or not they held an MBA or other man-
agement degree). Because the common wisdom
is that ''business schools, not engineering staffs,
are [now] the favored sources of managerial ex-
pertise," 29 we were surprised at how hard it was
to find a company with a significant number of
managers of engineers who were not themselves
 639/1183

engineers. We now doubt the common wisdom

on this matterat least for the management of
engineers.

Our interviews cannot provide a complete pic-

ture of the way managers and engineers work
together. What they provide is a part of the pic-
ture different from that given by the scandals or
the existing management literature. Ours is a
study of technical decision making under nor-
mal conditions (or, at least, without the benefit
of hindsight that a disaster brings).
 640/1183

The picture is somewhat fuller than our method

of selecting companies and their absolute num-
ber suggests. Just over a third of our inter-
viewees (ten managers and eleven engineers)
had worked for at least one other employer first.
Several others had worked for another branch of
the same corporate family (in Germany, Japan,
or India). We encouraged these interviewees to
compare their present employer with their previ-
ous one or their employer's practices here with
its practices abroad. This gave us some insight
into a kind of company not officially
represented.

Some of the interviewees had worked for their

present employer for several decades, long
enough to see important changes in relations
between managers and engineers. We encour-
aged those interviewees to compare past and
present. These
 Page 130

comparisions imparted some sense of history to

what would otherwise have been a snapshot of
the present. 30

Evidence
 642/1183

My discussion of evidence has five parts. The

first compares the perspectives of engineers and
their managers. The second distinguishes three
kinds of company according to the criteria em-
phasized in engineering decisions. The third de-
scribes how engineering decisions are normally
made, noting differences related to kind of com-
pany. The fourth considers the effect of an open-
door policy, a code of ethics, and other devices
(including some not in the literature) on how en-
gineering decisions are made. The fifth de-
scribes a breakdown in the normal decision pro-
cess that our questionnaire uncovered, an undra-
matic form of what led to the Challenger
disaster.

Engineers and Managers:

Some Differences
 643/1183

Question 11 on the manager's questionnaire

("Are engineers good management material?")
and the identical question 12 on the engineer's
were designed to encourage interviewees to
compare and contrast managers' and engineers'
ways of doing things. We also expected answers
to question 5 on both questionnaires ("Is the
company's management trained or versed in the
company's technology?"), question 12* on the
manager's questionnaire ("What questions
should an engineer ask you to get the informa-
tion he needs . . .?"), and question 13* on the
engineer's ("What questions should a manager
ask you to get the information he needs . . .?") to
provide useful information about differences
between engineers and managers.
 644/1183

What we found was that the engineers and man-

agers interviewed were virtually unanimous in
the way they distinguished the engineer's per-
spective from the manager's. Both engineers and
managers agreed that some engineers could be
good managers, but they also believed that en-
gineers had to change (and that those who could
not would not make good managers). Three
sorts of change seemed to be involved (apart
from learning how to do budgets, fill out person-
nel reports, and the like).
 645/1183

First, an engineer must pay less attention to en-

gineering to be a good manager. "Letting go of
the hands-on-the-bench engineering was," for
one manager, "the most difficult part for me."
Another in a different company made the same
point: "An engineer [when he becomes a man-
ager] must look at the picture differently and de-
tach himself from the details of the job." An en-
gineer (in that company) made the same point:
"Engineers that can't wean themselves from the
engineering work make bad managers. . . . You
have to learn to let engineers do the engineer-
ing." The most negative comment about
engineer-managers came from an engineer in
another company: "No, engineers aren't good
management materialunless given specific train-
ing. Engineers have trouble giving up control
over every detail.''
 646/1183

Second, not only must engineers give up control

of engineering details, they must, as one man-
ager put it, "develop a broader horizon and look
at the big picture." For
 Page 131

another manager (at another company), that

broader horizon included learning "to think for-
ward, think about others, think in terms of hu-
man resources." Connecting the first change to
the second, another manager put it this way:
"We have to move from reaching the conclu-
sions to guiding the process which reaches the
conclusions."

The engineers could not have agreed more. One

suggested, "The engineer turned manager needs
to appreciate what it takes to implement his pro-
ject . . ., to take cost into account and . . . to
track performance on a weekly basis." Another
in a different company made the same point:
"He must learn to handle responsibility and
learn to get things done through his people. He
can't do it all himself."
 648/1183

Third and more fundamental, the manager must

not only widen his horizon but also change the
character of what he does. "Engineers like to
work with things," as one engineer noted, "[but]
managing is more a matter of people than
things." Or, as another engineer expressed it,
"Socially adept engineers make good managers.
Others should stay away from management."
Managers made the same point. One recalled, "I
had to become much more people sensitive."
Another observed, "You have to build effective
working relations with your people.''
 649/1183

Unfortunately, we did not ask in what ways an

engineer turned manager should not change.
Nonetheless, we did receive some relevant re-
sponses, most from managers. Here again, there
seemed to be a consensus. "[The manager]
shouldn't lose his technical touch," said one
manager. If he does, observed another in a dif-
ferent company, he will become "too superfi-
cial" and "no engineer goes to this type of man-
ager for help." "Technical understanding," ac-
cording to one engineer, "is crucial at times.
What's needed is a fine balance [between tech-
nical understanding and holding on to one's en-
gineering loves], and it is seldom found."
 650/1183

Although most companies at which we inter-

viewed provide some formal training for an en-
gineer turned manager, either in-house or more
often by paying tuition, the general opinion was
that the training was not much help (except in
handling personnel and technical business mat-
ters). A surprising number of both engineers and
managers answered question 11a (for managers)
or 12a (for engineers), "None" or "None, really,"
while others in the same company (often in the
same department) reported such training. One
engineer answered in a way that may explain
this apparent disagreement. Having answered,
"None to my knowledge," he added, "We have a
management training program, but it seems
pretty hokey, so I don't go to it." A manager in
the same company gave an answer that at once
suggests the vast scale of the company's efforts
and the great difficulty of the undertaking. Hav-
ing answered "No transition training," he went
on, "Well, we do have some supervisor
 651/1183

development courses. And an MBA program.

Role models. But that's about it. Nothing that
really prepares an engineer for the transition."
Perhaps an engineer in another company best
expressed the underlying difficulty: "Engineers
know their products but management is a trait."

For most managers we interviewed, the most

helpful preparation for managing was early ex-
perience at the edge of management, for ex-
ample, as group leader, together with a certain
amount of informal coaching. Most managers
seem to be trained "on the job." Those engineers
who can't change enough (or don't want
 Page 132

to)what one manager called "the scholar

type"never get beyond group leader. The rest get
more and more management responsibility (and
less time for engineering) until they become
full-fledged managers.
 653/1183

The transition from engineer to manager then, is

not primarily thought of as the acquisition of
technical knowledge an engineer can't expect to
understand. 31 A manager may indeed know
about matters an engineer does not because the
manager gives his attention to matters an engin-
eer does not (just as the engineer gives his atten-
tion to matters the manager does not). But the
manager's knowledge is in principle as easy for
the engineer to understand as it is for the tech-
nically trained manager to understand what the
engineer knows. The good engineering manager
differs from the bench engineer primarily in be-
ing able to do his engineering through other en-
gineers. So, according to this common under-
standing of management, an engineer and an en-
gineer turned manager should have no more
trouble communicating their respective (technic-
al) concerns to each other than one engineer or
manager has communicating them to another.
 654/1183

This common understanding may explain both

why we found so few nonengineers to be man-
aging engineers and why those few were con-
centrated in production. According to our inter-
viewees, production is that part of engineering
where experience, rather than technical training,
is most likely to be the decisive factor. Even so,
we noticed unusual friction between production
engineers and their nonengineer managers. At
the one company that did have nonengineers
managing engineers, one college-trained engin-
eer told us, "I have to explain to my own man-
ager in `baby talk' since he is not an engineer.
This is frustrating. I pull my hair out when he
repeats my recommendation to his manager
since he presents the recommendation incor-
rectly." One college-trained nonengineer man-
ager confirmed this description (while giving
the manager's side): "Sometimes engineers will
spoon feed me. Then I'll tell them to hurry up.
Or they'll water the information downyou know,
 655/1183

talk about apples and bucketsthen I'll tell them

to talk about engines. Engineers often don't
know how to talk to nonengineers."
 656/1183

What we derived from these interviews was not

so much an impression of a breakdown of com-
munications between engineers and their nonen-
gineer managers as of an inauspicious thinning
out of communication. A lot of important in-
formation seemed to be "lost in translation." We
found something similar in the one company at
which many of the managers were foreigners
struggling to perfect their English. Thus, one
American engineer gave us this example (after
making clear that he thought his manager was a
good engineer): "Let's say we discover a design
change is needed on a local part. I might make
the change myself and put it into operation and
then tell my manager. This is just a simpler way
to go. If I had an American manager, it would
be easier to explain the fine details and involve
him." Good technical communication is surpris-
ingly fragile.

Three Kinds of Company

 657/1183

The companies at which we interviewed seem to

be of two kinds: "engineer oriented" and "cus-
tomer oriented." To these two kinds must be ad-
ded a third, "finance oriented." Although none
of the companies at which we interviewed was
finance
 Page 133

oriented, we did hear about finance-oriented

companies from several interviewees when they
contrasted their present employer with a previ-
ous employer or the way their employer does
things now with the way it used to do them.
Finance-oriented companies seem to be different
enough from engineer- and customer-oriented
companies to be treated separately. 32

An engineer-oriented company is distinguished

by general agreement that quality is the primary
consideration (or, rather, the primary considera-
tion after safety). So, for example, in one such
company, an engineer volunteered, "It is com-
pany religion to seek perfection." A manager in
the same company was equally definite: "We
have overdesigned our products and would
rather lose money than diminish our reputation."
 659/1183

Such companies do not ignore cost but, as one

engineer put it, "Cost comes in only after quality
standards are met." They also do not ignore their
customers, but they are likely to take pride in
how often they say no to them. So, for example,
one manager at such a company told us, "If a
customer wants to take a chance, we won't go
along." An engineer at the same company told
us: "We do actually say no to customers. . . . We
refuse customer applications to exceed our rat-
ings in spite of these often being big-ticket items
where money losses can be significant. We will
negotiate with customers to move them within
our specifications. We very rarely budge from
this posture."
 660/1183

Such a company is not likely to maximize return

on investmentin the short term at least. But it
can be successful by another measure. Each of
the four companies we identified as engineer
oriented told us it had a large and growing share
of the markets in which it competed. Two were
closely held; three (including one of the closely
held) were large multinationals.
 661/1183

We do not call a company "engineer oriented"

because engineers in fact run it. Like Morton
Thiokol, all the companies at which we inter-
viewed had engineers (or "former engineers") at
all levels up to (and sometimes including) exec-
utive officers. Rather, what led us to call some
companies engineer oriented is that their way of
doing business closely fit the stereotype of en-
gineers as concerned primarily with safety and
quality (and of managers as differing from en-
gineers in their greater concern with customer
satisfaction and finance). The companies we call
engineer oriented were therefore ones in which
the engineers felt at home. What was surprising
was that the managers in these companies
seemed to feel exactly the same way.
 662/1183

Still, even in such a company, the expression

"take off your engineering hat and put on your
management hat" would not have been mean-
ingless (even ignoring personnel matters). The
engineers were likely to think the managers
"more cost oriented." Managers, on the other
hand, could still contrast the engineer's tendency
"to go into too much detail" with the manager's
tendency to be "too superficial[to] want only a
`go or no go' decision.''

The contrast with customer-oriented companies

is nonetheless substantial. For customer-oriented
companies, customer satisfaction is the primary
consideration (or, rather, the primary considera-
tion after safety). "The main objective," as one
engineer in such a company put it, "is meeting
the customer's requirements." A manager in the
same company gave this example: "If a particu-
lar batch can't meet
 Page 134

specs, we might call the customer, tell him what

we have and ask whether we should ship any-
way." In place of the engineer-oriented
company's internal standard of quality is the ex-
ternal standard of what the customer wants or is
willing to accept.
 664/1183

In such a company, the engineer's concern with

quality regularly comes into conflict with
management's concern to satisfy the customer.
Consider, for example, the question: Should we
substitute a cheaper material for a more expens-
ive one, making a part significantly less durable,
if the part's probable life is still significantly
longer than that of the machine into which it
will be put? Both engineer-oriented and
customer-oriented companies have to answer
such questions. In an engineer-oriented com-
pany, it will probably be understood as an en-
gineering question, that is, as a question about
how to define quality. In a customer-oriented
company, however, it will probably be under-
stood as a choice between engineering standards
and management standards, that is, as a choice
between quality ("lowering standards") and giv-
ing the customer what he wants ("a cost-effect-
ive solution" to his problem). So, even if the de-
cision is ultimately the same, the dynamics of
 665/1183

deciding will be different (in this respect at

least).

The finance-oriented company resembles the

engineer-oriented company in having an internal
standard of success but resembles the customer-
oriented company insofar as that standard is dis-
tinct from quality. For a finance-oriented com-
pany, certain business numbers (for example,
gross profit or return on investment) are the
primary considerations. Customer satisfaction
and quality are relevant only as means of max-
imizing those numbers. As one former employee
of a finance-oriented company put it, "[The] at-
titude [there] was `we get by with what the cus-
tomer cannot detect.'"
 666/1183

Finance-oriented companies tend to measure

success in tons produced, units out the door, or
other quantities rather than in ways explicitly
acknowledging quality or customer satisfaction.
Although we might expect engineers to prefer
such hard measures to quality or customer satis-
faction, all references to finance-oriented com-
panies were negative or at best neutral. One
manager recalled that "the production process
[there] was driven by a `units out the door' men-
tality which often inhibits quality and cost-ef-
fectiveness." Another manager recalled with ob-
vious pain being asked to make small adjust-
ments in test results (that is, as he saw it, falsify-
ing the data) so that a product could be said to
meet customer specifications and be shipped.
The standard of success in a finance-oriented
company seems to be much more foreign to en-
gineers than that of a customer-oriented
company.
 667/1183

Being a finance-, customer-, or engineer-ori-

ented company is not, like being male or female
in humans, a matter of being more or less per-
manently one or the other. We interviewed at
one customer-oriented company that seemed to
be consciously trying to become engineer ori-
ented. (The engineers reported these efforts with
a tone of "at last," while the managers were
plainly having difficulty adjusting to the new
demand for quality.) We also interviewed at
several companies that seemed to have gone
from finance oriented to customer oriented with-
in the last decade or so. We even interviewed at
two companies which we assigned to the
customer-oriented (rather than to the engineer-
oriented) category only after considerable dis-
cussion. What made these companies difficult to
classify was that one of their largest customers
was pressing them so hard for quality that they
themselves seemed to be
668/1183
 Page 135

uncertain whether they thought quality a mere

means to satisfy a major customer or something
good in itself.

The distinction between engineer-, customer-,

and finance-oriented company is probably best
thought of as a rough topology useful for organ-
izing the data presented here or as specifying
"ideal types" actual companies only approxim-
ate to varying degrees (with perhaps some de-
partments or divisions in the same company be-
longing to one type while others belong to oth-
ers). The distinction has no obvious connection
with that other ideal type, "the technology-driv-
en company." Most of the companies in our
sample probably qualify as technology driven.

Normal Decisions
 670/1183

Questions 3, 4, and 7 on the manager's question-

naire and questions 3, 4, and 6 on the engineer's
questionnaire were designed to tell us who made
the decisions and how they were made. Ques-
tion 8 on the manager's questionnaire and ques-
tion 10 on the engineer's questionnaire were de-
signed to tell us whether the interviewee ap-
proved of existing practice. In fact, questions 8
and 10 (about how decisions should be made)
often led interviewees to modify their descrip-
tion of how decisions were made. Occasionally,
questions 12 and 12* (for managers) and 13 and
13* (for engineers) also led to such
modifications.

Engineer-Oriented Companies
 671/1183

Some of our interviewees initially described

what sounded like a modern version of the staff-
line division between managers and engineers.
For example, one manager told us, "Managers
nearly always make the decisions"; another,
"Managers have the most weight." One engineer
put it this way: "[The engineer] gives the best
advice he can but it's their money." Another told
us that, in case of disagreement, "The boss typ-
ically wins."
 672/1183

However, such comments were largely contra-

dicted by what even these interviewees went on
to tell us about decision making in their com-
pany. For example, the same manager who told
us managers nearly always make the decision
also told us: "If an engineer has a good case, a
manager seldom, if ever, would overrulethat is,
if the engineer really feels it won't work.
However, a manager might step in regarding
costs, customer preferences, or some life cycle
strategythat is, something that is not absolutely
engineering in nature." In the same vein, the en-
gineer who told us the boss typically wins ad-
ded, "I haven't experienced this."
 673/1183

What in fact emerged from our interviews was a

process of "negotiation" (as one manager called
it) much more reminiscent of an academic de-
partment than an army barracks. Engineers' "re-
commendations" were often indistinguishable
from decisions. Managers generally "overruled"
engineers' recommendations only when nonen-
gineering reasons (such as cost or schedule)
seemed to outweigh engineering considerations.
Managers generally let the engineer do the en-
gineering. And even when they "overruled" an
engineering recommendation for nonengineer-
ing reasons, they did not literally overrule it. In-
stead, they presented the additional reasons to
the engineer and sought the engineer's concur-
rence, either by winning him over
 Page 136

with the new information or by seeking some

compromise. Consensus seemed to be the mark
of a good decision; outright overruling,
something to be avoided at almost any cost.
 675/1183

This process of seeking consensus (a better term

than "negotiation") seemed to rest on three as-
sumptions: (1) that disagreement about any en-
gineering or related management question is ul-
timately factual; (2) that when reasonable tech-
nically trained people with the same information
cannot reach consensus on a factual question,
there is not enough information for a good de-
cision; and (3) that, except in an emergency,
putting off the decision until there is enough in-
formation (or a better understanding of the in-
formation available) is better than making a bad
decision. Our interviews suggest that these as-
sumptions are shared by engineers (and
engineer-managers) at whatever kind of com-
pany they work. These assumptions, however,
are likely to be more potent in an engineer-ori-
ented company. There the priority given quality
gives engineering considerations a force they
cannot have when customer satisfaction or "the
numbers" carry more weight than quality.
 676/1183

Whether such considerations as quality or cus-

tomer satisfaction are literally factual is a philo-
sophical question we may ignore here. What we
mean by calling such considerations "factual" is
simply that experience teaches those who dis-
agree about such matters to expect to settle their
disagreements by further testing, other new in-
formation, or reconsidering information already
available. For our purposes, what is important is
that engineers and managers do expect to agree
on questions of safety, quality, customer satis-
faction, and cost even if they do not expect
agreement on anything else.
 677/1183

The power of these assumptions can be seen in

comments like the following. Asked whether he
and his engineers always see eye to eye, one
manager, having answered no, went on to ex-
plain: "There are different ways to approach a
problem. Young engineers are often inexperi-
enced and need to learn from their mistakes.
There are no real differences, though, on matters
of safety and qualitythese are pretty much black
and white." Asked how much weight an
engineer's recommendation should have, he re-
sponded 100% and added, "I've always reached
agreement with my engineers." Another man-
ager informed us that if a manager and engineer
disagree over a major technical decision "engin-
eers and managers go to a boss together . . . The
boss then decides. But we haven't had major
problems here."
 678/1183

Engineers sketched a similar picture. Asked how

engineering decisions were made in his com-
pany, one engineer responded, "I'm handed a
design and asked, `How do we produce this?'
Eventually I make a recommendation. My boss,
a supervising engineer, says yes or no. If he says
no, he gives reasons. If I'm not convinced,
there's no standoff; we just go out and test." The
boss seems to have no more weight in the de-
cision than the engineer. The ultimate arbiter is
another test. Another engineer, the one who said
he gave the best advice he could "but it's their
money," nonetheless reported that he and man-
agement "always see eye to eye in the end." In
fact, he had never been overruled.
 679/1183

This process of reaching consensus seems to

presuppose that engineers and managers have
the same information. Because openness about
technical and related business matters would
seem to be crucial to reaching such a consensus,
what engineers and managers at these compan-
ies report about technical communications is
 Page 137

important. Questions 9 and 10 (on the manager's

questionnaire) and question 8 (on the engineer's
questionnaire) were intended to tell us how open
communication of technical information was.
 681/1183

Managers at engineer-oriented companies were

unanimous that they never withheld technical in-
formation from their engineers. Though evenly
divided about whether their engineers ever with-
held information from them, managers never in-
dicated they thought their engineers' conduct a
problem. One manager's answer may explain
why. Having said that his engineers do some-
times withhold information "to cover up a mis-
take," he added: "Sometimes I need to ask ques-
tions to determine who made a mistake." Anoth-
er manager put the point more gently: "I believe
that engineers never intentionally withhold in-
formation, [but every] person tries to put his
best foot forward." In an engineer-oriented com-
pany, the natural tendency of engineers to with-
hold embarrassing information seems a small
impediment an experienced manager can over-
come with a few probing questions, not any-
thing likely to affect significantly the free flow
of information.
 682/1183

The engineers saw things a bit differently. They

generally agreed that their managers were open
with them. Only one thought there "have been
cases when the boss had information and did not
give itbut never knowingly." On the other hand,
none reported knowingly withholding informa-
tionexcept, significantly, in the company with
foreign managers and American engineers. For
one engineer, the problem was "the other way":
"Usually . . . I provide too much detail to my su-
periors. I have had to learn brevity. But there is
a fine line between too much and too little. I be-
lieve in open communications, and for that reas-
on I don't hold back." Another engineer, while
denying that he ever withheld information, did
admit that "lots gets lost in translation."
 683/1183

In any organization, the ultimate test of open-

ness is bad news. Our interviews at one
engineer-oriented company provided an ex-
ample of how bad news was handled. The ex-
ample gives some insight into how such a com-
pany remains engineer oriented even under the
market's constant pressure to pay more attention
to customers.
 684/1183

A manufacturer of motors for pleasure boats

asked the company to make a part for the
manufacturer's engine that would outlast the en-
gine under normal operating conditions but
would quickly wear out if the engine operated at
full power for very long. A part adequate for ex-
tended operation at full power would have been
much more expensive. Company policy was to
make parts so that they would outlast the engine,
however it was used. So, the engineer in charge
recommended against making the part. After
much back and forth along the chain of com-
mand, the engineer's superior decided to go
ahead, explaining the decision this way: "There
is no safety issue even if the motor fails. There
is no real quality issue, either. Pleasure boats are
never run at full power long enough for the part
to fail. Hence, the part will be cost-effective for
the use it will serve. I do, however, agree there
is at least the possibility of legal liability here
should the engine be misused. So, we must take
 685/1183

care to inform the customer of our concerns in

writing and require him to take full legal re-
sponsibility for the part."

A few years later, the customer sold out to

someone who made towboats, as well as pleas-
ure boats. The new owner promptly put the en-
gine on its towboats. The part would fail after
only a few hundred hours of towing. Legally,
the
 Page 138

company was in the clear. But, because its name

was on the part, it received some complaints.

We heard this story from both managers and en-

gineers. We heard it not only in the department
involved but also in other departments. Each
person who told the story treated it as a caution-
ary tale. The company had taken a risk it should
not have. No one wondered whether the profit
from the deal might have justified the risk or ar-
gued that satisfying the original customer ex-
cused it. The bottom line was that the decision
had harmed the company's reputation. What
could be worse than that? Here was an experi-
ence to learn from, not a skeleton to be locked
away in a closet and forgotten. As one engineer
predicted with evident pride, "We probably
won't do anything like that again."

Customer-Oriented Companies
 687/1183

Decision making in customer-oriented compan-

ies is similar to decision making in engineer-ori-
ented companies. Once again we heard echoes
of the staff-line division in interviews that even-
tually revealed a quite different process. For ex-
ample, a manager who first told us that "engin-
eers lay out options; managers choose," immedi-
ately corrected himself: "Well, managers choose
when the decision involves risks or resources.
Other decisions, purely technical ones, are really
for engineers." Similarly, the same engineer
who initially told us "The manager decides,"
later told us, "If I don't like a decision, I would
go to my boss. I could go to my boss's boss, too,
but I never had to. . . . Technical questions are
talked out."
 688/1183

The search for consensus was again central to

decision making. One engineer at a small com-
pany described decision making quite simply:
"We operate by consensus." A manager in a
large company described the process in greater
detail: "Engineers have high weight on technical
issues. The problem is integrating technical re-
commendations into company interest. Cost.
Marketing strategy. Change in technology. Etc.
It's important that the engineer's recommenda-
tion get out beyond the immediate group. When
he sees how his decision does not fit into the
large picture, he's likely to rethink it."
 689/1183

Despite the basic similarity between decision

making in engineer-oriented and customer-ori-
ented companies, we did notice four significant
differences. Customer-oriented companies
seemed to (1) assign greater importance to the
engineer's role as advocate; (2) place more em-
phasis on nonengineering considerations in de-
cision making; (3) be more explicitly concerned
with safety (even though the technology seemed
no riskier); and (4) have more difficulty main-
taining open communications. Let's consider
these in order.
 690/1183

1. In most of the customer-oriented companies

at which we interviewed, relations between indi-
vidual engineers and managers seemed as good
as at the engineer-oriented companies. Yet, the
managers repeatedly stressed the need for engin-
eers to "hammer" on their recommendations.
One manager at a small company thought that
"an engineer should be willing to go to the mat
if he feels strongly that quality is violated." A
manager at a large company agreed: "Engineers
should never be content to see their professional
judgment superseded. If there's a good reason
for
 Page 139

the manager's decision, the engineer should

agree. If the engineer doesn't agree, something
must be wrong. Everyone should keep talking."
 692/1183

The managers clearly thought of their engineers

as advocates of a point of view which, though
different from their own, had to be weighed
against their ownor rather, integrated with it.
There was no mystery about how the two points
of view differed. According to one manager,
"[satisfying] the customer's needs [involves]
three factors . . .: quality (which is a technical
matter), timing (which is a concern of sales);
and specs/cost." The engineers spoke for the
"technical." A manager at another company con-
trasted his role with the engineers' this way: "It
has to be decided where the line is on a specific-
ation. For example, how `perfect' does
something have to be. I occasionally have to ex-
plain, `Hey guys! It doesn't have to be abso-
lutely perfect.' . . . The customer's needs are the
most basic consideration." Another manager at
this company gave the same picture but in a
phrase familiar from the Challenger disaster:
"The most important factors in company
 693/1183

decisions are business issues: What does the

customer want? What are his expectations?
What can we do to optimize given time and
quality requirements? Often, it's time versus
quality. And then you have to decide which hat
to wearengineer's or manager's."
 694/1183

2. Engineers in most customer-oriented compan-

ies seemed to acceptor at least be resigned tothe
conflict between technical and business consid-
erations. As one engineer put it, "Cost issues are
constraints I can understand." There was,
however, one company in which the engineers
showed no such resignation. This was the
customer-oriented company that seemed to be
trying to become engineer oriented. Here, for
example, one engineer told us: "Technical ques-
tions get short-changed to make schedule. `We
can do it better,' I say, but my manager says,
`No time.'" Another engineer said with evident
disgust, "They'll sacrifice quality to get it out the
door," adding, "Why not do it right the first time
rather than taking a lot of time later to patch up
a system?"
 695/1183

3. In engineer-oriented companies, "safety" and

"quality" were mentioned in the same breath
(when "safety" was mentioned at all). That,
however, was not true in customer-oriented
companies. In customer-oriented companies,
safety had the same absolute priority it had in
engineer-oriented companies, but it was men-
tioned much more often. So, for example, the
same engineer who said quality was sacrificed
to get products "out the door" stressed that he
"never felt safety was being sacrificed.'' Many
engineers also told us that they should have the
"last word" on safety (even though they did not
claim the last word on anything else). Managers
agreed, "It's okay to overrule an engineer's re-
commendation on a business issue. But on
safety, exposure to dangerous materials, etc., the
engineer should have the last word."
 696/1183

4. Given the importance assigned consensus in

customer-oriented companies, open communica-
tions should be as important in such companies
as in engineer-oriented ones. Many of the man-
agers seemed to believe so. Indeed, generally,
they were more emphatic about being open with
engineers than the managers at the engineer-ori-
ented companies. Thus, one manager observed
(in response to question 9), "I never withhold
technical information. That's dumb." Another (at
another company), "Never. That's dangerous." A
third, "There's no need . . . We've got strict rules
on use of information."
 Page 140

Yet, in each customer-oriented company at

which some managers answered in this way,
others reported withholding information relev-
ant to technical decisions. For example, one
manager admitted: "I have withheld proprietary
information, for example, relating to prepara-
tions for a joint venture that might mean using a
different technology." Others, while denying
that they had withheld information, reported su-
periors withholding information from them: "I
should add," said one, ''that engineers are often
in the dark and are subject to last-minute sur-
prises. Our department last year was working on
existing products, things that were familiar. We
were not told about any new possibilities or any
new product challenges. We were provided only
with vague clues. I don't know why."
 698/1183

These managers also seemed more concerned

about engineers withholding information from
them than were their counterparts at engineer-
oriented companies. Thus, one manager repor-
ted, "Engineers tend to give me a rosier picture
than is factual just to continue getting my sup-
port. I try to counteract that by MBWA [man-
agement by wandering around]. This is a lot
more effective than formal performance re-
views." Another manager at the same company
stressed the dark side of such withholding: "Yes,
but it only happens when they don't know
enough to know what to tell. For example, now
and then, a guy gets into trouble and thinks he
can fix it himself. The result is I find out when
it's too late to helpand I get burned, too. That's
happened a couple of times in my career." A
manager at another company put it more suc-
cinctly: "Do they withhold information from
me? When they screw up, yes."
 699/1183

Yet, other managers at these same companies

denied that their engineers ever withheld tech-
nical information from them. One manager was
more cautious: "This is the toughest question on
the list. I've occasionally had the feeling there
was more there than I could see in the engineer's
report."

Engineers gave an equally mixed report on com-

munications. For example, one engineer told us
of a "recent survey" that indicated that "people
believe upper management holds back informa-
tion from the company," adding, "My current
manager does not withhold information from
me." An engineer at another customer-oriented
company admitted to the "feeling" his superior
was withholding technical information. Yet,
most engineers reported that they did not think
their managers withheld technical information
from them.
 700/1183

Interestingly, unlike engineers at engineer-ori-

ented companies, engineers in some customer-
oriented companies did report withholding in-
formation from managers. One observed, "I
have, but I'm not sure it was necessary. I have
withheld a theory or brainstorm until it was
tested to verify it positively. I have delayed bad
news in order to retest first." A group leader at
another company admitted, "I sometimes don't
tell my manager about a decision, if I am
already quite comfortable with it."
 701/1183

Technical communication in customer-oriented

companies thus seems to be somewhat less open
than in engineer-oriented companies. Given how
much these companies differed, the cause of that
tendency is probably complex. Still, two factors
are clearly relevant. First, the relatively greater
importance of business information in decisions
of customer-oriented companies seems likely to
change the nature of withholding such informa-
tion. Even if the same amount of business in-
formation were withheld in a customer-oriented
as in an engineer-oriented company, its with-
 Page 141

holding would be more likely to threaten con-

sensus in a customer-oriented company (where
it would be a more important part of the big pic-
ture). Second, the greater emphasis on the en-
gineer as advocate in customer-oriented com-
panies may itself tempt engineers to engage in
lawyerly tactics. But, whatever the cause, a
customer-oriented company that wants to decide
by consensus will, it seems, have to take more
care to keep information flowing than an other-
wise similar engineer-oriented one.
 703/1183

Perhaps this is the place to note that we found

little in relations between managers and engin-
eers that resembled the ruthless gang culture
Jackall reported. What explains that? At least
two factors may help explain the apparent dif-
ference between what Jackall reports and what
we report. One factor is that Jackall's description
may be true only of companies against which
our method of selection was biased. Our method
of obtaining interviews seems to have selected
in favor of "good companies"; Jackall's method
probably did not. Another factor may be that
what Jackall describes begins above the engin-
eering departments in the companies at which
we interviewed. Interestingly, we did find one
manager supporting this explanation (but
without any example from his experience). A
group leader responded to question 12* (in part):
"Higher managers often become involved in
company politics, however, and may comprom-
ise our engineering values. As managers go
 704/1183

higher up [here], their engineering values be-

come corrupted, in my opinion. No, I cannot
think of any precise example, I am a blank right
now. But, managers can become selfish. They
want to be promoted and will enhance this pro-
spect by focusing on high-visibility projects that
look good."

Finance-Oriented Companies

We do not have enough information in our inter-

views to conclude anything about the normal de-
cision making process in finance-oriented com-
panies. We do, however, have enough to offer
four related hypotheses:
1. Because finance-related information tends to
be centralized in a way customer-related inform-
ation is not, engineers in a finance-oriented com-
pany normally receive less information crucial to
company decisions than in a customer-oriented
company.
 705/1183

2. Because engineers receive less crucial inform-

ation, their recommendations will carry less
weight in finance-oriented than in customer-ori-
ented companies.

3. Because their recommendations carry less

weight in finance-oriented companies than in
customer-oriented companies, finance-oriented
companies are less likely to try to reach con-
sensus with their engineers.

4. Because finance-oriented companies are less

likely to try to reach consensus with their engin-
eers than customer-oriented companies are, they
are more likely both to compartmentalize de-
cision making and to treat engineering as a staff
function. As one manager remembered: "[At his
old finance-oriented employer], a report couldn't
leave the department without the co-signature of
a manager. . . . The engineering function can be
muzzled by a heavy-handed manage-
ment. . . . Management's pressure on engineers
sometimes results in low quality."
 Page 142

The Effect of Various Devices

 707/1183

The literature surveyed in the section "Relevant

Literature" recommended a number of devices
to improve communications within an organiza-
tion to reduce the chance that the organization
would do something wrong. What do our inter-
views tell us about those devices? We found
nothing like Waters's "ethical audit" or Raelin's
awards for bringing bad news. Though one com-
pany had a mentoring program for engineers of
the sort Raelin suggests, no interviewee men-
tioned it. Our interviewees did mention the fol-
lowing devices however: a code of ethics, ethics
training, open-door policy, ombudsman, and re-
duced compartmentalization (including such
things as a technical promotion ladder parallel-
ing the management ladder). Except for reduced
compartmentalization, none of these devices
was common to more than a few companies.
Some occurred at only one. In general, the large
companies were more likely to have adopted
some of these devices than the small were; the
 708/1183

customer-oriented companies more likely to

have done so than the engineer-oriented com-
panies. I have organized evidence concerning
these devices under three headingscodes, ap-
peals, and reduced compartmentalization. Ex-
cept for reduced compartmentalization, none of
these devices seems to have had much effect on
technical decision making. We also came across
one informal procedure, "bringing others in,"
and one formal procedure, "independent tech-
nical review," neither of which was mentioned
in the literature. I discuss them as well.

Codes
 709/1183

At four of the six customer-oriented companies

we interviewed, some of our interviewees
answered yes to question 4.b on both question-
naires: "Does your company have a code of eth-
ics?" Most of those so answering were man-
agers. Their answers were often qualified. And
they often disagreed in important details. Con-
sider, for example, these two answers from man-
agers in the same company. "We have a busi-
ness code of ethicsno gifts, etc." one told us,
"but we have nothing called a `code of ethics'
for engineers, nothing that would, for example,
provide guidance if someone orders an engineer
to change test results." Yet, another manager at
the same company thought not: "Well, there are
policies on . . . e.g. entertainment. But no formal
codeexcept [the CEO's] letters. Nothing written,
for example, on how a technical rep should act
in a customer's plant." In another company, one
manager informed us, "Yes, we have a code of
ethics. We're even going to get a lecture on it
 710/1183

from the legal department tomorrow. But it

doesn't affect engineering work. What matters is
the `spec book.''' Yet, another manager there (a
group leader) answered, "I think we have one,
but I'm not sure. I vaguely remember being giv-
en a pamphlet when I was hired that said
something about all this."

If that is how managers described their

company's code of ethics, what did the engin-
eers say? Most either told us that their company
had no code or responded in some such way as
this: "No real corporate code, just individual
standards. Maybe there's a code of ethics some-
where . . . but I don't know of it."

Because the corporate codes apparently have

little to say about engineering decisions, any
training in those codes could have little effect on
those decisions. We
 Page 143

are then in no position to judge the effect of ap-

propriate ethics training on the way engineers
and their managers make technical decisions. As
far as we could tell, no company has a code ap-
propriate for engineers.
 712/1183

No one at any company at which we inter-

viewed mentioned a professional code as a
guide to decisions, and, as far as we can tell, no
company at which we interviewed had ever cir-
culated or endorsed the code of any professional
society. Our interviews gave us no insight into
why that was so. Our interviews did, however,
suggest that one explanation should be ruled
out. Engineers, it might be thought, did not men-
tion professional codes because they had no
sense of themselves as professionals. Yet, some
of the engineers we interviewed clearly did
think of themselves as professionals. For ex-
ample, one engineer (a group leader at a
customer-oriented company) answered question
12: "Managers of engineers should provide sup-
port, not control. Engineers work on their own.
They like their work. They're professionals.
Look at the engineers here. They work forty-
five hours a week, even though they are paid a
forty-hour salary. They'll sacrifice to get the job
 713/1183

done right. Where they work is too cramped,

overcrowded, and dingy."
 714/1183

We may be left with a mystery, it seems. If the

corporate code of ethics has little to say to en-
gineers about their technical decisions, and few
seem aware of their profession's code of ethics,
why do engineers so uniformly become advoc-
ates for safety and quality? The traditional an-
swer of sociology is "socialization" (in profes-
sional school or on the job). Our interviews tend
to confirm this answer. For example, one engin-
eer at an engineer-oriented company explained
his commitment to safety and quality this way,
"I learned that attitude as part of my profession-
al training." He then added, "But . . . it is [the
company's] attitude too." In fact, most inter-
viewees who had an explanation for the emphas-
is on safety and quality referred only to com-
pany "norms," ''spec book," or other detailed en-
gineering standards developed within the com-
pany. And this was true even of engineers who
made it clear that they had to contend with
 715/1183

managers who routinely wanted to put customer

satisfaction ahead of quality.

Our impression is that the engineers' concern

with safety and quality is too ingrained for most
of them to have a good sense of its origin. What
they do have a good sense of is how pervasive
such concern is in the company in which they
work. This does not mean that, for example,
managers are not more concerned with other
things but that even managers recognize safety
and quality as central considerations. The com-
pany demonstrates this concern not so much
through general pronouncements as through rel-
atively strict adherence to thousands of minute
specifications.

Appeals
 716/1183

The small companies were so informal that re-

course to the "top" seemed routine. As one en-
gineer in an engineer-oriented company put it,
"What would be the point of such a policy [an
open-door policy]? I walk into the office of the
president and vice president every day." But
even this engineer thought it important to tell his
immediate supervisor first: "A manager doesn't
like to hear bad news from outside. So, I first
tell him and get advice." Yet, even in the small
companies, our interviewees could not recall
taking a technical question directly to the top.
 Page 144

Given their informality, it is not surprising that

the small companies seemed to have no formal
appeals procedure of the sort discussed in the
literature. The informality was enough. The
formal procedures were available only in larger
companies (and not always there). The most
common of these formal appeals was the open-
door policy. A subordinate dissatisfied with his
superior's decision could take it to his superior's
immediate superior who would hear him out,
might make subtle inquiries, and might even
change the decision if that seemed justified, all
things considered.
 718/1183

Though this procedure could in principle be

used for any problem, it was in fact much more
likely to be used for "personnel" than for "en-
gineering" problems. One manager in a
customer-oriented company described his
company's open-door policy in this way:
''Engineers can go to my boss and complain.
This happens sometimes, on personnel matters,
primarily. It's never happened on an engineering
question." An engineer gave a strikingly similar
description of the appeal procedure at another
customer-oriented company: "You can go to his
superior. I've only done this once or twice, but
more on personnel than on technical matters."
 719/1183

Only one company had something like a formal

ombudsman. A group leader explained the pro-
cedure: "There is a formal path to use in such
cases. . . . It is a strong way to express your dis-
agreement. It has not been used very often in my
recollection. I can think of just one case in
which an engineer used it. This was a case
where a product was being tested. The engineer
thought that the performance problem was due
to a screw that was not tightened all the way to
ground contact. I hadn't responded quickly
enough to his recommendation, he thought. So,
he used [this procedure] and I had to respond."
The procedure involved filling out a form and
placing it in a special box emptied once a day.
The form is delivered directly to someone in the
general manager's office. The person against
whom the complaint is made is then notified and
has a certain number of days to respond.
 720/1183

Few engineers or managers at this company

mentioned this procedure as a way to appeal en-
gineering decisions (just as few engineers or
managers made much use of any other formal
appeal procedure). Why? Fear of reprisal may
seem the most likely explanation. As one man-
ager put it: "Most managers don't mind. But
there are some around here who would do a guy
in for going over his head." There are, however,
two reasons to doubt this explanation. First, per-
sonnel appeals are less rare even though they
seem at least as likely to lead to reprisals. Se-
cond, few engineers we interviewed expressed
any fear of reprisals. As one group leader at the
same company explained, "Yes, this can ruffle
some feathers. But a manager who indulges in
reprisals doesn't last long."
 721/1183

Perhaps, then, a better explanation for the relat-

ive rarity of appeals is simply that both engin-
eers and managers work harder to reach agree-
ment on engineering questions than on person-
nel questions. They work harder because they
expect to reach agreement. Engineering ques-
tions are, as explained earlier, supposed to be
"factual" in a way personnel questions often are
not. As one manager explained why his
engineer-oriented company did not need a form-
al procedure for technical appeals: "There's no
appeal process. I can't imagine an engineer and
manager not being able to come away with a
solution." Given such expectations, going over a
manager's head is likely to suggest a criticism of
the manager's technical judgment, something
 Page 145

much more likely to ruffle the feathers of an

engineer-manager than a disagreement over a
personnel matter.
 723/1183

There is other evidence for this explanation.

Some interviewees reported an informal proced-
ure easily mistaken for an open-door policy. The
procedure had no name (and may in fact be no
more than a natural by-product of reduced com-
partmentalization). One engineer, at an
engineer-oriented company, described it this
way: "Policy is to discuss the [technical] prob-
lem with your boss. If you can't agree, the two
go up to the next level or bring in more people
who know about the problem. There's no written
policy as far as I know. That's just how we do
it." We found this procedure of "bringing in
more people" in customer-oriented companies
as well. For example, one engineer in such a
company told us, "If I had a [technical] concern
I didn't think was properly resolved, I would
write a note to my boss restating itwith copies to
lots of people, including [his boss's equivalent in
the next department over]. Writing such notes is
not all that uncommon."
 724/1183

This procedure of "bringing others in" seems to

differ from an open-door policy in at least three
important ways. First, it does not seem to be a
formal policy in any company at which we in-
terviewed. No one knew its origin. It was, as
one engineer said, "just how we do it." Second,
no interviewee suggested that bringing others in
would "ruffle feathers" (although several sug-
gested that using an open door would). No one,
it seemed, doubted the benefits of another per-
spective. Third, and perhaps most interesting,
bringing others in seems to be a procedure even
managers can use, for example, when their own
arguments cannot budge an engineer from a re-
commendation they don't like. Asked when an
engineer should have the last word, one man-
ager (a group leader) responded, "Last word?
You can always get a second opinion."

Reduced Compartmentalization
 725/1183

For a century at least, one characteristic of en-

gineering has been the large number of engin-
eers involved in any significant project. The tra-
ditional way to approach an engineering project
was for a senior engineer, the project manager,
to divide the project into small parts, assign each
part to a particular engineer (or engineering
group), send them off, and then assemble the
results as they became available. The engineers
would not be encouraged to coordinate their
work with one another. Coordination would be
the manager's job. Engineers might not even
know who else was working on their project. In
fact, they might have little idea how their small
project fit into the overall work. Especially in a
large company, very little information could
flow directly between engineers. The project
manager alone would know more than a small
part of the overall work. Engineers would have
no choice but to defer to the manager's
judgment.
 726/1183

Burns and Stalker called this form of organizing

work "mechanical." They found many instances
of it in the British companies they studied in the
late 1950s and early 1960s. That highly com-
partmentalized way of managing engineering
was also practiced in the United States. One en-
gineer we interviewed recalled work at a previ-
ous employer (more than ten years earlier):
"There I would often be assigned a job by a P.E.
I never saw and sent him a written report. Occa-
sionally the report came back
 Page 146

with written comments. Usually I had no idea

what happened to it." Though this way of man-
aging engineering may continue in the United
States, we found little evidence of it in our
interviews.

We may distinguish two aspects of compart-

mentalization: vertical and horizontal. Vertical
compartmentalization produces a strict hier-
archy, with one manager having a certain num-
ber of subordinates, each of whom answers only
to her. They cannot go over her head without
her consent. Horizontal compartmentalization
puts up barriers between individuals, groups,
and departments on the same level. For ex-
ample, an engineer might have to ask his
manager's permission before talking to an engin-
eer in another department about a technical mat-
ter, or he might simply have no way to know
who else is doing work he should know about.
 728/1183

As our discussion of appeals suggests, we found

relatively little vertical compartmentalization.
We did find significant horizontal compartment-
alization, however, between major functions
such as sales and design or development and
manufacturing, primarily in the large compan-
ies. So, for example, one engineer in manufac-
turing complained that the engineers in develop-
ment still "throw things over the wall to us"that
is, develop a product without consulting the
manufacturing people about how it is to be man-
ufactured. Sometimes the technically neat solu-
tion causes trouble in manufacturing.
 729/1183

Although we found significant horizontal com-

partmentalization, we also found that every
company at which we interviewed was trying to
reduce it. Answers to questions 12* (for man-
agers) and 13* (for engineers) suggest that man-
agers, rather than engineers, are generally lead-
ing this effort. For example, one manager com-
plained: "I want my engineers to see their job as
involving more than technology. How should
mills relate? Where does what we're doing fit in-
to the [company's] future? They need to ask
more integrative questionse.g., `Who reports to
whom?' or `Who can hold their feet to the fire?'"
A manager at another company gave a different
list in the same spirit: "What don't engineers ask
that they should? Cost? Quality? Time? Cycle
time? Design for assembly? They are ready to
run as soon as they see the specs."
 730/1183

The engineers, in contrast, tended to think their

managers more likely to fail to look into the
engineer's own compartment. For example, one
engineer wanted his manager to ask, "How thor-
oughly did you analyze the problem? Did you
shoot from the hip? How much data and factual
evidence did you collect? Is it repeatable? If you
had more time, what would you do differently?
If you're wrong, what are the ramifications?
What's your second best answer?" An engineer
at another company gave a similar list: "Man-
agers need to ask, How did you reach that de-
cision? What information did you use? Did you
cover all the bases? Substantiate. The question
managers are least likely to ask is, Is this date
realistic? Usually, I'm told a date to be done by,
not asked when I can be done. Often the date
isn't realistic. But no one seems to know
thatuntil the deadline is close."
 731/1183

Overall, what we found was a highly fluid de-

cision process depending heavily on meetings
and less formal exchange of information across
even department boundaries. Managers seemed
to have little control over what information
would reach their engineers. Indeed, they
seemed anxious to get their engineers to hook
up with
 Page 147

others on their own. Their only complaints were

about remaining compartmentalization, espe-
cially the parochialism of their own engineers.

Although we heard many complaints about re-

maining compartmentalization, we also heard a
few arising from attempts to reduce compart-
mentalization. For example, an engineer at an
engineer-oriented company answered question
13 ("If you had full control . . ., what would you
do differently") as follows: "I wouldn't show up
at a field meeting with so many engineers we
outnumber the customer." Such outnumbering
was, it seemed, a common consequence of send-
ing one engineer from each department likely to
be involved in a particular project. The most
common complaint of this sort was simply "too
many meetings."

Independent Review
 733/1183

Several companies at which we interviewed had

a "technical review" in which a project group,
section, or department had to defend its proposal
to a committee of experienced engineers (and
managers) from elsewhere in the organization.
These reviews seemed to vary considerably in
formality (as well as in other respects). The
most elaborate we found was the HAZOP (haz-
ard and operability) study used by Amoco
Chemical. Designed for a particularly unforgiv-
ing technology, HAZOP is probably too elabor-
ate for most engineering undertakings. Even so,
it provides a standard against which other com-
panies can measure their own review
procedures.
 734/1183

Amoco uses HAZOP to evaluate both proposed

and existing installations. Because these two
uses differ significantly, I discuss them separ-
ately, beginning with proposed installations. For
a proposed installation, a department works out
a complete plan (which, for Amoco, routinely
includes having the plan reviewed by opera-
tions, maintenance, and installation, who are
supposed to work as carefully as they would if
they, not the HAZOP study, were the last step
before construction of the installation would
begin).
 735/1183

Once a department does all it can, including re-

ceiving approval (and funding) through the or-
dinary process, a HAZOP team is appointed, in-
cluding a leader and a secretary. The team
should consist of engineers experienced enough
to "look at paper and know what that implies
[about how a plant will run]." (One manager set
the required experience level rather high, at
"twenty-thirty years," but we interviewed one
engineer with eight years of experience who had
already served as a HAZOP leader.) No one in-
volved in the original design is on the HAZOP
team. Following a "formalized procedure," the
team examines every aspect of the proposal,
identifies possible flaws in the design, and
makes recommendations as it sees fit. The sec-
retary takes down all recommendations, ulti-
mately sending one copy to those who de-
veloped the proposal and filing the other "down-
town" (that is, at Amoco's corporate offices).
This usually takes "one to four months.''
 736/1183

Once the HAZOP team completes its work, a re-

sponse team is appointed, including only one
member of the original HAZOP team. Appar-
ently, for some projects at least, the response
team may consist entirely of managers. It is sup-
posed to respond to each HAZOP recommenda-
tion. (There may be several hundred.) Its re-
sponse will also be filed downtown. Ordinarily
all, or almost all, the HAZOP
 Page 148

recommendations are incorporated into the ori-

ginal plan. Any rejections must be justified in
the written response. Once the review is com-
plete, the project proceeds. (We received no in-
dication what would happen if a recommended
change put a project over budget.)
 738/1183

That is the procedure for HAZOP review of a

proposed installation. HAZOP can also be used
to review an existing installation. The chief dif-
ference is that the result of such a HAZOP re-
view will be recommendations piled on one or
another engineer's desk beside ordinary work or-
ders with which they will have to compete. This
seems to be HAZOP's Achilles heel. Several
managers told us that had Union Carbide used a
HAZOP procedure to design its plant in Bhopal,
"Bhopal" would not now be a household word.
They could well be right. But, had HAZOP
come into use only after the Bhopal plant was
built, the HAZOP study might only have pro-
duced a series of recommendations which,
though accepted by everyone, would, at the time
of the disaster, still have been sitting on the desk
of an engineer too busy "putting out fires."
 739/1183

Nonetheless, we recommend something like a

HAZOP review even for existing operations. Al-
though the review cannot guarantee that
everything recommended will be done, it seems
likely at least to call attention to important flaws
in existing installations. It can set an agenda.
For most companies, perhaps, identifying seri-
ous problems in some such constructive way as
this is at least half of ensuring that the problems
will be resolved relatively quickly.

A Breakdown in Normal Communications

 740/1183

Company B (as I shall call it) is a large

customer-oriented manufacturer. As at the other
companies at which we interviewed, engineers
and managers at Company B generally worked
by consensus. Company B had no open-door
policy, ombudsman, or other formal appeals
procedure. While going over the boss's head was
generally considered a bad idea, the company
did have frequent "review meetings" at which
technical disagreements could be aired. These
seemed to provide an important forum for
"bringing others in." Bringing others in was also
done more informally. Like other large compan-
ies at which we interviewed, Company B was
working hard to improve communications
between engineering functions, especially
between development and manufacture. And,
like our other customer-oriented companies,
Company B had a code of ethics with little rel-
evance to engineering. Over all, then, Company
 741/1183

B would seem to differ in no fundamental way

from other companies at which we interviewed.

Yet, Company B clearly did differ. It seemed to

have a communications problem much like the
one Feynman reported at NASA. The evidence
for this claim may be divided into four categor-
ies. Company B differed from other companies
we interviewed in (1) the way managers and en-
gineers felt about each other, (2) the amount of
information managers withheld, (3) the promin-
ence of "top-down engineering," and (4) the way
management chose to encourage an important
development project.
 Page 149

Managers Versus Engineers

Unlike most of our interviewees, those at Com-

pany B had little doubt about whether they were
engineers or managers. Thus, one manager told
us, "I was an engineer until a few weeks ago.
[Then] I was promoted to chief engineer."
Though his former job as supervising engineer
also involved a good deal of managing, he did
not then consider himself a manager. Another
interviewee spoke with equal assurance even
though a demotion had not changed his job at
all: "I was a manager before a recent reorganiza-
tion flattened the organization a bit. Now, I'm
not, though I was a group leader before and after
reorganizationwith the same responsibilities as
before."
 743/1183

Why were interviewees at Company B so much

clearer than our other interviewees about wheth-
er they were engineers or managers? The answer
is that Company B made the distinction sharp
and important. As one engineer explained,
"[Company B] gives better benefits to managers
than engineers." Another engineer made the
same point, while suggesting one disadvantage
of making the distinction so sharp: "Around
here many things, including fringe benefits and
office space, treat engineers one way and man-
agers another. The differences between engin-
eers and managers are emphasized. They are in
separate camps." This engineer also referred to
engineers turned manager as "former peers." We
heard nothing like this at any other company.

Withholding Information
 744/1183

Engineers and managers at Company B agreed

that engineers did not withhold technical in-
formation at all or only did so just long enough
to double-check it. Their answers were similar
to those at other companiesexcept that there was
no mention of engineers trying to cover up mis-
takes. Managers at Company B were also like
managers at other companies in being unanim-
ous that they did not withhold technical inform-
ation from their engineers. Where Company B
really differed from other companies was in the
answers engineers gave to that question.
 745/1183

The engineers were virtually unanimous in re-

porting that their managers did withhold tech-
nical information from them. For example, one
engineer told us that he "frequently felt that they
didn't tell me the whole story." When we asked,
"When?" he responded: "Whenever I couldn't
reach their conclusion on the same facts."
Another engineer gave this list of technical in-
formation managers were withholding: "[In-
formation] about cost. Proprietary information,
too. That is, either information that's not directly
relevant to what I must do or is too sensitive to
risk leaking.'' Another engineer added this ex-
ample: "We are reasonably sure that there is a
potential big overseas buyer for the [new tech-
nology we are working on], but no one is level-
ing with us about it. We are in the dark on this
and I don't like it." Yet another engineer de-
scribed a different sort of withholding: "Some-
times they want us to be underinformed, maybe
so as not to prejudice us or they don't think we
 746/1183

need to know how bad a problem it ise.g., that

we had the same complaint before and thought
we fixed it. More often, it's not deliberate. They
just don't see
 Page 150

the relevance of the information, e.g., they have

divided a problem into parts that are too small."
 748/1183

What explains this difference in the way engin-

eers and managers answered the question about
whether managers withheld information? It is
worth recalling that at other companies we inter-
viewed, the managers stressed how important it
was for engineers to include business considera-
tions in their engineering decisions. At Com-
pany B, managers never made this point. In-
stead, the engineers did. Indeed, the engineers at
Company B seemed to accept the broad concep-
tion of engineering that managers at other com-
panies were encouraging their engineers to ad-
opt. But, at Company B, the managers did not
seem to share that conception. They withheld in-
formation that managers at other companies did
not withhold. They withheld it simply because
they did not think it relevant to technical engin-
eering decisions and seemingly without realiz-
ing that their engineers did not share their con-
ception of what engineers need to know.
 749/1183

Top-Down Engineering

"Top-down engineering" does not mean typical

management functions like setting a general de-
velopment strategy, standard of quality, or even
the timetable for a particular project. It refers to
something much more specific: management's
involvement in the details of engineering. At
other companies at which we interviewed, both
engineers and managers thought that managers
should leave the engineering to the engineers.
We heard that thought expressed at Company B,
too, but with this difference: The target was not
(as at other companies) primarily the low-level
manager clinging to his "former love." The tar-
get was management generally, especially upper
management.
 750/1183

Company B apparently has a history of "engin-

eering from the top down." An engineer em-
ployed there for almost a decade recalled, "It
used to be that engineers didn't count for much.
The manager ruled or overruled. This happened
too often in the past. We would be told, `The
data must be wrong.' Some management guys
don't have an open mind, but this happens a
small percentage of the time now. Engineers are
being more encouraged. We've become more in-
volved. There has been more delegation and
that's good."
 751/1183

While that engineer stressed how much worse

things used to be, others stressed how bad they
still were. One told us, "If my recommendations
fit a pet theory, then they are acknowledged. If
not, I have to make one hell of an argu-
ment. . . . There was this case with the
seal. . . . My manager said, `If you think about
it, this seal should really work.' But it didn't
work. Then the idea was changed from let's see
if it works to how much leakage is O.K. in using
the seal. This adds to the costs and the risks, but
we're going ahead with it anyway." Another en-
gineer told a similar story: "[Recently,] we were
looking at two nozzles for spraying fuel into a
cylinder. We could get a 10% improvement with
almost no cost one way; or a possible 20% im-
provement doing it another way, but at consider-
able cost in redesign and no guarantee it would
work. [I recommended the 10% improvement.]
Management, not my boss, but someone higher
up, decided to go for the 20%."
752/1183
 Page 151

The engineers at Company B were virtually un-

animous, "Managers here still try to do too
much of the engineering." We heard the same
thing from two newly promoted managers. One
described decision making in his company this
way: "Managers provide proper manpower and
tools and work with engineersexcept sometimes,
when there is a management-driven decision.
Then the manager gets into the engineering it-
self. That's bad." The other new manager de-
scribed the appeal process this way: "There is no
formal process. After all, technical disagree-
ments are hardly ever dramatic. If I'm unhappy,
I just keep trying to change my boss's mind. I
try to wear him down. [But] sometimes the heat
comes from on top. We are told, `Consider this
design. Look at it. Tell us in detail what you
think.' Then the process feeds back up. These
top-down things give us the most trouble."
 754/1183

What he meant by trouble became clear when he

answered our question about what he would do
differently if he had full control of engineering:
"[There] are too many projects initiated by top
management, from the vice president on down. I
would prefer this changed. It creates tension at
lower levels because of the mode of introduc-
tion. People are told to do this, to do that, that
this is what we need. This creates bad feelings
and destroys creativity. These top-down actions
can be very specific and detailed. They can take
the form of designs and actual sketches. `This is
what we want.' All this is out of control, in my
opinion."
 755/1183

Though the new managers complained about

top-down engineering as much as the engineers
did, the more experienced managers did not.
They did, however, answer question 3 (about
how decisions are made) in a way confirming
what the engineers told us (and suggesting the
managers were on the engineers' side). "De-
cisions are made top down," one manager
began, "[but] we would like to push decisions
down the organization." Another made the same
point: "[Engineers] play as big a role as we
think they can handle. This is easier said than
done. I try to push decisions as low as possible.
But you can't delegate then not stay in touch. I
like to have engineering decisions or firm re-
commendations made at the project manager,
group leader, or bench engineer level."

Unintentionally Discouraging Bad News

 756/1183

Company B undertook a major technological

initiative on which its future success may de-
pend. The initiative was not simply a research
and development project. The assembly line was
being prepared simultaneously. Because the pro-
ject involved a major leap in technology, many
parts of the project were crucial to its success.
Had any of these parts proved impossible to de-
velop, or impossible to develop in time or eco-
nomically, Company B would have had nothing
marketable for its large investment.
 757/1183

Both the engineers and managers we inter-

viewed were involved in work on one or another
part of this new technology (as well as on other
projects of more immediate concern). The en-
gineers (and new managers) mentioned work on
this new technology frequently, especially when
asked whether Company B took large risks. The
experienced managers never mentioned the new
technology. One even denied that Company B
ever took risks, even after we restated the ques-
tion to include
 Page 152

financial risks. We seem, then, to have un-

covered a disagreement between managers and
engineers at Company B similar to the one
Feynman reported at NASA. The engineers re-
cognized risks that the seasoned managers did
not.

Of course, the similarity is not perfect. The risk

at Company B concerned its financial safety, not
anyone's life or health. Engineers differed from
managers on what would seem to be a question
of business rather than engineering. And, unlike
the engineers at NASA, the engineers at Com-
pany B had not yet been proved right.
 759/1183

Such differences are irrelevant now, however.

Even if the engineers were wrong, Company B
would still have had a communications problem.
As one engineer noted, "[There is] too much ru-
mor around here. . . . I wish the managers here
would just admit the problems we're having with
[the new technological initiative] and tell us
how they hope to respond." Clearly,
management's message was not getting through.
 760/1183

There are at least two reasons to think the engin-

eers were right. The first is obvious. Whether a
new technology can work is itself a technical
question about which engineers are likely to be
better informed than anyone else. Or, as one
group leader at Company B put it, "Guys at my
level know all the problems. But there's a filter-
ing process upward. Lots of things we don't tell
unless asked." Insofar as the financial risk the
company was taking is itself a function of that
technical risk, the engineers would be likely to
have good information even about the business
risk involved. The second reason to think the en-
gineers were right is that management itself took
strong action to move the project along. Unfor-
tunately, the action management took seemed
likely to have the opposite effect. The action
also suggests how large a gap divided engineers
and senior management.
 761/1183

Company B called in a management consultant,

"Doctor Feel-Good" ("DFG"). DFG tried to spur
creativity through a motivation program for
managers and senior engineers. The engineers
who told us of DFG's program made it sound
like a series of pep rallies. Whatever the pro-
gram was in fact, the engineers tended to think it
sillyor an admission of management's despera-
tion. DFG clearly was hurting engineering mor-
ale (just the opposite of what management must
have intended). But, perhaps more serious, DFG
also seemed to be damaging communications
between engineers and managers, the very com-
munications upon which any technological
breakthrough depended. "[The] effect of
[DFG]," said one engineer, ''has been to make
engineers feel out of step when they report that
something won't work."
 762/1183

The flow of bad news upward had at least been

slowed. Senior management might well be the
last to know how bad things were; they might
not find out until it was too late to do anything
about it. Perhaps our interviews already show
the filtering process at work. Though the engin-
eers talked openly about the technical bottle-
necks, shortage of staff, and the business risks,
experienced managers did not. They were posit-
ive about what they were doing. If they were no
more open with their engineers and fellow man-
agers than they were with us, it is easy to see
how they might help create, however uninten-
tionally, an environment in which even engin-
eers would feel pressured to tone down bad
news. As time went on, these managers, and
those above them, would be more and more cut
off from what the ordinary engineer knew.
 Page 153

Here is Feynman's "game"with this difference.

The disaster was yet to come and might never
happen. The company might yet "luck out." But,
even if disaster struck, it would not produce a
public scandal, only a lot of red ink or, at worst,
the ruin of a good company. 33

Conclusions

Our first hypothesis was that the boundary

between engineer and manager would be relat-
ively clear in most companies because the staff-
line mode of organization would force the dis-
tinction to be made clear. In fact, we found al-
most no trace of the staff-line distinction. What
we found in its place was something much more
like the distinction made in universities between
faculty and administrators.
 764/1183

In most universities, senior administrators (pres-

ident, vice presidents, and deans) hold faculty
appointments. Many still do some teaching.
Ordinary faculty, on the other hand, do consid-
erable administrative work, whether as depart-
ment chair or through various departmental, col-
lege, or university committees. Faculty differ
from senior administrators only in degree
(though "administrative staff" are more like
what engineers call technicians). Some ordinary
faculty may be paid more than any administrat-
or, even the university president.
 765/1183

In most companies at which we interviewed, the

distinction between engineers and managers was
similarly one of degree. A bench engineer was
an engineer who spent most of his time at his
bench (like an "ordinary faculty member"). A
pure manager was an engineer who no longer
did any engineering himself. Especially in large
companies, there might be several grades of
engineer-manager. In general, the distinction
between engineer and manager did not seem to
determine pay, benefits, or weight in technical
decisions.
 766/1183

The one company that seemed to make the dis-

tinction between engineer and manager as
sharply as we originally expected did not seem
to have any more of a staff-line organization
than the other companies at which we inter-
viewed. Yet, though only for accounting pur-
poses, the sharpness of the distinction seemed to
hurt relations between its engineers and man-
agers, making engineers feel as if they and man-
agers belonged to "separate camps." This bad
feeling may have contributed to the poor com-
munications we found there.
 767/1183

Our second hypothesis was that engineers would

be primarily concerned with safety and quality
while managers would be primarily concerned
with costs and customer satisfaction. This hypo-
thesis was generally confirmed but in a way
suggesting that the concerns overlap more than
commonly thought. Managers in most compan-
ies usually paid more attention to costs and cus-
tomer satisfaction in their initial response to an
engineer's recommendation than the engineers
initially did. In all companies at which we inter-
viewed, however, decision was generally by
consensus, not by management fiat. Decision by
consensus required managers to inform engin-
eers about considerations of costs and customer
satisfaction they may have overlooked. No
doubt as a result of that, most engineers we in-
terviewed had a much better appreciation of
such business matters than we expected. Even
allowing for the fact that most managers we in-
terviewed were trained as engineers, decision by
768/1183
 Page 154

consensus seemed to have a corresponding ef-

fect on managers. They seemed to have a better
appreciation of engineering considerations than
we expected. Decision by consensus itself ap-
pears to be an important means of maintaining
good communications between engineers and
managers.

Our third hypothesis was that engineers would

tend to defer to management judgment because
management had ultimate responsibility for de-
cisions. This hypothesis derived from our as-
sumption that engineering would be treated as a
staff function (with no responsibility for de-
cision) while management would be a line func-
tion. Yet, the hypothesis was in fact independent
of that assumption. It could have been con-
firmed even if, as it turned out, engineering was
a line function. Engineers could still have
routinely deferred to management.
 770/1183

Our findings here are therefore significant in

their own right. Deference to management was
not what was expected of engineers. Quite the
contrary. Engineers were expected to "go to the
mat" on any question of safety or quality they
considered important. Even managers who ex-
pressly reserved the right to overrule an engin-
eering recommendation emphasized the need for
engineers to "hammer" at them anyway. Engin-
eers themselves expressed no deference to man-
agement on questions of safety. There they ex-
pected their recommendation to be "final." Only
on questions of quality, customer satisfaction, or
cost were they willing to let management have
the last wordand, even then, they were willing to
give management an ''earful" first. Here again
the analogy with decision making in a university
(where faculty "advise" but expect to have ad-
ministrators take their advice) seems much
closer than decision making in the military
 771/1183

(wherewe are toldofficers "command" and "sub-

ordinates" are expected to "obey").
 772/1183

Our fourth hypothesis was that the more hier-

archical organizations were more likely to suffer
a communications breakdown than the less hier-
archical. This hypothesis, like the previous one,
was derived from the assumption that the com-
panies at which we interviewed would have a
traditional (quasi-military) hierarchy. Though
their tables of organization made them look as
hierarchical as we assumed they would be, none
of the companies at which we interviewed was
in fact organized in that way. The small com-
panies were too personal for formal hierarchy to
matter much. Even in the large companies, the
use of consensus and bringing other people in
meant that individual managers could not con-
trol information or access in the way they would
have in a traditional hierarchical organization.
(And, in addition, the managers generally did
not want to.) Even the communications gap we
found in Company B did not result from hier-
archical organization but from a combination of
 773/1183

other factors, including too narrow a definition

of engineering considerations, too much inter-
ference from the top in the details of engineer-
ing, a failure to consult directly with those most
likely to know, and the use of motivational tech-
niques likely to discourage the reporting of bad
news. The absence or presence of a code of eth-
ics or formal appeal procedure seemed to have
little part in technical communications between
managers and engineers.

Our fifth hypothesis was that we could develop

a procedure for identifying a communications
gap between engineers and managers if one ex-
isted. We now have some support for this hypo-
thesis. Our open-ended interview identified
what seemed to be a serious communications
gap at one company (Company B). The
interviews
 Page 155

also provided us much useful information about

how engineers and managers generally work
together.

Our sixth hypothesis was that we could add to

the stock of procedures for preventing a commu-
nications gap or at least to procedures for help-
ing to eliminate such a gap once it appears. We
came across two, the informal bringing others in
and the formal technical review.

Recommendations

I (and the research group) believe this research

justifies the following recommendations:
 775/1183

1. Companies should try to soften the distinction

between engineer and manager as much as pos-
sible. Too sharp a distinction (as in Company B)
seems to create resentment that can interfere
with communication. Providing for a promo-
tional ladder for bench engineers parallel to
management's may help reduce the feeling that
managers are "above" engineers. Managers, es-
pecially, seem to welcome the possibility of
bringing in a senior engineer (that is, a "technic-
al person" with rank equivalent to "manager")
when they disagree with an engineer's recom-
mendation. Companies should also look for oth-
er ways to treat engineers and managers as pro-
fessional employees, differing only in specific
function and responsibilities (for example, by
avoiding differences in benefits based on classi-
fication as ''manager" or "engineer").
 776/1183

2. Engineers should be encouraged to report

bad news. Communication is most likely to
break down between engineers and managers
when procedures or other aspects of the work
environment discourage engineers from report-
ing bad news (for example, design problems).
Top-down engineering may be justified at times,
but it should be accompanied by on-site visits
with the bench engineers doing the work ("man-
agement by walking around"). Senior manage-
ment needs to remember how much bad news is
likely to get filtered out by several layers of
management. Senior managers should also be
wary of motivational techniques that discourage
bad news or otherwise inhibit the give-and-take
that is a precondition of decision by consensus
working well.
 777/1183

While on-site visits, especially informal surprise

visits with bench engineers, can undercut the au-
thority of mid-level managers, that is not a ne-
cessary consequence. Undercutting can be
avoided by open discussion of the rationale for
the visit, emphasis on the helping (rather than
the controlling) role of managers, and (when a
problem is discovered) a focus on solving the
problem rather than finding someone to blame.

3. Companies should check now and then for

signs of trouble in relations between managers
and engineers. Such trouble may not be obvious
to managers inside the company even if it is ob-
vious to the engineers there. How many subor-
dinates tell a superior more bad news than he
asks for? One way for senior management to
discover trouble is to meet informally with small
groups of bench engineers and ask. Another way
is to have outsiders interview engineers and
managers in the way we have.
778/1183
 Page 156

4. Companies should encourage both engineers

and managers to settle technical disagreements
by informally bringing other experts in. Com-
panies should also consider adopting an open-
door policy, ombudsman, or other formal ap-
peals procedure. Though such formal proced-
ures are seldom used to settle technical disagree-
ments between managers and engineers, they
nonetheless seem to help establish an environ-
ment in which even technical information flows
more freely.
 780/1183

5. Companies should look for formal procedures

that bring out bad news that might otherwise be
missed. The most effective procedure of this sort
we came across was Amoco Chemical's HAZOP
study. Though this procedure is probably too
elaborate for most companies (that is, those with
a less dangerous technology), it may provide a
useful ideal against which any company can
measure its own technical review procedures. Of
particular value, is that: (1) the reviewing body
consists entirely of engineers who, though hav-
ing the appropriate experience, have had no part
in developing the plans (or process) they evalu-
ate (and so, no built-in conflict of interest); (2)
the plans have to stand on their own (the drafters
not being there to defend them); and (3) all re-
commendations are put in writing, rejecting a
recommendation requires a written justification,
and both recommendation and rejection are kept
on file (thus assuring later accountability). Such
an independent review gives everyone directly
 781/1183

involved in a project considerably more incent-

ive than they would otherwise have not to play
down bad news in the early stages of a project.
At a minimum, however, we think companies
should encourage engineers to put their doubts
in writing and circulate them among all those
concerned.

6. Companies should not expect a general code

of ethics to have much impact on engineering
decisions. Any company that wishes to make
safety or quality more central in its engineering
decisions probably has to do so through specific
technical specifications. It may also find training
engineers in their profession's code helpful, be-
cause these codes are generally more specific
about problems engineers face than is a general
business code. Such training may also confirm
engineers in the belief that their employer wants
them to be advocates for engineering standards.
 782/1183

7. Companies should try to improve the way

they use bad news. Companies cannot learn
from their mistakes if they do not remember
them. In particular, companies should consider
including information about how parts failed in
technical manuals (or data bases) engineers use
or, at least, bring engineers together from time
to time to discuss failures they have learned
from.

8. Technical engineering courses should include

more about the place of cost, manufacturability,
and other business considerations in engineer-
ing. One manager in fact told us that, except for
the graduates of co-op programs, engineers fresh
out of college were poorly prepared to think
about the range of considerations routinely part
of good engineering. There seemed to be gener-
al agreement that engineering education is now
too narrow.
 783/1183

9. Engineers should be trained to make a case

for their recommendations. Ability to present
data clearly, orally or in writing, and the ability
to make arguments from the data seem to be es-
sential to participating effectively in decision by
consensus. Right now, engineers seem to have
to learn these skills on the job. They are,
however, skills any school of engineering can
teach.
 Page 157

10
Professional Autonomy:
A Framework for Empirical Research
 785/1183

Employed engineers sometimes claim that their

status as employees denies them the autonomy
necessary to be "true professionals." Such
claims also appear in important scholarly work.
For example, in The Revolt of the Engineers,
Edwin Layton observed: "Employers have been
unwilling to grant autonomy to their employees,
even in principle. They have assumed that the
engineer, like any other employee, should take
orders. . . . [But] the very essence of profession-
alism lies in not taking orders from an employ-
er." 1 What are we to make of the claim that pro-
fessionalism is inconsistent with being an em-
ployee (or, at least, with taking orders from an
employer)? Is it a conceptual truth (a deduction
from definitions) or an empirical one? How
might it be provedor disproved?
 786/1183

The purpose of this chapter is to answer these

questions by developing a conception of profes-
sional autonomy. We have no such conception
now (though some claim otherwise). What we
have instead are conceptions of personal (or
moral) autonomy applied to the workplace.
These conceptions help us understand neither
the specific contribution of professional con-
straints to personal autonomy nor the ways cor-
porate organization and professional responsib-
ility might be consistent.
 787/1183

The literature on autonomy may be divided into

three categories: (1) a general philosophical lit-
erature on "personal" autonomy, (2) a philo-
sophical literature explicitly concerned with
"professional autonomy"; and (3) a sociological
literature concerned with autonomy in the work-
place. These three literatures, though related in
principle, seem in practice to have grown up
largely independent of one another. Let us con-
sider them in order. Having seen what they have
to offer, we should be in position to develop our
own conception of professional autonomy and
see what research, if any, it suggests.
 Page 158

Personal Autonomy

Over the last thirty years, philosophers have de-

veloped a substantial literature on "personal
autonomy" in part at least because traditional
conceptions of liberty or freedom seemed not to
do justice to concerns about the effect that
brainwashing, hypnotic suggestion, advertising,
and other mind-altering techniques might have
on the moral authority of agents. Brainwashing,
for example, leaves us at liberty to do what we
wantbut, by distorting what we want, seems to
deprive the resulting decisions of something ne-
cessary for them to be accorded the respect to
which they would otherwise be entitled.
 789/1183

The locus of most of this work was political

philosophy (the autonomy of citizens given the
state's control of education, information, and
public discussion), medical ethics (the
autonomy of patients), and business ethics (the
autonomy of customers subject to advertising).
This work went on under the title "personal
autonomy" to distinguish it from two older sub-
jects: "political autonomy," which is concerned
with what makes a state or nation self-govern-
ing, and "moral autonomy," which is concerned
with the conditions for moral responsibility (or
moral goodness). The reason for distinguishing
personal autonomy from political autonomy is
self-evident: Individual persons are neither
states nor nations. The reason for distinguishing
personal autonomy from moral autonomy is not
self-evident.
 790/1183

On many conceptions of moral responsibility-

indeed, on mostthe conditions necessary for
moral autonomy (for example, rationality) are
also conditions necessary for personal
autonomy. The term personal autonomy must, I
think, be treated not as signaling a subject dif-
ferent from moral autonomy but as signaling a
different emphasis within the same subject.
Those concerned with personal autonomy are to
be understood as concerned primarily with pro-
tecting agents (whether in political philosophy,
medical ethics, or business ethics) from certain
undesirable influences, or with reasons for re-
specting their decisions, not with evaluating
their conduct morally. Those concerned with
moral autonomy are, on the other hand, to be
understood as concerned with such evaluation
rather than with protecting moral agents against
certain undesirable influences. For our purposes,
this difference in emphasis does not matter. I
 791/1183

therefore generally use personal autonomy for

both.

Conceptions of personal autonomy resemble

each other in treating the autonomy of an act as
dependent on some feature of the agent. They
differ only in what that feature is. For some, an
act is autonomous only if the agent is autonom-
ous (at the moment of the act). We may call
these agent-centered conceptions. For other con-
ceptions, an act is autonomous only if the desire
leading to the act is autonomous. We may call
these conceptions desire centered.
 792/1183

Agent-centered conceptions may say little (in-

deed, nothing) about "autonomous desires." For
example, on Gerald Dworkin's recent account, a
person is autonomous insofar as (and only inso-
far as) she has certain capacities ("the capa-
city . . . to reflect on [her] first-order prefer-
ences, desires, wishes, and so forth, and the ca-
pacity to accept or to attempt to change these in
light of higher-order preferences and values"). 2
Desires as such are neither autonomous nor no-
nautonomous; autonomy is a function of the
way the person decides. An act (or choice) is
autonomous only insofar as it results from the
exercise of these capacities of the agent.3
 Page 159

Desire-centered conceptions, in contrast, con-

sider a person to be autonomous (or to have per-
sonal autonomy) only if (or only insofar as) his
acts (or other choices) are autonomous, and con-
sider his acts to be autonomous only if, or only
insofar as, they derive in the appropriate way
from desires, motives, or the like that are them-
selves autonomous. Autonomy is primarily a
characteristic of desires and only derivatively of
acts or persons. So, for example, a person might
be autonomous under such a conception only in-
sofar as his desires are.
 794/1183

All desire-centered conceptions of autonomy

now available are, I think, either historical, hy-
pothetical, or structural. In historical concep-
tions, a desire is autonomous if it results from a
certain process or is not the result of certain pro-
cesses. For example, on this conception my
present desire to eat chocolate chip cookies
would be autonomous if it actually originated in
an appropriately reflective process but not if it
originated in hypnotic suggestion or early so-
cialization into a "chocoholic" family. 4
 795/1183

Hypothetical conceptions of autonomy differ

from historical conceptions in focusing on a
possible future rather than on the actual past.
For example, on the hypothetical conception I
have defended elsewhere, my desire for chocol-
ate chip cookies is autonomous if it would sur-
vive repeated and vivid exposure to all the relev-
ant factssuch as the weight I am likely to put on
and the unpleasantness of eventually having to
choose between a crash diet and a heart attack.
If the desire would not survive exposure to such
facts, it is not autonomous.5
 796/1183

For structural conceptions, a desire is autonom-

ous if it stands in a certain relationship to other
current desires of the person (whatever the
desire's actual history or possible future). For
example, according to Harry Frankfurt's ac-
count, a person's first-order desire is autonom-
ous if it is one with which that person actually
identifiesthat is, with which his other desires
mesh in a certain way.6 The history of the desire
is relevant, if at all, only insofar as it affects a
person's willingness to identify with the desire.
According to another structural conception, a
desire is autonomous if, and only if, it fits the
person's actual life plan.7
 797/1183

Though these conceptions of autonomyboth

agent centered and desire centereddiffer in many
ways, they share one striking feature that is rel-
evant here. According to none is the employer-
employee relation necessarily inconsistent with
autonomy. According to agent-centered concep-
tions, the crucial question is: Does the employer
leave the employee with the relevant capacities
to reflect on his desires and to accept or change
them based on higher-order desires? While
some hierarchical organizations may so regi-
ment employees that independent thought is im-
possible, that does not seem to be true of all or
even most organizations employing engineers.
Certainly, it was not true of the organizations
studied in chapter 9. According to desire-
centered conceptions, the crucial question is:
Does the employer instill desires in an inappro-
priate way, or instill desires that could not sur-
vive exposure to the facts, or instill desires with
which the employee cannot identify? Again,
 798/1183

while the answer may be yes for some employ-

ers, it seems to be no for othersperhaps most.
Few organizations seem to turn their employees
into automatons.

If the employees of a certain organization are

not automatons, they should be able to act
autonomously in their workplace just as they
can outside. That they are obeying orders is not
enough to show that they are not acting
autonomously. If, for
 Page 160

example, they did as they were told because

upon reflection they had concluded that the em-
ployer knew what she was doing (or that co-
ordination among many employees was more
important than getting any one decision right),
they might well be both autonomous and scru-
pulously obedient under any of the conceptions
of personal autonomy identified here.
 800/1183

That it might lead to this conclusion may ex-

plain why those concerned with professional
autonomy generally ignore the literature on per-
sonal autonomy. That literature seems to settle
where it should elucidate. That it seems to settle
too much does not, however, explain why those
concerned with professional autonomy use the
term "autonomy," much less justify its use. Why
talk of "autonomy" rather than "freedom," "con-
trol,'' or even just "responsibility"? To answer
that question, we must examine the literature of
professional autonomy.

Professional Autonomy
 801/1183

The literature on professional autonomy seems

to distinguish two related senses of professional
autonomy. In both, acts are more central than in
conceptions of personal autonomy. In one sense
of professional autonomy, "the organizational,"
autonomy is regulation by one's own profession,
rather than regulation by the "laity". In the other
sense, "the individual," autonomy is control of
one's own work, rather than control by client,
patient, employer, or the like. 8
 802/1183

Organizational autonomy is primarily a property

of the profession as a whole. A profession is
autonomous insofar as it has control over its
own code of ethics, standards for admission to
the profession (including licensure or certifica-
tion), and disciplinary procedures: "In practice,
autonomy [in this sense] exists when the leaders
of a profession define or regulate the nature of
the services offered in the following ways: they
control recruitment and certification of mem-
bers, and set the standards of adequate prac-
tice."9 Organizational autonomy is a close relat-
ive of political autonomy.
 803/1183

In the United States, no profession is fully

autonomous in this sense. Lawyers, for example,
are licensed by a state agency (usually, the state
supreme court) and at least the final stages of
discipline are in that agency's hands. Although
the American Bar Association does prepare a
"model" code of ethics, the states do not have to
adopt it, and those that do adopt it are free to
make changesand sometimes do. Lawyers are,
of course, much involved in the state's regula-
tion of lawyers, but that involvement is at the
state's pleasure, not the profession's.
 804/1183

Engineers in the United States differ from law-

yers in this respect. The state does not make li-
censing a precondition for the practice of engin-
eering within its jurisdiction. Generally, states
require licensing only for those engineers who
work as lawyers traditionally have, offering
technical advice to an unsophisticated public or
preparing documents for a public record. Those
engineers who work as employees of a manufac-
turer or other organization capable of judging
credentials need not be licensed, because they
do not serve the public directly but only through
an employer. This difference between lawyers
and engineers seems unrelated to the sense of
professional autonomy that interests us. Our
questions would remain even if the United
States licensed all engineers, as Canada and
Mexico do;10 a licensed engineer can be
 Page 161

an employee and, as an employee, have much

the same problems of autonomy as an unli-
censed engineer.

That brings us to the second sense of profession-

al autonomy, the individual. We may distinguish
at least three analyses of this sort of autonomy
in the literature.
 806/1183

One is clearly connected with personal

autonomy. For example, K.R. Pavlovic defines
the individual autonomy of engineers as "relat-
ive absence of restrictions on action . . . [and] of
coercion [when] the actor is the initiator of ac-
tion rather than simply the medium." 11 This
definition has three undesirable features. First, it
is largely negative ("an absence"), while profes-
sional autonomy seems to be something positive
(a capacity for a certain sort of action). Second,
because liberty is often thought of as the ab-
sence of restrictions on action, Pavlovic's defini-
tion seems to confuse autonomy with liberty.
Third, insofar as ''the initiator" requirement adds
anything to liberty, the addition is unexplained.
The definition thus seems to assume an analysis
of professional autonomy rather than to offer
one.
 807/1183

A second analysis is related to organizational

autonomy. For Paul Camenisch, for example, in-
dividual professional autonomy means "[ulti-
mate] assessment only by one's professional
peers, not by laypersons, even when the latter is
the professional's employer."12 Though this ana-
lysis rings true, it also rings hollow. Everything
depends on how we understand "ultimate assess-
ment." We have already seen that, according to
at least one understanding of "ultimate assess-
ment," no profession has individual autonomy
(or, at least, none has that the state licenses).
 808/1183

Camenisch did not explicitly discuss the ulti-

mate assessment of engineers. He might have
learned something from doing so. Whatever
might be true of licensed professions, unlicensed
professions like engineering seem to lack a non-
controversial way to understand "ultimate as-
sessment." Of course, like a lawyer or physician,
an engineer cannot practice his profession
without a client or employer. There is, then, a
clear sense in which the market provides the ul-
timate assessment of an engineer, as it does for
lawyers and physicians. If an engineer cannot
find a client or employer, he cannot practice his
profession.
 809/1183

There is, however, also clearly a sense in which

the ultimate assessment belongs not to the em-
ployer but to other engineers. An engineer can
have an employer and still not practice her pro-
fession. Even if she has a degree in engineering
and a job with "engineer" in the title, she might,
for example, still only be the building's janitor.
To practice engineering, she must actually be
engaged in what is engineering in more than
name. What that is (as I argued in chapter 3)
something for engineers to decide, however in-
formally, rather than for those who know little
or nothing about engineering. Here, then, is an
ultimate assessment that engineers must make,
an assessment which, though present in law and
medicine, tends to be concealed by state-oper-
ated licensing.
 810/1183

The law courts provide yet another possible ven-

ue for ultimate assessment. Whatever engineers
say, the courts can find an engineer's work to
fall below the standard of reasonable care, im-
pose huge damages, and so force the engineer
from the field. Courts can even force standards
of care on the profession as a whole and so de-
cide what shall, and what shall not, be engineer-
ing practice. The engineering profession, the
employers of engineers, and the clients of engin-
eers may be powerless to undo what a judge or
jury has done. A court's finding of malpractice
seems as ultimate
 Page 162

an assessment of an engineer's work as any. Law

and medicine are, of course, also subject to this
sort of ultimate assessment.

So, great care is required to work out an under-

standing of ultimate assessment that is both de-
fensible and relevant to engineersand, indeed,
relevant to most professions as they are in fact
practiced in the United States. My own intuition
is that nothing of value will come from the at-
tempt. We cannot understand what makes a cer-
tain assessment "ultimate," if any is, until we
understand individual (professional) autonomy.
Once we understand individual autonomy, we
need not worry about whether any assessment is
ultimate.
 812/1183

A third conception mirrors the sociological liter-

ature discussed next. For Kenneth Kipnis, pro-
fessional autonomy is "control over the condi-
tions or content of work." 13 This is an ex-
tremely demanding conception of professional
autonomy. Professional autonomy in this sense
seems to be inconsistent both with the normal
authority of employers and with the cooperation
and division of labor necessary to make any
large organization work. Yet, only one philo-
sopher, Mike Martin, has noted how demanding
this analysis is, and only three writersAdina
Schwartz and Heinz Luegenbiehl, as well as
Martinhave sought to provide a conception of
managerial control (and organizational coopera-
tion) consistent with leaving to professionals
enough control over the conditions or content of
work to preserve the professional's autonomy.14
Although Martin's piece is especially suggest-
ive, even it does not deal directly with our sub-
ject, "professional" autonomy (though it uses the
 813/1183

word). Martin argues, in effect, that ordinary

morality, not anything distinctly professional,
determines what the employer can demand
without threatening autonomy. This is, in es-
sence, the position of Schwartz and Luegenbiehl
as well. They therefore provide an answer to the
question, "How can an employer's authority be
consistent with an engineer's personal
autonomy?" They never reach our question,
''How can an employer's authority be consistent
with an engineer's professional autonomy?"
They do not even see that this might be a differ-
ent question.

Sociological Literature
 814/1183

In 1939, Vannevar Bush, MIT engineer and

soon-to-be mobilizer of American scientific tal-
ent during World War II, suggested: "We may
as well resign ourselves to a general absorption
as controlled employees, and to the disappear-
ance of our independence. We may as well con-
clude that we are merely one more group of the
population . . . forced in this direction and that
by conflict between the great forces of a civil-
ized community, with no higher ideals than to
serve as directed."15 This pessimistic suggestion
rests on three controversial assumptions.
 815/1183

First, there is the assumption that working in

large organizations is new to engineers (Bush's
"we"). In fact, as we have seen, most engineers
always worked in large organizations, beginning
with the military, going on to the railways, and
then into manufacturing. Most engineers were
always "controlled employees."16 Like the milit-
ary, the clergy, and the professoriate, engineer-
ing was never independent in the sense a profes-
sion of independent practitioners can be.17
Bush's statement seems much more appropriate
to physicians or lawyers today than to engineers
at any time in their history: Physicians and law-
yers, after almost two centuries during which
 Page 163

they practiced primarily as "free professions,"

are now becoming primarily employees of large
organizations, some headed by members of their
own profession and some not. 18

Second, there is Bush's assumption that this

"loss of independence" is something engineers
must "resign" themselves to, that it is plainly an
evil. We have long known that few engineers,
especially those with only a B.S., feel that being
free of supervision is important.19 Generally,
engineers accept supervision as part of what it is
to be an engineer (or, at least, an engineer en-
gaged in important work). The problem seems
to be to find the right kind of supervision.20
 817/1183

The question of control over what is done with

the engineer's work is, of course, a distinct ques-
tion. But here again engineers are often thought
to be at a disadvantage when compared with
such "true professionals" as lawyers or physi-
cians. The engineer's employer decides whether
to do what the engineer recommends while,
some say, the lawyer or physician can simply do
it. This supposition seems even more of a mis-
take than the supposition that lawyers and physi-
cians have complete control over what they do.
Like engineers, lawyers and physicians are
primarily advisors. Under ordinary conditions,
they can say what they like, insist on it as much
as they like, but if the client or employer says
no, they can do nothing. A will is only a draft
until the client decides it is good enough to sign.
A physician who forces a competent patient to
submit to treatment is guilty of criminal battery.
Whatever professional autonomy is, it cannot be
control over one's client or employer.
 818/1183

Third, Bush assumes that being "controlled em-

ployees" means that "we" can have ''no higher
ideals than to serve as directed"in other words,
that the status of employee is inconsistent with
professionalism. This, of course, is the assump-
tion this chapter is to help evaluate. The empir-
ical literature does not address it. Instead, it ad-
dresses autonomy defined in some such way as
this: "a condition in which the performer, rather
than someone else, determines the sequencing of
tasks which comprise the job and how long one
performs a given task before switching to anoth-
er."21 Although this sort of control is, to some
degree, a precondition of professional
autonomy, it is clearly different. Every profes-
sion must at times adjust to the schedule of oth-
ers. Even the physician may have to drop
everything and come to the hospital when the
baby is ready to be born. No one supposes such
emergencies to threaten his professional
autonomy even if he has such emergencies
 819/1183

several times a day.22 Hence, most empirical

work on the autonomy of engineers is not dir-
ectly relevant to their individual professional
autonomy.
 820/1183

One near exception seems to be the work repor-

ted in chapter 9. My colleagues and I inter-
viewed sixty engineers and managers in ten
companies. We found few engineers or man-
agers with absolute control over any significant
decisions, but we did find considerable room for
discussion and a strong tendency to seek con-
sensus. We certainly got the impression that en-
gineers were free to exercise their professional
judgment (one obvious interpretation of profes-
sional autonomy); indeed, some managers
seemed to go out of their way to stress the im-
portance of engineers "sticking to their guns."
Unfortunately, we were not then concerned with
professional autonomy, so, our observations are
no more than suggestive.23 What more we
should have done depends on what individual
professional autonomy is. We must now work
out a useful conception.
 Page 164

A Conception of Individual Professional

Autonomy
 822/1183

The preceding examination of the literature on

autonomy suggests three criteria of adequacy for
any conception of individual professional
autonomy. First, any adequate conception
should explain the relation between professional
autonomy and personal autonomy or, failing
that, explain the importance of professional
autonomy in some other way. Second, any ad-
equate conception of professional autonomy
should make the professional autonomy of an
individual employee an empirical question or,
failing that, explain why many people consider
it to be empirical. Third, any adequate concep-
tion of professional autonomy should suggest
ways to test the empirical content, if any, of
claims concerning professional autonomy; it
should yield a practical research program or,
failing that, explain why none is possible. A
conception meeting these three conditions
would bring together, as much as possible, the
 823/1183

philosophical, professional, and sociological lit-

erature concerned with workplace autonomy.

How do we proceed? First, I sketch what I hope

will be a relatively uncontroversial conception
of "acting as a member of a profession." Second,
I identify a conception of personal autonomy
suitable for our purposes. Third, I show that one
can act autonomously, in this sense while acting
as a member of a profession. Professional
autonomy is a special kind of personal
autonomy: acting as a member of a profession.
Fourth, I show that one can act autonomously as
an employed member of a professionthat em-
ployment and professional autonomy are not in
principle inconsistent.

Having done all that, in the next section, I sug-

gest some relatively easily conducted research
to tell us how much professional autonomy em-
ployed engineers actually have.
 824/1183

Profession

There are no professions of one. To act as a

member of a profession is to act as a member of
a group. What kind of group? For our purposes,
a profession is a number of individuals sharing
an occupation voluntarily organized to earn a
living by serving some moral ideal in a morally
permissible way beyond what law, market, and
ordinary morality require. 24 This definition is
important enough to deserve amplification.

Professions are voluntary. We must claim mem-

bershipwhether by seeking a license to practice
the profession, by applying for a job calling for
a member of that profession (for example, "en-
gineer"), or just by announcing membership in
that profession (for example, "I am an engin-
eer"). We can always leave a profession by giv-
ing up the license (if there is one), withdrawing
from practice, and ceasing to claim membership.
 825/1183

Professions are organizations. We are not mem-

bers just because we claim to be. We must also
meet certain minimum standards of competence
and conduct that the group treats as a condition
of membership. Professions vary in degree of
organization. Some have formal tests for admis-
sion, licensing bodies, disciplinary com-
 Page 165

mittees, and the like. Most have schools from

which would-be practitioners should graduate, a
written code of ethics, and various associations
speaking for the profession in certain contexts.
All recognize a distinction between those who,
in virtue of competence and conduct, belong to
the profession and those who, falling short in
some way, do not belong though they claim to
("quacks," "charlatans," ''impostors," and so on).
A profession is more than an occupation; it is a
discipline.
 827/1183

All professions have special, morally permiss-

ible standards for conducting the business of
members. Typically, the primary purpose of
these standards is to offer nonmembers (the cli-
ent, employer, or public) some benefit or protec-
tion beyond what law, market, or (ordinary, pre-
profession) morality require. 25 These standards
differ from profession to profession. So, for ex-
ample, while engineers undertake to use their
distinctive knowledge, skill, and judgment in
ways that hold the public health, safety, and
welfare paramount, lawyers do not. Professional
standards are always morally permissible; a
"profession of thieves" is no more a profession
than play money is money. Not every discipline
can be a profession (in this sense).26
 828/1183

Professions differ from charities, mutual assist-

ance societies, and other altruistic organizations
in being concerned with how members earn
their living. There is no profession of amateurs.
A profession is nonetheless not an ordinary
business or occupational organization. A profes-
sion differs from an ordinary business in being
for persons in a single occupation or, at least, in
a family of occupations sharing a common body
of knowledge, skill, and judgment. A profession
differs from a trade association, union, or other
occupational organization in having as its
primary purpose something beyond benefiting
its members. A profession differs from both
businesses and occupational organizations in be-
ing designed primarily to serve a certain moral
ideal in a certain way. Physicians organized to
serve health; lawyers, justice within the law; and
so on.
 829/1183

The "moral" in "moral ideal" is meant to ex-

clude nonmoral ideals (the ideal of prudence, for
example). A moral ideal is moral both in the
minimal sense of being morally permissible (as,
for example, stealing competently is not) and in
the stronger sense of being morally good (that
is, tending to support morally right conduct).
But a moral ideal is also moral in a much
stronger sense. A moral ideal is a state of affairs
which, though not morally required, is one that
everyone (that is, every rational person) wants
others to approach when possible, wanting that
so much as to be willing to reward, assist, or at
least praise such conduct if that were the price
for others to do the same. Moral ideals have a
claim on us that nonmoral ideals do not. Profes-
sions are, by definition, praiseworthy (in the
way voluntarily undertaking any laudable re-
sponsibility is) because each profession, by
definition, undertakes to serve a moral ideal.27
 830/1183

To be a member of a profession is, then, to be

subject to a special set of standards. To act as a
member of a profession is openly to carry on
one's business according to those standardsfor
example, to declare in word and deed, "I work
as an engineer [that is, as engineers are sup-
posed to work]." Can one be subject to such
standards and still be autonomous? Before we
can answer that question, we must decide what
we mean by "autonomous."
 Page 166

Autonomy

An act's voluntariness creates a presumption in

favor of its autonomy (on any plausible concep-
tion of personal autonomy). If we can voluntar-
ily make a promise without compromising our
(personal) autonomy, we should be able to join a
profession without compromising our
autonomyeven if membership involves, as
promising does, commitment to acts we would
not otherwise do. So, because professions are
voluntary organizations, the presumption must
be that anyone acting as a member of a profes-
sion acts autonomously. 28
 832/1183

But that is only a presumption. Voluntariness

does not guarantee autonomy. Even the content
of a commitment can undercut the presumption
of autonomy. So, for example, however volun-
tary, the promise to subject ourselves forever
and in all respects to the arbitrary will of an im-
moral person does not seem autonomous. Al-
though the promise renounces personal
autonomy forever, its effect is only a symptom
of the problem with its autonomy. The problem
is that the promise does not seem to be either the
act of an autonomous person or an act that an
autonomous desire would motivate. We find it
hard to imagine how an autonomous agent could
voluntarily make such a commitment (and mean
it). We doubt the autonomy of such a commit-
ment until we hear an unusual, and convincing,
explanation of why, appearances to the contrary
notwithstanding, the promise is autonomous. So,
we need to explain why voluntary membership
in a profession is not like making such a
 833/1183

promise. We need some way to distinguish

those voluntary commitments that are autonom-
ous from those that are not.29 What help can the
conceptions of personal autonomy we identified
provide?
 834/1183

Agent-centered conceptions have nothing to say

about what we can do autonomously. A promise
to subject ourselves forever and in all respects to
the arbitrary will of an immoral person would be
autonomous, according to agent-centered con-
ceptions, if we are autonomous; and all we need
to be autonomous is the capacity to reflect on
our desires, to change them in light of higher-or-
der desires, and to act accordingly. This result
points up two weaknesses in agent-centered
conceptions. First, agent-centered conceptions
do not require a tight connection between the act
(in this case, a promise) and the agent's capacit-
ies to reflect (whatever they may be). A commit-
ment is autonomous if the person making it is
autonomous, however careless or ill-informed at
the time. Second, agent-centered conceptions
seem to rely on a weak procedural notion of ra-
tionality. Rationality consists in having some ca-
pacity for self-criticism and correction. There is
no limitation on the content of the processhence,
 835/1183

no way to guarantee that we cannot autonom-

ously make even such an outrageous promise as
we are now imagining. Agent-centered concep-
tions of autonomy seem unsuited to our pur-
poses.30

What about desire-centered conceptions? All

desire-centered conceptions of personal
autonomy avoid this first weakness. Indeed,
they seem designed to do just that. All require a
tight connection between an autonomous desire
and the act. For an act to be autonomous, the
motivating desire must itself be autonomous.

Desire-centered conceptions differ from one an-

other both in how they determine which desires
are autonomous and in the relation they require
between desire and
 Page 167

act. Historical conceptions require a certain his-

tory (for example, one avoiding hypnotic sug-
gestion) for a desire to be autonomous; structur-
al conceptions require a certain peaceful relation
between the desire in question and others the
person has; and hypothetical conceptions require
that the desire be able to survive certain tests.
Historical and structural conceptions are alike in
requiring only that the desire motivating the act
be autonomous for the act to be. Hypothetical
conceptions may require more: for example, that
the choice of the act (as well as the desire) be
able to survive certain testssuch as vivid and re-
peated exposure to all relevant facts. Nonethe-
less, all desire-centered conceptions might share
with agent-centered conceptions their second
weakness, too weak a notion of rationality.
Whether they do depends on how the crucial no-
tion of autonomous desire is filled in.
 837/1183

All desire-centered conceptions are alike in hav-

ing certain "internal constraints" on what counts
as an autonomous desire: for example, "fitting
the agent's life plan". These internal constraints,
differing from conception to conception, may or
may not amount to ''procedural rationality."
Most desire-centered conceptions of personal
autonomy are also alike in lacking any provision
for "substantive rationality." So, for example,
most cannot declare nonautonomous the desire
to have one's arm removed just to have it re-
moved; some contingent fact about the person
determines whether that desire is or is not
autonomous. Only my hypothetical conception
is different in this respect. Unlike all other con-
ceptions (agent centered as well as desire
centered), mine requires some correspondence
between the desire and the external world. On
my hypothetical conception, the desire in ques-
tion must not be such as to weaken when
 838/1183

exposed to the relevant facts. 31 It must, in this

sense at least, be rational.

If we combine my conception with certain min-

imal assumptions about human psychology, we
can identify certain substantive constraints on
what can be autonomous. For example, suppose
that, all else equal, people prefer liberty to sub-
jection. Any desire to subject oneself to another
just to do itas in the autonomy-renouncing
promise we have imaginedshould, upon expos-
ure to the reality of such subjection, disappear
(or at least weaken enough to change the de-
cision). Because the desire is not rational, any
choice dependent on it would not be
autonomous.
 839/1183

Other conceptions do not give this result even

when supplemented in the way mine was. Com-
mon senselike preferring liberty to subjectionis
no guarantee that people will not sometimes act
foolishly, that they will not sometimes develop
desires inconsistent with their own common
sense, or that they will not otherwise fail to use
all the information available. Contingency dogs
our acts. My hypothetical conception escapes
such contingencies by requiring the act's motiv-
ating desire to be capable of passing a test that
depends on more than what the agent's psycho-
logy happens to be at a given moment. The de-
sire must be able to survive repeated, vivid ex-
posure to the world beyond the agent.
 840/1183

Because my hypothetical conception seems the

most demanding conception of personal
autonomy available, any argument relying on it
to show that one can act autonomously as a
member of a profession should withstand substi-
tuting one of the other, less demanding concep-
tions of autonomy. If we can explain how, ac-
cording to this conception, personal autonomy is
consistent with membership in a
 Page 168

profession, we should be able to do the same us-

ing any other plausible conception of personal
autonomy. So, in what follows, I understand
personal autonomy to mean autonomy according
to my hypothetical conception.

Autonomy in Professions

If acting as a member of a profession is to be

consistent with acting autonomously (as we are
now conceiving autonomy), the profession has
to be more than voluntary. It has to be a com-
mitment consistent with all the relevant facts.
What are facts? What makes them relevant?
 842/1183

By "facts" I mean, at least, (1) common sense,

what "everyone" knows (that people generally
prefer life to death, that depriving people of
food can kill, that people know that, and so on),
and (2) special knowledge actually available to
people who inquire (what lawyers know about
the law, what biologists know about genes, what
you can look up in a public record, and so on).
A fact is "relevant" to choice of an act if, and
only if, exposure to it would, all else equal,
strengthen or weaken the agent's resolve to do
the act (whether by arousing new desires or by
strengthening or weakening a desire already mo-
tivating the act). To have all the facts relevant to
a decision vividly and repeatedly before one is
to be fully informed.
 843/1183

A member of a profession can act autonomously

in that capacity if, and only if, his actual choice
accords with what he would have chosen if fully
informed and motivated only by rational de-
siresthat is, desires that have themselves sur-
vived exposure to all the relevant facts. "Full in-
formation," a standard generally too costly for
practice, is reasonable here because it applies
only to a hypothetical choice. We do not ask
anyone actually to make this choice under these
conditions. We ask only that, given full informa-
tion and other appropriate conditions, a psycho-
logical theory would tell us what the individual's
actual choice would be or, at least, what options
would be open. If we call such a choice rational,
our question has become: How could anyone ra-
tionally choose to act as a member of a profes-
sion? 32 This is not a hard question.
 844/1183

To be a member of a profession is to be subject

to standards of conduct. These are, as such, bur-
dens (because, all else equal, liberty is better
than subjection to a standard). If the standards in
question were arbitrary, being subject to them
would be little different from being subject to
the arbitrary will of an individual. It would be
an unredeemed burden. But all else is not equal
for the standards of one's profession. Those
standards are not arbitrary. They must, of
course, be morally permissible. They must, in
addition, be designed to serve the moral ideal to
which the profession is committedor they would
be, for that ideal, entirely arbitrary and so,
strictly speaking, not the profession's standards
at all. Someone who objects to acting according
to a standard she believes inconsistent with, or
just independent of, service to the moral ideal in
question does not object to acting as a member
of the profession. Instead, she objects, as a
member of the profession, to the claim that
 845/1183

acting as a member of the profession includes

acting according to that standard.

Members of a profession may be members be-

cause they want to serve the ideal in question.
They then have some commitment to whatever
is in fact necessary to serve that ideal. But, even
if someone is a member for less noble
reasonsfor
 Page 169

example, just to earn a livinghe has a commit-

ment to the standards in question. He has
entered this particular profession because of
what it is, hoping for the benefits that come
from claiming honestly to belong to it rather
than to another or none at all. He should (as a
fully informed rational agent) understand that
those benefits depend in part on maintaining
certain standards. Who would claim to be an en-
gineer, for example, if engineers were generally
thought incompetent or dishonest? He should
(as a rational agent) remain in the profession
only as long as the benefits exceed the costs by
enough to make membership the best option
available for him.
 847/1183

Each profession is a cooperative practice. Each

member bears certain burdens and expects other
members to do the same. If most do as they
should, the profession should have a good repu-
tation and each member should be better off act-
ing as a member of the profession than acting as
an individual. If too many shirk, however, the
practice produces no net benefit, and those who
bear their share of the burdens suffer most. Be-
cause each profession is a voluntary cooperative
practice, its standards have the same moral
claim on members that the rules of a morally
permissible game have on voluntary players. Vi-
olating a professional standard is a form of
cheating. Members of a profession are therefore
morally bound to act as their profession re-
quires. They may sometimes have a justification
or excuse for not doing as they should (just as
they might for breaking a promise), but their
profession will, like a valid promise, always
burden their conduct with obligations they
 848/1183

would not have but for their profession. The

burden is not onerous, however, because there is
always the morally permissible alternative of
leaving the profession.

For any individual, membership in a profession

may or may not be rational. But, given what
professions are, the voluntariness of member-
ship creates a strong presumption that member-
ship is rational for those who actually are mem-
bers. In turn, the rationality of membership
vouches for the autonomy of acting as a mem-
ber. We autonomously submit to the standards
of a profession, when we can, because submit-
ting to them belongs to the very practice gener-
ating the benefits that make membership ration-
ally attractive and because taking these benefits
without submitting to those standards is morally
wrong. When we can no longer submit, we
shouldand canquit. The autonomy one has as a
member of a profession is not merely
presumptive.
 849/1183

Employees and Professional Autonomy

An employer either hires a member of a profes-

sion as a member of that profession or hires him
as something else. If she hires him as something
else, he will not have to act "in a professional
capacity." He will not be practicing his profes-
sion. He will be like the lawyer hired to teach
tennis at a country club or the engineer who
ends up as a company's comptroller. Insofar as
what he does is not done in his professional ca-
pacity, he may ignore his profession's standards.
Any problem of autonomy he has is not a prob-
lem of professional autonomy.
 850/1183

If, however, a member of a profession is hired

as a member of that profession (for example, as
an engineer), he may have a problem of profes-
sional autonomy. He may because, as a profes-
sional, he is supposed to do as his profession
says while, as an employee, he is supposed to do
as his employer says. The two commitments
 Page 170

create a potential for conflictbut only a poten-

tial. Problems of professional autonomy can ac-
tually arise only if the employer orders the pro-
fessional to do something his profession forbids.
There is good reason to expect this not to hap-
pen often: Why hire someone as an engineer for
example, if you do not what him to work as en-
gineers typically do? If you simply want
someone to obey your orders, why not hire
someone without a profession (who would or-
dinarily be less expensive, too)?
 852/1183

While there is good reason to expect the em-

ployed professional not to have problems of pro-
fessional autonomy often, there is also good
reason to expect such problems sometimes: Just
as people sometimes prefer theft to honest labor,
so employers sometimes want professionals in
name but not in deed. For example, an employ-
er, who wants to ship an order in time may tell
the engineer in quality control to "work the
data" until the tests come out "right." In effect,
the employer wants to substitute her judgment
for the professional's while claiming the
professional's authority for it. Obeying such an
order would turn the professional into a marion-
ette. Such an order is inconsistent with profes-
sional autonomy.
 853/1183

But not all orders are like that. For example, the
typical specifications for an engineering pro-
ject"under 2,000 pounds, under $2,000, and un-
der twenty months"simply state a technical
problem. Nothing in them is inconsistent with
an engineer's professional autonomy. The engin-
eer can both act as an engineer should and do as
ordered. Similarly, even when an employer says,
"Drop everything and figure out what's wrong
with this windshield wiper," the employer's or-
der need not threaten professional autonomy. As
long as dropping everything is consistent with
standards of professional practice, the
employer's control over what is done and when
it is done is consistent with professional
autonomy. Of course, whether dropping
everything is consistent with professional stand-
ards depends on what those standards are.
 854/1183

If we now return to the quotation with which we

began this chapter, we can easily see that it con-
tains a serious mistake. "Employers," Layton
says, "have been unwilling to grant autonomy to
their employees, even in principle." 33 What I
just argued, in effect, is that employers must,
both in principle and in practice, grant employed
professionals professional autonomy. Employers
must because otherwise they cannot have the be-
nefit of employing professionals. "[Employers]
have assumed that the engineer . . . should take
orders," Layton continues, "[but] the very es-
sence of professionalism lies in not taking or-
ders from an employer.''34 What I just argued is
that the essence of professionalism consists in
part of taking orders, those consistent with act-
ing as a professional, and in part of not taking
orders, those inconsistent with acting as a pro-
fessional. To be a "true professional" is to act as
the employer orders insofar as the orders are
consistent with the profession's standards.35
 855/1183

Possible Research on Professional Autonomy So

Conceived

If we look again at the sociological literature

supposedly concerned with workplace
autonomy, we can see how small modifications
could convert much of it into research on pro-
fessional autonomy. The problem with the exist-
ing research is a relatively unrefined conception
of professional autonomy. Professional
autonomy (or
 Page 171

even autonomy in general) is not simply a mat-

ter of quantity (how many decisions) or mere
quality (decisions about order, timing, content,
or whatever). Some decisions matter more than
others; and professional standards determine
which those are. Research on professional
autonomy should, then, begin by using the
standards of the profession in question to identi-
fy which decisions matter professionally and
which do not. The relevant decisions may well
vary substantially from profession to profession
even in the same workplace. Any research on
the professional autonomy of engineers, for ex-
ample, must begin by distinguishing those de-
cisions engineers claim for themselves from
those they leave to management. An order to ig-
nore safety might violate an engineer's profes-
sional autonomy, when an order to ignore cost
would not.
 857/1183

Therefore, a questionnaire should begin by try-

ing to determine the standards the professional
takes herself to be subject to. The questionnaire
might ask, for example, "Which factors are your
professional responsibility?" There would then
follow a list to choose from (safety, health, qual-
ity, cost, beauty, and so on). The list should be
developed taking the profession's formal stand-
ards into account, but in fact the answers may
not mirror what is professionally required. Pro-
fessionals are not always as well informed as
they should be. How great the difference is
between ideal and reality is, for any particular
profession, an empirical question, one about
which it would be good to have more informa-
tion. Though what the profession requires is cru-
cial for determining whether a professional actu-
ally has professional autonomy, what the profes-
sional supposes her profession to require is cru-
cial for determining whether the professional
feels that she has such autonomy.
 858/1183

Later questions should ask how often the profes-

sional employee is ordered to do something in-
consistent with her professional responsibilities
(as well, of course, as how often she is overruled
or ignored, though overruling or ignoring is con-
sistent with professional autonomy but not with
professional control or job satisfaction). The
questionnaire should also ask what the profes-
sional can do when she is overruled or ignored
on a question she believes her professional judg-
ment should decide. Can she, for example, ap-
peal an order to higher authority? What happens
if she does? The answers to such questions
should reveal how much tension exists in fact
between acting as a member of the profession
and acting as an employee.
 859/1183

We have, it seems, now turned the question of

how much professional autonomy employed en-
gineers have into an empirical question for
which ordinary methods of social research
should be adequate. Indeed, researchers can
"ask" similar questions of "free professions."
Clients sometimes order even free professionals
to do things contrary to professional standards
(or, at least, ask such things in a way that makes
saying no difficult). Data comparing both the
frequency of such incidents among "free" and
"captive" professionals and the outcomes might,
or might not, show free professions to have
more professional autonomy than employed
professions.
 Page 172

EPILOGUE:
FOUR QUESTIONS FOR THE SOCIAL
SCIENCES
This chapter tries to draw some lessons from the
preceding ten chapters. In doing that, it identi-
fies four questions concerning engineering for
which I would like to have answers. The ques-
tions all seem to be of the kind that the social
sciences in general, if not science and techno-
logy studies (STS) in particular, could, and
should, be helping with. The four share at least
two other features as well. First, any of them
might arise while teaching engineering ethics,
advising engineers on ethical questions, or oth-
erwise engaging in "engineering ethics." Se-
cond, answering them would serve both practic-
al philosophy and professional practice.
 861/1183

The four questions suggest a paradox. Though

the social sciences should already be hard at
work on them, in fact they hardly touch the sub-
ject. Indeed, when I ask about engineering, so-
cial scientists generally refer me to books that
turn out to be about science (especially physics,
biology, and medicine) or technology (objects
and their consumers, processes, and victims).
Consider, for example, Donald Mackenzie's
very interesting Inventing Accuracy: A Historic-
al Sociology of Nuclear Missile Guidance, a
book often presented to me as decisive proof
that the social sciences are giving engineering
its due. This is certainly a book about engineer-
ing. Its subject is an engineering problem, the
development of specifications for what consti-
tutes a direct hit with a long-range missile; its
cast of characters is also largely engineers). Yet,
it is proof of my thesis, not evidence against it.
The words "engineer" and "engineering" are
(except for the title of its chapter 3) virtually
 862/1183

absent. For all the reader is told, the book has

nothing to do with engineering. But the index
has a substantial number of listings for "applied
science." No attempt is made to distinguish the
contribution of engineers to the debate from the
contribution of scientists (if any), politicians, or
the like. 1 Here is a classic illustration of the
way the profession
 Page 173

of engineering, central to modern technology,

seems to be almost invisible to social scientists,
even those who study technology.

Of course, a few social scientists have done

some useful work on engineering, especially, if
we include historians among social scientists. (I
will give some examples later.) The trouble is
that social scientists have done far too little (es-
pecially when compared to what they could do
or what they have done for the physical sci-
ences), what they have done has yet to form a
distinct field of study, and what little they have
done has not been designed to help with engin-
eering ethics. 2 So, this final chapter is in part a
plea for an overdue adjustment in focus.
 864/1183

I proceed in this way. First, I briefly explain

what I take engineering ethics to be, summariz-
ing the book's main argument. Next, I state and
discuss my four questions, making clear what
each asks, why the answer might be important
for engineering ethics as I have interpreted that
subject, what would constitute an adequate an-
swer, and how social scientists might help. Last,
I consider what barriers stand in the way of an-
swering the questions. Everything I say is both
sketchy and preliminary, an invitation to begin a
discussionprologue as much as epilogue.

Engineering Ethics
 865/1183

Engineering ethics is a kind of applied, or prac-

tical, philosophy. It is concerned with under-
standingand helping to resolvecertain moral
problems arising in the practice of engineering.
These problems can be approached in at least
five ways: philosophical, casuistic, technical,
social, and professional.3

The first three approachesthe philosophical, ca-

suistic, and technicalare alike in assuming that
engineers are held only to the same moral stand-
ards as nonengineers. Professional organization,
if it matters at all, matters as a mere
"expression" of fundamental moral concerns, as
an aid or barrier to doing what should be done
anyway, profession or no.
 866/1183

The philosophical and casuistic differ from the

technical in relying solely on the facts of a par-
ticular situation to transform general standards
into specific directives. They differ from each
other in the way they determine those general
standards. The philosophical appeals to some
moral theory (utilitarianism, Kantianism, virtue
theory, or the like) to determine (along with the
facts) what should be done.4 The casuistic ap-
peals instead to ordinary moral standards (either
explicitly, for example, by citation of a com-
monly accepted moral rule like "Don't kill" or
"Don't take unnecessary risks," or implicitly, by
comparison of cases, ormost oftenby some com-
bination of these).5
 867/1183

The technical approach differs from the philo-

sophical and casuistic insofar as it relies (instead
or in addition) on special "principles of [good]
engineering" derived from the "nature" of engin-
eeringprinciples of competence or skill. Those
special principles (together with ordinary moral-
ity and the specific facts of a situation) determ-
ine what should be done in that particular situ-
ation. The nature of engineering may be time-
less (a Platonic idea) or a product of history
(like the English language), but, at any time, it is
a given, not something subject to change in the
way
 Page 174

a statute or contract is. 6 These first three ap-

proaches commit one or another of the fallacies
identified in chapter 3.

The social approach to engineering ethics re-

sembles the technical in beginning with
something special about engineers or engineer-
ing. It differs from the technical in understand-
ing this specialness as, at least in part, a product
of social decision (the role society constructs for
engineers). For the social approach, the stand-
ards of engineering ethics derive not from the
nature of engineering as such but from a morally
binding "contract" with society or from society's
morally binding dictate. For the social approach,
standards of engineering ethics are more or less
arbitrary, that is, dependent on what society and
engineers together happen to agree on or what
society happens to decide.7
 869/1183

The professional approach, the last on my list,

resembles the social insofar as both recognize a
certain arbitrariness in what may turn out to be
"ethical." The professional approach differs
from the social in placing that arbitrariness in
the profession of engineering rather than in the
decisions of society as such or in an agreement
engineers make with society. For the profession-
al approach, society (like morality or the nature
of engineering) is generally a mere "side con-
straint," not the primary, or equal, party in de-
termining the content of engineering ethics.8 I
call this approach professional to emphasize the
distinctive place it assigns the profession. For
this approach, and for this approach alone, a
morally binding code of ethics, typically the
work of a professional society, is a central fact.9
This last is, of course, the approach I take.
 870/1183

This book noted early that "ethics" is ambigu-

ous. We can use it (1) as a mere synonym for or-
dinary morality, (2) as the name of a specifically
philosophical study (the attempt to understand
morality as a rational undertaking), or (3) as
special standards of conduct morally binding on
members of a group because they belong to that
group ("special" both because the standards do
not apply to everyone and because they go bey-
ond what law, market, and ordinary morality re-
quire). All five approaches to engineering ethics
can use ethics in the first two senses (though the
technical tends to ignore the second). But only
the social and professional approaches can use it
in the third sense; only they recognize engineer-
ing ethics as including special, morally binding
standards of conduct.
 871/1183

I distinguish these five approaches because dis-

tinguishing them allows us to see important
common ground and reveals some important
differences as well. All five recognize how im-
portant understanding what engineers do is to
understanding what they should do. None of the
approaches supposes that we can do much of in-
terest in engineering ethics without knowing a
good deal about engineering, especially about
what moral problems actually arise in practice
and what resources are available for resolving
them. "Practical philosophy," however ap-
proached, requires an understanding of the rel-
evant practice.
 872/1183

The five approaches nonetheless differ some-

what concerning what understanding is neces-
sary. For example, the philosophical approach
makes an understanding of the history of profes-
sional codes seem beside the point. For the
philosophical approach, the central relation is
between some moral theory and the situation of
the individual professional.10 For the profes-
sional approach, however, the history of the pro-
fession is much more important because that
history should give insight into
 Page 175

how to interpret the profession's code and that

code is central to determining what the profes-
sional should do. This difference concerning
what understanding is necessary may well affect
the questions those committed to a particular ap-
proach ask about engineering.

Commitment to the professional approach,

though evident throughout this book, should not
matter much for what I shall do now. Although I
do not expect everyone to agree that I am asking
just the right questions, I do hope that everyone
agrees that I am asking the right sort of ques-
tion. Social scientists who find another approach
more to their liking than the professional can
ask somewhat different questions and organize
their research accordingly. They will still serve
engineering ethicsin part at least by showing
how empirically fruitful the other approaches
can be.
 874/1183

What is Engineering?
 875/1183

The first question I would like the social sci-

ences to help answer is one with which this
book began: "What is engineering?" My attempt
to answer it took me into fields of history and
sociology all but deserted. Why is this question
on my list? By now, the answer should seem fa-
miliar. Because I take the professional approach,
engineering ethics is, for me, the ethics of en-
gineers, a special group. I therefore need to
know who is in the group (and so subject to its
special standards) and who is out (and so not
subject to them). Determining who is an engin-
eer and who is not is not easy, as chapter 3 made
clear. We can, of course, settle many cases intu-
itively. For example, someone with a B.S. in
mechanical engineering from an ABET-accred-
ited school (say, from IIT in 1975), licensed to
practice nuclear engineering, and with twenty
years of experience designing nuclear power
plants, certainly is an engineer (for purposes of
engineering ethics). And just as certainly, the
 876/1183

operator of a diesel locomotive or the janitor of

an apartment building, though called engineer,
certainly is not. But what about those called
software engineers or genetic engineers? They
generally do not have an ABET-accredited de-
gree but, unlike diesel operators or janitors, gen-
erally have an education similar to that of engin-
eers and do work similar to engineering. And
what about the chemist or physicist who does
work engineers also do and who, unlike soft-
ware engineers or genetic engineers, may well
work beside engineers strictly so called?
 877/1183

Anecdotal evidence makes me think that engin-

eers (strictly so called) are pretty clear about
who is an engineer and who is not. Generally,
software engineers and genetic engineers are
not; the chemists or physicists may be "adopted"
after demonstrating certain skills on the jobbut,
until then, they are outsiders, whatever title they
hold and whatever work they do. Couldn't soci-
ologists tell us as much about this line drawing
as they have about the line drawing between
science, non-science, and pseudo-science? 11

The question "Who is an engineer?" soon leads

to a question actually on my list of four, one of
interest to all approaches to engineering ethics:
"What is engineering?" Even those who do not
recognize engineers as having a profession dif-
ferent from science or technology, still need to
distinguish, for example, those moral problems
properly belonging in a course in engineering
ethics from those properly ex-
878/1183
 Page 176

cluded. Not all moral problems engineers face,

even in the workplace, are problems of engin-
eering ethics. Some are problems of business
ethics or just ordinary moral problems.
 880/1183

Though "What is engineering?" is a philosophic-

al question, historians can help answer it in at
least two ways. Walter Vincenti's What Engin-
eers Know and How They Know It provides a
good example of the first way. 12 By using the
techniques of historical studyespecially, the
gathering and analysis of documents to recon-
struct a sequence of eventsVincenti was able to
help us see engineers at work, to distinguish
their work from that of scientists, and so to help
us understand some differences between engin-
eering and science. We need more work on the
difference between engineering and science.13
We also need similar work to help us understand
the difference between engineering and architec-
ture, between engineering and industrial design,
between computer engineering and software en-
gineering, and even between chemical engineer-
ing and industrial chemistry (especially when, as
sometimes happens, engineers and chemists
seem to be doing the same job).
 881/1183

Vincenti's work is a series of what we might call

historical case studies. Sociologists can do
something similar using participant observation.
The Soul of a New Machine is almost an ex-
ample of what sociologists could do.14 I say "al-
most" because Tracy Kidder, while describing
the work of engineers in informative detail, nev-
er tries to figure out what distinguishes what
they do from what the people in software do (or
why the department he studied wanted engineers
while software did not).15
 882/1183

Historians can also contribute to our understand-

ing of engineering in a way sociologists cannot.
Historians can tell us about the historical pro-
cess of defining engineering.16 Layton's Revolt
of the Engineers17 is an important example of
what can be donebut, in my view, one that
leaves most of the work undone. Layton's focus
is on the United States before World War Iwhen
the United States was still a cultural backwater.
The crucial events in professional development
may well have occurred in France, England, or
Germany. After all, engineering came to the Un-
ited States from France, with American engin-
eering schools copying the French at least up to
the Civil War; engineering codes of ethics were
adopted in England almost a half century before
they were adopted in the United States. We need
a trans-Atlantic history of engineering's defini-
tion.18
 883/1183

In chapters 1 through 3 I argued that what such

a history would teach is that we cannot usefully
define engineering by genus and species, as
Artistotleans would, or even by what engineers
do or how they do it (their function or method).
The useful definitions treat engineers as mem-
bers of a certain historical community, the pro-
fession of engineering, and then define that
community in terms of certain historically de-
veloping criteria including education, experi-
ence, and commitment to certain ways of doing
certain things. Engineering is what engineersat
the time in questiontypically do that members of
other occupational groups don't. Engineering
has ethics built into it insofar as, and only inso-
far as, the engineering community has in fact
adopted special morally permissible standards
beyond what law, market, and ordinary morality
require. This, I believe, is what such a history
would show. But, without the right sort of his-
torical research, research careful to distinguish
 884/1183

the profession of engineering from various com-

petitors, including proto-engineers,
 Page 177

the best we can do is what I did, a philosopher's

reconstruction of what historians have so far
discovered.

What do Engineers do?

 886/1183

If we define engineering in some such way as I

suggest, we can distinguish engineers from
those doing similar work. Engineers are those
whom the engineering profession recognizes as
members (for example, for purposes of member-
ship in engineering societies"at the professional
level"). That way of identifying engineers would
allow us to ask what, if anything, engineersat a
certain momentcontribute that others do notthat
is, what significance their special standards of
conduct have for others. The answer may well
interest many in the history, sociology, and
philosophy of technology because it provides a
way to study what effect, if any, differences in
profession have on the technology members of
professions make or use.
 887/1183

Although a number of writers stress the import-

ance of engineering (or, rather, technology) to
science, I don't think anyone has much to say
about the importance of engineers in science
(engineers strictly so called). Yet, at most "sci-
ence" laboratories I have visited, including Ar-
gonne, engineers seem to outnumber scientists
(though accurate figures were hard to get). The
only exception was a Red Cross lab where there
were no engineers. There physicians (or, at
least, M.D.s) outnumbered the scientists
(Ph.D.s). What do engineers do in science labs?
How much science is the work of engineers?
Why?
 888/1183

The relevance of such questions for engineering

ethics was brought home to me by a Department
of Energy report of a whistleblowing incident at
Argonne. 19 The whistleblower was described as
a "metallurgist" butby training, experience, and
(it seemed) commitmentshe was a metallurgical
engineer.20 The report described an internal in-
vestigation which, while siding with the whis-
tleblower on matters of fact, clearly had trouble
understanding why he was taking the science is-
sues so seriously. I had less trouble understand-
ing. It seemed to me he was doing what a good
engineer should, though he might have done it
more politicly. He was giving safety more
weight than law, market, and morality requires,
more weight than scientists commonly do; he
was giving it the weight engineering requires.
 889/1183

Here then is one place where empirical work,

historical or sociological, might help settle a dis-
pute between philosophers. Both the philosoph-
ical and casuistic approaches to engineering eth-
ics assume that engineers are not subject to any
special standards, that their local situation or
function is decisive in determining what they
should do (and so, what they are likely to do).
Studies of cases like that at Argonne, if they
generally turned out as I believe they would,
might provide substantial empirical evidence fa-
voring those approachesthe technical, social,
and especially professionalthat identify engin-
eers as subject to special standards (whether of
competence or conduct).

How do Engineering Decisions get Made?

 890/1183

Most engineering goes on in large organizations,

governmental or commercial. Large organiza-
tions exist to do large jobs, doing them by divid-
ing them into manageable
 Page 178

parts. If these parts are too small, engineers as-

signed one of them could not determine what ef-
fect their work would have on the public health,
safety, or welfare or even on their employer.
Their work would be "bureaucratized," in one of
the uglier senses of that ugly word. If most en-
gineering work is bureaucratized in this sense,
engineering ethics must either be irrelevant to
most engineers or consist of matters tangential
to engineering as suchfor example, treatment of
other engineers. Engineering ethicsas now con-
stitutedpresupposes a world in which engineers
generally know what they do.
 892/1183

Here, then, is a sense of the question, "What do

engineers know?" quite different from
Vincenti's but as worthy of investigation. My
own research, reported in chapter 9, and much
anecdotal evidence as well, convinces me that
engineers generally have a pretty good idea of
who will use their work and how; they know
what they do. Their work is not, and generally
cannot be, bureaucratizedor, at least, not without
prohibitive waste. Yet, although there is a large
literature on business organization, relatively
little of it is on technical organizationand very
little on engineers in particular. 21 Worse, virtu-
ally no work has been done on decision rules in
technical decision making. Although most of the
organizational literature makes it sound like
managers decide and employees, including en-
gineers, either submit, perhaps dragging their
feet, exit, or blow the whistle, the research re-
ported in chapter 9 suggests something quite dif-
ferent, a process in which consensus is the rule,
 893/1183

engineers generally have the power of veto over

management decisions, and engineers are well
informed in part at least because information is
necessary to win their consent.

What can Engineers do?

Most engineers are employees, subject to ter-

mination of employment at the employer's will.
Some writers have concluded from this that en-
gineering is a "captive profession," that engin-
eers have little room for professional autonomy,
and that therefore there is little room for engin-
eering ethics.22 I have already suggested one
reason to think this view mistakenat least in
some organizations, those that decide engineer-
ing questions by consensus. But what about an
organization in which engineers are well in-
formed but decision is not by consensus? What
room for professional autonomy there?
 894/1183

Here we need philosophical work on the concept

of "professional autonomy" sensitive to the
problems of empirical research. Chapter 10 is an
example of what can be done, but one taking the
professional approach to engineering ethics.
Those who do not share my approach may want
to develop a philosophically defensible alternat-
ive. Alternatives would allow the social sciences
to tell us whether different, philosophically de-
fensible definitions of professional autonomy
lead to different answers to empirical questions
about how much professional autonomy engin-
eers actually havefor example, which organiza-
tions, if any, actually eliminate or drastically
confine the professional autonomy of engineers.
We would then have a better idea when talking
about the professional responsibility of engin-
eers makes sense (and what philosophical com-
mitments, if any, underwrite such talk).
 Page 179

Conclusion

In 1994 the philosopher Carl Mitcham published

an expansive, thoughtful, and informative book,
Thinking through Technology. In one small
corner, he argued that engineering ethics is part
of science and technology studies. 23 While I
agree that engineering ethics is properly a part
of science and technology studiesor, at least,
could beI think that, as a matter of fact, it is not.
I have two reasons for so thinking.
 896/1183

First, engineering ethics has so far developed as

a field of professional ethics. The professions
include many nontechnical professions,
everything from journalism to accounting, from
lawyering to nursingfields that science and en-
gineering studies would have trouble absorbing.
So, at best, engineering ethics would have to di-
vide its citizenship between two distinct fields.
So far, it has not even done that. Even in a field
such as medical ethics, where empirical research
is much more extensive than in engineering eth-
ics and where technology is also a major source
of ethical issues, the people doing professional
ethics are different from those doing science and
technology studies. Whatever can be said about
science and technology studies and engineering
ethics as abstract fields, as living research com-
munities, they were, and remain, largely separ-
ate. So, for example, typical journals in science
and technology studies (Technology and Cul-
ture; Science Studies; Science, Technology, and
 897/1183

Human Values; and so on) would not appear on

most lists of professional ethics journals; nor
would typical professional ethics journals (Busi-
ness and Professional Ethics Journal, Interna-
tional Journal of Applied Philosophy, Science
and Engineering Ethics, and so on) appear on
most lists of journals in science and technology
studies.24
 898/1183

My second reason for thinking that engineering

ethics is not now part of science and technology
studies is that those working in science and
technology studies are generally indifferent to
professional ethics. Thinking through Techno-
logy bears the subtitle "The Path between
Engineering and Philosophy" because Mitcham
saw himself as mediating between two ways of
doing philosophy of technology, the engineering
way and the "humanities" way. Neither way
pays much attention to engineering as a profes-
sion. The chief concern of both is technology
rather than engineering. Rarely does someone
on Mitcham's long list of contributors to one
way or the other of doing philosophy of techno-
logy get closer to engineering ethics than tech-
nology assessment or public policy. That is not
very close.
 899/1183

Will science and technology studies contribute

substantially more to engineering ethics than
they have so far? I don't know. But I hope they
will. To understand moral problems we must see
them in context. To understand problems of en-
gineering ethics, we must understand the engin-
eering context. Who among social scientists are
better placed than those working in science and
technology studies to describe, interpret, and
otherwise improve our understanding of the
context of engineering?

I said "hope," not "expect," because a large bar-

rier seems to stand in the way of science and
technology studies doing the work I am calling
for. Science and technology studies grew up
concerned with knowledge (for science) and
things (for technology). Only in the last two
decades have science studies come to focus on
 Page 180

research communities rather than on science.

Technology studies still seems focused on
things, processes, and knowledge, on techno-
logy rather than on specific technological pro-
fessions. I hope this book, whatever else it does,
serves as an invitation to social scientists, espe-
cially those in science and technology studies, to
consider studying the technological professions
as professions, especially, engineering, the most
important technological profession of them all.
 Page 181

APPENDIXES
 Page 183

Appendix 1
Questionnaire for Engineers
Explain project. Assure anonymity. Then ask:
Are you an engineer or a manager?

1. What is your professional background?

a. How did you come to work here?

b. What does your company do?

Example?

2. What do you do here?

3. How does your company make engineering

decisions? Can you give an example?

a. What part do engineers play in import-

ant design and operation decisions here?
 903/1183

b. What part do managers play in im-

portant design and operation decisions
here?

4. What are the most important factors determ-

ining company decisions on matters of
engineering?

a. Does your company take large risks in

its technical decisions? Why?

b. Does your company have a code of

ethics?

5. Is the management of the company trained or

versed in the company's technology? How cur-
rent do you feel they are?
 Page 184

6. Are your engineering recommendations being

acknowledged in such a way that you receive
assurance that they have been received and will
be acted upon in accordance with your state-
ments? Explain.

a. What review process is in place for an

engineer's concern?

b. Do you have, and participate in, a pro-

cess of technical design review with
your peers? With management on critical
design specifications?

7. Do you think there are any communications

problems between your supervisor and his su-
pervisors? Examples?

a. Do you ever find it necessary to with-

hold information from your superiors? If
so, explain?
 905/1183

b. Have you ever felt that your superiors

were not telling you the whole truth? If
so, explain.

9. Have you ever felt that safety or quality were

being sacrificed for reasons with which you did
not agree? If so, explain.

a. What would you do if you thought

safety or quality were being sacrificed?

10. On what issues do you think professional en-

gineers should be content to see their judgment
superseded? On what issues should the
engineer's judgment be the last word?

11. If you don't like what your immediate super-

ior is doing, what can you do about it?

a. Does your firm have a formal open

door policy? Is it used to appeal technic-
al decisions? How does it work?
 906/1183

12. Are engineers good management material?

Why or why not?

a. What transition training or coaching is

provided for an engineer promoted into
management?

b. In what important ways must a pro-

moted engineer change?

13. If you had full control over the engineering

work in your company, what would you do dif-
ferently? Why?

a. Are your engineering recommenda-

tions being affected by considerations or
pressures that deny you the opportunity
to provide the optimum solution to some
problem?
 Page 185

13* What questions should a manager ask you to

get the information he needs to make the right
decision? Which, if any, of these questions is a
manager least likely to ask?

14. Are there any questions we didn't ask that

we should have? Anything you want to add to
what you have already said?
The star following a number indicates the ques-
tion was added after interviewing began.
 Page 186

Appendix 2
Questionnaire for Managers
Explain project. Assure anonymity. Then ask:
Are you an engineer or a manager?

1. What is your professional background?

a. How did you come to work here?

b. What does your company do?

2. What do you do here?

3. How does your company make engineering

decisions?

a. What part do engineers play in import-

ant design and operation decisions here?
 909/1183

b. What part do managers play in im-

portant design and operation decisions
here?

4. What are the most important factors determ-

ining company decisions on matters of
engineering?

a. Does your company take large risks in

its technical decisions? Why?

b. Does your company have a code of

ethics? What part does it play in your
decisions?
 Page 187

5. Is the company's management trained or

versed in the company's technology?

a. How current do you feel they are?

b. Should managers have a technical

background?

6. Do you and your engineers always see eye to

eye on technical questions? If not, when not?
What happens?

7. How much weight does an engineer's recom-

mendation have?

a. Does an engineer's technical expertise

weigh as heavily as management consid-
erations in making decisions?

b. What review process is in place for an

engineer's concerns?
 911/1183

8. On what issues should professional engineers

(on staff) be content to see their professional
judgment superseded? On what issues, if any,
should the engineer's judgment be the last word?

9. Do you ever find it necessary to withhold

technical information from your engineers? If
so, explain.

10. Have you ever felt that your engineers were

not telling you the whole truth? If so, explain.

11. Are engineers good management material?

Why or why not?

a. What transition training or coaching is

provided for an engineer promoted into
management?

b. In what important ways must a pro-

moted engineer change?
 912/1183

12. If you had full control over the engineering

work in your company, what, if anything, would
you do differently? Why?

a. Are your recommendations now being

affected or colored by considerations or
pressures that deny you the opportunity
to provide the optimum solution to some
problem?

12* What questions should an engineer ask you

to get the information he needs to make the right
decision? Which, if any, of these questions is an
engineer least likely to ask?

13. Are there any questions we didn't ask that

we should have? Anything you want to add to
what you have already said?
The star following a number indicates the ques-
tion was added after the interview began.
Appendix 3
Interviewee Characteristics
Table 1: Total interviewed
Engineers Ma
Customer-Oriented Companies (6)
3
3
2
3
5
3
Engineer-Oriented Companies (4)
0
3
3
4
29
 914/1183

This table does not include three background interviews with

eers. Thecompany listed with no engineers had a dozen or so
ager in the first trialof the questionnaire.

Table 2: Employment history

Engineers (n = 29)
No. of employers
Only 1 18
2 or more 11
No. of yrs w/present employer
03 5
39 11
1019 8
20+ 0
unknown 5
The range for engineers was from one to eighteen years; for m
Table 3: Fields of engineering (determined by degree or, in its
Civil Chemical Electrical Mechanical
2 4 12 20
In addition: One engineer (not counted above) claimed degree
two others (alsonot counted above) claimed a B.S. in construc
Of the remaining six interviewees,two had degrees in chemist
eers) and one had an associate degree in qualityassurance. Th
bring the total to 60.
 Page 191

NOTES

Preface

1. Charles Harris, Michael Pritchard, and Mi-

chael Rabins, Engineering Ethics: Concepts and
Cases (Wadsworth: Belmont, Mass., 1995).

2. Walter G. Vincenti, What Engineers Know

and How They Know It (Johns Hopkins
University Press: Baltimore, 1990).

3. Samuel Florman, The Existential Pleasures of

Engineering (St. Martin's Press: New York,
1976); and The Civilized Engineer (St. Martin's
Press: New York, 1987).

Chapter 1
 917/1183

This chapter began as a GTE Lecture at the

University of Wisconsin Center/Fond du Lac,
October 13, 1992. I should like to thank those
present, as well as my colleagues, Wilbur Ap-
plebaum (history) and Sid Guralnick (civil en-
gineering) and my friend, Mike Rabins (Texas
A & M) for help in sorting through the issues
discussed here. The chapter appeared in a some-
what longer version and under the title ''An His-
torical Preface to Engineering Ethics," in
Science and Engineering Ethics 1 (January
1995): 33-48. Reprinted by permission.

1. See, especially, Paolo Rossi, Philosophy,

Technology, and the Arts in the Early Modern
Era, translated by S. Attanasio (Harper and
Row: New York, 1970). The Oxford English
Dictionary gives 1615 as the first known use of
the word "technology" in English.
 918/1183

2. David Noble claims that Jacob Bigelow, a

Boston physician who helped found the Mas-
sachusetts Institute of Technology (1865) was
instrumental in introducing "technology" (in its
modern sense) into general usage. To establish
his claim, Noble offers this quotation from
Bigelow's Elements of Technology (1829):
There has probably never been an age in which
the practical applications of science have em-
ployed so large a portion of talent and enterprise
of the community, as in the
 Page 192

present. To embody . . . the various topics which

belong to such an undertaking, I have adopted
the general name of Technology, a word suffi-
ciently expressive, which is found in some of the
older dictionaries, and is beginning to be revived
in the literature of practical men at the present
day. Under this title is attempted to include an
account . . . of the principles, processes, and no-
menclatures of the more conspicuous arts, partic-
ularly those which involve the application of sci-
ence, and which may be considered useful, by
promoting the benefit of society, together with
the emoluments of those who pursue them.
(David Noble, America by Design [Alfred A.
Knopf: New York, 1977] pp. 3-4)
 920/1183

My reading of this passage does not agree with

Noble's. Bigelow refers to a revival of the term
already beginning; hence, by his own admission,
Bigelow probably is not "instrumental" in its re-
vival. Further, Bigelow's own use of the term is
barely distinguishable from that found in the old
dictionaries to which he refers. The chief differ-
ence is the observation that some of the more
"conspicuous arts" he describes ''involve the ap-
plication of science." He still seems far from the
modern idea of technology as either inventions
or the "science of invention."

3. See, for example, Plato's Theaetetus (III:

172-173): "[A] philosopher is a gentleman, but a
lawyer is a servant. The one can have his talk
out, and wander at will from one subject to an-
other, as the fancy takes him . . . [but] the law-
yer is always in a hurry."
 921/1183

4. Hannah Arendt, The Human Condition

(Doubleday: Garden City, N.Y., 1959) p. 323n.

5. "If we assume that the Middle Ages ended

with the fifteenth century, then a simple count of
inventions made or adopted by Europeans dur-
ing the period confirms that it was, as regards
technics, more creative than any previous epoch
in recorded history." (Dictionary of the History
of Ideas: Studies of Selected Pivotal Ideas, ed-
ited by Philip P. Weiner [Charles Scribner's
Sons: New York, 1973] vol. 4, p. 359 [D.S.L.
Cardwell, technology]) I am, of course, still
comparing Greece's Golden Age with a similar
stretch of time during the Dark Ages. Were I to
compare the Dark Ages with the Hellenistic
Period, I would hedge my claim a bit (and begin
to worry about how to count inventions).
 922/1183

6. See, for example, Spencer Klaw, "The Fausti-

an Bargain" in The Social Responsibility of the
Scientist, edited by Martin Brown (Free Press:
New York, 1971) pp. 3-18.

7. Who is this "we"? Certainly, you and I, but

probably, as well, most inhabitants of the planet.

8. The modern prejudice against manual labor

seems to vary from place to place and time to
time. It is certainly less in the United States
today than, say, in France a century ago. So, for
example, it is today hard to imagine the events a
French mechanical engineer (or mechanic) of
the nineteenth century recalled. After church, he
struck up a conversation with a young woman
(with her mother standing by). When she found
out that he built steam engines for a living, she
shuddered:
 923/1183

"What! You work, you are therefore exposed to

all the filth that trade includes?" A bit vexed I re-
sponded, "But yes, miss, and I dare to believe
that none is apparent at this moment." The moth-
er turned her back and the eyes of my beautiful
neighbor fell on my well-ground hands, which
did not betray me, and she moved away. For her,
I was a plague-stricken person. (Eda Kranakis,
"Social Determinants of Engineering Practice: A
Comparative View of France and America in the
Nineteenth Century," Social Studies of Science
19 (February 1989): 5-70, at 13)

The whole paper is well worth reading both for

the contrast it draws between French and Amer-
ican practice and for the cultural explanation it
offers.

9. Though this description of science is no doubt

biased in favor of the natural or physical sci-
ences, it also applies (with a bit of stretching) to
the social sciences. The social sciences can
 Page 193

be practiced as an attempt to understand hu-

man society from the outside, that is, as a
part of nature. Of course, many social scient-
ists now consider such value-free science to
be impossible and its attempt likely to
mislead.

10. World Almanac (World Almanac: New

York, 1989) p. 158. This number must be taken
only as a rough approximation. The Labor De-
partment (three years later) set the number of
engineers at 1,519,000 (Occupational Outlook
Handbook [U.S. Department of Labor, Bureau
of Statistics, Washington, D.C., May 1992], p.
64); whereas the National Science Foundation
put the number at 2,849,800 (U.S. Scientists and
Engineers: 1988 Estimates, Surveys of Science
Resources Series, National Science Foundation,
NSF 88-322, p. 6). Apparently, it is not easy to
count engineers. Why?
 925/1183

11. "Principles of Medical Ethics" in Codes of

Professional Responsibility, edited by Rena A.
Gorlin (Bureau of National Affairs: Washing-
ton, D.C., 1986) p. 99.

12. The first of these modern professions was

the apothecaries, a profession now deceased,
which reorganized in 1815. The other liberal
professions followed only slowly, beginning
with the solicitors in the 1830s. See W.J. Read-
er, Professional Men: The Rise of the Profes-
sional Classes in Nineteenth-Century England
(Basic Books: New York, 1966) esp. pp. 51-55.
 926/1183

13. Thomas Percival, Medical Ethics; or A Code

of Institutes and Precepts Adapted to the Profes-
sional Conduct of Physicians and Surgeons
(1803). The word "institutes" suggests that
Percival's model here is in part at least jurispru-
dence. (Since the emperor Justinian's famous
textbook, "institutes" signals that a book so
titled is a textbook or summary of the law, the
rest of the title telling the particular jurisdiction,
for example, Coke's Institutes of the Laws of
England.) Percival in fact makes this connection
in his introduction, indicating that he originally
intended to call his text "medical jurisprudence."
About half the text is a summary of English law
a physician should know. Much of the rest con-
sists of ''precepts" (i.e., advice) rather than of
standards of conduct (ethics strictly so-called).
Medical ethics in the modern sense is actually a
small part of that seminal work.
 927/1183

14. See, for example, M. David Burghardt, In-

troduction to the Engineering Profession
(HarperCollins: New York, 1991) p. 26: "We
shall assume that wherever there was an inven-
tion or innovation, engineering was required."
Burghardt does not say the same about engin-
eers. Can there be engineering without engin-
eers? Or Billy Vaughn Koen, Definition of the
Engineering Method (American Society of
Engineering Education: Washington, D.C.,
1985) p. 26: "After 20 or 30 centuries, the en-
gineer learned how to correct this problem by
allowing the front axle to pivot on a king bolt as
stage three in the evolution of cart design."
Engineers three or four thousand years ago?
 928/1183

15. See, for example, Ralph J. Smith, Engineer-

ing as a Career, 3rd (McGraw-Hill: New York,
1969) p. 22: "It has been said that the history of
civilization is the history of engineering. Cer-
tainly it is true that the highly developed civiliz-
ations have all been noted for their accomplish-
ments in engineering." Substitute "building" for
engineering and there would be nothing to ob-
ject to. The same is true of scholarly works such
as Donald Hill, A History of Engineering in
Classical and Medieval Times (Croom Helm:
London, 1984). A careful researcher and writer,
Hill argues for his application of "engineer" to
ancient builders (rather than, as most writers do,
just assuming the application to be obviously
justified). Yet he soon admits that "classical and
medieval engineers did not have a quantified,
scientific basis for their designs" (p. 5), that they
lacked the formal training characteristic of mod-
ern engineers, and that they even lacked a full-
time occupation (p. 7). They did engineering
 929/1183

(read "building") as a sideline. So, whatever

they were, they were not a professionor even an
occupation.

16. Though we now translate corps du génie as

"corps of engineers," there is in fact no exact
English equivalent. While génie corresponds to
the English "gin" (as in "cotton gin") and per-
haps "jenny" (as in "spinning jenny") the French
has less suggestion of "engines." Perhaps the
best translation would be "corps of the con-
triver"though this lacks the sug-
 Page 194

gestion of magic (as in the English "genie").

Unlike the corps du sappeur ("the corps of
spaders") for example, the corps du génie
seems not to have taken its name from the
implements it used but to have given those
implements its name. This suggests an inher-
ent novelty in what it did.

17. See Frederick B. Artz, The Development of

Technical Education in France, 1500-1850
(MIT Press: Cambridge, Mass., 1966) p. 48; or
W.H.G. Armytage's history of engineering in
Britain, (misnamed) A Social History of Engin-
eering (Faber and Faber: London, 1961) pp. 96,
99.
 931/1183

18. Of course, many changes were made in the

engineering curriculum (as well as many experi-
ments with less demanding curriculum). Whole
new subjects, such as thermodynamics or elec-
tricity, were added, and geometry and trigono-
metry gave way to a second year of calculus.
For engineers, these may appear more than im-
provements in detailand, for many purposes,
they certainly are. Yet, for our purposes, such
differences between the French curriculum of
1799 and today's typical engineering curriculum
hardly affect the gap between what engineers
learn and what lawyers, physicians, or even ar-
chitects learn. Those who wish to understand
what distinguishes engineering from other pro-
fessions must pay attention to what engineers
have in common, especially over long stretches
of time, rather than what divides them.
 932/1183

19. Artz, Technical Education, 47-48. I should

perhaps warn that the members of this corps do
not seem to have been known as "civil engin-
eers" (ingénieurs civils) but as "road and bridge
engineers." The French seem to have reserved
"civil engineer" for engineers employed by
private persons. All members of a corps, wheth-
er of the corps du génie (militaire) or of the
corps des ponts et chaussées, were state em-
ployees. The English term ''civil engineer" may
have derived from a misunderstanding of the
French term (since the English, with a relatively
weak state, had no exact counterpart to the corps
des ponts et chaussées). Here is work for histori-
ans. Compare Kranakis, "Social Derterminants,"
esp. pp. 29-30.

20. Engineers also had some secret methods (for

example, Monge's descriptive geometry). Artz,
Technical Education, 106.
 933/1183

21. Ibid., 81-86.

22. But note Peter Michael Molloy's remark in

Technical Education and the Young Republic:
West Point as America's École Polytechnique
(unpublished dissertation, Brown University,
1975) p. 105: "From a description of the cur-
riculum, there should be no mystery for the
change in the School's name in 1795 from École
des Travaux Publics to École Polytechnique." It
remains a mystery to me.

23. Artz, Technical Education, 154-155.

 934/1183

24. Ibid., 160. I should, perhaps, say "this École

Polytechnique." By 1830, the École Polytech-
nique had become so devoted to mathematics
that (by American standards of the day) its
graduates generally did not seem to practice en-
gineering. This later École Polytechnique may,
then, provide an early example of the ability of
science education to crowd out engineering. See
Molloy, Technical Education, esp. pp. 119-130.
On the other hand, this interpretation may
simply be unfair (within the French context).
See Kranakis, "Social Determinants," esp. pp.
22-29, for reasons to think that the École Poly-
technique remained an engineering school
throughout the nineteenth century.

25. Artz, Technical Education, 160-161. I

should perhaps add that only practical diffi-
culties seem to have prevented all this from hap-
pening as early as 1802. Malloy is very good on
this.
 935/1183

26. Eugene Ferguson, "The Imperatives of

Engineering" in John G. Burke et al., Connec-
tions: Technology and Change (Boyd and
Fraser: San Francisco, 1979) pp. 30-31.
Ferguson's "imperatives" are, of course, Koen's
"heuristics."

27. For a detailed study of one of these proxy

measures that, in the end, had to be dis-
 Page 195

carded, see Walter G. Vincenti's discussion

of "stability" in What Engineers Know and
How They Know It (Johns Hopkins
University Press: Baltimore, 1990) pp.
51-108.

28. In the absence of government action, engin-

eering societies developed codes both to en-
hance efficiency (by promoting standardization)
and to maintain safety. These codes set stand-
ards for engineers to follow "voluntarily" and
provide "model codes" legislatures can adopt.
For more on this point, see chapter 7 (in this
volume). Since this activity has been criticized
as a usurpation of a governmental function,
Ferguson's complaint that engineers should have
done more of it seems unfairat least until he of-
fers a theory of which activities belong to gov-
ernment and which to private organizations or
individuals.
 937/1183

29. Not all engineering designs dumb down a

job. Many engineering designs "automate," that
is, eliminate the routine work of many while
creating technically sophisticated jobs for a few.
These designs may be regarded as the limiting
case of dumbing down or as an entirely different
way of doing without large numbers of highly
skilled workers. I don't think much turns on how
it is categorized.

30. For more on the military connections of en-

gineering, together with the connections
between the military and technology, see Barton
Hacker's, "Engineering a New Order: Military
Institutions, Technical Education, and the Rise
of the Industrial State," Technology and Culture
34 (April 1993): 1-27.

31. Artz, Technical Education, 162.

 938/1183

32. Is this a description or a prescription, a state-

ment of what civil engineering is or a statement
of what it should be? That is not a hard question
but it is one that should give philosophers a
reason to pay more attention to professions than
they generally do. Thedgold's definition is prob-
ably both descriptive of civil engineers gener-
ally and, for that reason, prescriptive for anyone
who wants to carry on his occupation under the
title "civil engineer." How can that be? That is a
question addressed here in chapters 4 and 10.

33. Accreditation Board of Engineering and

Technology, Code of Engineering Ethics, first
principle (1985). (emphasis added). Whereas
tense suggests description, context suggests pre-
scription. Here again we see the congruence of
description and prescription characteristic of
professions.
 939/1183

34. Compare Koen, Definition, esp. 63-65. Koen

rightly points out that engineers sometimes go
beyond science (what we now know) and some-
times ignore science (because truth is too ex-
pensive) and so at least part of the time cannot
be said to be applying science. This is, I believe,
an important point but one quite distinct from
the one I am making. Perhaps engineering dif-
fers from applied science in enough ways that
the interesting question is not how engineering
differs from applied science but why they were
ever thought to be the same.
 940/1183

35. All "exceptions" are in the electrical engin-

eering department. Whether they are truly ex-
ceptions is a matter about which I remain uncer-
tain. On the one hand, these professors of engin-
eering were educated as engineers; on the other,
they considered their location in an engineering
department to be an accident. They admit-
tedindeed, declaredthat they could just as well,
or even better, have been lodged in a physics de-
partment. They seem, therefore, to have lost
their identity as engineers along with their in-
terest in helping to make something useful. In
that respect at least, they constitute evidence for,
rather than against, my claim.
 941/1183

36. Compare Vincenti, What Engineers Know,

161: "Engineers are after a theory they can use
for practical calculations. . . . To obtain such a
theory they are willing, when necessary, to
forgo generality and precision . . . and to tolerate
a considerable phenomenological component.
Scientists are more likely to be out to test a the-
oretical hypothesis . . . or infer a theoretical
model."

37. For some historical background on this use

of "ethics," as well as my rationale for prefer-
ring it, see Michael Davis, "The Ethics Boom:
What and Why," Centennial Review 34
 Page 196

(Spring 1990): 163-186. For those who ask

why codes of professional ethics must be
morally permissible. The short answer is: If
not morally permissible, they cannot be mor-
ally binding; if not morally binding, they
would seem to be more accurately described
as an ethic, mores, ethos, custom, or practice
than as ethics (strictly speaking). Insofar as
this book proves how useful my way of un-
derstanding professional ethics is, it
provides close to a full answer to that
question.
 943/1183

38. I am, of course, assuming that American en-

gineers did not have an "unwritten code" before
then. That may seem a daring assumption. It is
not. Most "unwritten law" is in fact written. For
example, the "unwritten constitution'' of Eng-
land is recorded in royal charters, parliamentary
debates, case law, and even newspaper reports.
It is unwritten only in the sense that there is no
authoritative document such as the U.S. Consti-
tution. Humans have great trouble coordinating
what they do without putting expectations into
words and, in large organizations, without put-
ting those words on paper. So, the apparent ab-
sence of any written codes of engineering ethics
in the United States before 1900even in the un-
official form that Percival produced for English
physicians a century earlier or that Sharswell
produced for American lawyers in the 1830sis, I
think, decisive evidence against the existence of
any unwritten code.
 944/1183

39. Compare chapter 8 (in this volume).

40. Though I make this claim about what moral-

ity demands without argument, and in the belief
that it is both true and obvious, I should admit
that certain utilitarians, those moral theorists
who think morality consists in maximizing over-
all happiness or social utility, do not. Their the-
ory makes morality much more demanding,
which has proved a problem for their theory, not
a problem for ordinary moral agents.

Chapter 2
 945/1183

This chapter began as the first Annual Engineer-

ing Ethics Lecture, funded by GTE, at Wayne
State University, Detroit, Michigan, November
19, 1992. I would like to thank those present for
a useful discussion. I should also thank Mike
Rabins (mechanical engineering, Texas A & M)
and my colleagues Tom Misa (history) and Sid
Guralnick (civil engineering) for many helpful
comments on an earlier draft, and Bill Pardue
for helping to track down many of the refer-
ences given here.

1. See, for example, Lawrence P. Grayson, "The

American Revolution and the `Want of Engin-
eers,'" Engineering Education 75 (February
1985): 268-276.

2. See chapter 1 for a defense of this claim.

3.
 946/1183

Sylvanus P. Thayer was appointed director in

1817, by which time enrolment and teaching
staff had increased to 250 cadets and 15 profess-
ors covering mathematics, "engineering," and
natural philosophy, recently joined by Claude
Crozet (1790-1864) a graduate of the École Poly-
technique, who introduced the teaching of de-
scriptive geometry to the college, and in 1821
published the first textbook on the sub-
ject. . . . Thayer graduated from Dartmouth in
1807, and from the Military Academy in 1808.
He had studied military engineering develop-
ments in France and this influence was evident in
his reorganization of the curriculum and mode of
instruction at West Point. He used texts em-
ployed at the École Polytechnique, divided
classes into small sections, required weekly class
reports, and developed a grading system. (Ge-
orge S. Emmerson,
 Page 197

Engineering Education: A Social History [Crane,

Russak & Company: New York, 1973], pp.
140-141).

While Thayer seemed to dislike the overly the-

oretical approach the French took to engineering
education, his primary reason for not taking
over more of the curriculum of the École Poly-
technique seems to have been the relatively poor
preparation of American students (and the de-
sire, or necessity, not to make admission too
difficult).
 948/1183

4. Though Partridge did now and then teach a

course in civil engineering at West Point, he was
not a civil engineer (and, apparently, was barely
qualified to teach the course). Because he rejec-
ted most of the innovations introduced by his
successor as superintendent, Sylvanius Thayer,
the West Point Partridge tried to reproduce was
the pre-1817 version (which may have much to
do with the failure of Norwich to equal Thayer's
West Point in either quality or quantity of engin-
eers graduated). For more on this subject, see
Thomas J. Fleming, West Point: The Men and
Times of the United States Military Academy
(William Morrow: New York, 1969) esp. pp.
3-14, 34.
 949/1183

5. Daniel Hovey Calhoun, The American Civil

Engineer: Origins and Conflict (Technology
Press [MIT]: Cambridge, Mass., 1960) p. 45.
Compare James Gregory McGivern, First Hun-
dred Years of Engineering Education in the Un-
ited States (1807-1907) (Gonzaga University
Press: Spokane, Washington, 1960) pp. 38,
42-45; and Emmerson, Engineering Education,
141-142. I have not found the corresponding
figures for the number of military engineers
(though, given that the Army does not seem to
have been able to absorb all West Pointers, few
Norwich graduates could have found work as
military engineers during this period).
 950/1183

6. All the histories of engineering education

cited here ignore both the Virginia Military In-
stitute and the Citadel. Most also ignore both
Annapolis and the impact of naval engineers on
the development of mechanical engineering in
the land-grant schools after the Civil War. For
one who does not, see Monte A. Calvert, The
Mechanical Engineer in America, 1830-1910
(Johns Hopkins Press: Baltimore, 1967) esp. pp.
48-51. A surprising number of engineering
schools werelike Texas A&Mvirtually military
academies until the 1960s. Such facts lead me to
suspect that the relation between engineering
and military education was, until quite recently,
a lot closer than the histories of engineering
education indicate. That relationship might ex-
plain much about the characteristic attitudes of
American engineers in times pastand why some
of these may be fading (for example, engineers'
political conservativism).
 951/1183

7. McGivern, First Hundred Years, 50-51. See

also Ray Palmer Baker, A Chapter in American
Education: Rensselear Polytechnic Institute,
1824-1924 (Charles Scribner's Sons: New York,
1924) pp. 48-56.

8. Baker, American Education, 35, 44-46. But

about twenty-five of Rensselear's graduates
from the period before 1840 eventually became
engineers. Calhoun, American Civil Engineer,
45.

9.
Though Eaton [the school's first director] had in-
sisted that most colleges attempted to teach so
many subjects that they could teach none of them
well, and that Rensselaer should limit its activit-
ies primarily to the sciences, progress in them
had been so rapid that Greene [the new director
in 1847] concluded that it was again time [for the
school] to narrow its field. (Baker, p.p. 39-40).
 952/1183

Note that engineering is here considered part of

"the sciences."

10. Frederick B. Artz, The Development of

Technical Education in France, 1500-1850
(MIT Press: Cambridge, Mass., 1966) p. 267.
 Page 198

11. Emmerson states (without evidence) that

Rensselear's new curriculum was modeled not
on that of the École Polytechnique but on the
École Centrale [d'Arts et Metiers] of Paris. Em-
merson, 148, 153-156. McGivern says the same.
See McGivern, 59. But neither tries to explain
Rensselaer's "Polytechnic" (or what difference
Greene would have seen between these
institutions).
 954/1183

12. The British did, it is true, establish a school

of military engineering at Woolwich in 1741.
The school retained a number of notable applied
mathematicians who wrote some elementary
textbooks engineers found useful. Emmerson,
Engineering Education, 33. Yet, unlike West
Point, Woolwich seems to have had little influ-
ence on engineering generally, or on engineer-
ing education in particular, even in England, un-
til the second half of the nineteenth century (if at
all), that is, not until after talent replaced patron-
age as the primary means of gaining entry (and
something like the French curriculum was adop-
ted). Reader, Professional Men, 96-97. Compare
Artz, Technical Education, 261.
 955/1183

13. W.H.G. Armytage, Social History of Engin-

eering (Faber and Faber: London, 1961) pp.
160-161. Jonathan Williams, the first superin-
tendent of West Point, observed in 1802: "To be
merely an Engineer . . . is one thing, but to be an
Officieur du Génie is another. I do not know
how it happened but I cannot find any full Eng-
lish Idea to what the French give to the profes-
sion." Quoted in Peter Michael Molloy, Tech-
nical Education and the Young Republic: West
Point as America's École Polytechnique,
1802-1833 (unpublished dissertation, Brown
University, June 1971) 241-242. The irony, of
course, is that (as I explained in chapter 1) when
the term "engineer" was brought into English, it
was brought in to name people who had special
skills similar to those that distinguished the
French officieur du génie from the architects,
millwrights, and the like that the English already
had. Because the English (and Americans) were
not yet able to copy the French method of
 956/1183

educating engineers, "engineer" in English could

not carry the same import as officieur du génie.
Perhaps today, the term Williams so felt the
need for would be "professional engineer" (or
"degreed engineer''). Molloy is very good on
American backwardness in understanding en-
gineering. See, esp., pp. 425-463.

14. For a more or less complete listing of the

dozen or so "engineers or quasi-engineers"
available for public works in the United States
before 1816, see Calhoun, American Civil
Engineer, 7-23. Calhoun is also good on what in
American ways of doing things made it hard for
even these few to find employment.

15. McGivern, First Hundred Years, 15-23.

 957/1183

16. Charles F. O'Connell, Jr., "The Corps of

Engineers and the Rise of Modern Management,
1827-1856," in Military Enterprise and Techno-
logical Change, edited by Merritt Roe Smith
(MIT Press: Cambridge, Mass., 1985) pp. 95-96.

17. Merritt Roe Smith, "Army Ordnance and the

`American System' of Manufacturing,
1815-1861," in Military Enterprise, pp. 40-86.

18. James Kip Finch, The Story of Engineering

(Doubleday: Garden City, N.Y., 1960) pp.
262-265. This Erie Canal school is, then, a
throwback to the first days of the corps du
genie. See chapter 1 (in this volume).

19. Finch, Story, 267-269.

20. Ibid, 268-269.

 958/1183

21. O'Connell, in Military Enterprise, 100-106.

Note the initial resistence of the civilians to
army-style standardization.

22. Smith, in Military Enterprise, 77-78. See

also David A. Hounshell, From the American
System to Mass Production, 1800-1932: The
Development of Manufacturing Technology in
the United States (Johns Hopkins University
Press: Baltimore, 1984).

23. Edwin T. Layton, Jr., The Revolt of the

Engineers (Case Western Reserve University
Press: Cleveland, 1971) p. 3. The Census used
the term "civil engineer." Layton believes that
 Page 199

term would, at that time, probably have in-

cluded mechanical engineers (and, indeed,
all other nonmilitary engineers). My guess is
that "civil engineer" probably excluded most
engineers (or proto-engineers) in mining and
manufacture, who did not call themselves
engineers (and certainly would not have
called themselves "civil engineers"). Note,
for example, the fields listed for the
Lawrence Scientific School at about this
time. This disagreement with Layton is, non-
etheless, probably a quibble. Before the
Civil War, the number of these other engin-
eers was probably small compared to the
number of civils. Here, though, it would be
good to have more information.
 960/1183

24. The phrase in quotes is Steve Goldman's.

See "The Social Captivity of Engineering" in
Critical Perspectives on Nonacademic Science
and Engineering, edited by Paul Durbin (Lehigh
University Press: Bethlehem, Pa., 1991) pp.
121-146. But the sentiment seems to be wide-
spread. See, for example, David Noble, America
by Design (Alfred A. Knopf: New York, 1977).
Noble's nostalgia for the lost shop culture seems
to confuse inventing in general, which, indeed,
can exist in small, and even isolated, organiza-
tions, and engineering (which is a special kind
of inventing: centralizing, standardizing, and so
on), which probably cannot. The shop culture,
however admirable, seems to lose out to engin-
eering in certain environmentscapitalist or not.
(For example, engineers had much the same role
in the Soviet Union as in the United States.)
Noble contributes to our understanding of what
might give engineering an advantage over shop
culture (while giving that advantage a cast more
 961/1183

sinister than necessary on the facts even as he

presents them).

25. An ASCE was actually founded in 1852, its

membership almost entirely in New York City.
But, like other attempts at organizing engineers
before the Civil War, that ASCE seems to have
died out within a few years. The connection
with the ASCE of 1867 is tenuous, another ex-
ample of professions trying to add to their lin-
eage. For a bit more on this, see Layton, Revolt,
28-29.

26. For some of this history, see also Bruce Sin-

clair, A centennial History of the American Soci-
ety of Mechanical Engineers, 1880-1980
(University of Toronto Press: Toronto, 1980);
Terry S. Reynolds, 75 Years of Progress: A His-
tory of the American Institute of Chemical
Engineers (American Institute of Chemical
Engineers: New York, 1983).
 962/1183

27. A. Michal McMahan, The Making of a Pro-

fession: A Century of Electrical Engineering in
America (Institute of Electrical and Electronic
Engineers: New York, 1984) chap. 11.

28. Perhaps the first branching came even earli-

er, with the split between artillery and military
engineering. The roots of the two words"engine"
(from Latin ingenium for a natural ability or
genius) and "artillery" (from Latin ars for skill
or art)suggests how close their relationship ori-
ginally was.
 963/1183

29. James Kip Finch, A History of the School of

Engineering, Columbia University (Columbia
University Press: New York, 1954) pp. 65-66.
This was also the time when the faculty of the
school was renamed "the Faculty of Applied
Science." French engineering seems to have
grown from a single seed, but the American
looks much more like three trees that grew into
one (the French civil and military engineers, the
German mining and metallurgical "engineers,"
and the American and perhaps English mechan-
ical "engineers") branching even as they
combined.

30. Layton, Revolt, 3.

31. Compare Billy Vaughn Koen, "Toward a

Definition of the Engineering Method," Engin-
eering Education 75 (December 1984):
150-155.

32. Vincenti is very good on this subject.

 964/1183

33. Consider, for example, the expression "rock-

et scientist." In fact, there are virtually no rocket
scientists. Almost everyone associated with the
design, development, testing, deployment, and
operation of rockets is an engineer. Whatever
success rocketry has
 Page 200

had is largely due to engineers. "Rocket sci-

entists" should not be getting credit for any
of it.

34. Engineers also have a tendency to claim suc-

cesses for engineering whether or not the person
responsible was in fact an engineer by training.
It is this tendency that leads engineers to claim,
for example, that the builder of an Egyptian pyr-
amid or the inventor of the cotton gin was an en-
gineer. There is, it seems to me, considerable
unfairness in claiming the successes for engin-
eering while (as generally happens) blaming the
failures on others, for example, "managers,"
"tinkerers," "technicians," or "scientists.'' I have
therefore tried to develop a more even-handed
concept of engineering.
 966/1183

35. The only exception I know of is recent: in

some "software engineering," where engineers
(or other programmers) directlyon a computer-
construct programs for computers. They do not
write instructions for human beings (even in the
indirect way engineers in research and develop-
ment do) except insofar as they prepare the ne-
cessary documentation. Through their computer,
software engineers actually give directions dir-
ectly to "mechanical workers." Of course, as
chapter 3 demonstrates, there are other reasons
to wonder whether software engineering is en-
gineering at all. For our purposes now, however,
it is enough to point out that even these engin-
eers, if that is what they are, must, while in-
structing machines, take into account the human
environment in which the machines operate.
Their technical knowledge, like that of most en-
gineers, still includes much about how people
and things work together.
 967/1183

36. Compare Calhoun, American Civil Engin-

eer, 77: "The engineer role was specialized out
of the executive role." Even an engineer work-
ing in research and development is engaged in
developing instructions for production of some
safe and useful physical systemif she is working
as an engineerhowever many steps may stand
between the original research and the final
product. Koen's otherwise intelligent discussion
of design seems to miss entirely the role of
design as instruction to others.
 968/1183

37. Several professors of engineering have told

me that this is now changing, that engineers are
increasingly working in groups bringing togeth-
er engineers from different fields. This may be,
but my own interviews with working engineers
did not reveal much integration, even in re-
search. Here we have an empirical question
about which it would be good to have more in-
formation. But, whatever turns up, I am sure that
engineers will not for many years achieve the
integration of fields commonplace in law offices
or hospitals.
 969/1183

38. For example, the most scientific of the major

engineering societies, the Institute of Electrical
and Electronic Engineering (IEEE), is the only
one to forget that it had a code of ethics (redis-
covering it in the 1970s only after it had written
a new one). For a bit more on this, see Michael
Davis, "The Ethics Boom: What and Why,"
Centennial Review 34 (Spring 1990): 163-186,
esp. pp. 173-174. The IEEE's recent efforts in
ethics seem to signal an important change. But
do they? It would be interesting to have a de-
tailed analysis of what is really going on.

39. For example:

 970/1183

Much time is wasted in our colleges and technic-

al schools over higher mathematics. Every engin-
eer will have to agree with me that the cases
where the use of the higher calculus is indispens-
able are so few in our practice, that its study is
not worth the time expended upon it, and we
have the highest authority for saying unless its
use is constantly kept up we become too rusty to
use it at all. Unless the student possesses ex-
traordinary genius for mathematics, I would limit
its study to the ordinary analysis. (Thomas C.
Clarke, "The Education of Civil Engineers,"
Transactions of the American Society of Civil
Engineers 3 (1875): 557 [quoted in McGivern,
First Hundred Years, 113]).
 Page 201

Because, even now, I hear practicing engin-

eers make this point, I must wonder whether
teaching calculus (two years of it at present)
may not have more to do with shaping the
mind (or "weeding out" a certain sort of
mind) than with imparting the calculus itself.
It is easy to imagine programs in which
much of calculus is a technical elective and
the remainder integrated (in practical form)
into engineering science courses themselves.
For a longer and more critical discussion of
the calculus requirement (one rather hard on
it) see Sally Hacker, Doing it the Hard Way
(Unwin: Boston, 1990) pp. 139-154.

40. For an interesting discussion of this debate,

though largely limited to mechanical engineer-
ing, see Calvert, Mechanical Engineer, 63-85.

41. Calhoun, American Civil Engineer, 45.

 972/1183

42. Ibid., 50-53.

43. The practical success of West Point is easy

to underestimate. Consider, then, what was said
by Francis Wayland, president of Brown
University, 1827-1855. Near the end of his term,
which included bringing engineering to Brown,
he observed enviously that "the single academy
at West Point, graduating annually a smaller
number than many of our colleges, has done
more toward the construction of railroads than
all our one hundred and twenty colleges united."
Quoted in McGivern, First Hundred Years, 91.

44. Ibid., 152-154.

45. Calvert, Mechanical Engineer, 203.

46. McMahan, Making of a Profession, 33-43.

 973/1183

47. Compare: "The Society would have been a

small one and of limited influence had its mem-
bership been restricted to the type of consulting
or creative engineer alone. The factory engineer
is more and more a manager of men. . . . The en-
gineer must be what he is often called, a busi-
nessman." Frederick R. Hutton (1907) Secretary
of the American Society of Mechanical Engin-
eers. Quoted in Layton, Revolt, 37.

48. Of course, some engineering societies, espe-

cially in their early years, admitted into mem-
bership persons who, though not school-trained,
were "in responsible charge" of engineering
work for a number of years. The criterion was, it
should be noted, not simply "being in charge"
but being in ''responsible charge" for a certain
length of timelong enough, presumably, for the
person to show that he could do the job. And,
even this criterion looks more like a political
compromise than a natural definition.
 974/1183

49. Compare Layton, Revolt, esp. 58-60.

50. See, esp., ibid., 25-52. This vagueness may

explain why (like the original ASCE) at least
one twentieth-century engineering society, the
short-lived American Association of Engineers,
allowed architects to join. See Peter Meiksins,
"Professionalism and Conflict: The Case of the
American Association of Engineers," Journal of
Social History 19 (Spring 1983): 403-421, esp.
406. The exclusion of rank-and-file workers
may indicate a class bias, but I think it indicates
more than that. Many people who called them-
selves engineers, for example, train drivers or
scientific tinkerers, would have seemed ignorant
of much engineers had in common, even engin-
eers who came up through the ranks. What
Layton in fact reports is, I think, part of the pro-
cess by which "engineer" came to mean in Eng-
lish what it did in French (and what Williams
understood by officieur du génie).
 975/1183

51. For a hilarious example of how too much

emphasis on "science" can interfere with the
practice of engineering, see Bruce Seely, "The
Scientific Mystique in Engineering: Highway
Research at the Bureau of Public Roads,
1918-1940," Technology and Culture 24 (Octo-
ber 1984): 798-831. Note also Edna Kranakis's
description of the decline of French engineering
during the nineteenth century, "Social Determin-
ants of Engineering Practice: A Comparative
View of France and America in the Nineteenth
Century," Social Studies of Science 19 (Febru-
ary 1989): 5-70.
 Page 202

52. McGivern, First Hundred Years, 65. Note

that engineering here means what we now call
civil engineering. Though grouped with civil en-
gineering, both what we call mining (and metal-
lurgical) engineering and mechanical engineer-
ing are not conceived as engineering (properly
so-called). Here is further evidence that we
should be more cautious about thinking of en-
gineering as "fragmenting" during the nine-
teenth century. As I would tell the story, higher
education played a crucial part in giving engin-
eering a unity it did not originally have in the
United States (and might never have achieved
otherwise). In this regard, it is worth noting that
early civil engineers seem generally to have
failed at both mechanical engineering and min-
ing. Calhoun, American Civil Engineer, 82-87.

53. McGivern, First Hundred Years, 64-69.

 977/1183

54. Ibid., 79-82. Compare the history of the

École Polytechnique after 1804.

55. See Bruce Seely, "Research, Engineering,

and Science in American Engineering Colleges:
1900-1960," Technology and Culture 34 (April
1993): 344-386); and Lawrence P. Grayson, "A
Brief History of Engineering Education in the
United States," Engineering Education 68
(December 1977): 246-264, esp. 257-261.
 978/1183

56. Engineers were, of course, aware quite early

that engineering had a creative aspect. But other
aspects of engineering, especially the drudgery
of drafting and calculating, may have meant that
few engineers actually got to be "creative." If so,
then computers may have shifted dramatically
the balance between drudgery and creativity;
that shift may, in turn, partially explain the cur-
rent emphasis on design. But what explains the
decline of "shop training"? (Even would-be em-
ployers do not seem to want engineering schools
to prepare students for the shop floor.) Has en-
gineering changed in some fundamental way in
this century (or has industry)?

57. For more, see chapter 10 (in this volume).

58. This is quite clear in, for example, Walter G.

Vincenti, What Engineers Know and How They
Know It (Johns Hopkins University Press: Bal-
timore, 1990).
 979/1183

59. For others who have noted the sad state of

our understanding of engineering, see James K.
Feibleman, "Pure Science, Applied Science,
Technology, Engineering: An Attempt at Defin-
itions," Technology and Culture 2 (Fall 1961):
305-317: M. Asimov, "A Philosophy of Engin-
eering Design," in Friedrich Rapp, ed., Contri-
butions to a Philosophy of Technology (Reidel:
Dordrecht, Holland, 1974) pp. 150-157; George
Sinclair, "A Call for a Philosophy of Engineer-
ing,'' Technology and Culture 18 (October
1977): 685-689; Taft H. Broome, Jr. "Engineer-
ing the Philosophy of Science," Metaphilosophy
16 (January 1985): 47-56; Paul T. Durbin,
"Toward a Philosophy of Engineering and
Science in R & D Settings," in Paul Durbin, ed.,
Technology and Responsibility (Reidel:
Dordrecht-Holland, 1987) pp. 309-327.
 980/1183

60. This is, of course, not intended as a defini-

tion of "profession" but merely as a sketch of
one, one adequate for our purposes now. For
more of what I mean by "profession," see
chapters 4 and 10, and some of my other works
on the subject: "The Moral Authority of a Pro-
fessional Code," NOMOS 29 (1987): 302-337;
"The Use of Professions," Business Economics
22 (October 1987): 5-10; "Vocational Teachers,
Confidentiality, and Professional Ethics," Inter-
national Journal of Applied Philosophy 4
(Spring 1988): 11-20; "Professionalism Means
Putting Your Profession First," Georgetown
Journal of Legal Ethics (Summer 1988):
352-366; "Do Cops Really Need a Code of Eth-
ics," Criminal Justice Ethics 10 (Summer/Fall
1991): 14-28; "Science: After Such Knowledge,
What Responsibility?," Professional Ethics 4
(Spring 1995): 49-74.
 981/1183

61. Quoted in McGivern, First Hundred Years,

106. At the same place, he offers similar ex-
amples from the American Institute of Mining
(1873) and the American Society of Mechanical
Engineers (1880).

62. Grayson, "Brief History," 254.

 Page 203

63. Ibid., 258. Today that organization is the

Accreditation Board of Engineering and Tech-
nology (ABET).

64. This claim will seem controversial only to

those, mostly sociologists and those who defer
to them, who wish to equate "profession" with
"skilled occupation" (or with "licensed skilled
occupation"). There are at least two reasons to
reject this equation. First, members of a profes-
sion are usually at pains to claim that they be-
long to a profession, not just a skilled occupa-
tion. The equation makes their claim false by
definition, leaving the question why anyone
would say such a thing. Second, as we shall see,
ethical standards do seem to give considerable
insight into talk about "profession.''

65. For more on this, see Michael Davis, "The

Ethics Boom."
 983/1183

66. For a good (if somewhat jaundiced) account

of this period, with its effects both on industry
and engineering, see Noble, American by
Design.

67. The electrical engineers seem to have had

the greatest difficulty here (an eight-year pro-
cess). See McMahan, Making of a Profession,
112-117.

68. For an enlightening discussion of the ways

people change, see Mortimer R. Kadish, The
Ophelia Paradox: An Inquiry into the Conduct
of Our Lives (Transaction: New Brunswick,
N.J., 1994).

Chapter 3
 984/1183

I should like to thank Helen Nissenbaum, Ilene

Burnstein, and Vivian Weil for helpful com-
ments on my first draft of this chapter. A short
version appeared as "Defining Engineering:
How to Do It and Why It Matters," Journal of
Engineering Education 85 (April 1996): 97-101;
a full version (under the present title and, des-
pite the date, a year later) appeared in Philo-
sophy and the History of Science: A Taiwanese
Journal 4 (October 1995): 1-24. Reprinted by
permission.

1. Gary A. Ford and James E. Tomayko, "Edu-

cation and Curricula in Software Engineering,"
Encyclopedia of Software Engineering, vol. 1
(John Wiley & Sons: New York, 1994) p. 439.
 985/1183

2. "In the 1991 Computer Society [of the Insti-

tute for Electrical and Electronic Engineers]
membership survey, over half (54 percent) of
the current full members polled indicated that
they consider themselves software engineers, as
did 40 percent of the affiliate members." Fletch-
er J. Buckley, "Defining software engineering,"
Computer 2 (August 1993): 77.

3. There are no hard numbers for software en-

gineers (though I have heard estimates as high
as three million worldwide). The claims made
here merely constitute my compilation of the
opinions of those who seemed to have the best
chance of being right.
 986/1183

4. See, for example, Mary Shaw, "Prospects for

an Engineering Discipline of Software," IEEE
Software (November 1990): 15-24. Though she
begins this intelligent article with a definition of
engineering and devotes much of its body to the
history of engineering, her topic is really the
growth of disciplines generally. She could have
written much the same article, using law, medi-
cine, or even auditing, rather than engineering,
as the paradigm of a disciplinewith more clarity
about what she was doing.

5. This definition, the work of the National Re-

search Council's Committee on the Education
and Utilization of the Engineer, appears in
Samuel Florman, The Civilized Engineer (St.
Martin Press: New York, 1987) pp. 64-65.

6. Compare the more elegant Canadian

definition:
 987/1183

The "practice of professional engineering" means

any act of planning, designing, composing, eval-
uating, advising, reporting, directing or super-
vising, or managing any of
 Page 204

the foregoing that requires the application of en-

gineering principles, and that concerns the safe-
guarding of life, health, property, economic in-
terests, the public welfare or the environment.
(Emphasis added)

Canadian Engineering Qualifications Board,

1993 Annual Report (Canadian Council of Pro-
fessional Engineers: Ottawa, 1993) p. 17. The
report contains no definition of engineering
principles.

7. Similar problems arise for "genetic engineer"

and might arise for other "engineers," for ex-
ample, "social engineers." (This problem of
definition is, of course, not limited to engineers:
lawyers are no more successful defining the
practice of law or doctors the practice of
medicine.)
 989/1183

8. The IEEE defines software engineering as

"application of a systematic, disciplined, quanti-
fiable approach to the development, operation
and maintenance of software: that is, the applic-
ation of engineering to software." Buckley, "De-
fining Software Engineering," 77. This defini-
tion (or, rather, that "that is") begs the question
whether the systematic, disciplined, and quanti-
fiable approach in question is an application of
engineering to software or the application of a
different discipline. Not all systematic, discip-
lined, and quantifiable approaches to develop-
ment, operation, and maintenance are necessar-
ily engineering. Indeed, that software is primar-
ily not a physical but a mathematical (or lin-
guistic) system at least suggests that engineering
principles have only limited application.

9. Florman, The Civilized Engineer, 65-66.

 990/1183

10. I charitably ignored the "or" in "mathematics

and/or the natural sciences." There never was a
time that the training of engineers did not in-
clude a good deal of both mathematics and the
physical sciences (at least chemistry and phys-
ics). If software engineers do not generally have
similar training in the physical sciences, no
amount of training in mathematics will fill the
gap between them and the great body of engin-
eers strictly so called.

11. It is perhaps worth noting that engineers do

in fact produce beautiful objects, for example,
the Brooklyn Bridge or the typical computer's
circuit board. Nonetheless, engineers are not
artists in the way architects are. For engineering,
beauty is not a major factor in evaluating work;
utility is.
 991/1183

12. Compare Fletcher J. Buckley, "Background

to the Motion [to have the IEEE CS Board of
Governors appoint an ad hoc committee to initi-
ate the actions to establish software engineering
as a profession] (April 15, 1993)":
In 483 B.C., Xerxes, King of Persia and Media,
as part of his campaign to conquer Greece,
ordered two floating bridges to be constructed
across the Hellespont to provide passage for his
army from Asia to Europe. After the bridges
were completed, a storm arose and the bridges
were destroyed. Xerxes had the engineers killed
and another set of bridges constructed, thus
demonstrating at that time, the existence of
standards of personal accountability for profes-
sionals working in their field of competence.
 992/1183

This passagelike its twin in Buckley, "Defining

software engineering," 76is remarkable for its
misunderstanding of both engineering and pro-
fessions. Buckley has, of course, no reason to
call the builders of Xerxes bridge engineers
rather than bridge builders, no reason even to
describe them as professionals rather than
skilled men. He certainly overlooks whether the
bridge's failure was due to incompetence or to
forces beyond any builder's competence to man-
age at the time. Like a similar story about hav-
ing the sea flogged, this one seems to be more
about the arbitrariness of Persian rulers than
about the standard of accountability to which
anyone would want to be held. Its place in a mo-
tion concerned with organizing software engin-
eering as a "profession" is therefore (at best)
inauspicious.

13. Compare, for example, the Roeblings, father

and son, both engineers (by today's
993/1183
 Page 205

standard definition) with the millwrights, in-

dustrialists, and other contemporary bridge
builders most of whom would today not be
allowed to design or build bridges. Were
these other "technologists," self-taught and
relatively slapdash, as much engineers as the
Roeblings because much of what they built
worked?

14. Consider, for example, L.A. Belady, in

"Foreword," Encyclopedia of Software Engin-
eering, p. xi: "[The] term software engineering
expresses the continued effort to put program-
ming into the ranks of other engineering
disciplines."
 995/1183

15. Engineers, especially civil engineers, like to

count the Roman builders among their profes-
sion. When asked why, they usually point out
how enduring the Roman roads, aqueducts,
theaters, and other constructions proved to be.
This answer seems to me to offer evidence
against their thesis as if it were evidence for it.
Engineers are fond of the saying, "An engineer
is someone who can do for one dollar what any
fool can do for ten"or, as ABET put it more pro-
saically, "Engineering is the profession in which
a knowledge of the mathematical and natural
sciences gained by study, experience, and prac-
tice is applied with judgment to develop ways to
utilize, economically, the materials and forces of
nature for the benefit of mankind." That the Ro-
man builders built so much that outlasted their
empire by more than a thousand years at least
suggests that they spent where an engineer
would have saved. When we recall that none of
Rome's great builders made a career of building
 996/1183

but instead, oversaw public works one year and

a province's government the next, we must con-
clude that though great builders, they could nev-
er qualify for admission to an engineering soci-
ety at the "professional level." They may have
"functioned as engineers" (however anachronist-
ically), but they were not members of the pro-
fession (or even employed in the underlying
occupation).

16. For a spirited defense of this mistake, see

John T. Sanders, "Honor Among Thieves: Some
Reflections on Professional Codes of Ethics,"
Professional Ethics 2 (Fall/Winter 1993):
83-103. If the article's title does not show what
is wrong with equating competence with profes-
sion, the article's suggestion that we consider the
mafia to be a profession should.
 997/1183

17. Note that Florman, generally so astute about

engineering, endorses this equation of conscien-
tiousness with ethicalness. Florman, The Civil-
ized Engineer, 104.

18. Are either of these provisions, or any other,

unique to engineering, an expression of its es-
sential nature? I know of none. A particular pro-
fessional code seems to me to involve a distinct-
ive reworking of general moral ideals to fit cer-
tain conditions and aspirations, distinctive but
not necessarily unique.

19. In the United States, the codes date from the

second decade of this century. In Great Britain,
they came almost a half century earlier. The
counterpart for these codes didn't appear on the
continent of Europe, or in other civil law juris-
dictions, until well after World War II. Why?
 998/1183

20. For more details, see Lawrence P. Grayson,

"A Brief History of Engineering Education in
the United States," Engineering Education 68
(December 1977): 246-264.

21. See chapter 4 (in this volume) for a defense

of this response. Meanwhile, note that the
IEEE's code of ethics applies only to "IEEE
members." It is a code of ethics for members of
a technical society, notlike the codes of ABET
or the National Society of Professional Engin-
eers (NSPE) codethe code of a profession.
Indeed, I would attribute its shortening over the
years to the attempt to cover a membership in
which the proportion of engineers is declining
and both the number and kinds of nonengineers
are increasing. Generally, codes of ethics grow
with experience; shrinkage is therefore a sign of
trouble.

22. Shaw, "Prospects," 22.

 999/1183

23. Michael S. Mahoney, "The Roots of Soft-

ware Engineering," CWI Quarterly 3 (December
1990): 325-334, at 326.

24. Ibid., 327.

 Page 206

25. For example, some practicing engineers, un-

til recently, encouraged schools of engineering
to reduce the academic requirements for a de-
gree in favor of more "shop experience." Yet, at-
tempts to take the shop-experience approach
very far seem to produce foremen rather than
engineers. Apparently, the very abstractness for
which the practitioners criticized engineering
education contributed to success as engineers
even when (as the practitioners correctly noted)
the specific skills taught (for example, advanced
calculus) generally went unused. Why?

26. Some electrical engineering departments of-

fer degrees of this description, for example, in
"computer engineering, software option."
 1001/1183

27. There is, of course, more than one code of

ethics for U.S. engineers. This may suggest that
engineering in the United States is not one pro-
fession but several. That suggestion should not
be embraced. Of the three major codes usually
mentioned on such occasions, the IEEE code is
not a professional code at all; it applies not to
engineers (as a professional code should) but to
IEEE members (whether engineers or not). Be-
cause it also contains nothing more demanding
than the other codes and nothing inconsistent
with them, we may ignore it here. The other two
major codes do apply to engineers as such, dif-
fering only in detail (with the NSPE code gener-
ally being somewhat less demanding). Because
the NSPE seems to have developed its code with
state enforcement in mind, I think it reasonable
to treat the ABET code as the basic professional
code (especially because most engineering soci-
eties endorse it). So, when I speak of the
 1002/1183

engineer's code here, it is the ABET code that I

intend.

28. See, for example, John D. Musa, "Software

Engineering: The Future of a Profession," IEEE
Software (January 1985): 55-62. Musa presents
software engineering as a profession independ-
ent of engineering (though his use of the term
engineering suggests the opposite).

29. Compare Shaw, "Prospects," 21:

Unfortunately, [the term "software engineering"]
is now most often used to refer to life-cycle mod-
els, routine methodologies, cost-estimation tech-
niques, documentation frameworks . . . and other
techniques for standardizing production. These
technologies are characteristic of the commercial
stage of evolution`software management' would
be a much more appropriate term.

Chapter 4
 1003/1183

This chapter began as the first third of Engineer-

ing Codes of Ethics: Analysis and Applications,
a "module" prepared with Heinz Luegenbiehl in
1986 for a series published by IIT's Center for
the Study of Ethics in the Professions under a
grant from the Exxon Education Foundation (the
same series in which chapter 7 appeared).
Though this module was never published, a
shorter and substantially different version ap-
peared as "Thinking Like an Engineer: The
Place of a Code of Ethics in the Practice of a
Profession," Philosophy and Public Affairs 20
(Spring 1991): 150-167. Reprinted by permis-
sion. I should like to thank the series' Advisory
Panel, Heinz Luegenbiehl, the editors of Philo-
sophy and Public Affairs, and those who
listened patiently to one version or another for
much useful advice.
 1004/1183

1. David E. Sanger, "How Seeing-No-Evil

Doomed the Challenger," New York Times, June
29, 1986, sec. 3, p. 8.

2. The Presidential Commission on the Space

Shuttle Challenger Disaster (U.S. Government
Printing Office: Washington, D.C, 1986). v. I, p.
94. The preceding narrative is based on testi-
mony contained in that volume (esp. pp.
82-103).

3. William H. Wisely, "The Influence of Engin-

eering Societies on Professionals and Ethics" in
Ethics, Professionals, and Maintaining Compet-
ence: ASCE Professional Activities Com-
 Page 207

mittee Specialty Conference, Ohio State

University, Columbus, Ohio, 1977 (Americ-
an Society of Civil Engineers: New York),
1977, pp. 55-56.

4. See, for example, A.G. Christie, "A Proposed

Code of Ethics for All Engineers," Annals of
American Society of Political and Social
Science 101 (May 1922): 99-100.
 1006/1183

5. What is the origin of the term "bench engin-

eer"? I have encountered two guesses. One at-
tributes the term to a bitter analogy with galley
slaves, who rowed their life away chained to a
bench. The other guess involves a more pleasing
analogy with scientists, especially physicists and
chemists, who worked at "benches" with their
lab equipment around them. For scientists, a
"bench scientist" is a real scientist; those scient-
ists who devote themselves to supervision, to
meetings, and so on are no longer doing science
but administration. Neither of these analogies is
appropriate for engineering. On the one hand,
except for draftsmen (who did work side by side
in large rooms, seldom leaving the drafting
board) few engineers seem to spend even a ma-
jority of their day in one place. They have tech-
nicians to supervise, "fires to put out," and meet-
ings to go to. Both movement and administra-
tion are more central to engineering than to
science.
 1007/1183

6. William H. Wisely, "The Influence of Engin-

eering Societies on Profesionalism and Ethics"
in Engineering Professionalism and Ethics
(Robert E. Kreiger: Malabar, Fl. 1983) p. 33.

7. Andrew G. Oldenquist and Edward E. Slow-

ter, "Proposed: A Single Code of Ethics for All
Engineers," Professional Engineer 49 (May
1979): 8-11.

8. Note, for example, the quotation from A.G.

Christie at the beginning of this chapter; or Mor-
ris Llewellyn Cooke, "Ethics and the Engineer-
ing Profession," Annals of the Association for
Political and Social Science 101 (May 1922):
68-72, esp. 70.

9. See, for example, W.J. Reader, Professional

Men: The Rise of the Professional Classes in
Nineteenth-Century England (Basic Books:
New York, 1966) esp. pp. 51-55.
 1008/1183

10. Recall Thredgeld's famous definition (cited

in chapter 1): "[The] profession of civil engineer
[is] the art of directing the great sources of
power in nature for the use and convenience of
man."
 1009/1183

11. For further defense of this theory of profes-

sion, see Michael Davis, "The Moral Authority
of a Professional Code," NOMOS 29 (1987):
302-337; "The Use of Professions," Business
Economics 22 (October 1987): 5-10; "Vocation-
al Teachers, Confidentiality, and Professional
Ethics," International Journal of Applied Philo-
sophy 4 (Spring 1988): 11-20; "Professionalism
Means Putting Your Profession First," Geor-
getown Journal of Legal Ethics 2 (Summer
1988): 352-366; ''Do Cops Really Need a Code
of Ethics," Criminal Justice Ethics 10 (Summer/
Fall 1991): 14-28; "Science: After Such Know-
ledge, What Responsibility?," Professional Eth-
ics 4 (Spring 1995): 49-74; and "The State's Dr.
Death: What's Unethical about Physicians Help-
ing at Executions?" Social Theory and Practice
21 (Spring 1995): 31-60.
 1010/1183

12. Compare Michael Davis, "The Special Role

of Professionals in Business Ethics," Business
and Professional Ethics Journal 7 (1988):
83-94.

13. Devotees of decision theory will instantly

recognize the convention in question as the solu-
tion to the coordination problem commonly
known as the prisoner's dilemma. I avoid the
term here because it seems wholly out of place
when there are no prisoners and when the choice
posed is far better than a dilemma. Like many
other technical terms of decision theory,
"prisoner's dilemma" seems more likely to mis-
lead those not familiar with it than to grant
insight.
 1011/1183

14. I hope this appeal to fairness raises no red

flags, even though the principle of fairness has
been under a cloud ever since the seemingly
devastating criticism it received in Robert No-
zick, Anarchy, State, and Utopia (Basic Books:
New York, 1974). I limit my use to obligations
generated by voluntarily claiming benefits of a
cooperative practice that are otherwise not avail-
able. Most attacks on the principle of fairness
are on the "involuntary benefits" version. See,
for example, A. John Simmons, Moral Prin-
ciples and Political Obligations (Prince-
 Page 208

ton University Press: Princeton, N.J., 1979)

pp. 118-136. And even those attacks are
hardly devastating. One can either refine the
principle, as Richard Arenson did in "The
Principle of Fairness and Free-Rider Prob-
lems," Ethics 92 (July 1982): 616-633; or, as
in Michael Davis, "Nozick's Argument for
the Legitimacy of the Welfare State," Ethics
97 (April 1987): 576-594, show that
Nozick's original criticism, and most sub-
sequent criticism, depends on examples that,
on careful examination, fail to support the
criticism.
 1013/1183

15. I do not claim that the engineers treated

safety as paramount because they knew what the
ABET code said. When you ask a lawyer about
a professional code, she is likely to tell you she
studied the ABA code in law school and, claim-
ing to have a copy around, will produce it after
only a few minutes of searching her desk or
bookshelves. When you ask an engineer the
same question, he is likely to tell you that his
profession has a code while admitting that he
never studied it and that he has none around to
refer to. He may even admit to never having
seen a copy. Yet, anyone who has spent much
time working with engineers knows that they do
not treat safety in the same way managers do
(hence Mason's plea to "take off your engineer-
ing hat"). The engineers' code of ethics seems to
be "hard-wired" into them. Interestingly, engin-
eers are not the only professionals for whom the
written code seems to play so small a part. For
another example, see Michael Davis,
 1014/1183

"Vocational Teachers, Confidentiality, and Pro-

fessional Ethics," International Journal of Ap-
plied Philosophy 4 (1988): 74-90.

16. I do not claim that Lund would explain his

decision in this way. Indeed, as I suggest in
chapter 5 (in this volume), I think his explana-
tion would be quite different, though no less
troubling.

17. For criticism of this analysis, though one

that misunderstands it, see Nigel G.E. Harris,
"Professional codes and Kantian duties" in Eth-
ics and the Professions, edited by Ruth F. Chad-
wick (Avebury: Aldershot, England, 1994) pp.
104-115.

18. For a good summary of these other contrib-

uting causes, see Diane Vaughn, The Challenger
Launch Decision (University of Chicago Press:
Chicago, 1996).
 1015/1183

19. Philosophers will note that this is not one of

the four standard senses of responsibility
(capacity-responsibility, liability-responsibility,
causal-responsibility, or role-responsibility);
nevertheless, it seems to me to be a legitimate
sense. I should like to thank Jeff McMahan for
pointing it out to me (in another context).

Chapter 5
 1016/1183

An early version of this chapter was read at the

Center for the Study of Ethics in Society,
Western Michigan University, Kalamazoo,
October 27, 1988, under the title "Keeping
Good Apples from Going Bad." A later version,
under the present title, was read at the Philo-
sophy Colloquium, Illinois Institute of Techno-
logy, March 30, 1989 and, somewhat revised,
published as "Explaining Wrongdoing" in the
Journal of Social Philosophy 20 (Spring/Fall
1989): 74-90. Reprinted by permission. I would
like to thank those present at these two events
for their encouragement and criticism. I would
also like to thank Paul Gomberg for his careful
reading of the penultimate draft and Fay Sawyi-
er for helping me to see microscopic vision for
the first time.
 1017/1183

1. My wrongdoers, for example, are not like the

"hard men" in Jack Katz, Seductions of Crime:
Moral and Sensual Attractions of Doing Evil
(Basic Books: New York, 1988). Katz's crimin-
als do indeed seem to will evil.

2. "Price Fixing and Bid Rigging in the Electric-

al Manufacturing Industry," Administered
Prices: Hearings before the Senate Subcommit-
tee on Antitrust and Monopoly of the Committee
on the Judiciary, pt. 27, 1961, 16652.
 Page 209

3. "After the Fall: Fates Are Disparate for Those

Charged with Inside Trading," Wall Street
Journal, November 18, 1987, p. 22.

4. For example, 46% of all aggravated assaults

and 52% of all burglaries go unreported. Crime
in the United States (Federal Bureau of Investig-
ation, U.S. Department of Justice: Washington,
D.C.:, 1980) pp. 20, 23.

5. See chapter 3 (in this volume).

6. Compare Lawrence Kohlberg, Essays on

Moral Development (Harper and Row: San
Francisco, 1981).

7. Compare Chester A. Barnes, The Functions of

the Executive (Harvard University Press: Cam-
bridge, Mass., 1938) p. 276:
 1019/1183

It is apparent that executives frequently fail. This

failure may be ascribed in most cases, I believe,
to inadequate abilities as a first cause, usually
resulting in the destruction of responsibility. But
in many cases it may be inferred that the condi-
tions impose a moral complexity and a moral
conflict not soluble. Some actions which may
within reason appear to be dictated by the good
of the organization as a whole will obviously be
counter to nearly all other codes, personal or
official.

8. See, for example, Jan Elster, Ulysses and the

Sirens, rev.ed. (Cambridge University Press:
New York, 1984), for a good treatment of weak-
ness of will.
 1020/1183

9. Mike W. Martin, Self-Deception and Morality

(University Press of Kansas: Lawrence, Kansas,
1986). Martin defines self-deception in a way
consistent with modern psychology, producing a
conception of self-deception likely to be useful
in practice. Most work on this subject is, in-
stead, concerned with logical puzzles created by
assuming a unitary self that must both know and
not know at the same time.
 1021/1183

10. For a convenient survey of the literature on

self-deception that brings out the range of men-
tal states that might be included within that ca-
pacious term, see Alfred Mele, "Recent Work
on Self-Deception," American Philosophical
Quarterly 24 (January 1987): 1-17. For discus-
sion of the related phenomenon of shifting re-
sponsibility, see Stanley Milgram, Obedience to
Authority: An Experimental View (Harper and
Row: New York, 1974). I did not discuss the hy-
pothesis Milgram's work might suggest because,
as far as I can see, no one ever offered to take
responsibility for Lund's actions. Lund's claim
that he had "no choice" is not an appeal to the
authority of othersthough it is something equally
troubling. After all, what can be more obvious
than that Lund had a choice? He could simply
have said "no" and taken the consequences.
 1022/1183

11. I would like to thank Vivian Weil and Mi-

chael Pritchard for helpful criticism of one or
another early version of the explanation of mi-
croscopic vision. They are, however, not re-
sponsible for any errors that remain.

12. I derive this information from a video of

Boisjoly's appearance before Caroline
Whitbeck's engineering design course, "Com-
pany Loyalty and Whistleblowing: Ethical De-
cisions and the Space Shuttle Disaster" (January
7, 1987), especially his answers to student
questions.

13. For a bit more on the differences between

the way managers and engineers approach risk,
see chapter 9 (in this volume).
 1023/1183

14. Of course, we are here assuming that a well-

run organization would divide its decisions into
two categories, engineering decisions and man-
agement decisions, with engineers having the
last word in one category and managers having
the last word in the other. Another possible ar-
rangement is to recognize that the two categor-
ies overlap far too much to allow such a division
in responsibility and to require managers and
engineers to reach a consensus. For more on this
possibility, see chapter 9 (in this volume).
 Page 210

15. This is the standard view of relations

between managers and engineers that chapter 9
challenges. I state it here without endorsement.

16. Literature on the Challenger disaster gener-

ally assumes that management's way of dealing
with risk was clearly wrong. For a rare and
thoughtful attempt to make management's case,
see William Starbuck and Frances Milliken,
"Challenger: Fine-Tuning the Odds Until So-
mething Breaks," Journal of Management Stud-
ies 25 (July 1988): 319-340.
 1025/1183

17. In fact, our price fixer had no memory of

ever seeing the policy. Price Fixing, 16152. For
a somewhat different version of this story (in-
cluding the claim that he must have seen the
policy) see James A. Waters, "Catch 20.5: Cor-
porate Morality as an Organizational Phenomen-
on," Organizational Dynamics 6 (Spring 1978):
3-19. Waters emphasizes "organizational
blocks" to proper conduct rather than the normal
processes that concern me. But I believe nothing
I say here is inconsistent with what he says.
Wrongdoing in a complex organization is likely
to have many contributing causes. Waters and I
differ only in being interested in different con-
tributing causes. It is, of course, an empirical
question whether either of us is even partly right
(though one very hard to test decisively with the
information we have or are likely to get). I
would say the same about Saul W. Gellerman,
"Why `Good' Managers Make Bad Ethical
 1026/1183

Choices," Harvard Business Review 64 (July-

August 1986): 85-90.

18. The same is true of business ethics. Business

professors who limit themselves to technical
matters do not simply fail to do good. However
unintentionally, they actively contribute to the
wrong their students do, if they eventually do
wrong. They help to blind their students to
something the students might otherwise see.

19. See, for example, Robert M. Liebert, "What

Develops in Moral Development?," and Mordi-
cai Nisan, "Content and Structure in Moral
Development: An Integrative View" in Moral-
ity, Moral Behavior, and Moral Development,
ed. by William M. Kurtines and Jacob L.
Gewirtz (John Wiley & Sons: New York, 1984)
pp. 177-192, 208-224.
 1027/1183

20. Caroline Whitbeck, "Teaching Ethics to

Scientists and Engineers," Science and Engin-
eering Ethics 1 (July 1995): 299-308.

Chapter 6
 1028/1183

Early versions of this chapter were presented at

the Neil Staebler Conference, Institute of Public
Policy Studies, University of Michigan, Ann Ar-
bor, February 17, 1988; at Aquinas College,
Grand Rapids, Michigan, September 21, 1989;
and at the Mechanical Engineering Bi-Weekly
Seminar Series, Western Michigan University,
Kalamazoo, Michigan, October 3, 1989. I would
like to thank those present, as well as my col-
league Vivian Weil, for helping me see the
many sides of whistleblowing. I would also like
to thank the editor of Business and Professional
Ethics Journal for his helpful comments and
some useful references. Originally published in
Business and Professional Ethics Journal 8
(Winter 1989): 3-19.
 1029/1183

1. For a summary of what is or could be offered,

see Martin H. Malin, "Protecting the Whis-
tleblower from Retaliatory Discharge," Journal
of Law Reform 16 (Winter 1983): 277-318. For
some suggestion of how ineffective that protec-
tion is, see Thomas M. Devine and Donald G.
Aplin, "Whistleblower ProtectionThe Gap
Between the Law and Reality," Howard Law
Journal 31 (1988): 223-239; and Rosemary
Chalk, "Making the World Safe for Whis-
tleblowers," Technology Review 91 (January
1988): 48-57.

2. The literature describing the suffering of

whistleblowers is, of course, large. For a schol-
arly summary, see Myron Peretz Glazer and
Penina Migdal Glazer, The Whistleblowers: Ex-
posing Corruption in Government and Industry
(Basic Books: New York, 1989). There is, in
contrast, little about how the organization either
suffers or benefits. Why?
1030/1183
 Page 211

3. The holder of a "professional position" is

more likely to become a whistleblower than an
ordinary employee. See, for example, Marcia P.
Miceli and Janet P. Near, "Individual and Situ-
ational Correlates of Whistle-Blowing," Person-
nel Psychology 41 (Summer 1988): 267-281.

4. Compare Chester A. Barnes, The Function of

the Executive (Harvard University Press: Cam-
bridge, Mass., 1938).

5. For a good discussion of the problems of de-

fining "whistleblowing," see Frederick Elliston
et al., Whistleblowing Research: Methodologic-
al and Moral Issues (Praeger: New York, 1985)
esp. pp. 3-22, 145-161.
 1032/1183

6. Even this definition should be used with cau-

tion. In most organizations, there are "ordinary"
channels the use of which does not offend and
"extraordinary" channels the use of which of-
fends. Sometimes we can only determine that a
channel is extraordinary by using it. Those using
an extraordinary channel are treated as whis-
tleblowers (and, indeed, are often so labeled
even when they are not whistleblowers accord-
ing to thisor any other standarddefinition). Sim-
ilarly, the dispute between a whistleblower and
her organization may in part be over whether
her objection is a moral rather than a technical
one (everyone agreeing that if the objection is
moral, she would be justified). But, because
they think the objection is not a moral one, they
consider her a "disgruntled employee," not a
whistleblower. I do not intend what I say here to
turn on how we resolve such difficult cases. For
a good summary of the recent literature of defin-
ition, see Marian V. Heacock and Gail W.
 1033/1183

McGee, "Whistleblowing: An Ethical Issue in

Organizational and Human Behavior," Business
and Professional Ethics Journal 6 (Winter
1987): 35-46. See also Michael Davis, ''Some
Paradoxes of Whistleblowing," Business and
Professional Ethics Journal 15 (Spring 1996):
3-19.

7. I have in mind especially the response to

whistleblowers within academic institutions.
See, for example, Bruce W. Hollis, "I Turned in
My Mentor," The Scientist 1 (December 14,
1987): 1-13.

8. William Shakespeare, Anthony and Cleopatra

(Act II: sc. 5).

9. Robert Jackall, Moral Mazes (Oxford

University Press: New York, 1988), esp.
105-112, 119-133.
 1034/1183

10. Why this asymmetry? One reason may be

that inaccurate whistleblowing is less likely to
make news. Newspapers, police departments,
and senior managers are constantly receiving
"tips" that don't pan out. These are not news.
Another reason inaccurate whistleblowing re-
ceives little attention may be that reliably de-
termining that a particular whistleblower is inac-
curate can be difficult. The whistleblower's
evidence may establish only a presumptive case
against an organization. The organization may
not be able to reply in full without revealing
proprietary information or violating the privacy
of other employees, leaving outsiders no way to
know that the whistleblower is mistaken. Or, the
organization in question may not be able to
make a determination without great expenseand
therefore may never bother. Much whistleblow-
ing seems enveloped in the organizational equi-
valent of what Clausewitz called "the fog of
battle." If we knew more about cases of
 1035/1183

inaccurate, mistaken, or otherwise flawed whis-

tleblowing, perhaps our assessment of the over-
all good effect of whistleblowing would change.
Perhaps whistleblowing, like tyrannicide, is so
likely to hit the wrong target that it cannot in
practice be justified. This is a subject about
which we need to know more.

11. See, for example, Dick Polman, "Telling the

truth, paying the price," Philadelphia Inquirer
Magazine, June 18, 1989, pp. 16ff.

12. For an interesting analysis of this traditional

view of organizational authority (and related is-
sues) see Christopher McMahan, "Managerial
Authority," Ethics 100 (October 1989): 33-53.

13. I owe this observation to Thomas Devine. I

found no research that confirms or disconfirms
it.
 Page 212

14. For a procedure I doubt will do much good,

see Theodore T. Herbert and Ralph W. Estes,
"Improving Executive Decisions by Formalizing
Dissent: The Corporate Devil's Advocate,"
Academy of Management Review 2 (October
1977): 662-667. Dissent is likely to be more ef-
fective if the dissenter is not viewed as "just go-
ing through the motions" and likely to be more
common if not the job of just one personor so it
seems to me. But here is another question about
which we need to know more.

15. Compare James Waters, "Catch 20.5: Cor-

porate Morality as an Organizational Phenomen-
on," Organizational Dynamics 6 (Spring 1978):
3-19.

16. Moral Mazes, for example, 105-112.

 1037/1183

17. These are, of course, matters of what is now

often called "culture." For a good discussion,
see Charles O'Reilly, "Corporations, Culture,
and Commitment: Motivation and Social Con-
trol in Organizations," California Management
Review 31 (Summer 1989): 9-25.

18. This claim is defended in chapter 5 (in this

volume). See also M. Cash Matthews, "Ethical
Dilemmas and the Disputing Process: Organiza-
tions and Societies," Business and Professional
Ethics Journal 8 (Spring 1989): 1-11.

19. Michael Davis, One Social Responsibility of

Engineering Societies: Teaching Managers
About Engineering Ethics, Monograph #88-WA/
DE-14 (American Society of Mechanical Engin-
eers: New York, 1988).
 1038/1183

20. Perhaps the best example of such a person is

Roger Boisjoly (if we can count his testimony
before Congress as whistleblowing). The warn-
ings Boisjoly gave on the night before the Chal-
lenger exploded were (though technically accur-
ate) in the bloodless language in which engin-
eers generally communicate. He never said, for
example, "This decision could kill seven human
beings." How might things have gone had Bois-
joly (or anyone else present) said something of
that sort when NASA pressured Thiokol to ap-
prove a launch? It is a hard question, to be sure,
but one that at least suggests the potential power
of language at the moment of decision. For de-
tails, see The Presidential Commission on the
Space Shuttle Challenger Disaster (U.S.
Government Printing Office: Washington, D.C.:
June 6, 1986).

Chapter 7
 1039/1183

This chapter began as pages 1-26 of Paula

Wells, Hardy Jones, and Michael Davis, Con-
flicts of Interest in Engineering (Kendall/Hunt
Publishing Company: Dubuque, Iowa, 1986) a
module in the Series in Applied Ethics funded
by the Exxon Education Foundation. Reprinted
by permission. I would like to thank the staff of
IIT's Center for the Study of Ethics in the Pro-
fessions and the Advisory Panel of the Series for
help both in formulating the original project and
in carrying it to completion, and Michael
Pritchard for several perceptive comments on
the published version.

1. American Society of Mechanical Engineers v

Hydrolevel Corporation, 456 U.S. 556 (1982).

2. Ibid., 559.
 1040/1183

3. Voluntary Industrial Standards: Hearing be-

fore the Senate Subcommittee on Antitrust and
Monopoly of the Committee on the Judiciary,
94th Cong., 1st sess. 1975, 153-214 at 173.

4. Ibid., 174.

5. Ibid., 176, 184-185.

6. ASME v. Hydrolevel, 561-562.

7. Ibid., 563.

8. The critical sentence reads: "If a means for re-

tarding control action is incorporated in a low-
water fuel cutoff, the termination of the retard
function must operate to cutoff the fuel supply
before the boiler level falls below the visible
part of the water guage glass." That sentence re-
placed: "It should be carefully noted that regard-
less of the design of any automatic
 Page 213

low water cutoff, the intent of the first sen-

tence in paragraph HG-605(a) is that such
low water fuel cutoff devices function so
that the fuel supply shall be actually stopped
when the surface of the water falls to the
lowest visible part of the water gauge glass."
Senate, 188, and ASME v. Hydrolevel, 130.
If (as it seems) there is no important differ-
ence (but clarity) between these two sen-
tences, why should anyone be concerned
about James's part in substituting one for the
other?

9. Priscilla S. Meyer, "Knocking the Competi-

tion: How Rival's Use of `Industrial Code' Re-
port Created Problems for a Tiny Company,"
Wall Street Journal July 9, 1974, p. 44.

10. Voluntary Industrial Standards, 213.

11. ASME v. Hydrolevel, 564.

 1042/1183

12. It is perhaps worth pointing out that this is a

controversial assumption. The appellate court
described Hardin's conduct as "fraud, a willful
and knowing misrepresentation of the code"
(without offering any additional evidence). Hy-
drolevel v. ASME, 125. And the legal counsel
for the American National Standards Institute
lumped James with Hardin as "two renegades"
(again, without any additional evidence). Willi-
am H. Rockwell, "Hydrolevel Decision as Ap-
plied to Antitrust Violations of Standards Mak-
ing Organizations," Perspectives on the Profes-
sions 3(3) 1983: 3-5. On the other hand, in 1975
ASME claimed there was nothing to what
Hardin and James did beyond the mere "appear-
ance of wrongdoing." Nancy Rueth, "A Case
Study,'' Mechanical Engineering 97 (June
1975): 34-36, at p. 36. That was still ASME's
position a decade later. See, for example,
Charles W. Beardsley, "The Hydrolevel CaseA
Retrospective," Mechanical Engineering 106
 1043/1183

(June, 1984): 66-73; or Rockwell, "Hydrolevel

Decision," 5.

13. There are, of course, some respects in which

natural duties are not absolute. I will point out
two later.

14. Voluntary Industrial Standards, p. 179.

15. Hydrolevel v. ASME, 123.

16. ASME v. Hydrolevel, 559; and Beardsley,

72.

17. Tekla S. Perry, "Antitrust Rule Chills Stand-

ards Setting," IEEE Spectrum 11 (August 1982):
52-54.

18. Compare Voluntary Industrial Standards, p.

214, where ASME's attorney (Mr. Stanton)
makes a similar point.

19. Ibid., 205.

 1044/1183

20. Ibid., 175.

21. Ibid., 211.

22. Ibid., 206.

23. Ibid., 210.

24. Ibid., Compare James's comment, ibid., 190.

25. Ibid., 211.

26. In what follows, I use the current Code of

Ethics (1995). The relevant provisions are simil-
ar to those of the code in force during
1971-1972 in most relevant respects (though the
format is much different).

27. Senate, 192, 211-212; and Hydrolevel v.

ASME, 126.
 1045/1183

28. See ASME v. Hydrolevel, 571 n. 8, for evid-

ence that James' employer thought James so in-
fluenced (or, at least, was willing to defend
James' unpaid activities within ASME on that
basis).

29. These are published in batches several times

a year in P.E. Professional Engineer, the official
publication of the NSPE. These opinions have
also been collected (up till 1990) in six volumes
under the title Opinions of the Board of Ethical
Review (National Society of Professional Engin-
eers: Washington, D.C.).

30. Voluntary Industrial Standards, p. 211.

 Page 214

31. Compare Bernard Gert, Morality: A New

Justification of the Moral Rules (Oxford
University Press: New York, 1988).
 1047/1183

32. There is a family of consequentialist

views"rule utilitarianism"which holds that one
should generally follow rules ("rules of thumb,"
"prima facie rules," or the like) rather than al-
ways decide how to act by considering the con-
sequences case by case. The idea is that the
rules should be designed so that generally fol-
lowing them maximizes good consequences in
the long run. We may ignore this refinement be-
cause all forms of rule utilitarianism either fit
the definition of moral rules given here or suffer
from the same lack of information about con-
sequences as any other attempt to determine
what Hardin and James did wrong solely by
considering the consequences of their acts. See
David Lyons, Forms and Limits of Utilitarian-
ism (Oxford University Press: Oxford, England
1965).
 1048/1183

33. For more on my understanding of all this

(including my reasons for denying that I am op-
erating with a form of rule utilitarianism) see
Michael Davis, "The Moral Legislature: Moral-
ity without an Archimedean Point," Ethics 102
(July 1992): 303-318.

34. This is an old point about the self and its in-
terest, but one that may deserve more stress than
I give in the text. Consider someone so honest
that he could not live with himself if he behaved
dishonestly. For him, a dishonest action would
be irrational (as well as, and because it is, im-
moral). It would, in other words, be contrary to
his self-interest, given the kind of self he is. For
the virtuous, virtue really is a reward.
 1049/1183

35. Logicians might want to claim that every

rule with its exceptions can be rewritten as a
rule without exceptions. For example, "Don't
kill, except in self-defense or defense of the in-
nocent" might just as well be written "Don't
non-defensively kill." Whatever the clumsiness
of such rewriting, the logicians are formally
right. They are, however, morally wrong. The
"Don't____, except____" form of moral rules
captures an underlying logic the exceptionless
form does not. All else equal, killing people re-
quires justification, justification that has to fit
the killing under one of the exception clauses.
That is not true of not killing. Those who obey
the main clause of a moral rule need no justific-
ation; those who violate the main clause do,
even if what they do comes under one of the
exceptions.
 1050/1183

36. We are, of course, assuming, that engineers

can, as moral agents, only be bound by a code
insofar as the code is itself interpreted in a way
consistent with morality. Although that assump-
tion should be uncontroversial, it has not always
been. See, for example, Benjamin Freedman, "A
Meta-Ethics of Professional Morality" in Moral
Responsibility and the Professions, edited by
Bernard Baumrin and Benjamin Freedman
(Haven Publications: New York, 1983) pp,
61-79; or Alan H. Goldman The Moral Founda-
tions of Professional Ethics (Rowman and Lit-
tlefield: Totawa, N.J. 1980), who argues that
some professionals (for example, judges) are ex-
empt from certain moral constraints while acting
in their professional capacity. (Note, however,
that Goldman does not argue that engineers are
exempted in this way.) For what seems to me a
decisive refutaton of this "separationism," see
Alan Gewirth, "Professional Ethics: The
 1051/1183

Separatist Thesis," Ethics 96 (January 1986):

282-300.

37. The original version of this definition, like

some engineering codes, included a phrase ("or
to perform some other service for him or her")
suggesting that engineers do something beyond
offer professional judgment and that conflict of
interest might arise when they are providing that
other service. I cannot deny that engineers occa-
sionally do work not involving judgment, but I
have yet to think of a case in which such work
involves a conflict of interest. Understanding
conflict of interest requires recognizing the fun-
damental importance of judgment in the very
concept.
 1052/1183

38. Michael Pritchard, "Conflict of Interest:

Conceptual and Normative Issues," Academic
Medicine 71 (December 1996): 1305-1313, sug-
gested that the "interest" in conflicts of interest
should be interpreted as limited to something we
might "pursue, act in behalf of, or act for the
sake of." Ibid., 1309. Mere wants, desires, or
other circumstances tending to interfere with
Page 215
 1054/1183

competent judgment should not be counted

as capable of creating a conflict of interest.
Pritchard's suggestion would, I think, have
to be taken if "conflict of interest" were a
term in which the parts preserved their
meaning (as they do in "conflicting in-
terests"). But, in fact, it is idiom, carrying a
meaning more or less independent of its
parts. Utility, not etymology, may therefore
shape what interpretation we give it. There
are, I think, at least two reasons not to allow
"interest" to be confined to interests (strictly
so called). First, practice is not so neat.
Note, for example, that section II.4(a) of the
NSPE code recognizes "any business associ-
ation, interest, or other circumstance which
would influence or appear to influence their
judgment'' as requiring disclosure as "known
or potential conflicts of interest." Second, it
is not clear what the practical advantage of
limiting "interests" to interests strictly so
 1055/1183

called would have. Presumably, the "other

circumstances" should be disclosed anyway.
Must we have separate (but otherwise paral-
lel) rules for them?

39. For a more extensive defense of this analys-

is, see Michael Davis, "Conflict of Interest,"
Business and Professional Ethics Journal 1
(Summer 1982): 17-27; and Michael Davis,
"Conflict of Interest Revisited," Business and
Professional Ethics Journal 12 (Winter 1993):
21-41.

Chapter 8
 1056/1183

The first version of this chapter was presented to

a session of the First World Congress of
Biomechanics, La Jolla, California, August 31,
1990; a much enlarged version was published as
"Codes of Ethics, Professions, and Conflict of
Interest: A Case Study of an Emerging Profes-
sion, Clinical Engineering," Professional Ethics
Journal 1 (Spring/Summer 1992): 179-195. I
would like to thank those few present at La
Jolla, especially one of my co-panelists,
Caroline Whitbeck, for asking the right ques-
tions. I should also like to thank the three re-
viewers at PEJ for extensive comments on the
second version. They are, of course, not re-
sponsible for a number of minor revisions I have
made since.
 1057/1183

1. For more about clinical engineering, see Mi-

chael J. Shaffer and Michael D. Shaffer, "The
Professionalization of Clinical Engineering,"
Biomedical Instrumentation and Technology
(September/October 1989): 370-374; and
Pamela Saha and Subrata Saha, "Ethical Re-
sponsibilities of the Clinical Engineer," Journal
of Clinical Engineering 11 (January/February
1986): 17-25.
 1058/1183

2. Compare John Kultgen, Ethics and Profes-

sions (University of Pennsylvania Press: Phil-
adelphia, 1988) p. 216. Although I agree with
Kultgen's experimentalism ("Every code must
be treated as a hypothesis to be tested and adap-
ted while following it") I emphatically reject his
Cartesianism ("A rational code would contain
the results individuals would have reached
themselves if they had reasoned objectively long
enough on an adequate base of experience"). As
I shall try to show later, a code of professional
ethics necessarily involves certain public con-
ventions (much as do standards of safety or reli-
ability). What matters most is that members of
the profession in question apply the same stand-
ard (not which of several morally permissible
standards is applied). This is not to say that the
convention chosen does not matter, only that no
amount of "objective reasoning" can substitute
for a coordinated decision. A professional stand-
ard need not represent a preexisting consensus;
 1059/1183

it may in fact create that consensus (much as a

promise can create an agreement where none ex-
isted before). A profession's code of ethics is the
solution of a coordination problem, the sort of
practical problem no individual can solve alone.

3. For more on conflict of interest, see chapter 7

(in this volume).

4. This "assurance" is, of course, merely a psy-

chological fact. Whether it corresponds to real-
ity, whether one can govern one's judgment as
much as one believes one can, is not easily de-
termined. Determination cannot be left to the
engineer's own judgment, because that is in
 Page 216

question. (The ancients Greeks had a saying

relevant here: "Whom the gods would des-
troy, they first make mad.") Outsiders, other
engineers or other employees of Big Bill,
might well doubt the engineer's ability to
govern his judgment. Indeed, the engineer's
problem arises in large part because that is a
reasonable judgment on the evidence avail-
ableand he has no way to add evidence that
would change that judgment.

5. Does this provision make sense? Can anyone

be "sure" of not being influenced? Certainly,
provided the decisions in question do not in-
volve judgment (for example, because there is
only one drug, device, or appliance that could be
prescribed). Financial interest cannot create a
conflict of interest where it cannot affect
judgment.
 1061/1183

6. Here I am following the new IEEE code (ad-

opted August 1990).

7. In what sense is this standard higher and not

just different? It is higher in at least two senses.
First, it is higher in the sense of "more demand-
ing." One satisfies the lower standard in satisfy-
ing the higher and then does something more.
Second, it is higher in the sense of "morally bet-
ter." People who satisfy this higher standard de-
serve praise they would not deserve for satisfy-
ing the lower standard. These two senses are,
though related, not the same. We can at least
imagine higher standards in the first sense ("new
heights in torturing'') that are not higher in the
second.
 1062/1183

8. The NSPE's Code of Ethics is, of course, not

the NSPE's last word on this (or any other) eth-
ical question. As explained in chapter 7 (in this
volume), NSPE's BER regularly issues opinions
on questions like that posed here. Indeed, it
deals with questions very close to this one. See,
for example, BER 69-13 and BER 71-6, which
seem to explain why the NSPE's code now sets a
standard lower than other engineering societies
do. Opinions of the Board of Ethics Review (Na-
tional Society of Professional Engineers: Wash-
ington, D.C.) collected in six volumes to date.

9. For a fuller statement of this argument (but in

the context of lawyering) see Michael Davis,
"Professionalism Means Putting Your Profes-
sion First," Georgetown Journal of Legal Ethics
2 (Summer 1988): 341-357.

10. If this claim seems to need more defense, re-

call the argument of chapter 3 (in this volume).
 1063/1183

11. For more on this use of codes, see Heinz C.

Luegenbiehl, "Code of Ethics and the Moral
Education of Engineers," Business and Profes-
sional Ethics Journal 2 (Summer 1983): 41-61.
But note that I do not agree with Luegenbiehl in
thinking of a code of professional ethics as mere
"guidelines." See also Michael Davis, "Who
Can Teach Workplace Ethics?" Teaching Philo-
sophy 13 (March 1990): 21-36.

12. John Ladd, "Collective and Individual Moral

Responsibility in Engineering: Some Questions"
in Beyond Whistleblowing: Defining Engineers'
Responsibilities, edited by Vivian Weil (Center
for the Study of Ethics in the Professions,
Illinois Institute of Technology: Chicago, 1983)
pp. 102-103.
 1064/1183

13. Like ordinary morality, professional ethics

does have an external aspect as well. Most pro-
fessionals are ethical in part because they do not
want to suffer the (justified) criticism, circum-
spection, or boycott that unprofessional conduct
invites. What distinguishes both ethics and or-
dinary morality from (mere) law is that this ex-
ternal "sanction" is only one reason; there is an
internal "sanction" as welland this internal sanc-
tion is generally good enough most of the time
to sustain obedience.

14. Caroline Whitbeck, "Teaching Ethics to

Scientists and Engineers," Science and Engin-
eering Ethics 1 (July 1995): 299-308.
 1065/1183

15. Engineering, even more than medicine,

seems to work by consensus. (See chapter 9 [in
this volume].) Yet, unlike medicine, those writ-
ing on engineering seldom note this tendency (in
print at least) and never consider what signific-
ance it might have for understanding engineer-
ing. For some idea of the problems decision by
consensus raises for medicine (and
 Page 217

may raise for engineering as well) see the

entire August 1991 issue of Journal of Medi-
cine and Philosophy.
 1067/1183

16. Codes of ethics are sometimes criticized for

(as one of the Professional Ethics Journal's re-
viewers put it) "papering over" differences with
"vague language." I have four objections to this
criticism: First, the criticism seems to assume
that language can be (absolutely) precise. That
is certainly a mistake. Linguistic expressions
differ from one another only in degrees of
vagueness (or, what comes to the same thing,
degrees of precision). Second, the criticism
seems to overlook the alternative to the ''vague
language" in question. Given the differences al-
legedly "papered over," the alternative to the
vague language in question would seem to be no
language at all, that is, less precision than is in
fact possible. The alternative to papering over
differences would thus seem to be "magnifying"
them. That hardly seems preferable. Third, the
criticism seems to assume that there is
something wrong with using language of a cer-
tain degree of precision when no agreement on
 1068/1183

anything more precise is possible. That too

seems a mistake. Pushing precision beyond what
is now possible can be expensive. The achieve-
ment may not be worth the expense. Given the
practical purpose of any code of ethics, the
prudent approach must, it seems to me, be to
state what can be stated at the time. Such a state-
ment does not paper over disagreements. The
disagreements are not concealed but simply left
as they were. Each member of the profession is
free to interpret the language agreed to as seems
right to her, to act accordingly, to answer for
what she has done, and so to contribute to a fund
of common experience out of which a more pre-
cise statement may in time grow. Fourth, the
criticism seems to understate the essential role
of language in all this. Everyone does, or at least
should, understand that any documentwhether
code of ethics, table of tolerances, or even
private lettercannot be taken to express any
single state of mind in its author (or authors).
 1069/1183

The document says what it says, whatever its

author actually intended. If the author was fool-
ish, careless, or simply unlucky, the document
may well say more, less, or even something rad-
ically different from what she intended. Like
other acts, linguistic acts can misfire. Interpreta-
tion is not a matter of reading off the clear (or
unclear) intention of an author. It is, rather,
working with a text according to certain more or
less definite procedures. A document is not a
mere vessel transmitting well or poorly the in-
tentions of its author. Interpretation begins
where the author stopped.

Chapter 9
 1070/1183

This chapter began as a project funded by a

grant from the Hitachi Foundation of America
and carried out under the direction of a seven-
member panel of academics and practitioners in
the Chicago area (the chapter's "we"). The panel
included Thomas Calero (Business, IIT) Mi-
chael Davis (Center for the Study of Ethics in
the Professions, IIT) Robert Growney (Corpor-
ate Vice President, Motorola) David Krueger
(Director, Center for Ethics and Corporate
Policy) Elliot Lehman (Chairman, Fel-Pro) and
Lawrence Lavengood (Business, Northwestern
University). Vivian Weil (Center for the Study
of Ethics in the Professions, IIT) chaired the
panel. Calero, Davis, and Krueger conducted in-
terviews at the following companies: Fel-Pro In-
corporated, Omni Circuits, Bosch Corporation,
W.E. O'Neil Construction Company, Motorola,
Inland Steel Company, Navistar, Amoco Chem-
ical Company (two sites) Hitachi Automotive
Products (USA) and Cummins Engine. We
 1071/1183

would like to thank these companies for their

help both in setting up the interviews and in
making sure they went smoothly. We should
also like to thank the following people for
providing comments on the first draft of this re-
port: Diana Stork (Business, University of Hart-
ford) Deborah Johnson (Department of Science
and Technology Studies, Rensselaer Polytechnic
Institute) Peter Whalley (History, Loyola
University of Chicago) and Steven Shortell
(Business, Northwestern University). One or an-
other summary of this chapter was presented at
the National Society of Professional Engineers
Annual Meeting
Page 218
 1073/1183

(Industry Practices Division) Charleston, South

Carolina, January 21, 1992; at the National Con-
ference on Ethics and the Professions,
University of Florida, Gainesville, February 1,
1992; and at a seminar sponsored by the Depart-
ment of Mechanical Engineering, Texas A & M
University, March 12, 1992. The discussions
that followed provided welcome confirmation of
our results. A short version of this chapter was
published as "Technical Decisions: Time to Re-
think the Engineer's Responsibilities?," Business
and Professional Ethics Journal 11 (Spring/
Summer 1992): 41-55; a longer version as
"Ordinary Technical Decision-Making: An Em-
pirical Investigation," in Responsible Commu-
nications: Ethical Issues in Business, Industry,
and the Professions, edited by James A. Jaska
and Michael S. Pritchard (Hampton Press:
Cresskill, N.J., 1996) pp. 75-106; and a com-
plete version as Michael Davis, ''Better Commu-
nication Between Engineers and Managers:
 1074/1183

Some Ways to Prevent Many Ethically Hard

Choices," Science and Engineering Ethics 3
(April 1997): 171-212. Reprinted by permission.

1. Henry Petroski, "The Iron Ring," American

Scientist 83 (May-June 1995): 229-232.

2. Note the crucial "seem" in this sentence. The

issue of probabilities here is more complex than
Feynman (or those he interviewed) indicates.
For more on that complexity, see William H.
Starbuck and Frances J. Milliken, "Challenger:
Fine-Tuning the Odds until Something Breaks,"
Journal of Management Studies 25 (July 1988):
319-340.

3. Richard Feynman, "An Outsider's Inside

View of the Challenger Inquiry," Physics Today
(February 1988): 26-37, esp. 34.

4. Ibid.

5. Ibid.
 1075/1183

6. For a good technical description of the

"game" to which Feynman refers, see Trudy E.
Bell, "The fatal flaw in Flight 51-L," IEEE
Spectrum (February 1987): 36-51. Compare
David A. Bella, "Organizations and Systematic
Distortion of Information," Journal of Profes-
sional Issues in Engineering 113 (October
1987): 360-370.

7. Feynman, "Outsider's Inside View," 34.

8. Robert Jackall, Moral Mazes: The World of

Corporate Managers (Oxford University Press:
New York, 1988) pp. 112-119.

9. Albert Shapero, Managing Professional

People (Free Press: New York, 1985).
 1076/1183

10. Joseph A. Raelin, The Clash of Cultures:

Managers and Professionals (Harvard Business
School Press: Cambridge, Mass., 1986); and
"The Professional as the Executive's Ethical
Aidede-Camp," Academy of Management Exec-
utive 1 (August 1987): 171-182.

11. Raelin, "Executive Aide-de-Camp," 1987.

12. Chapter 5 (in this volume).

13. James A. Waters, "Catch 20.5: Corporate

Morality as an Organizational Phenomenon,"
Organizational Dynamics (Spring 1978): 3-19.

14. Ibid. 11.

15. Harold Henderson, "McGregor v. the NRC:

Why Did the Nuclear Regulatory Commission
Fire One of Its Toughest Plant Inspectors?,"
Reader (Chicago) Friday, July 22, 1988, pp. 1ff.
 1077/1183

16. Brian Urquhart, "The Last Disaster of the

War," New York Review of Books, September
24, 1987, pp. 27-30; and Thomas Petzinger,
"Hangar Anger: Mechanic's Woes Show How
Safety Became a Big Issue for Eastern," Wall
Street Journal, June 9, 1988, pp. 1ff.

17. Waters, 1978; and Waters, "Integrity Man-

agement: Learning and Implementing Ethical
Principles in the Workplace," in Executive In-
tegrity, edited by Suresh Srivastva et al. (Jossey-
Bass: San Fransciso, 1988).

18. Raelin, The Clash of Cultures, 246-263.

19. See, for example, Chris Argyris and Donald

Schön, "Reciprocal Integrity: Creating Condi-
tions That Encourage Personal and Organiza-
tional Integrity" in Executive Integrity,
 Page 219

edited by Suresh Srivastva et al. (Jossey-

Bass: San Francisco, 1988) pp. 197-222; and
Gerald E. Ottoson, "Essentials of an Ethical
Corporate Climate" in Doing Ethics in Busi-
ness, edited by Donald G. Jones (Oelgesch-
lager, Gunn and Hain: Cambridge, Mass.,
1982) pp. 155-163.

20. The only exception we found is Bruce F.

Gordon and Ian C. Ross, "Professionals and the
Corporation," Research Management 5
(November 1962): 493-505.
 1079/1183

21. Perhaps the most noteworthy exceptions are

the very tentative studies by Bart Victor and
John B. Cullen, "The Organizational Bases of
Ethics Work Climates," Administrative Science
Quarterly 33 (March 1988): 101-125; and Alan
L. Wilkins and William G. Ouchi, "Efficient
Cultures: Exploring the Relationship Between
Culture and Organizational Performance," Ad-
ministrative Science Quarterly 28 (September
1983): 468-481.

22. Tom Burns and G.M. Stalker, The Manage-

ment of Innovation (Tavistock Publications:
London, 1966). I would like to thank Peter
Whalley for pointing out this book.

23. Barry A. Turner, Man-Made Disasters (Lon-

don: Wykeham Publications, Ltd., 1978) esp.
pp. 17-30, 57-67, 120-125, and 189-199.
 1080/1183

24. Merrit R. Smith, ed., Military Enterprise

and Technological Change (MIT Press: Cam-
bridge, Mass., 1985) esp. pp. 11-14 87-116.
 1081/1183

25. This, of course, is not the only way in which

to use the terms "staff" and "line" in business.
Most frequently, perhaps, these terms are today
used to distinguish between the historically old-
est functional units of a business (production
and sales) and the more recent (personnel, legal,
accounting). On this version of the distinction,
engineering might be either a staff or a line
function (depending on the history of the com-
pany). Often too, the staffline distinction is used
to contrast those functions that contribute (more
or less) directly to the bottom line ("profit cen-
ters") with those that contribute only indirectly
("service functions''). On this version, some en-
gineering functions (for example, operations and
perhaps research) would be line functions while
other engineering functions (for example, qual-
ity control or safety) would be staff functions.
This diversity in the way the staff-line distinc-
tion is made today may itself signal that the
 1082/1183

original use no longer fits most American

businesses.

26. Rosalind Williams, "Engineering's Image

Problem," Issues in Science and Technology 6
(Spring 1990): 84-86.

27. The companies selected had business con-

nections with one of the corporate members of
our panel, with one of our two ethics centers, or
with our sponsor (or even with some combina-
tion of these).

28. We had one interviewer on only three occa-

sions, one for a whole day when there was no
other way to schedule the interview, and twice
for part of an afternoon when one interviewer
had to leave early.

29. Williams, "Engineering's Image Problem,"

84.
 1083/1183

30. For more information about the characterist-

ics of those interviewed, see Appendix 3 in this
volume.

31. It is perhaps worth noting that no one men-

tioned a commonplace of academic criticism,
the need for engineers turned managers to learn
to live with ambiguity. What explains the si-
lence of our interviewees on this point? One
possibility is that, as practitioners, they have
already had to get used to ambiguity. Another
possibility is that the crucial transition is not
between engineering and technical management
but between technical and nontechnical manage-
ment. Here is a question that invites further
research.
 1084/1183

32. No member of our working group is alto-

gether satisfied with the names we gave these
three kinds of company. Our only defense is
that, after far too much discussion, we could not
do better.

33. For an apparently analogous case ending in a

half-billion-dollar write-off at General
 Page 220

Electric, see Thomas F. O'Boyle, "Chilling

Tale: GE Refrigerator Woes Illustrate the
Hazards in Changing a ProductFirm Pushed
Development of Compressor Too Fast,
Failed to Test Adequately," Wall Street
Journal, Monday, May 7, 1990, pp. 1 ff.

Chapter 10

An earlier version of this chapter was published

as "Professional Autonomy: A Framework for
Empirical Research," in Business Ethics
Quarterly 6 (October 1996): 441-460. Reprinted
by permission. I would like to thank the Nation-
al Science Foundation for grant SBR-9320166
under which that article was written.

1. Edwin Layton, The Revolt of the Engineers

(Case Western Reserve University Press: Cleve-
land, 1971) p. 5.
 1086/1183

2. Gerald Dworkin, The Theory and Practice of

Autonomy (Cambridge University Press: New
York, 1988) p. 22.

3. Thomas Scanlon, "A Theory of Freedom of

Expression," Philosophy and Public Affairs 1
(Winter 1972): 204-226; Adina Schwartz,
"Autonomy in the Workplace" in Just Business:
New Introductory Essays in Business Ethics, ed-
ited by Tom Regan (Random House: New York,
1984) pp. 129-166; Joseph Raz, "Autonomy,
Toleration, and the Harm Principle," in Issues in
Contemporary Legal Philosophy: The Influence
of H.L.A. Hart, edited by Ruth Gavison (Oxford
University Press: New York, 1987) 313-333;
Stanley I. Benn, A Theory of Freedom (Cam-
bridge University Press: Cambridge, 1988) esp.
chapters 8 and 9; and Diana T. Meyers, Self, So-
ciety, and Personal Choice (Columbia
University Press: New York, 1989).
 1087/1183

4. John Christman, ed., "Introduction," The In-

ner Citadel (Oxford University Press: New
York, 1989) p. 9.

5. Michael Davis, "Brandt on Autonomy" in Ra-

tionality and Rule-Utilitarianism, edited by
Brad Hooker (Westview Press: Boulder, Colo.,
1993) pp. 51-65; and Irving Thalberg, "Hier-
archical Analyses of Unfree Action," Canadian
Journal of Philosophy 8 (June 1978): 211-226.
But the most famous hypothetical conception of
autonomy is probably Kant's. For Kant,
autonomy is acting in accordance with those
maxims one can (without contradiction) will to
be universal laws. Immanuel Kant, Foundations
of the Metaphysics of Morals, 2d, edited by
Lewis White Beck (Macmillan/Library of the
Liberal Arts: New York, 1990) pp. 63-73. One
need not actually will the maxim to be a univer-
sal law. It is enough that one can.
 1088/1183

6. Gerald Dworkin, "Concept of Autonomy" in

Science and Ethics, edited by Rudolph Haller,
(Amsterdam: Rodopi Press, 1981) pp. 203-213.
Harry G. Frankfurt, "Freedom of the Will and
the Concept of a Person," Journal of Philosophy
68 (January 1971): 5-20.

7. Robert Young, "Autonomy and the Inner

Self," American Philosophical Quarterly 17
(January 1980): 35-43.

8. Kenneth Kipnis, "Professional Responsibility

and the Responsibility of Professions" in Profits
and Professions: Essays in Business and Profes-
sional Ethics, edited by Wade L. Robinson, Mi-
chael Pritchard, and Joseph Ellin (Humana
Press: Clifton, N.J., 1983) p. 16.

9. Arlene Kaplan Daniels, "How Free Should

Professions Be?," in The Professions and Their
Prospects, edited by Eliot Freidson (Sage:
Beverly Hills, Cal., 1971), p. 39.
 1089/1183

10. Actually, it would probably be better to say

"as Canada and Mexico try to do." I was told by
officers of engineering societies in both Canada
and Mexico that many engineers in both coun-
tries who work in large companies are unli-
censed. They will get into trouble if they are
publicly identified as engineers (for example, in
a newspaper article, during a television inter-
view, or even on company letterhead) but not
otherwise. So, the contrast with American prac-
tice is not nearly as sharp as it has seemed (and
will probably become less sharp as the three
countries move toward economic union).
 Page 221

11. K.R. Pavlovic, "Autonomy and Obligation:

Is There an Engineering Ethics?" in Ethical
Problems in Engineering, 2nd ed., vol. 1, edited
by Albert Flores (Center for the Study of the
Human Dimensions of Science and Technology:
Troy, N.Y., 1980) p. 90.

12. Paul F. Camenisch, Grounding Professional

Ethics in a Pluralistic Society (Haven: New
York, 1983) p. 30.

13. Kipnis, "Professional Responsibility," p. 16.

 1091/1183

14. Mike W. Martin, "Professional Autonomy

and Employers' Authority," in Profits and Pro-
fessions: Essays in Business and Professional
Ethics, edited by Wade L. Robinson, Michael
Pritchard, and Joseph Ellin (Humana Press:
Clifton, N.J., 1983) pp. 265-273; Adina
Schwartz, "Autonomy in the Workplace," and
Heinz C. Luegenbiehl, "Computer Profession-
als: Moral Autonomy and a Code of Ethics,"
Journal of Systems Software 17 (1992): 61-68.

15. Layton, Revolt, 7.

16. See chapters 1 and 2 (in this volume).

17. Daniel Hovey Calhoun, The American Civil

Engineer (Technology Press-MIT: Cambridge,
Mass., 1960) esp. pp. 182-199.
 1092/1183

18. Cf. Stephen J. O'Connor and Joyce A. Lan-

ning, "The End of Autonomy? Reflections on
the Postprofessional Physician," Health Care
Management Review 17 (Winter 1992): 63-72;
George J. Agich, "Rationing Professional
Autonomy," Law, Medicine and Health Care 18
(Spring-Summer): 77-84; and John Child and
Janet Fulk, "Maintenance of Occupational Con-
trol: The Case of Professions," Work and Occu-
pations 9 (May 1982): 155-192.

19. Robert Perrucci and Joel E. Gerstl, Profes-

sions without Community: Engineers in Americ-
an Society (Random House: New York, 1969) p.
119.
 1093/1183

20. J. Daniel Sherman, "Technical Supervision

and Turnover Among Engineers and Techni-
cians: Influencing Factors in the Work Environ-
ment," Group and Organization Studies 14
(December 1989): 411-421; Steven P. Feldman,
"The Broken Wheel: The Inseparability of
Autonomy and Control in Innovation within Or-
ganizations," Journal of Management Studies 26
(March 1989): 83-102; and Bernard Rosen-
baum, "Leading Today's Professionals,"
Research-Technology Management (March-
April 1991): 30-35.
 1094/1183

21. Gene F. Brady, Ben B. Judd, and Setrak

Javian, "The Dimensionality of Work
Autonomy Revisited," Human Relations 43
(1990): 1219-1228, esp. p. 1220; Paul E. Spect-
or, "Perceived Control by Employees: A Meta-
Analysis of Studies Concerning Autonomy and
Participation at Work," Human Relations 39
(November 1986): 1005-1015, esp. p. 1006;
Jiing-Lih Farh and W.E. Scott, Jr., "The Experi-
mental Effects of `Autonomy' on Performance
and Self-Reports of Satisfaction," Organization-
al Behavior and Human Performance 31
(1983): 203-222, esp. p. 205; Patrick B. Forsyth
and Thomas J. Danisiewicz, "Toward a Theory
of Professionalization,'' Work and Occupation
12 (February 1985): 59-76, esp. p. 60; Peter
Meiksins, "Science in the Labor Process: Engin-
eers as Workers" in Professionals as Workers:
Mental Labor in Advanced Capitalism, edited
by Charles Derber (G.K. Hall: Boston, 1982) pp.
 1095/1183

121-140, esp. p. 131: and even Layton, Revolt,

5.

22. Martin, "Professional Autonomy," is good

on this point.

23. See also Robert Zussman, Mechanics of the

Middle Class: Work and Politics Among Amer-
ican Engineers (University of California Press:
Berkeley, 1995), p. 222.

24. That "beyond" no doubt includes "compet-

ence." But I do not think competence the heart
of the matter. For someone who does, see John
T. Sanders, "Honor Among Thieves: Some Re-
flections on Professional Codes of Ethics," Pro-
fessional Ethics 2 (Fall/Winter 1993): 83-103.
 1096/1183

25. I say "typically" because a few professionsor

quasi-professionshave an impersonal ideal. For
example, science, at least on some conceptions,
serves no client, employer, or public but the
truth. What distinguishes professions from other
occupations is not service to others as such but a
moral ideal, defensible in part by the way
serving it benefits others. The truth
 Page 222

of science, though an impersonal object of

service, remains a morally good object of
service (just as justice, health, and safety
are) because the truths of science are import-
ant to us all, whether practically important
(as much of physics, chemistry, and biology
are) or just intellectually important (as much
of astronomy, etymology, and anthropology
are).
 1098/1183

26. Must "profession" be used in this way? That

depends on what is meant by "must." Both the
dictionary and ordinary usage allow for other
ways of using the word. So, if the question asks
about what usage requires, the answer is cer-
tainly no. If, however, the question asks how I
intend to use the word, or what usage I consider
more helpful in this context, the answer is yes. I
believe this way of using the word ''profession"
captures the project of the professions better
than any other. (This is, of course, an empirical
claim, one to be tested by asking members of
professions, especially those who have thought
most about what their profession means to them,
to choose among this definition and the
alternatives.)
 1099/1183

27. Voluntarily undertaking to serve a moral

ideal is, of course, not without its moral risks.
Like promising, it opens one to crititicism to
which one would not otherwise be open, the cri-
ticism that comes when one fails to do what one
has undertaken. But it also provides a basis for
further praise, the praise due one who has lived
up to his commitment.

28. We are, of course, assuming that this "any-

one" includes only sane adults of at least ordin-
ary intelligence, in other words, the sort of
people professions typically admit to practice.
 1100/1183

29. This problem has its counterpart in political

philosophy: Can one owe allegiance to law and
still be morally autonomous? For a sample of
the arguments against any consistency between
legal obligation and moral autonomy, see Robert
Paul Wolff, In Defense of Anarchism (Harper
and Row: New York, 1970) esp. pp. 3-19. Al-
though our question is easier to deal with than
the politicalbecause membership in a profession
is voluntary in a way subjection to law is notit is
worth noting that one major approach to making
legal obligation and moral autonomy consistent
is social contract theory, which tries to under-
stand subjection to law as if it were as voluntary
as membership in a profession. The real volun-
tariness of professions does, however, change
significantly what is necessary to preserve
autonomy. Compare the solution I offer here
with my response to Wolff, Michael Davis,
"Avoiding the Voter's Paradox Democratically,"
 1101/1183

Theory and Decision 5 (October 1974):

295-311.

30. Indeed, agent-centered conceptions probably

have a different purpose in view, to distinguish
those who have a right to autonomy we are
bound to respect from those who have no such
right. A right to autonomy is, typically, a right
to have primarily self-regarding decisions re-
spected whatever their apparent merit. It is not
surprising then that agent-centered conceptions
do not suit our purposes; the decisions of profes-
sionals are, typically, not primarily self-regard-
ing (and are not supposed to be). Compare
Dworkin, The Theory and Practice of
Autonomy, p. 19: "I am not trying to analyze the
notion of autonomous acts."
 1102/1183

31. Kant scholars may object that I am being un-

fair to Kant. They might be right. Kant has a no-
tion of "contradiction with a system of nature,"
that may provide a substantive test. Kant,
Foundations, 39. I decline to use Kant here be-
cause so many have found this notion too ob-
scure to be helpful and because it seems less de-
manding than the test I propose.

32. For a defense of the equation of rationality

with autonomy, see Davis, "Brandt on
Autonomy" (and the work of Richard Brandt
cited there).

33. Layton, Revolt,. 5.

34. Ibid.
 1103/1183

35. We can now offer an analogous analysis of

moral autonomy: Moral autonomy consists of
being able to do as morality requires (when that
ability includes both having the appropriate de-
sires and having the capacity to act on them).
So, one can both submit to law and be morally
autonomous, as long as the law does not require
anything morality forbids. This analysis of mor-
al autonomy makes the relationship between
moral and personal autonomy hard to sort out.
For example, do I have personal autonomy
whenever I can act as morality
 Page 223

requires or must I have other capacities as

well (such as, say, the ability to look after
my own interests)?

Epilogue

I read the first draft of this chapter, under the

title "Questions for STS from Engineering Eth-
ics," at a session of the Society for the Social
Study of Science Annual Meeting, Charlottes-
ville, Virginia, October 22, 1995. I would like to
thank those present, both audience and other
panelists, but especially Vivian Weil, for many
helpful comments.
 1105/1183

1. Donald Mackenzie, Inventing Accuracy: A

Historical Sociology of Nuclear Missile Guid-
ance (MIT Press: Cambridge, Mass., 1990).
Another recent book often offered to me as
evidence that the social sciences are already
studying engineering is Robert J. Thomas, What
Machines Can't Do: Politics and Technology in
the Industrial Enterprise (University of Califor-
nia Press: Berkeley, 1994). Engineers (so de-
scribed) do now and then have walk-on parts,
but there is no attempt to study their contribu-
tion systematically, much less to consider who
among the managers is in fact an engineer (op-
erating as such). Indeed, the focus of this book
seems to be the machinists on the shop floor.
 1106/1183

2. What Steve Woolgar hailed as "The Turn to

Technology in Social Science Studies," Science,
Technology, and Human Values 16 (Winter
1991): 20-50, has so far not reached engin-
eeringexcept for those who equate technology
with engineering. That equation is quite com-
mon. Consider, for example, an early piece,
James K. Feibleman, "Pure Science, Applied
Science, Technology, Engineering: An Attempt
at Definition," Technology and Culture 2
(1961): 305-317. Engineering actually receives
no definition or, indeed, hardly a mention after
the titleand the only extended discussion of en-
gineering concerns "Roman engineers.'' The Ro-
mans called those guys "builders" ("architects").
"Engineer" is (as we learned in chapter 1) a rel-
atively recent coinage; reference to ancient en-
gineers should at least come with a justification
(and quotation marks to signal the anachron-
ism). Feibleman's errors have been repeated for
more than thirty years.
 1107/1183

3. In the context of medical ethics, I callfor ob-

vious reasonsthe technical "the therapeutic." Mi-
chael Davis, "The State's Dr. Death: What's
Unethical about Physicians Helping at Execu-
tions?" Social Theory and Practice 21 (Spring
1995): 31-60.

4. For an example of the philosophical approach

to engineering ethics, see Nigel G.E. Harris,
"Professional codes and Kantian duties" in Eth-
ics and the Professions, edited by Ruth F. Chad-
wick (Avebury: Aldershot, England, 1994) pp.
104-115.

5. For a good example of the casuistic approach,

see Ken Alpern, "Moral Responsibility for
Engineers," Business and Professional Ethics
Journal 2 (Winter 1983): 39-48; or Eugene
Schlossberger, The Ethical Engineer (Temple
University Press: Philadelphia, 1993).
 1108/1183

6. For a good example of the technical ap-

proach, see the use Mike W. Martin and Roland
Schinzinger make of the concept of engineering
as social experimentation in Ethics in Engineer-
ing, 2nd ed. (McGraw-Hill: New York, 1989);
or Timo Airaksinen, "Service and Science in
Professional Life" Ethics and the Professions,
edited by Ruth F. Chadwick (Avebury: Alder-
shot, England, 1994), pp. 1-13.

7. I can't think of a clear case of a philosopher

using the social approach in engineering ethics.
I list it here because it seems to pop up regularly
in discussions with engineers. For a philosopher
who used it in medical ethics, see Robert M.
Veatch, "Medical Ethics and the Grounding of
Its Principles," Journal of Medicine and Philo-
sophy 4 (March 1979): 1-19.
 1109/1183

8. This is not to deny that there are, now and

then, moments resembling direct negotiation
between society and engineering; it is merely to
acknowledge how rarely society,
 Page 224

whether through government or through

newspaper editorials or other non-govern-
mental pressures, takes an active part.

9. For a text in engineering ethics that takes the

professional approach, see Charles Harris, Mi-
chael Pritchard, and Michael Rabins, Engineer-
ing Ethics: Concepts and Cases (Wadsworth:
Belmont, 1995).

10. For an example of how hostile to codes a de-

votee of the philosophical approach can be, see
John Ladd, "Collective and Individual Respons-
ibility in Engineering: Some Questions" in Bey-
ond Whistleblowing: Defining Engineers' Re-
sponsibilities, edited by Vivian Weil (Center for
the Study of Ethics in the Professions, Illinois
Institute of Technology) pp. 90-113.
 1111/1183

11. For a rare (but welcome) example of what

could be done, see Peter Whalley, "Negotiating
the Boundaries of Engineering: Professionals,
Managers, and Manual Work," Research in the
Sociology of Organizations 8 (1991) 191-215.

12. Walter G. Vincenti, What Engineers Know

and How They Know It (Johns Hopkins
University Press: Baltimore, 1990).

13. For another good example of what can be

done, see Bruce Seeley, "The Scientific Mys-
tique in Engineering: Highway Research at the
Bureau of Public roads, 1918-1940," Techno-
logy and Culture 25 (October 1984): 798-831.

14. Tracy Kidder, The Soul of a New Machine

(Little Brown: Boston, 1981)
 1112/1183

15. I would offer the same qualified praise for

Kathryn Henderson's work, for example, "Flex-
ible Sketches and Inflexible Data Bases: Visual
Communication, Conscription Devices, and
Boundary Objects in Design Engineering,"
Science, Technology, and Human Values 16
(Autumn 1991): 448-473.

16. Interestingly, at least some sociologists, not-

ing this advantage of the historians, simply ad-
opted their methods. See, for example, Peter
Meiksins, "The `Revolt of the Engineers' Recon-
sidered," Technology and Culture 29 (1986):
219-246. One of the good features of science
and technology studies is that disciplinary
boundaries remain relatively unimportant. So,
"historian" must be read here as "someone func-
tioning as a historian" rather than as "someone
of that profession."
 1113/1183

17. Edwin Layton, The Revolt of Engineers

(Case Western Reserve University Press: Cleve-
land, 1971)

18. For a rare example of what such work might

look like, see Eva Kranakis, "Social Determin-
ants of Engineering Practice: A Comparative
View of France and American in the Nineteenth
Century," Social Studies of Science 19 (Febru-
ary 1989): 5-70.

19. Report of Investigation into Allegations of

Retaliation for Raising Safety and Quality of
Work Issues Regarding Argonne National
Laboratory's Integral Fast Reactor Project (Of-
fice of Nuclear Safety, U.S. Department of En-
ergy: Washington, D.C., December 1991).
 1114/1183

20. Although he had a Ph.D. in metallurgy, his

bachelor's degree was in metallurgical engineer-
ing (Colorado School of Mines, 1978) and his
job description at Argonne was "associate en-
gineer and experimenter." Report of Allegations,
19: 7. The report does not make clear whether
his graduate training and other job experience
was in engineering or science (though what it
does say is at least consistent with his graduate
training being in an engineering department).

21. Among the most important exceptions are

Robert Perrucci and Joel E. Gerstl, Profession
without Community: Engineers in American So-
ciety (Random House: New York, 1969); much
of the work of Edward W. Constant II; and
Robert Zussman, Mechanics of the Middle
Class: Work and Politics Among American
Engineers (University of California Press:
Berkeley, 1995).
 1115/1183

22. See, for example, Richard DeGeorge, "Eth-

ical Responsibilities of Engineers in Large Or-
ganizations," Business and Professional Ethics
Journal 1 (1981): 1-14.
 Page 225

23. Carl Mitcham, Thinking through Techno-

logy: The Path between Engineering and Philo-
sophy (University of Chicago Press: Chicago,
1994) pp. 103-105.

24. While the editorial boards of many of these

journals do mix scholars in professional ethics
with those in science and technology studies, I
take that fact to indicate original hopes rather
than present reality. The present separation of
the fields is, I think, due not to a failure of the
journals' founders to understand the connection
between professional ethics and science and
technology studies but a failure to realize that
vision in daily practice. A choice of editor-in-
chief here, the weight of submissions there,
slowly turned journals with a distinct vision into
journals much like others in their disciplinean
interesting subject for a monograph or two on
the sociology of the social sciences.
1117/1183
 Page 227

BIBLIOGRAPHY
Agich, George J. "Rationing Professional
Autonomy," Law, Medicine and Health Care 18
(Spring-Summer 1990): 77-84.

Airaksinen, Timo. "Service and Science in Pro-

fessional Life" in Ethics and the Professions, ed.
Ruth F. Chadwick (Avebury: Aldershot, Eng-
land, 1994), 1-13.

Alie, Raymond E. "Professionalization in a Ra-

tionally Managed Corporate Organization,"
Journal of Management 8 (1982): 65-81.

Alpern, Ken. "Moral Responsibility for Engin-

eers," Business and Professional Ethics Journal
2 (Winter 1983): 39-48.
 1119/1183

American Society of Mechanical Engineers v.

Hydrolevel Corporation 456 U.S. 556 (1982).

Arendt, Hannah. The Human Condition

(Doubleday: Garden City, N.Y., 1959).

Arenson, Richard. "The Principle of Fairness

and Free-Rider Problems," Ethics 92 (July
1982): 616-633.

Argyris, Chris, and Donald Schön. "Reciprocal

Integrity: Creating Conditions That Encourage
Personal and Organizational Integrity" in Exec-
utive Integrity, ed. Suresh Srivastva et al.
(Jossey-Bass: San Francisco, 1988), pp.
197-222.

Armytage, W.H.G. A Social History of Engin-

eering (Faber and Faber: London, 1961).
 1120/1183

Artz, Frederick B. The Development of Technic-

al Education in France, 1500-1850 (MIT Press:
Cambridge, Mass., 1966).

Asimov, M. "A Philosophy of Engineering

Design" in Contributions to a Philosophy of
Technology, ed. Friedrich Rapp (Reidel:
Dortrecht-Holland, 1974), pp. 150-157.

Baker, Ray Palmer. A Chapter in American

Education: Rensselear Polytechnic Institute,
1824-1924 (Charles Scribner's Sons: New York,
1924).

Barnes, Chester A. The Functions of the Execut-

ive (Cambridge, Mass.: Harvard University
Press, 1938).

Bayles, Michael. "Against Professional

Autonomy," National Forum 58 (Summer
1978), 23-26.
 1121/1183

. Professional Ethics (Wadsworth: Belmont,

Cal., 1981), pp. 133-138.
 Page 228

Beardsley, Charles W. "The Hydrolevel CaseA

Retrospective," Mechanical Engineering 106
(June 1984): 66-73.

Bell, Trudy E. "The Fatal Flaw in Flight 51-L,"

IEEE Spectrum (February 1987): 36-51.

Bella, David A. "Organizations and Systematic

Distortion of Information," Journal of Profes-
sional Issues in Engineering 113 (October
1987): 360-370.

Benn, Stanley I. A Theory of Freedom (Cam-

bridge: Cambridge University Press, 1988).

Brady, Gene F., Ben B. Judd, and Setrak Javian.

"The Dimensionality of Work Autonomy Revis-
ited," Human Relations 43 (1990): 1219-1228.

Broome, Taft H., Jr. "Engineering the Philo-

sophy of Science," Metaphilosophy 16 (January
1985): 47-56.
 1123/1183

Buckley, Fletcher J. "Defining Software Engin-

eering as a Profession," Computer 26 (August
1993): p. 76-78.

Burghardt, M. David. Introduction to the Engin-

eering Profession (HarperCollins: New York,
1985).

Burke, John G. et al. Connections: Technology

and Change (Boyd and Fraser: San Francisco,
1979).

Burns, Tom and G.M. Stalker. The Management

of Innovation (Tavistock: London, 1966).

Calhoun, Daniel Hovey. The American Civil

Engineer: Origins and Conflict (Technology
Press-MIT: Cambridge, Mass., 1960).

Calvert, Monte A. The Mechanical Engineer in

America, 1830-1910 (Johns Hopkins University
Press: Baltimore, 1967).
 1124/1183

Camenisch, Paul F. Grounding Professional

Ethics in a Pluralistic Society (Haven: New
York, 1983), pp. 29-3

Canadian Engineering Qualifications Board,

1993 Annual Report (Canadian Council of Pro-
fessional Engineers: Ottawa, 1993).

Chalk, Rosemary. "Making the World Safe for

Whistleblowers," Technology Review 91 (Janu-
ary 1988): 48-57.

Child, John, and Janet Fulk. "Maintenance of

Occupational Control: The Case of Professions,"
Work and Occupations 9 (May 1982): 155-192.

Christie, A.G. "A Proposed Code of Ethics for

All Engineers," Annals of American Academy of
Political and Social Science 101 (May 1922),
pp. 97-104.
 1125/1183

Christman, John. ed. The Inner Citadel (Oxford

University Press: New York, 1989).

Cooke, Morris Llewellyn. "Ethics and the

Engineering Profession," Annals of the Americ-
an Academy for Political and Social Science
101 (May 1922): 68-72.

Crime in the United States (Federal Bureau of

Investigation, U.S. Department of Justice:
Washington, D.C., 1980).

Daniels, Arlene Kaplan. "How Free Should Pro-

fessions Be?" in The Professions and Their Pro-
spects, ed. Eliot Freidson (Sage: Beverly Hills,
Cal., 1971.

Davis, Michael. "Avoiding the Voter's Paradox

Democratically," Theory and Decision 5 (Octo-
ber 1974): 295-311.
 1126/1183

. "Brandt on Autonomy" in Rationality and

Rule-Utilitarianism, ed. Brad Hooker
(Westview Press: Boulder, Colo., 1993), pp.
51-65.

. "Conflict of Interest," Business and Profession-

al Ethics Journal 1 (Summer 1982): 17-27

. "Conflict of Interest Revisited," Business and

Professional Ethics Journal 12 (Winter 1993):
21-41.

. "Do Cops Really Need a Code of Ethics,"

Criminal Justice Ethics 10 (Summer/Fall 1991):
14-28.
 Page 229

. "The Ethics Boom: What and Why," Centenni-

al Review 34 (Spring 1990): 163-186.

. "The Moral Authority of a Professional Code,"

NOMOS 29 (1987): 302-337.

. "The Moral Legislature: Morality without an

Archimedean Point," Ethics 102 (July 1992):
303-318.

. "Nozick's Argument for the Legitimacy of the

Welfare State," Ethics 97 (April 1987): 576-594.

. One Social Responsibility of Engineering Soci-

eties: Teaching Managers About Engineering
Ethics, Monograph #88-WA/DE-14 (American
Society of Mechanical Engineers: New York,
1988).

. "Professionalism Means Putting Your Profes-

sion First," Georgetown Journal of Legal Ethics
(Summer 1988): 352-366
 1128/1183

. "Science: After Such Knowledge, What Re-

sponsibility?," Professional Ethics 4 (Spring
1995): 49-74.

. "Some Paradoxes of Whistleblowing," Busi-

ness and Professional Ethics Journal 15 (Spring
1996): 3-19.

. "The Special Role of Professionals in Business

Ethics," Business and Professional Ethics
Journal 7 (1988): 83-94.

. "The State's Dr. Death: What's Unethical about

Physicians Helping at Executions?," Social The-
ory and Practice 21 (Spring 1995): 31-60.

. "Technical Decisions: Time to Rethink the

Engineer's Responsibilities?," Business and Pro-
fessional Ethics Journal 11 (Fall-Winter 1992):
41-55.
 1129/1183

. "Thinking like an Engineer: The Place of a

Code of Ethics in the Practice of a Profession,"
Philosophy and Public Affairs 20 (Spring 1991):
150-167.

. "The Use of Professions," Business Economics

22 (October 1987): 5-10.

. "Vocational Teachers, Confidentiality, and

Professional Ethics," International Journal of
Applied Philosophy 4 (Spring 1988): 11-20

. "Who Can Teach Workplace Ethics?," Teach-

ing Philosophy 13 (March 1990): 21-36.

DeGeorge, Richard. "Ethical Responsibilities of

Engineers in large Organizations," Business and
Professional Ethics Journal 1 (1981): 1-14.
 1130/1183

Devine, Thomas M., and Donald G. Aplin.

"Whistleblower ProtectionThe Gap between the
Law and Reality," Howard Law Journal 31
(1988): 223-239

Durbin, Paul T. "Toward a Philosophy of Engin-

eering and Science in R and D Settings," in
Technology and Responsibility, ed. Paul Durbin
(Reidel: Dordrecht, Holland, 1987), pp.
309-327.

. ed. Critical Perspectives on Nonacademic

Science and Engineering (Lehigh University
Press: Bethlehem, Pa., 1991).

Dworkin, Gerald. "Concept of Autonomy" in

Science and Ethics, ed. Rudolph Haller (Ams-
terdam: Rodopi Press, 1981), pp. 203-213.

. The Theory and Practice of Autonomy (Cam-

bridge University Press: New York, 1988).
 1131/1183

Elliston, Frederick, et al. Whistleblowing Re-

search: Methodological and Moral Issues (Prae-
ger: New York, 1985).

Elster, Jan. Ulysses and the Sirens, rev. ed.

(Cambridge University Press: New York, 1984).

Emmerson, George S. Engineering Education:

A Social History (Crane, Russak and Company:
New York, 1973), pp. 140-141.

Farh, Jiing-Lih, and W.E. Scott, Jr. "The Experi-

mental Effects of `Autonomy' on Performance
and Self-Reports of Satisfaction," Organization-
al Behavior and Human Performance 31
(1983): 203-222.

Feibleman, James K. "Pure Science, Applied

Science, Technology, Engineering: An Attempt
at Definitions," Technology and Culture 2 (Fall
1961): 305-317.
 Page 230

Feldman, Steven P. "The Broken Wheel: The

Inseparability of Autonomy and Control in In-
novation within Organizations," Journal of
Management Studies 26 (March 1989): 83-102.

Feynman, Richard. "An Outsider's Inside View

of the Challenger Inquiry," Physics Today
(February 1988): 26-37.

Finch, James Kip. The Story of Engineering

(Doubleday: Garden City, N.Y., 1960).

. A History of the School of Engineering,

Columbia University (Columbia University
Press: New York, 1954).

Fleming, Thomas J. West Point: The Men and

Times of the United States Military Academy
(William Morrow: New York, 1969).

Florman, Samuel. The Civilized Engineer (St.

Martin Press: New York, 1987).
 1133/1183

. The Existential Pleasures of Engineering (St.

Martin's Press: New York, 1976).

Ford, Gary A., and James E. Tomaykoo Encyc-

lopedia of Software Engineering (John Wiley
and Sons: New York, 1994).

Forsyth, Patrick B., and Thomas J. Danisiewicz.

"Toward a Theory of Professionalization," Work
and Occupation 12 (February 1985): 59-76.

Frankfurt, Harry G. "Freedom of the Will and

the Concept of a Person," Journal of Philosophy
68 (January 1971): 5-20.

Freedman, Benjamin. "A Meta-Ethics of Profes-

sional Morality" in Moral Responsibility and the
Professions, ed. Bernard Baumrin and Benjamin
Freedman (Haven: New York, 1983), pp. 61-79.
 1134/1183

Gellerman, Saul W. "Why `Good' Managers

Make Bad Ethical Choices," Harvard Business
Review 64 (July-August 1986): 85-90.

Gert, Bernard. Morality: A New Justification of

the Moral Rules (Oxford University Press: New
York, 1988).

Gewirth, Alan. "Professional Ethics: The Separ-

atist Thesis," Ethics 96 (1986): 282-300.

Glazer, Myron Peretz, and Penina Migdal

Glazer. The Whistleblowers: Exposing Corrup-
tion in Government and Industry (Basic Books:
New York, 1989).

Goldman, Alan H. The Moral Foundations of

Professional Ethics (Rowman and Littlefield:,
Totawa, N.J., 1980).
 1135/1183

Gordon, Bruce F., and Ian C. Ross. "Profession-

als and the Corporation," Research Management
5 (November 1962): 493-505.

Gorlin, Rena A. ed. Codes of Professional Re-

sponsibility (Bureau of National Affairs: Wash-
ington, D.C., 1986).

Grayson, Lawrence P. "The American Revolu-

tion and the `Want of Engineers,'" Engineering
Education 75 (February 1985): 268-276.

. "A Brief History of Engineering Education in

the United States," Engineering Education 68
(December 1977): 246-264.

Hacker, Barton. "Engineering a New Order:

Military Institutions, Technical Education, and
the Rise of the Industrial State," Technology and
Culture 34 (April 1993): 1-27.
 1136/1183

Hacker, Sally. Doing It the Hard Way (Unwin:

Boston, 1990).

Harris, Charles, Michael Pritchard, and Michael

Rabins. Engineering Ethics: Concepts and
Cases (Wadsworth: Belmont, Mass., 1995).

Harris, Nigel G.E. "Professional Codes and

Kantian Duties" in Ethics and the Professions,
ed. Ruth F. Chadwick (Avebury: Aldershot,
England, 1994), pp. 104-115.

Heacock, Marian V., and Gail W. McGee.

"Whistleblowing: An Ethical Issue in Organiza-
tional and Human Behavior," Business and Pro-
fessional Ethics Journal 6 (Winter 1987): 35-46.

Henderson, Harold. "McGregor v. the NRC:

Why Did the Nuclear Regulatory Commission
Fire One of Its Toughest Plant Inspectors?,"
Reader (Chicago), July 22, 1988, pp. 1ff.
 Page 231

Henderson, Kathryn. "Flexible Sketches and In-

flexible Data Bases: Visual Communication,
Conscription Devices, and Boundary Objects in
Design Engineering," Science, Technology, and
Human Values 16 (Autumn 1991): 448-473.

Herbert, Theodore T., and Ralph W. Estes. "Im-

proving Executive Decisions by Formalizing
Dissent: The Corporate Devil's Advocate,"
Academy of Management Review 2 (October
1977): 662-667.

Hill, Donald. A History of Engineering in Clas-

sical and Medieval Times (Croom Helm: Lon-
don, 1984).

Hollis, Bruce W. "I Turned in My Mentor," The

Scientist 1 (December 14, 1987): 1-13.
 1138/1183

Hounshell, David A. From the American System

to Mass Production, 1800-1932: The Develop-
ment of Manufacturing Technology in the Un-
ited States (Johns Hopkins University Press:
Baltimore, 1984).

Jackall, Robert. Moral Mazes: The World of

Corporate Managers (Oxford University Press:
New York, 1988.

Kadish, Mortimer R. The Ophelia Paradox: An

Inquiry into the Conduct of Our Lives (Transac-
tion: New Brunswick, N.J., 1994).

Kanigel, Robert. "The Endangered Profession-

al," Johns Hopkins Magazine 40 (June 1988):
17-43.

Kant, Immanuel. Foundations of the Metaphys-

ics of Morals, 2d. ed. Lewis White Beck (Mac-
millan/Library of the Liberal Arts: New York,
1990).
 1139/1183

Katz, Jack. Seductions of Crime: Moral and

Sensual Attractions of Doing Evil (Basic Books:
New York, 1988).

Kidder, Tracy. The Soul of a New Machine

(Little Brown: Boston, 1981).

Kipnis, Kenneth. "Professional Responsibility

and the Responsibility of Professions" in Profits
and Professions: Essays in Business and Profes-
sional Ethics, ed. Wade L. Robinson, Michael
Pritchard, and Joseph Ellin (Humana Press:
Clifton, N.J., 1983), pp. 9-22.

Klaw, Spencer. "The Faustian Bargain" in The

Social Responsibility of the Scientist, ed. Martin
Brown (Free Press: New York, 1971).

Koen, Billy Vaughn. Definition of the Engineer-

ing Method (American Society of Engineering
Education: Washington, D.C., 1991).
 1140/1183

. "Toward a Definition of the Engineering Meth-

od," Engineering Education 75 (December
1984): 150-155.

Kohlberg, Lawrence. Essays on Moral Develop-

ment (Harper and Row: San Francisco, 1981).

Kranakis, Edna. "Social Determinants of Engin-

eering Practice: A Comparative View of France
and America in the Nineteenth Century," Social
Studies of Science 19 (February 1989): 5-70.

Kultgen, John. Ethics and Professionalism

(University of Pennsylvania Press: Philadelphia,
1988), pp. 274-306.

Kurtines, William M., and Jacob L. Gewirtz.

ed., Morality, Moral Behavior, and Moral
Development (John Wiley & Sons: New York:
1984).
 1141/1183

Ladd, John. "Collective and Individual Moral

Responsibility in Engineering: Some Questions"
in Beyond Whistleblowing: Defining Engineers'
Responsibilities, ed. Vivian Weil (Center for the
Study of Ethics in the Professions, Illinois Insti-
tute of Technology: Chicago, 1983), pp.
102-103.

Larson, Magali Sarfatti. The Rise of Profession-

alism (University of California Press: Berkeley,
1977).

Layton, Edwin. The Revolt of the Engineers

(Case Western Reserve University Press: Cleve-
land, 1971).

Luegenbiehl, Heinz C. "Code of Ethics and the

Moral Education of Engineers," Business and
Professional Ethics Journal 2 (Summer 1983):
41-61.
 Page 232

. "Computer Professionals: Moral Autonomy

and a Code of Ethics," Journal of Systems Soft-
ware 17 (1992): 61-68.

Lyons, David. Forms and Limits of Utilitarian-

ism (Oxford: Oxford University Press, 1965).

Mackenzie, Donald. Inventing Accuracy: A His-

torical Sociology of Nuclear Missile Guidance
(MIT Press: Cambridge, Mass., 1990).

Mahoney, Michael S. "The Roots of Software

Engineering," CWI Quarterly 3 (December
1990): 325-334.

Malin, Martin H. "Protecting the Whistleblower

from Retaliatory Discharge," Journal of Law
Reform 16 (Winter 1983): 277-318.
 1143/1183

Matthews, M. Cash. "Ethical Dilemmas and the

Disputing Process: Organizations and Societ-
ies," Business and Professional Ethics Journal 8
(Spring 1989): 1-11.

Marcus, Afred A. "Professional Autonomy as a

Basis of Conflict in an Organization," Human
Resource Management 24 (Fall 1985): 311-328.

Martin, Mike W. "Professional Autonomy and

Employers' Authority" in Profits and Profes-
sions: Essays in Business and Professional Eth-
ics, ed. Wade L. Robinson, Michael Pritchard,
and Joseph Ellin (Humana Press: Clifton, N.J.,
1983), pp. 265-273.

. Self-Deception and Morality (University Press

of Kansas: Lawrence, Kan., 1986).

Martin, Mike W., and Roland Schinzinger. Eth-

ics in Engineering, 2nd ed. (McGraw-Hill: New
York, 1989).
 1144/1183

McGivern, James Gregory. First Hundred Years

of Engineering Education in the United States
(1807-1907) (Gonzaga University Press:
Spokane, Wash., 1960).

McMahan, Christopher. "Managerial Author-

ity," Ethics 100 (October 1989): 33-53.

McMahan, A. Michal. The Making of a Profes-

sion: A Century of Electrical Engineering in
America (Institute of Electrical and Electronic
Engineers: New York, 1984).

Meiksins, Peter. "Professionalism and Conflict:

The Case of the American Association of Engin-
eers," Journal of Social History (Spring 1983):
403-421.

. "The `Revolt of the Engineers' Reconsidered,"

Technology and Culture 29 (1986): 219-246.
 1145/1183

. "Science in the Labor Process: Engineers as

Workers" in Professionals as Workers: Mental
Labor in Advanced Capitalism, ed. Charles Der-
ber (G.K. Hall: Boston, 1982), pp. 121-140.

Meiksins, Peter F., and James M. Watson, "Pro-

fessional Autonomy and Organizational Con-
straint: The Case of Engineers," Sociological
Quarterly 30 (1989): 561-585.

Mele, Alfred. "Recent Work on Self-Decep-

tion," American Philosophical Quarterly 24
(January 1987): 1-17.

Meyer, Priscilla S. "Knocking the Competition:

How Rival's Use of `Industrial Code' Report
Created Problems for a Tiny Company," Wall
Street Journal, (July 9, 1974) p. 44.

Meyers, Diana T. Self, Society, and Personal

Choice (New York: Columbia University Press,
1989).
 1146/1183

Miceli, Marcia P., and Janet P. Near, "Individual

and Situational Correlates of Whistle-Blowing,"
Personnel Psychology 41 (Summer 1988):
267-281.

Milgram, Stanley. Obedience to Authority: An

Experimental View (Harper and Row: New
York, 1974).

Mitcham, Carl. Thinking through Technology:

The Path between Engineering and Philosophy
(University of Chicago Press: Chicago, 1994).

Molloy, Peter Michael. Technical Education

and the Young Republic: West Point as
America's École Polytechnique (unpublished
dissertation, Brown University, June 1971).

Musa, John D. "Software Engineering: The Fu-

ture of a Profession," IEEE Software (January
1985): 55-62.
 1147/1183

Noble, David. America by Design (Alfred A.

Knopf: New York, 1977).
 Page 233

Nozick, Robert. Anarchy, State, and Utopia

(Basic Books: New York, 1974).

Pavlovic, K.R. "Autonomy and Obligation: Is

there an Engineering Ethics?" in Ethical Prob-
lems in Engineering, 2nd ed., vol. 1, ed. Albert
Flores (Center for the Study of the Human Di-
mensions of Science and Technology: Troy,
N.Y., 1980), pp. 89-93.

Percival, Thomas. Medical Ethics; or a Code of

Institutes and Precepts Adapted to the Profes-
sional Conduct of Physicians and Surgeons
(1803).

Perrucci, Robert, and Joel E. Gerstl, Professions

without Community: Engineers in American So-
ciety (Random House: New York, 1969).
 1149/1183

Perry, Tekla S. "Antitrust Rule Chills Standards

Setting," IEEE Spectrum 11 (August 1982):
52-54.

Petroski, Henry. "The Iron Ring," American

Scientist 83 (May-June 1995): 229-232.

Petzinger, Thomas. "Hangar Anger: Mechanic's

Woes Show How Safety Became a Big Issue for
Eastern," Wall Street Journal, June 9, 1988, pp.
1ff.

Polman, Dick. "Telling the truth, paying the

price," Philadelphia Inquirer Magazine, June
18, 1989, pp. 16ff.

[The] Presidential Commission on the Space

Shuttle Challenger Disaster (U.S. Government
Printing Office: Washington, DC, 1986).
 1150/1183

"Price Fixing and Bid Rigging in the Electrical

Manufacturing Industry," Administered Prices:
Hearings before the Senate Subcommittee on
Antitrust and Monopoly of the Committee on the
Judiciary, 78th Congress 1st session pt. 27,
1961.

Pritchard, Michael. "Conflict of Interest: Con-

ceptual and Normative Issues," Academic Medi-
cine 71 (December 1996): 1305-1313.

O'Boyle, Thomas F. "Chilling Tale: GE Refri-

gerator Woes Illustrate the Hazards in Changing
a ProductFirm Pushed Development of Com-
pressor Too Fast, Failed to Test Adequately,"
Wall Street Journal, May 7, 1990, pp. 1ff.

Occupational Outlook Handbook (U.S. Depart-

ment of Labor, Bureau of Statistics: Washing-
ton, D.C., May 1992).
 1151/1183

O'Connor, Stephen J., and Joyce A. Lanning,

"The End of Autonomy? Reflections on the
Postprofessional Physician," Health Care Man-
agement Review 17 (Winter 1992): 63-72.

Opinions of the Board of Ethical Review (Na-

tional Society of Professional Engineers: Wash-
ington, D.C.).

Oldenquist, Andrew G., and Edward E. Slowter,

"Proposed: A Single Code of Ethics for All
Engineers," Professional Engineer 49 (May
1979): 8-11.

O'Reilly, Charles. "Corporations, Culture, and

Commitment: Motivation and Social Control in
Organizations," California Management Review
31 (Summer 1989): 9-25.
 1152/1183

Ottoson, Gerald E. "Essentials of an Ethical

Corporate Climate" in Doing Ethics in Business,
ed. Donald G. Jones (Oelgeschlager, Gunn and
Hain: Cambridge, Mass., 1982), pp. 155-163.

Raelin, Joseph A. "An Anatomy of Autonomy:

Managing Professionals," Academy of Manage-
ment Executive 3 (August 1989): 216-228.

. The Clash of Cultures: Managers and Profes-

sionals (Harvard Business School Press: Cam-
bridge, Mass., 1986.

. "The Professional as the Executive's Ethical

Aide-de-Camp," Academy of Management Ex-
ecutive 1 (August 1987): 171-182.

Raz, Joseph. "Autonomy, Toleration, and the

Harm Principle" in Issues in Contemporary
Legal Philosophy: The Influence of H.L.A. Hart,
ed. Ruth Gavison (Oxford University Press:
New York, 1987), 313-333.
 1153/1183

Reader, W.J. Professional Men: The Rise of the

Professional Classes in Nineteenth-Century
England (Basic Books: New York, 1966).
 Page 234

Report of Investigation into Allegations of

Retaliation for Raising Safety and Quality of
Work Issues regarding Argonne National
Laboratory's Integral Fast Reactor Project (Of-
fice of Nuclear Safety, U.S. Department of En-
ergy: Washington, D.C., December 1991).

Reynolds, Terry S. 75 Years of Progress: A His-

tory of the American Institute of Chemical
Engineers (American Institute of Chemical
Engineers: New York, 1983).

Rockwell, William H. ''Hydrolevel Decision as

Applied to Antitrust Violations of Standards
Making Organizations," CSEP Perspectives on
the Professions 3(3) (1983): 3-5.

Rosenbaum, Bernard. "Leading Today's

Professionals," Research-Technology
Management (March-April 1991): 30-35.
 1155/1183

Rossi, Paolo. Philosophy, Technology, and the

Arts in the Early Modern Era, trans. S. Attanas-
io (Harper and Row: New York, 1970).

Rueth, Nancy. "A Case Study," Mechanical

Engineering 97 (June, 1975): 34-36.

Saha, Pamela, and Subrata Saha, "Ethical Re-

sponsibilities of the Clinical Engineer," Journal
of Clinical Engineering 11 (January/February
1986): 17-25.

Sanders, John T. "Honor Among Thieves: Some

Reflections on Professional Codes of Ethics,"
Professional Ethics 2 (Fall/Winter 1993):
83-103.

Sanger, David E. "How Seeing-No-Evil

Doomed the Challenger," New York Times, June
29, 1986, sec. 3, p. 8.
 1156/1183

Scanlon, Thomas. "A Theory of Freedom of Ex-

pression," Philosophy and Public Affairs 1
(Winter 1972): 204-226.

Schlossberger, Eugene. The Ethical Engineer

(Temple University Press: Philadelphia, 1993).

Schwartz, Adina. "Autonomy in the Workplace"

in Just Business: New Introductory Essays in
Business Ethics, ed. Tom Regan (Random
House: New York, 1984), pp. 129-166.

Seely, Bruce. "Research, Engineering, and

Science in American Engineering Colleges:
1900-1960," Technology and Culture 34 (April
1993): 344-386)

. "The Scientific Mystique in Engineering:

Highway Research at the Bureau of Public
Roads, 1918-1940," Technology and Culture 24
(October 1984): 798-831.
 1157/1183

Shaffer, Michael J., and Michael D. Shaffer,

"The Professionalization of Clinical Engineer-
ing." Biomedical Instrumentation and Techno-
logy (September/October 1989): 370-374.

Shapero, Albert. Managing Professional People

(Free Press: New York, 1985).

Shaw, Mary. "Prospects for an Engineering Dis-

cipline of Software," IEEE Software (November
1990): 15-24.

Sherman, J. Daniel. "Technical Supervision and

Turnover Among Engineers and Technicians:
Influencing Factors in the Work Environment,"
Group and Organization Studies 14 (December
1989): 411-421.

Simmons, A. John. Moral Principles and Polit-

ical Obligations (Princeton University Press:
Princeton, N.J., 1979).
 1158/1183

Sinclair, Bruce. A Centennial History of the

American Society of Mechanical Engineers,
1880-1980 (University of Toronto Press:
Toronto, 1980).

Sinclair, George. "A Call for a Philosophy of

Engineering," Technology and Culture 8 (Octo-
ber 1977): 685-689.

Smith, Merritt Roe. ed., Military Enterprise and

Technological Change (MIT Press: Cambridge,
Mass., 1985).

Smith, Ralph J. Engineering as a Career, 3rd

ed. (McGraw-Hill: New York, 1969).

Spector, Paul E. "Perceived Control by Employ-

ees: A Meta-Analysis of Studies Concerning
Autonomy and Participation at Work," Human
Relations 39 (November 1986): 1005-1015.
 1159/1183

Starbuck, William, and Frances Milliken,

"Challenger: Fine-Tuning the Odds until So-
mething Breaks," Journal of Management Stud-
ies 25 (July 1988): 319-340.
 Page 235

Thalberg, Irving. "Hierarchical Analyses of Un-

free Action," Canadian Journal of Philosophy 8
(June 1978): 211-226.

Thomas, Robert J. What Machines Can't Do:

Politics and Technology in the Industrial Enter-
prise (University of California Press: Berkeley,
1994).

Turner, Barry A. Man-Made Disasters (Wyke-

ham: London, Ltd., 1978).

U.S. Scientists and Engineers: 1988 Estimates,

Surveys of Science Resources Series, National
Science Foundation, NSF 88-322.

Urquhart, Brian. "The Last Disaster of the War,"

New York Review of Books, September 24,
1987, pp. 27-30.
 1161/1183

Vaughn, Diane. The Challenger Launch De-

cision (University of Chicago Press: Chicago,
1996).

Veatch, Robert M. "Medical Ethics and the

Grounding of Its Principles," Journal of Medi-
cine and Philosophy 4 (March 1979): 1-19.

Victor, Bart, and John B. Cullen, "The Organiz-

ational Bases of Ethics Work Climates," Admin-
istrative Science Quarterly 33 (March 1988):
101-125.

Vincenti, Walter G. What Engineers Know and

How They Know It (Johns Hopkins University
Press: Baltimore, 1990).

Voluntary Industrial Standards: Hearing before

the Senate Subcommittee on Antitrust and
Monopoly of the Committee on the Judiciary,
94th Cong., 1st sess. 153-214 (1975).
 1162/1183

Waters, James A. "Catch 20.5: Corporate Mor-

ality as an Organizational Phenomenon," Or-
ganizational Dynamics 6 (Spring, 1978): 3-19.

. "Integrity Management: Learning and Imple-

menting Ethical Principles in the Work-place" in
Executive Integrity, ed. Suresh Srivastva et al.
(Jossey-Bass: San Fransciso, 1988).

Weiner, Philip P., ed. Dictionary of the History

of Ideas: Studies of Selected Pivotal Ideas,
(Charles Scribner's Sons: New York, 1973).

Whalley, Peter. "Negotiating the Boundaries of

Engineering: Professionals, Managers, and
Manual Work," Research in the Sociology of
Organizations 8 (1991) 191-215.

Whitbeck, Caroline. "Teaching Ethics to Scient-

ists and Engineers," Science and Engineering
Ethics 1 (July 1995): 299-308.
 1163/1183

Wilkins, Alan L., and William G. Ouchi, "Effi-

cient Cultures: Exploring the Relationship
between Culture and Organizational Perform-
ance," Administrative Science Quarterly 28
(September 1983): 468-481.

Williams, Rosalind. "Engineering's Image Prob-

lem," Issues in Science and Technology 6
(Spring 1990): 84-86.

Wisely, William H. "The Influence of Engineer-

ing Societies on Professionals and Ethics" in
Ethics, Professionals, and Maintaining Compet-
ence: ASCE Professional Activities Committee
Specialty Conference, Ohio State University,
Columbus, Ohio, 1977 (American Society of
Civil Engineers, New York, 1977), p. 51-62.

. "The Influence of Engineering Societies on

Professionalism and Ethics" in Engineering
Professionalism and Ethics (Robert E. Kreiger:
Malabar, Fl., 1983), p. 28-37.
 1164/1183

Wolff, Robert Paul. In Defense of Anarchism

(Harper and Row: New York, 1970).

Woolgar, Steve. "The Turn to Technology in

Social Science Studies," Science, Technology,
and Human Values 16 (Winter 1991): 20-50

World Almanac (World Almanac: New York,

1989)

Young, Robert. "Autonomy and the Inner Self,"

American Philosophical Quarterly 17 (January
1980): 35-43.

Zussman, Robert. Mechanics of the Middle

Class: Work and Politics among American
Engineers (University of California Press:
Berkeley, 1995).
 Page 237

INDEX

AAE (American Association of Engineers), 45,

201n50

AAES (American Association of Engineering

Societies), 110

ABET (Accreditation Board of Engineering and

Technology), 46-47, 48, 54, 110, 175, 206n27

AEC (American Engineering Council), 45.

See also ABET, ECPD

Absolutism, moral, 100-101.

See also ethics, duty-based

 1166/1183

AIEE (American Institute of Electrical Engin-

eers), 45.

See also IEEE

apothecaries, 193n12

applied science, 8, 15, 27, 28

AMA (American Medical Association), 8, 24,

109.

See also physicians

architects

and engineers, 9, 10-11, 14, 22, 28

among Greeks, 5, 34

among Romans, 223n2

Argonne National Laboratory, 177

 1167/1183

ASCE (American Society of Civil Engineers),

22, 29, 45, 199n25

ASEE (American Society of Engineering Edu-

cation), 29

ASME (American Society of Mechanical Engin-

eers), 45, 85-96, 105-106, 110, 202n61

ASME v. Hydrolevel, 85-88, 89, 105-106

autonomy

individual, 161

moral, 158

organizational, 160

personal, 158-160

political, 158

professional, 160-162, 168-169

 1168/1183

B&O (Baltimore and Ohio Railway), 21

Barnes, Chester A., 209n7

BART (Bay Area Transit Authority), 78, 123

BER (Board of Ethical Review), 46, 96-98.

See also NSPE

biomechanical engineer, 107

Bhopal, chemical plant, 148

Boisjoly, Roger, 67, 73, 212n20

British Institute of Civil Engineers, 15, 49

Buckley, Fletcher J., 204n12

 Page 238

business ethics, cases, 123-124, 210n18

Bush, Vannevar, 162-163

civil engineers, 15, 194n19, 202n52

in France, 10

and other engineering fields, 22

and Roman builders, 205n18

in the United States, 18-19, 23, 45, 198n23

Camenisch, Paul, 161

Challenger disaster, 43-44, 63-65, 67-68, 123

and communication, 120-121, 125

importance of, 42
 1170/1183

and whistleblowing, 73

chemical engineer, 107-108

chemist, 15

chemistry, 23

code, engineering, 85-86.

See also code of ethics

code of ethics, 18, 217n16

and business, 71, 122

and engineering, 45-60, 142-143

and medicine, 8-9,

and professions, 29-30, 37, 108-109,

111-115

conflict of interest, defined, 83, 101-102

 1171/1183

conscience, 35-36, 97-98

consequentialism, 88, 89-90

corps des pont et chaussés, 10, 11

corps du génie, 10, 193n16

corps of engineers, 10, 193n16

curriculum, engineering, 36-37, 194n18

and practice, 20, 25, 200n39, 206n25

relation to profession, 29

and sciences, 20

See also engineering science

deontology, 88-89, 98-101

 1172/1183

discipline, viii, 72, 112

Dworkin, Gerald, 158, 222n30

Eaton, Amos, 25, 197n9

École Polytechnique, 11, 15, 18-19, 194n22

and Rensselaer, 20

and West Point, 196n3

ECPD (Engineering Council for Professional

Development), 29, 46.

See also ABET

engineer, defined, 115

engineering

curriculum (see curriculum, engineering)

 1173/1183

defined, 32-34

design in, 28, 114

ethics, vii, 16-17, 29-30, 35, 37

imperatives of, 12-15

method (see method, engineering)

science, 23, 38

societies in, 23, 25, 38

See also philosophy of engineering

engines of war, 9

England, 20, 49.

See also British Institute of Civil Engineers

EPA (Environmental Protection Agency), 13

Erie Canal, 21, 25

 1174/1183

ethics

codes of, 16, 36, 39, 142-143

defined, vii, 16

See also consequentialism, deontology,

medical ethics, morality, relativism

fairness, principle of, 207n14

Feibleman, James K., 223n2

Ferguson, Eugene, 12-14

Feynman, Richard, 120-121, 218n2

Florman, Samuel, ix, 205n17

Frankfurt, Harry, 159

G
 1175/1183

GE (General Electric)

and information gap, 219n37

and practical engineering, 25-26

and price-fixing, 62, 65, 68-69, 123

Gewirth, Alan, 214n36

Green, Melvin, 91-92, 97

Greene, Benjamin Franklin, 25

Harvard University, 20, 27

HAZOP (Hazard and Operability), 147-148, 156

Hydrolevel. See ASME v. Hydrolevel

 Page 239

IEEE (Institute for Electrical and Electronic

Engineers)

codes of ethics, 46-47, 48, 109, 200n38,

205n21

not a professional society, 206n27

and software engineers, 203n2

interpretation, 56-58, 115

invention, 6

Jackall, Robert, 75, 122, 141

K
 1177/1183

Kant, Immanuel, 220n5

Kidder, Tracy, 176

Koen, Billy Vaughn, 193n14, 193n26, 194n34,

200n36

knowledge workers, 3

Kranakis, Edna, 192n8

Kultgen, John, 215n2

Ladd, John, 112-113, 224n10

Lawrence Scientific School (Harvard), 27

lawyers, 29, 48, 49, 52, 208n15

and autonomy, 160

and free professions, 22, 162-163

 1178/1183

and microscopic vision, 66

unity of their profession, 24-25

Layton, Edwin, 27, 157, 170, 176

Louis XIV, 10, 14, 18

Lund, Robert, 43-44, 121

as engineer, 51-52, 53-59

and microscopic vision, 63-65, 67-68, 70-71,

208n16

Luegenbiehl, Heinz C., 152, 216n11

Mackenzie, Donald, 172

McDonald's, 14

management, 122-123, 125

 1179/1183

and Challenger disaster, 44, 63-65, 66-67,

120-121

engineering as, 24, 26-27

and risk, 67

and whistleblowing, 76, 77, 79

Martin, Mike, 65, 162, 209n9, 223n6

Massachusetts Institute of Technology, 27

medical ethics, 8-9, 223n7

medicine. See physicians

Meiksins, Peter, 224n16

method, engineering, 12-15, 38

and ethics, 114-115

and risk, 67
 1180/1183

and science, 10-11, 27

microscopic vision, 65-68

Milgram, Stanley, 209n10

Molloy, Michael Peter, 194n22, 198n13

moral blindness, 72

morality, 16-17, 53-54, 55, 107, 112-113. See

also moral rules

moral relativism, 88, 91-98

moral responsibility, 158

moral rules, 91, 99-100

NASA (National Aeronautical and Space

Agency), 59, 67, 125.
 1181/1183

See also Challenger

NSPE (National Society of Professional Engin-

eers), 46-47, 48, 92-98, 110, 205n21, 206n27

Noble, David, 191n2, 199n24, 203n66

Norwich University, 19, 25, 197n4

occupation, 28-29, 35, 49, 50, 112

officieur du génie, 198n13, 201n50

open door policy, 80, 81

Partridge, Alden, 19, 197n4

Pavlovic, K.R., 161

Percival, Thomas, 9, 193n13

 1182/1183

Persia, 204n12

philosophy, defined, ix, 6

philosophy of engineering, viii

philosophy of professions, vii

physicians, 8-9, 24-25, 47, 49, 162-163, 216n15

and free professions, 22

and medical ethics, 8-9

and microscopic vision, 66

and science, 177

prisoner's dilemma, 207n13

@Created by PDF to ePub