Michael Scriven

THOMASJ. BATA LIBRARY

TRENT UNIVERSITY
 Digitized by the Internet Archive
in 2019 with funding from
Kahle/Austin Foundation

https://archive.org/details/reasoningOOOOscri
REASONING
 REASONING
Michael Scriven

Trers? University library

PETERBOROUGH, ONT.

McGRAW-HILL BOOK COMPANY

New York St. Louis San Francisco Auckland Bogota Diisseldorf

Johannesburg London Madrid Mexico Montreal New Delhi Panama Paris
Sao Paulo Singapore Sydney Tokyo Toronto
 -Bern, s 3B

REASONING
Copyright © 1976 by McGraw-Hill, Inc. All rights reserved.
Printed in the United States of America. No part of this publication
may be reproduced, stored in a retrieval system, or transmitted, in any
torm or by any means, electronic, mechanical, photocopying, recording
otherwise, without the prior written permission of the publisher.
67890 DODO 7832109
This book was set in Times Roman by Rocappi, Inc.
The editors were Jean Smith and James R. Belser;
the cover was designed by Albert M. Cetta;
the production supervisor was Charles Hess.
The drawings were done by Vantage Art, Inc.
R. R. Donnelley & Sons Company was printer and binder.

Library of Congress Cataloging in Publication Data

Scriven, Michael.
Reasoning.
Includes index.
1. Reasoning. I. Title.
BC177.S37 160 76-49937
TSBN 0-07-055882-5
 To Mary Anne Warren
whose inspiration and support
caused this book to be
and benefited what it became

330440
 Contents

The Aims of the Book ix

To the Student xi
To the Instructor xiii

Chapter 1
The Nature of Reasoning 1
Chapter 2
Teaching and Learning Reasoning 12
Chapter 3
The Foundations of Argument Analysis 29
Chapter 4
Developing the Concepts of Argument Analysis 54
Chapter 5
Clarifying Meaning 101
vii
vlll CONTENTS

Chapter 6
The Finer Points of Argument Analysis 152
Chapter 7
Special Types of Argument 196
Chapter 8
Extensions and Ramifications 233

Index 241
 The Aims of the Book

1 To improve your skill in analyzing and evaluating arguments and

presentations of the kind you find in everyday discourse (news media, discus¬
sions, advertisements), textbooks, and lectures.
2 To improve your skill in presenting arguments, reports and instruc¬
tions clearly and persuasively.
3 To improve your critical instincts, that is, your immediate judgments
of your attitudes toward the communications and behavior of others and
yourself, so that you consistently approach them with the standards of rea¬
son and the attitude of reasonableness.
4 To improve your knowledge about the facts and arguments relevant
to a large number of important contemporary issues in politics, education,
ethics, and several practical fields.

These are practical aims but not narrow practical aims: the third one, in
particular, is very far-reaching and requires a whole shift of values for most
of us. So this is intended to be a powerful, as well as a practical, book using

ix
X THE AIMS OF THE BOOK

practical, everyday examples of the kind that a citizen, especially a citizen-

student, runs into all the time.
The text starts with a short discussion of reasoning itself, then explores
the details of this approach, and then gets down to practical procedures for
improving one’s reasoning skills. The discussions and examples always begin
at a rather elementary level, but they get into harder material fairly quickly.
If you are, or are hoping to be, a teacher or a lawyer, a scientist or an
executive, what you learn here will be professional skills for you—vital pro¬
fessional skills. But for any citizen, they are also essential skills.
 To the Student

You ve just read what this book tries to do. It does it in some unusual ways,
which are discussed in Chapter 2, as soon as we’ve talked about the general
nature of reasoning in Chapter 1. But two special features of the book need
to be called to your attention immediately. The first is that there is a sub¬
stantial prize fund for the best criticism and/or suggestions for improving
the book—a minimum of $250 in the first year after publication and more if
the book sells well. I believe that consumers not only should, but can, im¬
prove products at least as much as the experts. At least half, and quite likely
all, the awards will go to student suggestions—and their originators will be
acknowledged in later editions.
The second point to keep in mind from the beginning is that the book is
set up to allow you to move fast through any material that seems too easy
for you. Believe me, it’s there only because someone appeared to need it; but
it won’t get in your way if you remember that the “A quizzes,” at the end of
each chapter, are self-testing devices. You can skim any chapter quickly and
then test yourself on the A quiz. If you get all but one or two answers
correct, and can see why you’re wrong on the others (or can give good

xl
xll TO THE STUDENT

reasons for thinking I’m wrong—in which case, write and tell me), then
spend a little time thinking about the C quiz questions and move on to the
next chapter. If you do worse than this, read the chapter carefully and try
the B quiz (the “basic assignment quiz”), which your instructor will grade.
While doing this more careful reading, check your comprehension by writ¬
ing titles for the numbered sections in the space to the right of the number.
(If you look at the last couple of chapters in the text, you’ll see that the
sections already have titles. See if you can do a similar job for the sections in
the early chapters; it’s a good basis for summarizing or outlining them.)
The first chapter is a bit more philosophical than the rest of the book.
The second is more pedagogical than the rest. Be sure to try the third before
you feel you’ve got into the wrong meeting.

Michael Scriven
 To the Instructor

This book provides a basic structure on which you can elaborate or from
which you can diverge in many ways. It can be assigned entirely as back¬
ground reading for all the members of a class, or for those having difficulty
in basic argument analysis skills; or the first two chapters can be assigned for
background reading and the rest used as a text; or the entire book, plus
some extra reading, can make a very full semester’s course.
The simplest procedure for its use is outlined in the preceding section
(“To the Student”). Students should be assigned one (or more) chapters and
asked to do the basic checkup, self-grading quiz at the end (Quiz A) for their
own information. If they pass this (with no more than one or two errors),
they should then do the B quiz for you, and also one or more of the ques¬
tions in the C quiz. If you give them a free choice in C, they can suit their
own interests better, but you’ll have to discuss in class, or make written
comments on, a number of different questions. That’s good if you want to
spend substantial time on that chapter. If you want to move faster, assign a
particular question from the C quiz, and you’ll only have to talk about or
grade a single question. For example, you may want to avoid getting deeply

xlll
xlv TO THE INSTRUCTOR

into questionable issues about the nature of reason (Chapter 1) before get¬
ting into skill training. You could assign just one (or no) question from Quiz
1- C. But it is important to have each student turn in either Quiz 1-B or Quiz
2- B, preferably both, because you need to know whether you’re dealing with
some students with real reading deficiencies, an increasingly serious prob¬
lem. One reason why Chapters 1 and 2 are there is to enable you to diagnose
that problem fast. The B quiz in each chapter is to provide feedback to you
(and possibly also to your students). If you need to be sure they’re not over
their heads, ask them to do the quiz and turn it in. It is designed for either
anonymous or signed use, fast grading, and for either handback or retention.
A number of novel teaching procedures that you may find worth adopt¬
ing are suggested at various places throughout, but especially in Chapter 2.
No doubt you’ll find that some of them don’t suit your style of teaching or
are unsatisfactory in other ways, and you’ll have ideas of your own. I’d
appreciate hearing about both your reactions and your own ideas. In any
case, a glance at Chapter 2 may interest you in the pedagogy of this ap¬
proach, and Chapter 3 will give you a preview of the basic content. Different
levels of student ability can be accommodated in the way described in the
instructions to the student. Various interesting topics are raised in the “ad¬
vanced” quiz on each chapter, enough to hold your better students’ atten¬
tion, in most cases. If those options are still not enough, you will probably
have a dozen others occur to you—don’t forget that I value your suggestions
enough to have set up a prize fund for them. A proportion of the royalties
from this book will always, as they should, go into that fund for rewarding
those who help to improve the book, teachers as well as students.
I believe reasoning is the most worthwhile, though one of the hardest,
subjects to teach. It would be absurd to suppose I alone can teach it half as
well as we can together, with users (students and teachers) helping to do it
better, and the author serving as a coordinator—and as a contributor, too.
No other approach really makes sense. Future editions of the book will be
steadily and, I expect, extensively redone in the light of our joint experience.
No suggestion, however radical, will be dismissed without careful consider¬
ation and, when possible, a careful reply. Those adopted will of course be
acknowledged, whether or not rewarded.
There are a number of fascinating technical issues involved in the prob¬
lem of approaching the teaching of reasoning, whether as an introduction to
other philosophy courses or science courses or composition or law school, or
just as a general-purpose tool for the thinking person. At some point, it may
seem worth discussing these issues at length in a separate teacher’s guide,
but for the moment let me make two points that seem crucial. The first is
that the evidence from educational psychology seems entirely overwhelming
with respect to one point, namely, that so-called “transfer of learning” or
“generalization” always turns out to be less than educators had previously
 TO THE INSTRUCTOR xv

supposed. It is not so long ago that Latin was justified because it was sup¬
posed to make people think more clearly. Then the same was said of mathe¬
matics. Then of science. But the fact is that no significant transfer from the
standard approaches to any of these subjects to the analysis of problems
outside the subject matter they cover has ever (to my knowledge) been dem¬
onstrated, whereas the incredible incompetence of a brilliant mathematician
in handling business affairs, of a physicist expounding on politics or psychol¬
ogy? has been too often authenticated to be treated as aberrant. (The ex¬
treme example of this was a study of teaching children how to tell time
which revealed they couldn’t transfer their mastery with clock faces with one
design of numerals to even modest success when the numerals were
changed.)
It follows that one has to view with great skepticism the very idea that
formal logic is likely to help improve reasoning skill. What it improves is
skill in doing formal logic. The syllogism was probably nearer to reality
(though not to comprehensiveness) than the propositional calculus, but not
near enough to make it useful in handling the average editorial or columnist
today. It’s not incidental that there are ferocious unresolved issues even
within the theory of the syllogism and in interpreting the connectives of the
propositional calculus.
But even if those issues were somehow settled to the satisfaction of the
specialist, the difficulty of using the results surfaces in another way, which
can be put as follows. The use of any calculus to handle problems that
surface in reality (in natural language) involves three steps. Encoding the
original problem into its formalized representation; computing or transform¬
ing it, using the formalized version of the problem; and decoding that, i.e.,
translating it back into the original language or real-world terms. This is
what we do when we use math to solve problems of making change or
designing planes; and this is what we have to do in applying formal logic.
The attraction of the whole procedure lies in the reliability of the transfor¬
mations made within the calculus. But unless the reliability of the encoding
and decoding steps is at least as good as that of the transformation steps, we
may gain nothing from the use of the calculus. And the problem with formal
logic is that the encoding step, particularly, is just about as debatable (in
anything but trivial arguments where there’s no need to use the calculus) as
the assessment of the original argument. A similar argument shows why the
attempt to make a Newtonian science out of psychology has been such a
dismal failure. And it shows why computerization often fails to solve applied
problems.
It is for these reasons that I believe the only way to improve reasoning
skills is by staying very close to real examples. One can use checklists; simple
devices for display and abbreviation; paradigm examples; and handy labels.
xvl TO THE INSTRUCTOR

But a calculus is a diversion one can’t afford; it is a combine-harvester when

we need a carving knife.
One might suppose that the preceding argument establishes a strong
case for the “fallacies approach.” It might, except that the fallacies generally
turn out not to be fallacies—unless one builds into the identification process,
and hence into the labels, all the skills needed for analysis without the taxon¬
omy of fallacies. In that case one has made it a formal approach, and the
encoding (i.e., diagnosing) step has become the tricky one. There are resid¬
ual elements of the fallacies approach in what we do here—we do use some
pejorative names for certain types of error, but only as a way to wind up, or
put a seal on, the analysis. The hard work comes first.
I sometimes think one can best spotlight the gap between formal logic
and real reasoning by pointing out that almost every real argument involves
assumptions, but that, as far as I know, there has never been an even mod¬
erately successful attempt to analyze the concept of an assumption. (The
enthymeme approach is not even moderately successful as a pragmatic de¬
vice.) Without such an analysis, effective criticism of an argument, or an
arguer, is hopelessly crippled. I am sure the analysis I give has faults (and I
hope to hear of them from you), but at least it seems to help in handling
most arguments.
There are a dozen other examples; the status of the “paradoxes of
material (or strict, or formal) implication” is one. These have been called
paradoxes, “paradoxes,” and logical truths. In the context of practical rea¬
son, they are just symptoms of incorrect analysis, no more, no less. It is
merely an error to suggest that a false (or necessary false) proposition im¬
plies every proposition; it’s not paradoxical, quasi-paradoxical, or true. It’s
even false to say that a proposition implies itself. No better example of the
lack of interest in reasoning, as opposed to interest in formal systems, can be
given than the efforts to substitute formal simplicity for practical utility in
the development and assessment of modern symbolic logic. Halfway (though
hardworking) efforts ranging from those of C. I. Lewis to those of Anderson
and Belnap simply do not tackle the need to distinguish between the rela¬
tions (1) “guarantees the truth of,” (2) “is a good reason for believing,” and
(3) “allows the derivation of.” The truth of p guarantees the truth of p, but
it sure isn’t a good reason for believing it. The falsity of p allows the deriva¬
tion of q (in the propositional calculus), but neither guarantees its truth nor
provides a good reason for believing it. This book is about good reasons, not
repetitions or transformations. It’s just a start on what shouldn’t be but is,
almost an untouched subject.
I now turn to some particular issues, logical, terminological, and peda¬
gogical.
On terminology, I have given up the fight on “valid,” which I believe is
now used in the scientific literature as well as in common parlance to cover
 TO THE INSTRUCTOR xvii

the truth of the premises as well as the soundness of the inferences in an

argument (as in “valid reason”). However, I do not support the broader
usage: I simply put the weight on “sound inferences” versus “true premises.”
But I draw the line at “infers” used to mean “implies” and I spend some
time teaching the distinction. I abandon efforts to be strict about symboliz¬
ing the use/mention distinction with quotation marks, since if Quine can’t
get it right in his text. I suspect it’s not an exact distinction. Instead, I use
quotes as they are normally used, for quotation for reference which is not
otherwise obvious, or for “raised eyebrows.” Although I distinguish “deduc¬
tive from “inductive" in the usual way, I do not argue for the distinction as
being sharp, and nothing in the analytical procedures depends on deciding
whether a raw argument is deductive or inductive, since most of them can be
reconstructed either way. There’s essentially no technical jargon in this text;
the instructor will have to judge whether the loss is serious, and can always
add some on a handout sheet or two.
Most of the pedagogical devices are discussed at length in Chapter 2,
since I think students should understand and discuss them, and they make
good subjects for arguments and analysis. Aids for the instructor will be
added in an instructor’s guide if there are later editions, including extra
quizzes, tear-out answer sheets and templates for rapid correction, com¬
ments on the B and C quiz questions, and more references. Two procedures
worth mentioning here are having the students correct one another’s quizzes
at the start of a class in order to get fast feedback to you and to them about
progress; and having them give their quizzes anonymously, which reduces
anxiety about a possible prejudice on your part originating with their poor
early performance. To do this, you simply have them turn in the paper with
a code number (any six-figure number they select) instead of a name, and
you put the papers on the desk in numerical order at the next class, leaving
five minutes early so they can recover their papers in your absence. If my
students do well, I encourage them to start signing their papers so as to build
up a buffer against an unluckly performance on the midterm and final exams
(which have to be signed). And so on; many of you have more ingenious
teaching procedures, I’m sure.
I believe in pretests on the first day of class, and the last example in
Quiz 2-B will serve quite well for that purpose. There are several examples
from the later chapter quizzes that will serve as posttests, and you can then
compare performances to see if you think the course teaches something
useful.
Question 4 on Quiz 2-B is the only one on which we have good, recent
national data; the only (!) flaw is that the “official” answers—indeed, the
alternatives offered—are not persuasive. This is the National Assessment of
Education Progress question, and the official answers are: 4-1, d; 4-2, c; and
4-3, e. The percentages getting the “correct” answers, in the population of
xvlll TO THE INSTRUCTOR

seventeen-year-olds, in 1970-1971, were 45, 24, 29. Young adults (ages

twenty-six to thirty-five), scored about the same (46, 29, 31). The latter
group chose “I don’t know” 8 percent, 3 percent, and 30 percent of the time.
On very simple comprehension tests, with obvious correct answers, these
groups still failed about one-third of the time. Your students will virtually all
do better than this, of course, but the baseline data is useful in counseling
the weaker ones.

ACKNOWLEDGMENTS

Special thanks are due to Bob Ennis and Mary Anne Warren, who have
shared the teaching of my extension course Speed Reasoning; to Bob Cun¬
ningham and Bob Struckman, who used a pilot version of this course at the
University of San Francisco, as did Mary Anne Warren at Sonoma State; to
Brian Holm and Larry Wright, who helped with the Analytical Reasoning
course at the University of Indiana; and to Douglas Gasking, who taught
the first course in logic I ever attended and never forgot. This edition incor¬
porates many benefits from Alan Walworth’s and Brian Holm’s critical
reading. On the production side, many thanks to Judy Simmons for an
outstanding job in preparing the original typescript and to Sondra Harris for
modifying it.

Michael Scriven
 Chapter 1

The Nature of Reasoning

1-1

This first chapter is not intended to improve your skill in reasoning as such,
but to give you some general understanding of what it is and what it isn’t.
Consequently, the basic tests on this chapter and the next, which lay out the
special procedures of this text, are straight tests of your ability to compre¬
hend what you read. And that’s the prerequisite skill in reasoning, so it’s
extremely important that you test yourself on it. Very few people read care¬
fully, and starting in on argument analysis when you haven’t even got clear
what the argument says is a waste of time. Moreover, there’s no sharp line
between reading and reasoning, because reading with understanding re¬
quires that you see at least “obvious” implications of what’s being said,
which means making inferences from what’s being said, which is reasoning.
If you look at reading tests, you’ll see that some of the items are really tests
of reasoning; and so they should be. So the tests on this chapter are not only
tests of your straight reading care, but also tests of this very basic kind of
reasoning. Here’s an example. Suppose you read a long newspaper article

1
2 CHAPTER 1

which states that Prime Minister Gandhi has imposed press censorship in
India. One question to test your reading skill might be: Does freedom of the
press still exist in India (at the time of the article)? The obvious answer is,
no. But notice that the question tests understanding and not mere rote mem¬
ory, because freedom of the press is a technical phrase which does not occur
in the original. You have had to do a bit of translating and comparing of the
language in the question with the language in the original in order to answer
the question. That process of transforming and comparing language is the
most basic kind of reasoning process; and it’s also part of what’s referred to
as understanding the original material. So even basic quizzes at the end of
Chapters 2 and 3 are partially tests of reasoning itself as well as tests of a
prerequisite for reasoning, namely, reading carefully. About one-third of the
answers to questions that I use in a doctoral examination for school princi¬
pals are incorrectly directed because the principal has not read the question
carefully. You may find that a useful by-product of this text is an improved
ability to read carefully—and when it comes to reading contracts of sale or
descriptions of a referendum or motion on which you have to vote, careful
reading is not an exercise in nit-picking, though it may strike you that way at
times when you’re doing the exercises.
So reasoning begins with a careful reading—reading with understand¬
ing. But it goes on to more complicated matters like giving and analyzing
arguments. It is the process of systematically working toward the solution of
a problem, toward the understanding of a phenomenon, toward the truth of
the matter. And that’s when it becomes, not quite the exclusive property, but
the main power, of humanity.

1-2

Reasoning is the only ability that makes it possible for humans to rule the
earth and to ruin it.
All the other alleged distinctions between us and the other life forms on
the planet turn out to be illusory. Of course, they can see and hear and move
and smell better than we can. They have and use languages that do the
simple tasks of language, but not languages subtle and strong enough for
reasoning in any explicit sense. They have tools, but not tools as powerful as
reason. They make tools but not tools as strong as an argument. They have
standards of behavior which they enforce by systems of rewards and punish¬
ment; but they lack systems of law and codes of ethics because their stan¬
dards are not so precisely formulated, so general, so sensitive to the special
case as good codes can be; their systems are not rewritten in the light of
reasons for reform, are not able to be as fast-acting and flexible as a rea¬
soned code can be. Many of our own systems of law are far from fully
THE NATURE OF REASONING 3

reasoned, of course; we are here talking of what is possible, proper, and

sometimes done.

1-3

Reasoning isn t all done with language, but that’s how it’s usually conveyed
and mostly how it’s taught, and certainly how it’s written and thus best
recorded; and, in fact, it’s a help for this text rather than a limitation if we
make a virtue of necessity and emphasize the linguistic approach. For you
don’t want to be able just to reason because reasoning helps you to work out
correct answers for yourself. You also want to be able to persuade others of
those answers, or of the weaknesses in their own answers because you often
need their help or permission, or want to help them. “Let us reason to¬
gether’’ is an invitation to communication, not to independent meditation.

1-4

It’s essential to a democracy that its citizenry be both independently capable

of reasoning about the issues that confront it and able to use the social force
of reason to persuade one another, so as to reach a social solution that can
then be enacted with good support. To help with this second function, we
put a great deal of stress on communicating arguments, on reasoning with
others—much more than most books on logic do. But this isn’t an introduc¬
tion to “rhetoric” in the cynical sense (“mere rhetoric”) or debating skills as
such; its principal commitment is to improving the reader’s ability to give
and identify sound arguments (and to criticize unsound ones). Not everyone
is persuaded by sweet reason, but not everyone is persuaded by being
knocked on the head, either. There are moral advantages—respect for the
right of others to make up their own minds as well as political and legal
ones—for the use of reason.

1-5

But above all, there’s one supreme advantage for the use of reason, privately
or publicly. Reasoning is the best guide we have to the truth. That doesn’t
mean you should never listen to your “inner voice,” your instincts, or to
authorities; it just means that you ought to use reason to decide when to
listen to them. They are sometimes good indicators of truth, sometimes not.
Reason is the only way to tell and even it won’t tell you for sure, when to
listen to instinct or authority. Which is why it’s more important than either.
There’s nothing unreasonable about trusting your own negative judgment
about the honesty of a salesperson, even if you can’t formulate exactly why
4 CHAPTER 1

you’re dubious. But there would be something unreasonable about continu¬

ing to trust that instinctive judgment if it turns out to be wrong most of the
time. Reason is the overall best indicator of the way to the truth, which is
quite different from saying that having explicit reasons, being able to give a
complete sound argument, is always necessary to justify a judgment or an
action. You, or your teachers, or experience, may well have trained your
instincts to be a good guide to the truth.

1-6

How can one be sure that reason is the best guide to truth? Isn’t that just a
dogma? We can be sure simply because anything that turns out to be a
reliable guide to truth—for example, the instincts of an experienced auto
mechanic—is what reason tells you to follow unless something better turns
up. Reason operates on the principle of “If you can’t lick them, join them.”
If you can't find an explicit and virtually bulletproof argument that will
settle an issue, fall back on whatever you have that looks better than noth¬
ing; it’s still a reason. In short, of course, you have to use a little reason to
decide whether you have something else—like an intuition, a clue, or an
expert—that is better than nothing. So, though you’re not always in the
position of having an explicitly statable reason or argument for your deci¬
sion, action, or conclusion, you should always at least have an indirect argu¬
ment for it, that is, an argument (a reason) for heeding whatever you’re
using instead of a direct argument. Why? Because you want and need the
truth, whether it’s about diets, insurance policies, drugs, the use of violence,
or long-distance cruising in sailboats, and that’s all that reason is—the best
indicator of truth.
Suppose someone claims to have come up with a rival to reason, a new
magical approach to knowledge. Suppose this person turns out to be able to
predict earthquakes or diagnose illness or prescribe therapy better than all
the experts and scientists we have at the time. What does reason tell you to
do? It demands that you accept the new expert. How did you come to this
decision? By comparing the track records of the old experts with the record
of the new prophet. Which was, of course, the reasonable thing to do. You
didn t abandon reason. You just used it to abandon some old prophets, some
superceded indicators of the truth. Those who offer alternatives to reason
are either mere hucksters, mere claimants to the throne, or there’s a case to
be made for them; and of course, that case is an appeal to reason. Reason is
always the ultimate court of appeal—which is not to say that explicit direct
reasoning is always the best basis for judgment.
 THE NATURE OF REASONING
5

1-7

You won t find anything in this book that endorses the old prophets, except
where their track record is still the best. It’s nearly thirty years since I
founded the first university-based society for the study of extrasensory per¬
ception in the Southern Hemisphere, and I have found the rigidity of estab¬
lished science to be so extreme as to refute every claim for the objectivity of
the scientific method as advertised by most of those same scientists in their
introductory lectures. Closed-mindedness and prejudice are about as ram¬
pant among scientists as anywhere else. True scientific method is open-
minded. self-critical, flexible. Scientists are, in short, not as reasonable as
they would like to think themselves. The great scientists are often true excep¬
tions, they are nearly always attacked by their colleagues for their revolu¬
tionary ideas, not by using the standards of reason, but just by appealing to
the prejudices then current. Being reasonable takes great skill and great
sensitivity to the difference between “well-supported” and “widely ac¬
cepted. It also takes courage, because it seldom corresponds to being popu¬
lar.

1-8

So don’t confuse reasoning with calculating or measuring or ignoring emotions

or appealing to authority. Sometimes it involves these things and sometimes it
rejects them. Basically, it involves working out, as carefully as you can, the
best answer you can find, using whatever has value for that purpose and
resisting the temptations of the impostors. Not easy at all, but there’s no
alternative way to the truth, no short cuts until they are certified by reason.
And even then, you have to keep recertifying them, not letting them rest on
their laurels.

1-9

What can this book do to help? It can teach you a basic strategy for the
analysis of arguments, and hence for their presentation. It can teach you a
dozen special tactics for handling traps that are easy to fall into (fallacies), or
special situations in reasoning (like debate). It can teach you some general
skills that apply to several interesting areas of reasoning, particularly scien¬
tific and moral reasoning. It can, incidentally, teach you a great deal about
a number of controversial issues. But above all, it attempts to convey to you
an attitude, an approach, a skill—the skill of reasoning, the attitude of rea¬
sonableness. Neither comes easily; the first requires much practice, and the
second is a way of life and requires a lifetime of practice. If you can read
6 CHAPTER 1

these words, you can manage both, and nothing is more important to you
and to your neighbors on this planet.

GENERAL INSTRUCTIONS FOR QUIZZES

The A quiz is for self-testing and self-teaching. Use it as follows:

1 Take a plain sheet of paper and cover the whole of the A quiz: this
is the “slider” sheet.
2 Then move the sheet down, a line at a time, until you’ve seen the
whole of the first question, and before even looking at the alternative possi¬
ble answers (most of the early A quizzes are multiple-choice), ask yourself
whether you know the answer. If you do, say it to yourself or out loud if
you’re by yourself; better still, write it on the slider sheet. That is, make
yourself formulate it completely; don’t just settle for “having a feeling,” or
thinking you know it, or “having a general idea” of what it is. It’s essential
to write your answer on the slider sheet if you’re not dead sure it’s correct
because making yourself do that will often make you conscious of weak¬
nesses in the answer.
3 If you’re not sure what the answer is, move the slider down till you
can see the first answer. Is that the correct answer? (Or is it one correct
answer?) If you think so, circle the letter a that comes before it. You do not
need to see the other answers to decide this. If you're not sure, slide the
paper down a bit further till you can see the next answer. Think about it. Is
it a—or the—right answer? If so, circle the letter b that identifies it. Notice
that none, all, several, or just one of the alternatives may be correct; it’s a
weakness in the usual multiple-choice tests that they are unrealistic in that
they give you immediately a set of alternative answers without asking you to
try to think of them first and that they tell you that one of these is correct
and only one. That doesn’t correspond to reality at all. In reality, you usu¬
ally have to think up possible answers yourself, and no one is around to tell
you whether you’ve hit on a good one, or several good ones, or none. Some¬
times looking or asking around, or reading ahead a little, will give you some
further clues. Our procedure is pretty close to matching that real situation. If
the right answer is none of those spelled out, then circle the last alternative
(“Other”) and write in the correct answer.

Now, pull the slider down a bit further and you’ll come to the correct
answer(s) and (usually) a short explanation of why it is, or they are, correct
and the others aren’t. If the explanation isn’t enough, turn back in the chap¬
ter to where this point is discussed and reread that part. If that isn’t enough,
complain to the instructor or to me and we’ll change the question or add to
the text. Keep a copy of your complaints (and any other suggestions you
have for improving the text) so that you can submit them as an entry in the
Reasoning Prize competition.
THE NATURE OF REASONING 7

The B quiz is an assignment quiz. It's usually to be handed in, though it

can also be corrected in class, either by yourself or by someone else, while
the instructor reads out and discusses the correct answers. Handle this test in
the same way as the A quiz if you want to get the most out of it, but this
time you'll be handing in the slider sheet, so you have to write the answers
on it. Of course, you can correct them as and when the uncovering of more
of the quiz leads you to a better answer. But you’ll learn a great deal more if
you try to answer before looking ahead, and a reasonable person will there¬
fore exercise or develop the strength of mind to do it this way.
The C quiz is an advanced assignment quiz (or it may be used as a set of
questions for enriching the classroom discussion or your own reflections). It
will sometimes open up some of the deeper philosophical issues connected to
the topics in the chapter, or it may give more difficult examples on the same
issues as the B quiz.

QUIZ 1-A

1 “Reasoning is a particular skill in using thought and language, and it is profi¬

ciency in this skill, not the mere possession of language, which distinguishes the
human species from the others on this planet.”
a True (according to this text)
b False (according to this text)
c Other (explain)

Answer

Other species do possess very highly developed languages (the porpoises, most nota¬
bly), but so far, they have not appeared to enable the species to perform even mod¬
erately complex reasoning tasks such as humans perform by the age of ten. If they do
pass on information to one another on which they act (e.g., get out of a trap), as killer
whales appear to do, one might call this a very simple kind of reasoning. But it’s
nothing like what is involved in, say, working out how a complicated toy operates
when the instructions on the box aren’t very clear—easy for most nine-year olds. So
the quote seems correct and the first answer (alternative a) is the right one. Note that
wolves systematically solve problems, for example, how to spring a cunningly laid
trap, and it’s useless to insist that they aren’t reasoning when doing it. We’re merely
better at reasoning on most problems. There are many we can solve that no other
species can, and a major part of our skill is the capacity to reason with a linguistic
representation of the problem. It’s a difference of degree, not of kind, but it’s a very
large difference of degree.

2 Reasoning has the function of helping us achieve the answers to our own individ¬
ual problems, like which career to try for, which automobile to buy. But it also
has another function, which is where the use of publicly accessible (that is, spo¬
ken or written) language to set out reasons becomes essential. That function
involves:
8 CHAPTER 1

a Reflection on the individual’s situation in the Universe

b Using arithmetic
c Writing poetry
d Persuading others by argument
e None of the above, but rather (fill in your preferred answer):

Answer
Meditation about our place in the universe is traditionally associated with Eastern
religions and is often highly visual or feeling-oriented rather than verbal. There are
speculations by astronomers on the same subject which do require language, but the
example of the Eastern philosopher or sage shows that language isn’t essential for
option a, which is what the question asks. So that alternative isn’t right.
Arithmetic is being used when we count the number of items in a box and
multiply that by the number of boxes to get the total number of items in a shipment.
All of that can be done in your head without any “external” use of language, so b is
wrong.
Writing poetry requires the external use of language, but it’s not necessary to
set out reasons.
Persuading others can be done with a gun, which is nonlinguistic, but when
done with arguments, it does involve the “publicly accessible use of language to set
out reasons,” so that’s the right answer.

3 Dictators with complete power would not have to bother to persuade others, by
reasoning, of the truth of any conclusions, since they could force compliance.
More open forms of government, such as democracies, must, by contrast, rely
heavily on:
a Legislation
b Sympathetic understanding of the needs of the people
c Public proclamations
d Public presentation of arguments
e None of the above, but instead:

Answer

The first and third answers could just as well originate from a dictator “laying down
the law” as from a democracy. Elence, they are not procedures that a “more open”
form of government must use.
The second alternative is certainly a nice ideal for democracies and a useful
support for democratic governments. But it’s beside the point of the “contrast” re¬
quested in this question, which is obviously about the use of rational persuasion.
(True statements aren’t necessarily right answers.) The appropriate response is the
fourth one, and the point is extremely important since it establishes a link between
our political commitments and our educational obligations. The lack of curriculum
content (in the high school particularly, but also in most colleges) devoted to reason¬
ing about current political issues (as opposed to reasoning about old answers to old
problems—history—or reciting creeds and constructs without real understanding of
them) shows that we have not yet appreciated this point.
 THE NATURE OF REASONING 9

4 The process of systematically and skillfully considering and formulating argu¬

ments is a necessary step that must precede coming to any belief that can be
called:
a Reasonable
b Rational
c The result of reasoning
d The result of thinking
e Not intuitive

Answer

It’s often perfectly reasonable to believe that the temperature is above the freezing
point just from the feel of the air, without going through any process of argument. So
a won’t do, and b amounts to much the same, since the term “reasonable” means
much the same as “rational,” though one might say that the former is a little more
warm-blooded than the latter. One talks of a reasonable person, meaning someone
who will act reasonably, listen to reason, be moderate when moderation is appropri¬
ate, and so on. A rational person is someone with the capacity to reason and who
uses that capacity, wherever it may lead. But the difference is usually insignificant.
The process described in the question is the reasoning process and hence the
third option is correct, but the fourth describes a much wider class of beliefs includ¬
ing the result of irrational and fantastic thought processes, so it’s wrong.
What’s the opposite of intuitive thinking? One would normally say “delibera¬
tive,” and that’s pretty much what is described in the question, so the fifth answer
might also be checked. But perception or memory or believing an authority fall in
between, and if you thought of one of them, you should not have checked e.

Note Don’t be lulled into thinking that one or more of the answers
given must be correct. Sometimes you will have to fill in one of your own.
But by now you have an idea of what we’re looking for in your thinking.
You shouldn’t have much difficulty in Quiz 1-B. You may also find that your
instructor has better answers than I do for some of these quizzes. This isn’t
basic mathematics, where almost everything is cut and dried; it’s something
more important—more widely useful and more fundamental to our think¬
ing.

QUIZ 1-B

Put a name or a code number (depending on your instructor’s request) on your slider
sheet, which you should head “1-B.” (The code number can be any six-figure number
you think of—but make a note of it in your text.)

1 Observations of chimpanzee groups in Tanzania and at Yerkes make it clear they

have a highly developed social code. In their relation to this code, they are:
a No different from humans
10 CHAPTER 1

b Different in that the code can’t be taught as easily as if they had expressed it
in language
c Different in that they can’t change it quickly and advantageously as a result of
consultation, discussion of new circumstances, and promulgation
d Different in that no system of penalties for infractions of the code is known to
members of the group
e None of the above, but

2 “You can’t reason without the use of language.” Does the text commit itself to
this view?
a Yes
b No
c Other (explain)

3 “Verbal persuasion is another name for public reasoning.”

a Yes
b No (give an example that supports your answer)
c Other (explain)

4 The crucial general contrasts that are conceded to be legitimate in Chapter 1 are
between:
a Reason and instinct
b Reason and cause
c Reason and unreason
d Reason and feelings
e Reason and creativity

5 There is no arbitrariness in adherence to the standards of reason because:

a There is no alternative
b The alternatives are unreasonable
c Adherence is required only if you’re looking for the truth, and whatever turns
out to be a guide to that is what reason requires you to use
d Alternative approaches are legitimate only insofar as they are approaches to
something besides truth
e Other (explain)

6 “There’s a sharp line between reasoning and reading.”

a Yes
b No
c Other (explain)

QUIZ 1-C

1 How would you tell whether dolphins or giant insects on an alien planet actually
have a language in the same sense as we do?

2 Do sophisticated computers have a language in that sense?

THE NATURE OF REASONING 11

3 Computers calculate (don’t they?). Do they reason?

4 a Could one have any good reasons for complaint about government by a to¬
tally rational dictator?
b Could one express them?

5 There are times when one’s instincts seem to point in a different direction from
one’s reason. How might you describe such a situation other than as a clear
counterexample to the claim that instinct and reason are compatible?

6 What is the difference between “counterintuitive” and “irrational”?

 Chapter 2

Teaching and Learning

Reasoning

In a textbook on reasoning, there ought to be some signs that the author has
done some reasoning about textbooks on reasoning. You’ve already done
one or more quizzes which illustrate some slightly novel approaches. Let’s
look at some general aims and special procedures of this text that you may
find interesting.
Remember, this chapter has three functions: It introduces you to some
unusual features and aims of the book that might otherwise seem puzzling;
it will teach you something about the issues and controversies over various
approaches to grading and testing, useful knowledge to have as you go on in
higher education; and it provides a basis for testing your ability for “reading
with understanding,” or “basic reasoning,” as we’ve called it.

2-1

The first unusual feature of the text is the very fact that it does treat “read¬
ing with understanding” as the basic kind of reasoning. One often encoun¬
ters students these days who have never been obliged to read meticulously, a

12
TEACHING AND LEARNING REASONING 13

process that involves thinking out the significance and implications of what
one reads. Such students will have understandable difficulty with a conven¬
tional logic or rhetoric course just because they aren’t reading carefully. It’s
important tor the instructor to be able to diagnose this weakness as the cause
of their difficulties, rather than whatever technical content the usual course
incorporates. So the quizzes on Chapters 1 and 2 are essential diagnostic
tools for an instructor.

2-2

It’s unusual for texts to discuss what texts (and courses) in the relevant
subject should do. But students should be encouraged to take a critically
constructive attitude toward texts as well as their content, if their learning is
to have practical consequences. Of course, they always do this to some ex¬
tent—they’re quite willing to pass a general judgment on whether a text is
good, bad, or indifferent. But writing a good text is the result of applied
science, or a practical art. As we’ll see in this chapter, there are many weak¬
nesses in most approaches to teaching reasoning, at least some of which
we’re making an effort to avoid here. I’m sure we’ll make plenty of other
errors, which is why I’m offering $250 in prizes for the best criticisms and
suggestions received in the first year of publication. This chapter should help
you to think critically (to reason better) about texts and courses in general,
and hence should have some value beyond this course; but at least it should
make you a more sophisticated critic and user of this book.

2-3

The first chapter may have seemed a little abstract to you. It’s really not so.
Here are some of its practical implications in terms of the contents and
arrangement of this book. First, the emphasis on communication, on the
crucial importance of the social activity of reasoning, means that we can
scarcely afford any jargon at all. For you can’t restrict yourself socially to
having serious conversations or arguments only with people who happen to
have read this text or, for that matter, any other logic or rhetoric text. We
will use, and try to use carefully, the substantial slice of the English vocabu¬
lary that already exists for the assessment and expression of argument, of
reasoning. We’ll use technical terms or symbols very rarely, and when we do,
it will be just as a shorthand that you can easily translate straight back into
English as you read, and on which you will not be tested.

2-4

Since a main aim of reasoning is to reason with others like yourself, your
success shouldn’t be judged solely by an instructor who, after years of expe-
14 CHAPTER 2

rience, is probably able to decipher your meaning even when it’s very poorly
expressed. Hence, at the very least, you should have another student, per¬
haps a roommate or friend who isn’t taking the class, grade some of your
assignments (not just the multiple-choice ones, but some of the more ad¬
vanced ones). Your instructor may want to have you do this as an assign¬
ment, as follows: You may be asked to (1) do one of the exercises, (2) get it
graded by someone not in the class, with comments explaining the grade, (3)
add your own comments in reply, and (4) submit all this to the instructor,
who will determine a grade “for the record.” Or you may do this in class,
perhaps at the beginning of a class when everyone has brought in some
assignments; in that situation, the instructor may ask you to pass your as¬
signment to the person behind you or two seats to your left (far enough away
so you aren’t tempted to “help” with the grading), each of you then grading
someone else’s answers, with a few comments, and signing the grade. (It
doesn’t matter whether you did the same assignment or not; in fact, it’s
sometimes more realistic if you didn’t.) Then the assignments might be re¬
turned to the author for reflection and (written) reaction, and then handed
in. The graders of an assignment will then be graded individually on (1) how
well they graded it, as well as on (2) how well they did their own assign¬
ments, and on how well they reacted to the grades and comments on their
own work.
There are two reasons for doing this kind of exercise. First, it makes
each of you remember that you’re ultimately writing for your peers, not for
a professor. Your fellow students have to be able both to understand your
presentations and criticisms (your writing, for openers!), and to be persuaded
by them. Second, it makes you begin to think about the practical aspects of
evaluating the arguments and criticisms of other real people, not just exer¬
cises in a book. You’ll find you will be considerably more careful to explain
your points fully when you remember that someone else in the class is going
to read your argument. And many of you will find that your appreciation of
the task and of the comments of the instructor will improve considerably.
It’s easy to feel that “It’s obvious what I meant, even if I didn’t say it exactly
right,” when there’s just an instructor to argue with; but if one of your
classmates also couldn’t see what you “meant,” you have to think more
seriously about expressing your meaning better.
After a fair amount of this kind of practice, the instructor may feel
confidence in the class’s ability to grade assignments and may do this once in
a while without reviewing all the papers; that way, each student can get
almost instant feedback on the latest assignment before it is discussed in
class. This will often greatly increase your appreciation of the discussion.
If you happen to be using this text on your own, remember that you can
still get much of the benefit of the peer-grading system just described by
having someone else in your family, a friend, a coworker, or a roommate
TEACHING AND LEARNING REASONING 15

grade your assignments for you every now and again. It can be interesting
tor them because most of the topics of the arguments are pretty interesting
in themselves, and perhaps also because it’s interesting to play the role of the
teacher (which is not so easy to do well as one might think).

2-5

A fatal flaw about most texts in this area is that they only give you examples
for criticism that deserve criticism. That seems sensible enough, but it’s a
mistake. What is the skill we are trying to improve? It is not just the capacity
to distinguish, without help, between defective arguments and sound argu¬
ments. Long ago I discovered that a number of students in a reasoning
course, who had become quite good at critiquing all the arguments I set
them, ran into tremendous trouble when I asked them to go out and find a
defective argument in a current newspaper, magazine, or book. About a
quarter of them couldn’t find one at all because I had spent no time on
teaching the skill of discriminating between good and bad arguments; I’d
spent all the time on teaching them how to criticize arguments which they
already knew were bad in one way or another, which meant that all they had
to do was select the right kind of error out of the list they knew about. In
short. I'd converted the task of choosing between good and bad into the task
of choosing among sixteen kinds of bad, which isn’t the realistic task at all.
That’s very like the basic flaw in the usual type of multiple-choice exam
where you only have to pick the best answer of that set, rather than think up
the right answer, without a list of candidates, as in real life. Letters to the
editor in newspapers don’t come marked “Unsound—decide in which way.”
But exercises in logic books are usually like that—they present sets that you
know are unsound in one way or another.
In fact, the situation is usually worse than that, because usually the
exercises come at the end of a chapter on one particular kind of error, so the
reader has to apply only what has just been learned. When I first caught on
to this, I tried interspersing a few examples of good arguments amongst
exercises at the end of a lesson. The results were amazing; people discovered
the most astonishing and nonexistent errors in them. Desperate to answer
the (unwritten) call to condemn, they laid charges that were patently absurd
and spent disproportionate time to establish them. Obviously the “hidden
agenda” of the course were more influential than what I thought I was
teaching. Something had to be done.
In this book we use “scrambled” exercises as one of three ways to
handle this and some other problems. Although there are always some exer¬
cises at the end of a chapter to give you a chance to apply what you have
just learned, there may be some examples which have nothing wrong with
them (as far as I can tell), some which illustrate points from earlier chapters.
16 CHAPTER 2

and even some which deal with topics which are as yet undiscussed. I don't
say there will be such examples in each set, because that would tell you more
than you’d know in real life. I just say there may be some like that. So don’t
be sucked in by the context; treat each just like examples of argument that
you’d see outside of a logic/rhetoric/reasoning text.

2-6

Moreover, the “progress” questions—as opposed to the “practice” ones—

usually don’t have any cues at the beginning. Students often say, “Well, you
have to tell us what you want us to do with these examples.” Nobody tells
you that when you listen to the radio commentators or read a political
circular just before an election. You have to treat them “as is appropriate”;
something along the spectrum from tossing them into the literal or mental
garbage can to saying “Right on.” To make sure which treatment is called
for, you’ll usually have to do some analyzing, and in a text on reasoning,
you’ll have to set that analysis out as your justification for your conclusion,
just as you’d have to in real life if you were trying to persuade someone in
discussion or to write a report on a proposal. So the instructions before the
progress questions are minimal; maybe nothing at all, maybe “Comment as
seems appropriate.” That’s part of your task—deciding what’s appropriate.
In fact, you ought to be able to withstand the pressure of misleading instruc¬
tions. If you are told to criticize a perfectly good argument, are you confi¬
dent enough in your judgment to be able to withstand that pressure? You
should be, since you’ll often encounter hints or instructions that are at least
that misleading in real contexts. So watch out for deliberately misleading
“hints” later in the book.

2-7

Apart from scrambled and uncued or miscued exercises, our third move
toward realism involves requests for “field research”; you will be asked to go
out and find examples, perhaps of a certain type, to criticize. There’s no
question that this proves to be one of the hardest things for most students.
But of course it is also the most valuable, because it is nearest to the real use
of reasoning.

2-8

When talking about test questions, one kind was probably most puzzling to
you. Why would we ever use examples which bear on material not yet
covered? How else would we find out that you can already handle that
material and hence that we can drop, or drastically cut, that section? And
TEACHING AND LEARNING REASONING 17

how better can we persuade you that there is something useful in that later
chapter than by having you find yourself in difficulties when trying the type
of example it covers? Texts and courses need to offer these opportunities to
justify or improve themselves. The extreme case of this—hard for instructors
(and textbook writers) to accept—is the “challenge” system, where students
can take the final exam(s) without taking the course and get credit for the
course if they pass them. Taking an exam made of B and C quizzes, before
sitting through the course or before reading the text, should either result in
your passing and getting the credits, or failing and seeing that the course or
text (which gets most students through the exams) has some justification.

2-9

There’s something else this text does that isn’t often done to the same extent,
which is to give some examples of typical, usually real, bad attempts at
analyzing arguments and other statements. Texts are full of neat examples of
the correct way to do the things they’re teaching, but one doesn't always
catch on to that by being inspired by the instructor’s performance. It helps
quite a bit to look at how people blow it, not the people who commit falla¬
cies, but the people (you) who are trying to prove that someone else commit¬
ted a fallacy. You’ll thus get some practice in evaluating other students’
work even if the peer-grading procedure suggested earlier isn’t used.
There’s a school of thought in educational technology which believes
that good texts should never exhibit examples of bad work because students
tend to remember errors just about as well as they remember the right
approach. There are times when that may be correct (in writing sophisti¬
cated programmed texts, for example), but students often say that no single
experience was as illuminating to them, when perplexed about their lack of
progress, as that of seeing criticism of an attempt by another person at the
same problem.

2-10

In commenting on the examples of good and bad work illustrating this book,
an “absolute” standard of grading is used, either Pass/Not Pass (as on the
basic multiple-choice quizzes) or, with the harder examples requiring ex¬
tended analysis, A-F. This marking system simply means that a correct and
completely adequate answer for practical purposes gets an A; one with mi¬
nor blemishes, a B; one that does an adequate basic job but misses some
important points, a C; a marginal one—definitely unsatisfactory, with seri¬
ous omissions or errors but still getting hold of a reasonable proportion of
the important points, a D; and anything below that, an F or Not Pass. Those
standards are the realistic ones, the informative ones, and they have nothing
18 CHAPTER 2

to do with age or effort or background or IQ or how other people do on the

same question. Of course, there will be some subjectivity on the instructor’s
or author’s part in deciding between an A and B, and so on, but the point is
that I’m not supporting the use of “curves” or “norms” or any other “relativ-
izing” of the grade given to you during the course, since that just misleads
you; it isn’t the useful, relevant information for you. You can be the best in
the class, but still not be doing the job well and should only get a C; that’ll
tell you there’s some way to go before you’ve mastered reasoning skills. And
you could be the worst in the class and still deserve nothing less than an A.
Remember that an A refers to practical standards; very often, a few lines of
answer is all it takes to get the A—it’s just a matter of hitting the right nail
on the head, and showing that you have.
On easy questions, like the basic quizzes, we just talk about passing or
not passing; there’s no real way to distinguish on an A-F scale there. We
don’t mislead you into supposing you’re doing A work just because you pass
a basic quiz; but we also don’t downgrade you to a C, when you couldn’t do
any more than hit the right answer. Sometimes we use the fancier kind of
multiple-choice questions where all, none, or some of the alternatives given
could be correct, and you may have to fill in the right answer; on some of
those the quality of work can be high enough to get you into the A-F scale.
Now none of this has much to do with the final grades that your in¬
structor “gives you” (i.e., gives the registrar in your name) on the course.
The instructor might then take into consideration the age or background
level of the students, the amount of time (load) they have for the course, the
amount of personal help they get, and so on. In particular, it might be
appropriate to take into account the standards of the college or department.
Why should you be penalized for taking this important course when you can
take some Mickey Mouse course somewhere else and get a more-or-less
guaranteed A? And, for that matter, why should your instructor, and the
department which is offering this crucial course, be penalized by the admin¬
istration (as they normally would be) if the enrollment in this course drops
when it gets around that the standards are tough? (They may be penalized
by having the number of positions cut, which may mean your instructor’s
job.) These considerations have led to “grade inflation” in the past few years
that has made grades from some colleges and/or departments meaningless.
But this situation can be changed. The system I use in my courses
simply puts the responsibility for controlling grade inflation where it be¬
longs—on any administration which uses enrollment instead of quality of
content and teaching as the basis for rewarding and punishing departments
and teachers. Throughout the course, while dealing with the students, I
allocate grades as devices for passing back useful information to them, using
the absolute standards just described. When it comes to sending in grades to
the registrar, I merely adjust them upward (if necessary) so that they have
TEACHING AND LEARNING REASONING 19

the same average as the college average. To be a little more exact, I consider
two other matters. First, I make the adjustment only when I am convinced
that the class members have worked as hard as it’s reasonable to expect
them to work, given their overall load. (I find that they usually work harder
than that, since they’re quickly convinced of the importance of the course.)
Second, if I can get a look at their entering grade-point average, I can make
sure that if they are already an above-average group, they will have their
grades kept above the college average. It goes without saying that where, as
sometimes happens, they perform better on the absolute standards than they
would on the corrected-average basis, their grades are not down-rated.
Now one can go even further than this. In what I call the “ruthless
counterattack’’ response to grade inflation, one would adjust grades to the
average of the highest-grading department (or large course) in the college,
unless there is good evidence that that department or course is either enroll¬
ing geniuses or out-teaching everyone else on campus.
Of course, it is the administration or the joint efforts of the faculty and
administration that ought to be controlling runaway grade inflation. The
easiest way is to get faculty support for some very simple limitations, such as
a maximum of roughly one-third A’s, or two-thirds A’s and B’s combined, in
courses with twenty or more enrolled, unless they can qualify as an excep¬
tion on the basis of exceptionally talented students or exceptional teaching
effectiveness.
In order that students should not be unduly discouraged by their early
grades, especially their pretest grades, on the absolute scale, I usually issue a
correction sheet showing the general picture I expect, based on prior experi¬
ence. This sheet might look, in part, as follows:

If you got a D+ on this test, you can expect that putting in a reasonable
amount of work on the course will at least lead you to improve to about the C +
or B — level by the end of the term, and the grade-adjustment process will then
mean that you get a B to B + as your official term grade. You have a good
chance of doing better than this, especially if you work hard; you are unlikely to
do worse except by goofing off.

And so on with other grades.

Someone who gets an F— on the pretest (or a clear Not Pass on the
basic quizzes on Chapters 1 and 2), might, however, deserve some special
counseling. Such students may be in over their heads or hyperanxious, and
in some cases should be counseled out of the course.

2-11

A feature of some importance is that I’ve tried pretty hard to use examples
for discussion which are interesting in themselves, so that (1) you learn
20 CHAPTER 2

something about a number of worthwhile topics “on the side”; for example,
this chapter takes up issues about teaching and grading student work (and
teachers); (2) you have a sense of applying your developing reasoning skills
to cases which matter.
Naturally this means that the topics will be more controversial and
perhaps a little harder than the artificial ones we could easily make up to
illustrate the points. That’s better than becoming a master of trivial exam¬
ples and having no idea how to apply your skills to real examples, to prob¬
lems where your career or health or society is at stake. Reasonableness
hardly gets a real test if it isn’t tested under pressure, which means emotional
pressure, which means controversial issues.
For much the same reason, I occasionally use terms which, while
they’re not technical jargon, won’t be in every reader’s vocabulary. You’ll be
able to pick up their sense from the context and thus increase your reading
vocabulary a bit even if, occasionally, you have to look them up in a diction¬
ary. They will be terms worth knowing for other purposes than reading this
book.

2-12

There are several modest innovations in the approach to testing. You’ve

already noticed that the quizzes, the A quizzes especially, are designed to do
some teaching as well as to help you and the instructor find out how thor¬
oughly you understand the material. This is one illustration of a general
theme of the text, namely that testing and evaluation—in particular, critical
evaluation—should be done and regarded as a benefit, not an assault. It’s
something you have to learn to value—not necessarily enjoy—if you value
anything else. For it is an essential means to improvement. One of the hard¬
est lessons for some students to learn in a course on reasoning is that they
are often wrong and have to be told so. It’s easy to see how you can be
wrong in an elementary calculus course; in the first place, you’ve never had
the subject before and can’t be blamed for not learning it instantly; in the
second place, it’s hard; in the third place, the teacher has a degree in math
and that proves special skill; in the fourth place, one can prove what the
answer is by an impeccable and watertight proof.
Contrast a calculus course with a philosophy course, or a course like
this which involves the discussion of controversial issues on which you prob¬
ably have some strong opinions. You don’t like to admit your reasoning
could be at fault on issues where you already have committed yourself. You
knowr very well that expert opinion is divided on such matters. Your instruc¬
tors’ degrees—probably in philosophy or the humanities—don’t show
they’re almost certain to be right on such matters. And, indeed, you’re not
even convinced that there are any objective standards in the area. But, be-
TEACHING AND LEARNING REASONING 21

fore you start getting tough with the instructor, remember two things. First,
you may find that there is some subtle skill involved here, and if you stay
around a while, you may begin to see that your C —, when you expected an
A, was richly deserved, that you’ve found out why, and that you won’t make
that mistake again.
Second, and this is a plea from the heart, don’t forget that complaining
about your grade can have (and often does have two) extremely bad results
for you. It puts pressure on for grade inflation (it’s much easier to give
everyone an A than have to hassle about grades). And, if everyone gets an
A, most people, probably including you, are not being informed honestly
about the quality of their work, which is the most precious information you
can get from an instructor. Remember that the “in-course” grades are for
your benefit; don’t make them meaningless. (It’s another question whether
you should argue for restandardization of the grade before it goes to the
registrar.)
Another very bad result of complaining bitterly about your grade is that
it increases the pressure on your instructor to switch to teaching you some¬
thing that’s more like mathematics than practical reasoning. Most introduc¬
tory logic courses in the best universities today are simply introductions to
symbolic logic, which is essentially a branch of math. That puts the instruc¬
tors in a safe position because it’s a subject which they certainly know more
about than you and it’s one in which the answers can be proven right.
There’s just one catch. It has almost no known practical use.
So be nice to your instructor in this course—don’t force her or him to
give up on honest evaluation of student work and on practical courses.
Reasoning is something we’ve all done for years, and it is an ego shock to
accept the possibility you’re not as good at it as you thought. But it has been
my experience that most professors really are a good deal better at it than
most students, even though they will still be wrong at times. Stay with it;
you may learn some things about yourself and about how to stay cool when
criticized, as well as about how to reason better.

2-13

Remember, too, that there are no ideal teachers, there are only ideal combi¬
nations of teacher (or text) and student; the merits of texts or teachers
depend as much on you as on them. The quizzes here, for example, are only
potentially valuable; they’re only actually valuable if you use them to your
best advantage, which doesn’t mean with the least effort. The same holds for
your instructor. The question is not what he or she succeeds in stuffing into
you; it is what you can get from the teacher. You have to work at that. An
instructor (or text, or subject) may strike you at first as dull or as turning you
off because of some mannerisms or attitudes. The next question is whether
22 CHAPTER 2

you have the ability to overcome that initial reaction. You can usually bet
that your instructor (or the text) has some things to teach you that are worth
learning. Are you incapable of absorbing them? Sure, you’re entitled to try
to change the instructor by direct request, or by the student-evaluation sys¬
tem, or by complaint. (You’re being directly asked to change this text for the
better and rewarded for doing it.) But there’s also the question of whether
that’s the only direction where you should make an effort. Part of the effort
may better be put into improving your adaptability as a student: your ca¬
pacity to learn from books without five-color illustrations, to learn about
crucial subjects that happen not to be as groovy as some of the “touchy-
feelly” courses. A reasonable person always has to consider both alternatives
seriously, self-change as well as other change.

2-14

The text uses, or suggests that you or your instructor use, a number of
devices, memory aids, display techniques, and learning games. They can be
thought of as a more practical substitute for the technical terminology and
complex calculi of the traditional approach to logic, as devices that are
oriented more pedagogically than philosophically. For example, we use
checklists to remind you of the various steps you should be sure to cover in
argument analysis; we suggest you tape-record political debates on televi¬
sion so that you can replay them slowly for dissection at leisure. We recom¬
mend the use of multiple colors in setting out complex arguments for read¬
ability; the use of roommates as trial audiences; and the use of role playing
or simulation games to handle violently controversial material and other
debatable questions. This is the procedural side of the practical emphasis of
the course; the content emphasis on practicality is covered in previous re¬
marks and includes the use of realistic examples, of examples from which
you can learn about something besides reasoning, of scrambled exercises, of
“field trips,” and similar devices.

2-15

Finally, it is essential to understand one point above all others about the
approach in this book. The basic belief here is that thorough and decisive
analysis of practical prose, and particularly of arguments, is an extremely
sophisticated subject, not an elementary one. That doesn’t mean it’s fearfully
difficult for you to do, with appropriate training, but it does mean that the
theory behind it is very difficult. The study of the English language is an
extremely sophisticated subject. You can use it very well, but no one has yet
succeeded in providing an adequate grammar for it, let alone a comprehen¬
sive theory of its development or of the psychology of language learning. So
 TEACHING AND LEARNING REASONING 23

the emphasis of this book is on practice rather than theory, though there are
plenty of opportunities provided for excursions into theory, should you or
your instructor so choose.
The other consequence of this view of the study of reasoning as an
extremely sophisticated subject is again a practical one. The history of the
subject has consisted of accounts of people hopping quickly over the appar¬
ently simple, everyday examples of practical reasoning and moving on into
the interesting-looking theoretical woods surrounding it. As a result, they
have vastly underestimated the complexity of everyday reasoning. Analyzing
the reasoning in some comparatively short examples turns out to be at least
as hard as trying to analyze the "deep grammatical structure” of some short
sentences, a task which is now notorious for its difficulty. The practical
consequence of this is that you will often find that the discussion of an
everyday example is many times longer than the example itself, especially
when you are beginning to learn the techniques. Most books on reasoning-
texts in logic and rhetoric, for example—attempt to provide labels or calculi
to handle these examples swiftly. But those devices turn out to handle them
inadequately, in my view.
So you won't find your speed in logical analysis going up at first. It will
go down. When swimmers or tennis players come under the guidance of a
professional coach for the first time, they are often disappointed to find that
they have to change the way they were doing things, and they think the new
way feels more awkward, doesn’t work nearly as well. But the coach has
seen a basic flaw in the approach and knows that unless it is corrected, good
and quick performance will never be attained by the pupil. In the same way
here, although in the end we shall have sharpened up your logical wits and
your instinctive reasoning skills so that you can react much more quickly
and accurately than now, at first you'll have to go slower than ever before,
just to get the steps right. Be patient. When I first started work in a lumber
camp the axman who taught me said, “Learn to understand every piece of
timber first, before you attack it. If you understand it thoroughly, it will do
half the work for you.” The master diamond cutters in Amsterdam will
spend many hours studying a big stone in preparation for the few seconds it
takes to cleave it. You must learn to spend as much time as it takes to study,
and perhaps discuss, a key argument before you attack it. Speed builds
slowly.

QUIZ 2-A

1 Most textbooks on logic introduce a great deal of special terminology—jargon.

The major practical disadvantage of this, when you want to criticize arguments
put forward by your peers, is that:
a You may soon forget a new terminology.
 CHAPTER 2
24

b Different texts use different terminologies.

c The people with whom you are trying to communicate often haven t been to
college.
d Other (explain).

Answer
The first and second answers are true but they are just extra reasons against jargon,
not the major reason. Don’t forget them, though! You might get a later question on
a quiz which asks you to list several arguments against a “technical” approach to
logic. These aren’t the major reasons because, even if you didn’t forget, and even if
all the books used the same terminology, there would still be an overwhelming diffi¬
culty.
The third option may or may not apply to you, but it’s not the major problem,
because even if these people have been to college, it’s very unlikely they have taken
the same course and used the same text—and that’s the big disadvantage, which you
should have put in as d, Other, and should have explained.

2 “It seems pretty silly to ask someone who isn’t even taking a course to act like the
instructor—who presumably knows more (not less) than the people who are tak¬
ing the course—and grade one’s assignments. Isn’t this carrying the worship of
the know-nothings to an extreme?”
a For most courses this would be a good criticism.
b For courses whose main aim is to teach procedures for solving problems that
require knowing a technical subject, it would be a good criticism,
c For courses where the final result is an achievement which can be judged—not
taught—by the nonexpert, it’s an irrelevant criticism because the teacher’s
skills don’t automatically make the teacher a better judge of the product.
Cooking courses, handwriting, reading and speaking courses, drama and por¬
traiture courses, and even courses in the repair of motorcycles or fractured
limbs, should produce improvements which can be reliably detected by the
untrained observer/customer/patient.

Answer

We can surround the answer with a number of words, so that you won’t see it as soon
as you see the question, but the fact is that all three options are correct.

3 The quizzes at the end of the chapters in this book are often “scrambled” and/or
“uncued,” which means they may include which of the following?
a Tests on material from earlier chapters
b Requests to comment on or criticize perfectly satisfactory arguments
c Incomprehensible or chaotic material
d Arguments or passages of prose with no instructions as to what you should do
with them
e Tests of material from chapters you haven't yet read
TEACHING AND LEARNING REASONING 25

Answer

Let's begin by extending the question somewhat. The answer as it stands is easy,
since all these types of questions except c are examples of scrambled or uncued
questions: type c might or might not be, hence should not be checked. Can you now
state clearly the educational reason for including each of types a, b, d, and e? If so,
you've just done question 3 in Quiz 2-C.

QUIZ 2-B

1 The argument against using any substantial amount of special terminology in a

textbook on reasoning arises because of which of these functions of reasoning?
a Reasoning as a process for arriving at the most probable conclusion
b Reasoning as a process for persuading others of the probable rightness or
wrongness of given conclusions or arguments
c Reasoning as a process involving language

2 Knowing that someone else in the class or a roommate will grade some of your
assignments may have a good effect on you as you do the work, because:
a You know that the person will have to be able to read it and may not be as
good at reading other people’s poor handwriting as someone who does it all
the time.
b You’ll try to make your points understandable to someone who is nearer to
the eventual audience for your reasoning than a college professor,
c You’ll figure the person is pretty easy to con, so you won’t put too much effort
into the assignment.
d You want to make a good showing to peers as well as professors.

3 The trouble with setting quizzes that are strictly on the material covered in pre¬
ceding chapters is that:
a The student doesn’t know what kind of comments would be relevant,
b The teacher is not assisted in determining whether later material is superflu¬
ous.
c Retention of material from earlier sections is regularly tested,
d Material you run into in the real world is not usually full of clues that tell you
one particular chapter in this text is relevant.

4 Read the following passage and answer the questions that follow.

Until about thirty years ago, the village of Nayon seems to have been a self-
sufficient agricultural community with a mixture of native and sixteenth cen¬
tury Spanish customs. Lands were abandoned when too badly eroded. The
balance between population and resources allowed a minimum subsistence. A
few traders exchanged goods between Quito and the villages in the tropical
barrancas, all within a radius of ten miles. Houses had dirt floors, thatched
roofs, and pole walls that were sometimes plastered with mud. Guinea pigs ran
26 CHAPTER 2

freely about each house and were the main meat source. Most of the population
spoke no Spanish. Men wore long hair and concerned themselves chiefly with
farming.
The completion of the Guayaquil-Quito railway in 1908 brought the first
real contacts with industrial civilization to the high Inter-Andean Valley. From
this event gradually flowed not only technological changes, but new ideas and
social institutions. Feudal social relationships no longer seemed right and im¬
mutable: medicine and public health improved; elementary education became
more common; urban Quito began to expand; and finally—and perhaps least
important so far—modern industries began to appear, although even now on a
modest scale.

4.1 Why were there primitiveness and self-containment in Nayon before 1910?
a Social mores
b Cultural tradition
c Biological instincts
d Geographical factors
e Religious regulations
f I don’t know

4.2 By 1948 the village of Nayon was:

a A self-sufficient village
b Out of touch with the outside world
c A small, dependent portion of a larger economic unit
d A rapidly growing and sound social and cultural unit
e I don’t know

4.3 Why was Nayon originally separated from its neighbors?

a Rich arable land
b Long meandering streams
c Artificial political barriers
d Broad stretches of arid desert
e Deep rugged gorges traversed by rock trails
f I don’t know

5 Comment on the following news story in any way that seems appropriate.

MORE REFUSALS: U.S. SURVEY FINDS BIAS IN HOME LOANS

The federal agency that regulates institutions making the bulk of the nation’s
home loans reported yesterday that lenders refuse mortgages to blacks more
than twice as often as whites.
The Federal Home Loan Bank Board, which drew no conclusions from a
survey of lending practices in five cities, said that black home loan applicants
experience the highest rejection rate of any racial or ethnic group.
The bank board’s survey was the latest in a three-pronged effort launched
by major federal regulators of commercial and savings banks. The three agen-
TEACHING AND LEARNING REASONING 27

cies conducted the surveys as a prelude to drafting forms that would be used in
enforcing federal restrictions on discrimination.
The Federal Reserve Board reported in May that its six-area survey
showed minority applicants were rejected about twice as often for home loans.
Senator William Proxmire (Dem-Wis.) released last month comptroller of
the currency results which Proxmire said “strongly suggest that mortgage lend¬
ers are discriminating against blacks and other minorities.”
The comptroller’s study, which broke down rejections and acceptance
rates by income level, showed that even among applicants with assets in excess
of $20,000, blacks were rejected 21.6 per cent of the time and whites 14.1 per
cent of the time.
The Home Loan Bank Board, which regulates savings and loans but in¬
cluded all types of home loan institutions in its time survey, said blacks were
rejected 18 per cent of the time and whites 8 per cent of the time.
The Home Loan Bank Board study covered 53,705 loan applicants in
Buffalo, Chicago, San Antonio, San Diego and Washington, D.C.
Overall, 84 per cent of the applications were approved and nine per cent
turned down. The rest of the applications were either withdrawn or no action
was taken on them.
Whites had the highest acceptance rates, 85 per cent, and the lowest rejec¬
tion rate, eight per cent. Blacks received acceptance in 77 per cent of the cases
and rejection in 18.
For Asians the proportions were 83 per cent accepted and ten per cent
rejected. American Indians were accepted 83 per cent of the time and rejected
13 per cent of the time. The acceptance rate for Spanish-speaking applicants
was 81 per cent, the rejection rate 12 per cent.
The report also showed that men of all races were accepted 84 per cent of
the time and rejected nine per cent. For women the acceptance-rejection rates
were 81 per cent and 11 per cent.

QUIZ 2-C

1 Under what general conditions is technical terminology justified in a college or

high school subject? Someone defending a technical approach to logic would
probably make the best case by arguing—how?

2 a The Chinese cultural revolution involved replacing a large proportion of the

faculty at colleges with lay people. A number of “alternative schools” in the
United States have done the same. Set out the pros and cons of this move and
relate it to peer-grading as described in this chapter. (Developing a good
statement of pros may take a little research.)
b If you know what the peer-review system is in Big Science in the United
States, relate peer-grading to it.

3 See the answer to Quiz 2-A3.

28 CHAPTER 2

4 “The double standards approach to grading is deceptive packaging, too—to the

registrar. It is incompatible with a proper professional approach to teaching and
faculty responsibility.” Discuss.

5 “Giving misleading instructions to students on quiz questions is a violation of the

trust relationship that should exist between student and teacher and is therefore
unprofessional conduct.”

6 a Would or wouldn’t it be unfair to good students to impose “grades quotas” of

the kind suggested in the text (not more than one-third A’s, and two-thirds A’s
and B’s combined, with two possible bases for making exceptions)?
b What are significant differences between this proposed system and the old
“curve” system?
 Chapter 3

The Foundations of
Argument Analysis

After some preliminaries we will get into the main business of this chapter,
which is to set out the basic approach to examples of reasoning that you
wish to assess, or to the production of good arguments that you wish to use
in persuading other people. Later chapters, one or two of them longer, but
some shorter, will expand on some of the more difficult steps in this basic
checklist of procedures, and will also develop special techniques for han¬
dling particular types of arguments, such as very long arguments, debates,
and certain types of moral argument.

3-1

Reasoning is not only concerned with arguments, however. Arguments are

just one of the products of reasoning and one of the objects to which reason¬
ing is applied. But it is applied to more than arguments; in general, its
relevance is to what we might call “practical” prose, by contrast with (most)
“literary” prose. For example, the way in which instructions are written, say
for the operation of a complex piece of equipment like a high fidelity pream-
29
30 CHAPTER 3

plifier or Dolbyized receiver, can be criticized from a logical point of view, in

terms of whether the sequence of steps is one that corresponds to a sequence
of possible actions in operating or assembling the device. An extended in¬
struction manual is much the same as a textbook; and logical criticism of a
text (such as this one) is one example of applied reasoning. The reports
written by correspondents describing items of major news importance
around the globe are often such that they cannot be understood by a reason¬
able person because they would make sense only if certain circumstances
were true, although they are said not to be true in the opening paragraphs of
the report. (To be fair to the news services, the incomprehensibility of these
reports is often due to editing at the local level.) Spotting that kind of defect
is an exercise of the reason.
There are, in fact, substantial parts of the body of literary prose which
are subject to assessment in terms of their rationality, reasonableness, and
similar qualities. This is most obvious in something like a detective story,
where what is supposed to be produced by the end of the story is (typically)
a proof that a particular individual was guilty; if it isn’t a proof by the
standards of reason, then the story is flawed as a literary work just because
it fails to meet the standards of good reasoning. In science fiction, the prob¬
lem of maintaining consistency while describing an imaginary world is one
that has defeated more than one budding author.
So the range of impact of the type of critical analysis with which we are
concerned in this text is considerably wider than that of arguments as such.
Nevertheless, setting up a list of procedures for argument analysis is a good
beginning, because almost all the other circumstances in which we apply
logic (i.e., the standards or principles of good reasoning) are ones where we
can formulate the essential point of the material we are assessing as an
argument. In the remaining cases, the point is essentially that what we are
doing is applying the standards of reason to the material, or to the individ¬
ual, or to the group in question, and the punch in these standards, indeed the
one basic criterion to which they all refer, is the avoidance of inconsistency.
Criticism of arguments depends crucially upon appeal to the need to avoid
inconsistency, as we shall see, and the criticism of other types of practical
prose in rational terms depends on consistency in order to avoid incompre¬
hensibility.
We will begin with straight argument analysis and quickly show how it
can be extended to cover a good deal of prose. For example, much advertis¬
ing prose is intended to persuade you of a certain conclusion, although if
you look at it in grammatical terms, it does not have the grammatical form
of an argument. Finally, we will say something about the rational assess¬
ment of types of prose that are simply not reducible to explicit or implicit
arguments.
THE FOUNDATIONS OF ARGUMENT ANALYSIS 31

3-2

So the basic threat behind the rational analysis of practical prose is the
threat of inconsistency. In this section, let’s look at the question of why we
should try to avoid inconsistency, why the threat is a serious one; and then,
in the next section, we will look at the question of how inconsistency is
brought to bear in the analysis and improvement of reasoning.
The avoidance of inconsistency is crucial simply because consistency is
a requirement of communication. If you want to talk to people, if you want
to convey information to them, if you want to learn something from them,
indeed if you want to be able to learn something from the world itself—to
perceive something, or know it, or remember it—there is no alternative to
respecting the requirement of consistency. Look at a simple example. Sup¬
pose a friend said to you that the house where the party is on tonight “is one
you can’t possibly miss because it’s the only one on the street with a red roof
and without a colored roof.” The sentence is meaningless to you, gives you
no information which will help you find the house. It’s as if you asked what
the weather was like outside of a man who just came into the basement of
the library where there aren’t any windows, and he replied, “It’s pouring
rain and as dry as a desert.” You don’t know what he means. Of course, one
makes every effort to find a meaning in apparent contradictions and incon¬
sistencies; if your friend had said that the party was on at a house which is
“in Berkeley and not in Berkeley,” you would probably have thought to
yourself that it was on the borderline between Berkeley and a neighboring
community. If the man coming into the library had said that “it is raining
and not raining,” you perhaps would have thought that he was trying to
convey the fact that there were showers, or a very heavy mist with drops of
condensation falling through it. But, in hard-core contradictions, you can’t
“make sense of them,” and consequently you can’t learn anything from
them. Of course, the individual words make sense; it is not as if somebody
had uttered a sentence in gobbledygook or just a series of noises without
meaning. But the sentence as a whole doesn’t have any sense because it is
inconsistent, contradictory. The whole point of conveying information is
that one is describing a situation which is the case, and is thereby excluding
certain situations which might conceivably have been the case. If you are
told that a certain situation is the case and also that it isn’t, the contrast
between what is really the case and what isn’t has evaporated and you no
longer know what you are being told.
The avoidance of inconsistency is thus essential for communication.
The sort of inconsistency we have been talking about so far is what is some¬
times called “logical inconsistency”; it is an inconsistency which can be
detected by anyone who speaks the language, without any other special
32 CHAPTER 3

knowledge. But, of course, exactly the same point would apply with respect
to what we might call “factual inconsistency.” Supposing somebody tells
you that the party is going to be in a house that is in Berkeley and in San
Francisco; then, if you know the geographical relationship of these two
areas, you know that that description is inconsistent. If, on the other hand,
the description had been “in Berkeley and in Albany,” you could make some
sense out of it since these are adjacent cities and the house could well strad¬
dle the line between them. It is a little less than precise to describe the house
as being in both these cities—it would be more precise to say that part of it
is in one of the cities and another part is in the other. But, without nit¬
picking too much, you can make sense of the description. On the other hand,
given the geographical fact that there is no point of Berkeley which is less
than several miles away from any point of San Francisco, you can’t possibly
make sense out of the description of the house as being in both these places.
Factual inconsistency is really logical inconsistency between what is said
and an unstated extra fact.
A still weaker type of inconsistency, really a borderline case which
might more accurately be described as “improbability,” or “implausibility,”
or “quasi-inconsistency” than as strict inconsistency, arises when something
is described as having two properties which are almost certainly incompati¬
ble. For example, one might describe something as being red all over and
also green all over, a statement that would certainly be confusing for most
people. But there are certain types of shot silk (or opals) which are red all
over when viewed from one angle and green all over when viewed from a
slightly different angle, and people who know this fact might immediately
think that you were talking about silk or opals. A really good Russian alex¬
andrite, a rare gemstone, will appear dark green in daylight and dark red by
artificial light. So here we have a kind of borderline case where the descrip¬
tion has the appearance of logical inconsistency but might possibly be ex¬
cused on the ground that it can be, so to speak, “given a sense” by people
who know about some of these odd situations.
The main point of argument presentation is to show that some kind of
inconsistency or implausibility is involved in accepting the premises of the
argument and rejecting the conclusion. Arguments are meant to be persua¬
sive; and they succeed in being persuasive if they begin with assertions that
the listener or reader is known to accept, and if they continue by showing
that acceptance of those assertions (which are called the premises of the
argument) requires the acceptance of the conclusion. And what “requires”
means here is simply that it would be inconsistent in one of the senses just
described, to accept the premises and reject the conclusions.
The power of an argument depends upon two things; first, that it
should start off with premises that are known to be true or can be shown to
be true; and that it should proceed to demonstrate or exhibit the way in
THE FOUNDATIONS OF ARGUMENT ANALYSIS 33

which these premises “force” one to accept the conclusion or conclusions.

The "force" here is the force of wishing to avoid contradiction, or inconsis¬
tency of a weaker kind (“contradiction” is another name for “logical incon¬
sistency ’). In criticizing an argument, as opposed to proposing one, what we
try to do is to show that there is no inconsistency between accepting the
premises and rejecting the conclusion. (Or, of course, we may criticize an
argument by showing that even the premises aren’t true.)
So both criticism and presentation of arguments are duels in which the
weapon is logic and its cutting edge is inconsistency or contradiction.
Let’s look at a very simple example. Someone might be arguing that the
true cost of a particular boat that his friend is thinking of buying is nearly
twice as much as the friend has stated. The argument would consist of
pointing to a number of indirect costs, such as licensing, docking, mainte¬
nance, insurance, and the interest cost on the loan involved (or the lost
interest from investing the cash that was used to buy the boat outright).
Each of these assertions about indirect costs is a premise, and it can be
challenged directly or conceded. If it is conceded, the function of the argu¬
ment is to show that the total sum of all these indirect and overlooked costs
is approximately the same as the amount said by the friend to be the cost of
the boat. In working out that sum, of course we use the rules of arithmetic
(which are transformations based on avoiding inconsistency). In identifying
the indirect costs, we use various factual sources, possibly including our own
experience and the words of experts. We are really putting up a proposition
to our friend which consists in telling him either to point out something
wrong with our factual premises or something wrong with the laws of arith¬
metic; or else to concede that there is something inconsistent between his
cost estimate and ours. If that inconsistency exists, and if he concedes that
the indirect costs to which we are pointing really occur, he “must” concede
that he is wrong. His conclusion is really inconsistent with the real facts of
the situation, which are exhibited in the premises and conclusion of our
argument. In this particular case, the step from the premises to the conclu¬
sion we drew relied only on the extremely reliable laws of arithmetic. It
could therefore qualify as a case of “logical deduction.”
In another situation where the steps from our premises to our conclu¬
sion might involve certain rather plausible assumptions, say, about the way
in which taxation is likely to go up or the cost of repairs is likely to rise, we
would have to say that our conclusion was only “made probable” by our
premises. That type of argument is sometimes called an “inductive” argu¬
ment, by contrast with the “deductive” one given earlier. (It is sometimes
also called a “probabilistic” or—somewhat inaccurately—a “statistical” ar¬
gument.) The difference is not really very important from the point of view
of practical reasoning because exactly the same choices are open to the
respondent. The opponent must rebut either the premises or the chain of
34 CHAPTER 3

reasoning that takes us from those premises to the suggested conclusion.

That chain of reasoning may be merely a matter of rules of language or of
arithmetic, which are normally thought of as bulletproof (which would make
the argument deductive), or they may involve claims that are somewhat less
certain, possibly laws of science or generalizations based on experience (in
which case the argument is inductive). You don’t need to memorize the terms
“inductive” and “deductive”; we mention them only because you may run
across them in some of your background reading. A slight juggling of the
premises (by adding some unstated ones) and the conclusions can always
convert an inductive argument into a deductive one without any essential
loss of the “point of the argument,” so the distinction isn’t one you would
want to build very much on; and. to make matters worse, some of the most
respected professional logicians in the country today think there aren’t any
pure examples of deductive argument anyway. What we need to keep in
mind is that a given argument normally involves some premises and some
steps to get from these premises to the conclusion, and that if it’s going to be
attacked, either the premises or the steps (which are known as the infer¬
ences) have to be attacked. Premises are attacked in the straightforward way
in which one attacks any factual claim, that is, by producing evidence that it
isn’t true. The inference, on the other hand, can’t always be attacked in just
that way. A great part of this book is concerned with the assessment of
inferences, either with the intent to improve their clarity and reliability or to
expose their weaknesses.
So the weapon behind logical analysis, behind the appeal to reason, is
nothing more serious than the sharp edge of inconsistency. If people have no
interest in communication, no interest in the acquisition of information, no
interest in the truth, then there would be no reason for them to concern
themselves with avoiding inconsistency, and the practice of reasoning would
have no function. But, since we are all pretty much interested in survival,
indeed in improving our welfare or that of others, and since, in order to
achieve those aims, we have to determine what is the case and what is not
the case—for example, what sort of diet is healthy and what sort of medicine
is helpful and what sort of clothing will last and what kind of people can be
trusted—then we are all in the business of trying to determine the truth
about something or other, and hence we are all in the business of reasoning.
At least, we “should” be, meaning that if we are not, the price we pay is not
mere abstract inconsistency, but a simple practical loss of the things we
value.

3-3

It’s obvious that you have to speak consistently in order to convey informa¬
tion to others. Be sure that you understand why consistency is also necessary
THE FOUNDATIONS OF ARGUMENT ANALYSIS 35

for you yourself to know anything at all. If you believe that there is a God
and also that there isn t, you don’t know what you believe. And certainly
you can’t be said to know (and you can’t yourself say that you know) either
one of these propositions. Knowledge is possible only in the absence of
inconsistency in your beliefs just because “knowing” contradictory things
only shows you haven't yet worked out which things to believe.
One can make (some kind of) sense out of claims like “I know there’s a
God and that there isn’t,” that “human beings have free will and yet don’t,”
and so on. The sense that one makes out of such remarks is that there is some
sense in which one of these claims is true and another sense in which it is not.
That’s just a way of saying that you’ve converted each claim into two claims,
and that the denial of one of them isn't the denial of the other. In short, you
have salvaged consistency in order to be able to make sense of the claim to
knowledge.

3-4

You will have noticed in the last couple of sections that we frequently use a
rather long-winded way of expressing the fact that reasoning can be used in
the service of criticism or in the service of persuasion or communication. It
is time for us to look carefully at this double function of reasoning, and once
we have understood it, we will be able to avoid this rather verbose way of
putting points so as to make clear that we are not talking about just the
critical function of reason or just the persuasive function of reason.
People often imagine that rationality and logic are essentially negative
or critical enterprises. It is extremely important to understand why this isn’t
so. There are two quite separate arguments against this “negative” view of
logic. The first is clear enough in what’s been said earlier; logic or reasoning
is the means whereby we reach new conclusions, gain new knowledge, un¬
cover new and important facts. When you sit down to work out exactly what
it would cost you to own a boat, you are engaged in a logical process, you
are reasoning about the matter, and the result that you come to, if you are
good at this kind of reasoning, is a discovery, a new and important piece of
information, for you. So reasoning is a constructive and creative activity that
leads us to new knowledge. Moreover, it is not only with respect to our own
knowledge that it’s important; as we previously stressed, reason is a means
of persuasion, a means that is uniquely important in that it does not deny
the right of others to make up their own minds. It is not brainwashing or
propaganda or threat, but simply the production of considerations which—
in their own interest—should lead them to agree with you (assuming that
you have put forward a sound argument), and thereby come to new knowl¬
edge themselves. So logic is an important part of the communications web of
a just society.
36 CHAPTER 3

But there is a completely independent reason for rejecting the view that
there is some kind of contrast between rationality and creativity, between
the use of logic and the use of intuition or imagination. Let's look more
closely at the procedure for criticizing an argument. The argument consists
in a set of premises which are said to “imply,” that is, “lead to,” a conclusion
or set of conclusions. Suppose that we are concerned only with the criticism
side of logical activities. Even here the creative element is inescapable. Let’s
look at an example. Suppose that somebody is arguing that the intrusion of
herbicides into the ecology of the Snake River Valley area in Idaho has been
the cause of the drastic decline in the population of birds of prey, especially
the golden eagle, in that area. The evidence the person presents is simple.
We have the records of naturalists going back fifty years, which show this
area to have carried an exceptionally heavy population of birds of prey,
ranging from the golden eagle and the bald eagle down to the smaller fal¬
cons. We know that, in a relatively short period of time, corresponding
almost exactly to the period of introducing chemical herbicides and power¬
ful pesticides into the farming communities nearby, there has been a very
serious decline in the number of young hatching and the total population of
adults.
Now, what is the exact force of this argument? Why does the speaker
suppose that these facts “oblige” you to accept the explanation that she
believes follows from them? The force of the argument lies in the fact that it
is very difficult to think of any other explanation of the phenomenon of the
decline in the population of birds of prey, and since we believe that these
situations don’t occur without some explanation, it looks as if you are forced
to accept the speaker’s.
But you are forced to accept her argument only if you cannot think of
any other. If you can think of another explanation that is comparably plau¬
sible, then, until it is ruled out, she does not have a case. The process of
trying to think of alternative explanations of a set of facts—the premises in
an argument—is an entirely creative process. It is exactly the process which
the great original scientist goes through in coming up with a novel theory.
There are no precise rules to guide one in such a search, and it requires
imagination nurtured by a rich and varied experience to generate the novel
hypothesis here. So the very process of criticism necessarily involves the
creative activity of generating new theories or hypotheses to explain phe¬
nomena that have seemed to other people to admit of only one explanation.
As you begin to think about the example we have been discussing, a
number of other possible explanations may occur to you. For example,
farmers in that area may well have shifted the type of crop they were har¬
vesting to a crop that provides less adequate sustenance for the small birds
and mammals that are at the lower end of the food chain from the birds of
THE FOUNDATIONS OF ARGUMENT ANALYSIS 37

prey. There might have been a shift from grain crops to, say, soy beans,
which won't support the same number of rodents, finches, and other small
creatures. So, although human beings are in some sense responsible, it is
only in the sense that their previous activities maintained an unnaturally
high level of predator population, and the new crops have resulted in this
decline to approximately the level that existed before the first settlers began
farming those plains. Naturally, one would have to do some investigation to
find out whether this alternative explanation is really a live candidate, but
even thinking of such a possibility shows that the issue is not quite so closed
as the arguer supposes.
To sum up our point, then, it is a great mistake to suppose that imagi¬
native and creative thought processes are somehow separable from critical
and logical ones, since to do good criticism, one must in fact be capable of
the imaginative and creative production of new ideas.
And the other side of the coin is that creative or imaginative abilities
are useless if they are not combined with the possession of critical skills,
since one could not recognize a creatively generated solution to a problem
unless one had the critical skills necessary to identify it as a solution to that
problem. It’s easy to generate new theories of gravitation or new ways to
transport people faster than the speed of light if you are not very critical
about whether your ideas will work! Critical skills go hand in hand with the
creative ones; creativity is not just a matter of being different from other
people, it’s a matter of having a different idea that works as well as, or better
than, previous ideas. If originality just meant novelty, it would be of no
value; it means novelty and validity.

3-5

Now that we are about to begin serious argument analysis, it may be appro¬
priate to review our use of a number of terms that we have been using until
now. We have already made clear that we treat the terms “rational” and
“reasonable” as being more or less equivalent. They are also roughly equiv¬
alent to “logical.” But the term “logic,” although it is, in informal speech,
more or less equivalent to one sense of “reason” (one can equally well say
that “logic dictates that we believe so and so” or “reason dictates that we
believe so and so”), also has a more technical sense. The technical subject of
logic, as the term is employed in philosophy departments for example, refers
to what we might more usefully call “formal logic” or “symbolic logic.” This
may range from a discussion of the rather elaborate and very ancient tradi¬
tion of syllogistic logic, which began with Aristotle, to more recent develop¬
ments, particularly to modern symbolic logic which developed into a very
substantial subject in its own right in the twentieth century. It can be re-
38 CHAPTER 3

garded, however, as essentially a part of mathematics; it is an extremely

precise and formal discipline, and not one that can be readily, if at all,
applied to the analysis of everyday argument. When we talk in this book
about logic, we simply mean the discipline or principles of careful and sys¬
tematic reasoning. We are concerned not with the abstract type of logic but
with the more practical species.
There is another distinction which is connected to these terms and
which should be brought in here. We often talk about an argument being
“logically sound” in order to contrast it with its being “factually sound.” To
say that the argument is logically sound is to say that the reasoning in it is
sound, that the inference(s) from the premises to the conclusion(s) is/are
sound, without saying anything about whether the premises are themselves
sound. To say that it is factually sound is to say that the premises are in fact
true. As we have mentioned before, if you want to criticize an argument, you
can do so by focusing on two points; the truth of the premises, or the
goodness of the inference. These correspond to the question of whether the
argument is factually sound and the question of whether the argument is
logically sound. Determining the truth of the premises is the main task of the
subject-matter disciplines in science and the humanities; it isn’t our main
task here. Our focus is particularly on the reasoning process that moves one
from premises to conclusions. And it is quite appropriate that this should be
referred to as the “logic of the argument.” One will often encounter looser
uses of all these terms, but these few remarks should indicate the principal
differences that we have in mind when we refer to them in this text.
Another pair of terms that need to be used with some care are “impli¬
cations” and “inferences.” The implications of a statement are, of course,
what it implies, i.e., what follows from it or is supposed to follow from it. At
the most subtle end, the implications may amount merely to what the state¬
ment suggests or is intended to suggest; at the strict end, they mean just
those things that necessarily or logically follow from it. Inferences are the
steps from premises to conclusions, but the term is also used to refer to the
conclusions themselves; for example, one talks about someone making cer¬
tain inferences from the clues that are available at the scene of the crime.
This is essentially equivalent to saying that the person is drawing certain
conclusions from that evidence, or is seeing certain implications in it. But
there is one distinction between “infer” and “imply” that is grammatically
essential. Premises imply conclusions, they do not infer them. People infer
conclusions (from premises, evidence, and other things), which is the same as
saying that they “draw” the conclusions. And this activity of inferring is not
at all the same as that of implying conclusions, which people can do in other
circumstances. For example, when somebody—in a suitable context—says
in a belligerent tone that Jones got his doctorate at Harvard, one may take
THE FOUNDATIONS OF ARGUMENT ANALYSIS 39

the speaker to be “implying" that Jones knows what he is talking about on

the subject under dispute. The speaker is not inferring anything, but wants
you to infer that. We will run through some exercises at the end of this
chapter which will clarify this for you. But it is important to introduce you
to these distinctions now because, in the next section, we will be using these
terms with some care. Now let's get down to the paydirt.

3-6 THE SEVEN STEPS IN ARGUMENT ANALYSIS

What we will do in this section is first to list the Seven Steps, and then
explain them in a little more detail so that you will have a general picture of
the approach. In subsequent sections, we will go into considerably more
detail about the various elements and skills that are involved in each of the
Seven Steps, and apply them to various special types of argument. But no¬
tice from the very beginning that there isn’t anything highly technical or
particularly hard to understand about these steps. The difficult thing is to
follow them carefully and skillfully. And acquiring that care and skill is a
matter of practice and still more practice. Nobody gets to be good at mathe¬
matics without a lot of practice; nobody gets to be good at physical skills
without a lot of practice; and nobody gets to be good at reasoning without a
lot of practice. You have had a lot of practice in casual reasoning, and what
this course tries to do is to convert that practice into a considerably more
sophisticated, systematic, and effective process. Learning a new jargon or
some technical calculus would give you a sense of having learned something,
but it wouldn’t teach you something useful for the kind of task we are
focusing on in this book; the assessment and presentation of everyday argu¬
ments. So remember that the trick here is to do a familiar task better, not to
learn a new terminology for talking.
The Seven Steps are as follows:

1 Clarification of Meaning (of the argument and of its components)

2 Identification of Conclusions (stated and unstated)
3 Portrayal of Structure
4 Formulation of (Unstated) Assumptions (the “missing premises”)
5 Criticism of
a The Premises (given and “missing”)
b The Inferences
6 Introduction of Other Relevant Arguments
7 Overall Evaluation of this argument in the light of 1 through 6.

Let’s elaborate on each of these steps, so that we can follow them more
fully.
40 CHAPTER 3

1 Clarify Meaning

This step includes clarifying the meaning of:

Terms
Phrases
Sentences
Suggestions or implications
Arguments

Method

a Read most or all of the argument or passage under consideration

before trying any clarification.
b Replace unknown terms by reference to a dictionary,
c Rewrite any unclear parts, using clearer language,
d In particular, identify vague or ambiguous terms that you suspect
the argument is “exploiting”; e.g., by shifting from one meaning to another.
Translate the clause or sentence containing each occurrence of these terms
into other language that conveys the correct meaning of the term in each
context. That will show up any shifts in meaning.
e Write out any important unstated but intended implications or sug¬
gestions of the premises, the conclusions, and the argument as a whole.
(What’s it trying to get across that isn’t actually spelled out?)
f Ask yourself if you really understand how everything fits together. In
other words, have you a “feeling” for the argument or passage as a whole
(even if you don’t accept it)? Don’t let any hostility you may have for the
position expressed mislead you into misrepresenting the argument—say, by
making it more stupid than it already is (you think).
g Look over results of a through e and criticize the passage for unclar¬
ity where appropriate. Most of Step 1 is laying ground for later analysis. But
this part, g, is a component of your final criticism.

Remember

a The “meaning” of an argument (or word, or other expression) is not

what the arguer intended but what he or she said, taken as a native speaker
of the language would hear it.
b Still, you want to make the best guess at the arguer’s intended
meaning, and we can take account of context. In one context, “Dogs bite”
may mean “All dogs bite”; in another, “Most do”; and in another, “Dogs
sometimes bite.” That is, the meaning of words or phrases isn’t to be found
in those words all by themselves. Look at the context; if the speaker is
present, ask for clarification. If not, treat the words as having their usual
meaning.
 THE FOUNDATIONS OF ARGUMENT ANALYSIS 41

2 Identify Conclusions (Stated and Unstated)

Method

a Some of the unstated conclusions turn up in Step 1(e) while you’re

trying to get the meaning straight. Set them out now: write them in below
the passage of text, or fit them in (perhaps in the margin) where they come
in. Are there any more, perhaps unintended but unavoidable ones? Get
them all stated clearly and fairly. Which are the most important ones? Is
there one main conclusion? (There usually is.)
b To locate the stated conclusions, look for indicator words like
therefore, because, “so, and “thus,” and for placement cues such as
location at the end of a paragraph. (These cues are by no means always
reliable: you also have to depend on your sense of the meaning of the
passage as a whole.)
c Notice that there may be several conclusions in the argument, each
building on the previous ones. And a passage may also contain several en¬
tirely separate arguments.
d Within any one argument, try to decide if that argument has a main
conclusion (or conclusions) and if the others can be ranked as to their impor¬
tance. We can call the second group “secondary” conclusions, some of
which may still be quite important, others more or less incidental.

3 Portray Structure

That is, set out the relationships between conclusions and premises, in the
parts of the passage that are arguments. You’ve already identified the con¬
clusions. Now you just need to ask yourself what assertions are being put
forward to support each of these conclusions. These are the premises. Typi¬
cally, there will be other material in the passage that is neither a premise nor
a conclusion. It may be instructions, rhetoric, repetition, flourish, or other
statements. The following procedure is unnecessary for very simple argu¬
ments, and it should be applied to very long ones a page or paragraph at a
time.

Method

a Number each separate assertion; note that one sentence may contain
several assertions—for example, this one contains two, as well as an instruc¬
tion. Put square brackets at the beginning and end of each assertion, and
number it in the margin or above the line of type.
b Do not give a number to repetitions of the same assertion; repetition
often occurs for good reasons (as a reminder) as well as bad (not recognized
as repetition).
 42
CHAPTER 3

c Do not number irrelevant statements (“asides”). Remember that

your judgments of irrelevance or repetitiousness are crucial to your evalu¬
ation of the argument, and you must be ready to defend them. ’
d Do give a number to the implicit conclusions you first located in 1 (e)
and 2(a).
e Set out the relationships between the relevant assertions in a tree
diagram like the one shown here. It is read downward on the page.

(4)

If 1 and 2 are claims put forward to support 3 and are not themselves
supported by any other assertion, and if 3 is supposed to support 4, but not
vice versa, the diagram looks like the illustration. If 4 might be an unstated
conclusion, you might put it in parentheses, as shown.
f For a balance of considerations” argument, where we say that 1 2
and 3 suggest the conclusion 5, “despite” 4 (which points the other way) use
symbolism as shown in this diagram.

,© + © + © - ©
--V-J

©
g Sometimes, you can set the structure out on a single line eg(l +
2 + 3 + 4) -> 5 or 1 2 (3 + 4). The arrow then stands for “implies ”
Sometimes the suggestion is made that, for example, (1 +2) imply (3 + 4)
and are implied by them; then use a double-ended arrow, thus; (1 + 2) ^ (3

h Terminology: If statement 1 implies statement 2, we can also say 2

follows from 1, or ‘is a consequence” of 1, or that we can infer 2 from 1
, "Tr" t0 Say * mferS 2'' statements imPly but can’t infer; people can
do both (but not at the same time). 7 F H
i While doing this, begin to look for places where there are significant
uns ated assumptions (“missing premises”). You can locate them by adding
circles to the tree diagram with letters in them at the appropriate places
THE FOUNDATIONS OF ARGUMENT ANALYSIS 43

(a is an assumption that is
needed to support the inference
from 1 and 2 to 3)

To formulate them exactly, see the next section.

4 Formulate Unstated Assumptions

The most difficult part of reconstructing an argument is fair and clear for¬
mulation of the “missing premises,” i.e., unstated assumptions. You must
distinguish between:

a The arguer’s assumptions, what he or she consciously assumed or

would accept as an assumption if asked.
b The minimal assumptions of the argument: whatever is, logically
speaking, necessary to make it possible to get from the premises to the
conclusion of the arguer.
c The optimal assumptions, usually stronger claims than b which are
logically adequate and independently well-supported. (We’ll give examples
of all these soon.)

5 Criticize Inferences and Premises

Criticism of expression is already covered in 1(g).

Criticizing an inference from statement 1 to statement 2 means criticiz¬
ing the claim that 1 supports 2. You do not need to know whether 1 is true
or not in order to consider whether it supports 2. You just have to ask, if 1
were true, wouldn’t 2 then have to be true, or at least very likely be true?
(Understanding this point is also the key to testing a hypothesis, for when we
say, “If Jones did kill Mrs. Robinson, he would have had to run a mile in
five minutes to be in the restaurant by 9:10 p.m.,” we’re not saying he did or
that he didn’t, but we are suggesting that it’s reasonable to infer from the
claim that he did it to a certain conclusion. By checking on whether he could
run this fast, we are testing the hypothesis that Jones was the murderer.)
Criticizing a premise requires that, if the argument is going to be any
good as a way of marshaling support, the forces it calls up had better be
strong, i.e., the premises must be reliable. When the premises are technical
claims, you aren’t expected to comment on them in the course of logical
analysis. When they are definitions or analyses subject to logical criticism, or
matters of common knowledge, you are expected to assess them.
Good criticism of an argument requires that you look at both the reli¬
ability of the inference and the reliability of the premises. You might think
that there’s no point in looking at the inference if the premises are false. But
your criticism of the premises may be either in error or fairly easily met by
 44 CHAPTER 3

minor modification; you must guard yourself against this by covering both
types of criticism. Good criticism also involves selective attack; first attack
the main conclusions (via the premises and inferences that bear on them),
and spend less time on the others. And attack with your strongest weapons
first; do not start by making picky points, following the order of the state¬
ments in the original. Start in on the key weaknesses: start with your strong¬
est criticisms. Strong criticisms are those that could not be met except by
extreme modification or complete capitulation.
Criticism strategy involves the key move of “counterexampling.” It ap¬
plies to many types of premise and all types of inference, and it is an exercise
in imagination. Here’s an outline of the procedure that you can refer back to
dater. It may be hard to follow in this brief summary, but we’ll explain it
with examples in the next section.

a To counterexample a generalization—say, “All A’s are B’s” or “Any

A is B”—you think of indubitable cases of A that are definitely not B. (It is
irrelevant to think of B s that aren’t A's, since the claim wasn’t that all B’s
are A’s.)
b To counterexample a definition (“A means the same as B”), treat it
as a two-way generalization (i.e., “Any A must by definition be a B” and
Any B must by definition be an A”) and look for counterexamples in the
realms of possibility as well as actuality, since a definitional truth must hold
whenever the language can be clearly applied, not only where it has been
applied.
c To counterexample an inference, treat it like a one-way generaliza¬
tion. (It will be a definitional generalization in the case of strict deductions,
as in most mathematical inferences; a factual generalization in the case of
most scientific inferences.) That is, if the statement A is supposed to imply
statement B, try to think of cases where A would be true but B would be
either definitely false or unlikely.
d To counterexample an interpretation or analysis, treat it as an infer¬
ence from that which is interpreted to the interpretation itself, and handle it
as in c. (It may be intended as an equivalence, that is, a two-way inference.)

Remember

If you have extensively reconstructed an argument by filling in many miss¬

ing premises and conclusions, you will have done so partly by asking what it
would take to make up a good argument. Hence you often won’t find much
to criticize about the inferences in the reconstructed argument—your criti¬
cism will fall instead on the extra bolstering premises you had to add.

6 Consider Other Relevant Arguments

If you stopped after Step 5, you’d have a thorough critique and sometimes
that s all that s called for, but you wouldn’t know what to think yourself. For
THE FOUNDATIONS OF ARGUMENT ANALYSIS 45

to discover that a particular argument has some defects is not to discover

that it shouldn t be given some weight, perhaps a good deal. Perhaps enough
to act on. At this point, then, you must step backward and try to get a
perspective on the argument. First, ask yourself whether there are arguments
on the same issue which point in another direction, perhaps to the opposite
conclusion or to a somewhat different conclusion. (In the case of argument
from analogy, you may even find that the very same analogy can be viewed
somewhat differently and taken to support the opposite conclusion.) Next,
look for other arguments that support the same conclusion.

7 Overall Evaluation

Go back to your criticisms. How devastating are they? Could they be met by
modest modifications of the original material? Even when devastating, do
they cover all the original lines of argument? Look at the results of Step 6.
They not only should help you decide what you think but they also may help
you to see what the original argument was after. Have you overcriticized it?
Now, make your final judgment on the argument. Grade it, in several
dimensions if you like, but then make yourself give an overall grade. It’s a
cop-out not to. You must decide whether it does have force, and how much,
for you.

QUIZ 3-A

1 You read a newspaper report of an experiment which, the headline said, showed
that marijuana has damaging effects. In the experiment, a number of college
students (who were paid quite handsomely for participation) smoked twenty
joints (marijuana cigarettes, as if you didn’t know) a day and exhibited significant
loss of motor coordination, problem-solving ability, and so on. What is the rel¬
evance of argument analysis to this?
a It has no relevance, since the report merely tells us new facts,
b It has some relevance, because the question of whether grass is harmful is a
very controversial one.
c Nobody with any common sense smokes twenty joints a day, so the experi¬
ment isn’t relevant to normal use of pot.
d This experiment was supposed to show that something (a conclusion) follows
from some evidence, that is, that the report on it really constitutes an argu¬
ment, and hence it can be appraised by using the techniques of argument
analysis.

Answer

The first response applies to the first sentence describing the report, taken alone,
which says that marijuana had been claimed to have damaging effects. And it applies
to the second sentence, taken alone. But the second sentence is said to describe the
evidence which was given to support the first claim. Hence we have evidence and a
46 CHAPTER 3

conclusion which is said to follow—in short, an argument. So argument analysis is

relevant to the reported argument. You’re not criticizing the newspaper report (it
may or may not have been accurate), but you can criticize the argument on which it
reports.
The second response is wrong because the mere fact that a controversial issue is
involved doesn’t mean that an argument has been given. There are plenty of contro¬
versial claims (like “Marijuana is a dangerous and highly addictive drug”) which
argument analysis can’t help you with, as they stand, because they are mere state¬
ments. It’s only when someone begins to try supporting them, or attacking them, that
we have an argument to criticize.
The third option makes a very relevant point about this particular argument,
which would be worth mentioning if you were being asked to criticize it. But you
weren’t asked that; you were asked to say whether the newspaper report was some¬
thing to which argument analysis is relevant. Don’t rise to the bait of checking
responses just because they’re true, or relevant to something; the only issue is
whether they’re true and relevant to the question asked. This isn’t being picky, it’s
being logical.
The fourth response is exactly right.

2 “If you talk to the professional tea-tasters, you’ll find that they prefer Lipton’s.”
a This is a simple argument of the common form “If. . . , then . . . ,” where the
first clause contains the premises and the second the conclusion,
b This is a conditional statement (a prediction, in this case) which tells you what
would (supposedly) happen if you did something or other. Since it’s just a
statement, it can be challenged only on factual grounds, not for the quality of
the reasoning, i.e., on logical grounds.
c There’s a clear implication here, even though it’s not stated. The implication is
that Lipton’s is the best tea, presumably, in fact, the best tea for you (the
reader or listener or viewer—to whom this advertising is directed). So it is best
treated as an argument with a single stated premise and a single (unstated)
conclusion; hence argument analysis can be used,
d Other:

Answer

Arguments are often set out in the “If . . . , then . . .” form, although, to be precise,
this form represents only the reasoning step in an argument, that is, it asserts that the
premises of a particular argument do imply the conclusions. The argument itself,
stated in full, consists in first asserting that the premises are true, and then asserting
the reasoning step or the inference—that is, asserting that the premises imply a
particular conclusion—and hence, in light of this, affirming the conclusion. So an
argument is actually a little more complicated than an “If. . . , then . . .” statement.
But such a statement does express a key element in an argument (and sometimes is
used to express a whole argument). The element expressed is indeed the one that is of
particular importance to us in studying reasoning. However, this particular state¬
ment, beginning with “If . . . is not a claim about certain premises implying a
certain conclusion. It’s not a claim about implication at all. Supposing I say to you.
THE FOUNDATIONS OF ARGUMENT ANALYSIS 47

“If you would get out of the light for a minute, I could see what I'm doing.” I am
uttering a conditional or a hypothetical statement, just as I might say, “If we had
more money, we could afford two cars,” but it isn’t at all an implication statement, a
claim about something following from something else. “Getting out of the light”
makes something else possible, but it doesn’t make it necessary. Reasoning is all
about what “necessarily” (in either a strong or a weak sense) follows from something
else, and many “If . . statements are about what will or might in fact happen:
they’re often just predictions, which are statements of fact about the future. “Having
more money” makes possible the purchase of the second car; it doesn’t imply or
make necessary the purchase of the second car. “If you feel up to it, we’ll go home
through San Francisco so that we can pick up some of the inexpensive wine that we
like so much" is introducing a proposal, a possibility for us to consider. It isn’t
suggesting that “feeling up to it” requires that we go home through San Francisco in
order to avoid inconsistency. So the first alternative is not the right interpretation of
this little quotation.
We’ve really covered the second alternative as well in the preceding discussion;
this is, intrinsically, just a hypothetical statement—a statement about what would
happen if (on the hypothesis that) you did something or other. The only problem is
that the context of statements like this, which is obviously part of an advertisement,
is a good deal richer than the second alternative gives credit for. This is not just a
pure hypothetical statement, taken in context; it’s a hypothetical statement which is
meant to convey something else to you. Hence, the second alternative, which says
that it is “a conditional statement” is oversimplifying the matter, and the third alter¬
native answer is the correct one.
Don’t feel bad if you didn’t get this one right immediately, or if you find the
discussion of it to be not entirely convincing to you yet. You’re just beginning to
make a distinction for which you have to develop a feeling, and it will take a good
deal of practice before you feel confident about it. In the next chapter, we’ll get on to
some more examples that clarify this and related distinctions.

3 “There’s only one person, other than the victim, who was within 50 yards of the
place where the murder was committed at the time when it was committed. That
person is Royce. So there’s no question about who did it. Royce is the murderer.”

This is an example of your being “forced” to accept the conclusion in order to

avoid which of the following?
a Logical inconsistency (contradiction)
b Factual inconsistency (factual impossibility)
c Improbability
d Asserting that the victim’s violent death might have no explanation at all
e Having to produce another explanation that is more plausible

Answer

The type of argument is a scientific, a practical, or—as it is sometimes called—an

empirical argument. It rests upon what we know about the world, not upon the rules
of language. If you say that Jones is childless and hence cannot possibly be classified
 48
CHAPTER 3

as a parent, you re stating a consequence of the meaning of the terms themselves,

that is, a consequence of the rules of language. It would be a logical inconsistency to
assert that Jones was childless and a parent. But it isn’t a logical inconsistency to
assert that a murder can be committed by somebody more than 50 yards away from
the event. For example, somebody can be shot at a range of 800 yards, or killed with
an arrow at over 100 yards. Since we aren’t given any facts about the way in which
the murder was committed, we can only assume (to make sense out of the argument
at all) that it was committed in some way that one normally associates with close
contact, such as strangling or stabbing. But of course one can imagine ways in which
a body can be given the appearance of having been strangled by somebody close by,
although in fact the death was caused by somebody at a remote distance. These are
far-fetched possibilities, indeed; one can think of knives thrown by a crossbow, or
strangulation produced by a bola, the South American gaucho’s weighted cords, or
other way-out murder instruments. These are certainly improbable; but they are not
factually impossible. The line between what is inconsistent with the facts and what is
made unlikely by the facts is not a sharp one. Hence, either the second or third
alternative would be a justifiable answer here, provided that you were clear about
your reasons for picking the particular alternative you did pick. It is physically
impossible for a house to be in San Francisco and Berkeley at the same time given
the way those cities are separated at the moment, but it's only physically unlikely
(improbable) that a person could be in San Francisco at one moment and in Berkeley
60 seconds later. There are a number of rocket and jet-powered devices which could
reconcile those claims, though it’s extremely improbable that any ordinary citizen
would have access to them. J
Of course, all the preceding discussion rests upon the assumption that there
must have been some explanation of the victim’s violent death. Certainly one of the
pressures on you to accept a conclusion is your wish to avoid the possibility that
ere might be no explanation at all of the results. Hence, you should have checked
option d as one correct answer. Similarly, if you’re going to avoid accepting this
conclusion, you can do so only by producing a more plausible explanation than the
explanation that Royce did it, so you should have checked option e as well. How
does all the talk about explanations relate to inconsistency? In the following very
simple way: It is by discovering or inventing alternative explanations that you avoid
he inconsistency involved in saying, “There must be some explanation, but I won’t
accept the only possible one.” If there’s only one possible one, and there has to be
one that s true, then that one is it.

4 Which of the following (a) represent correct usage of the terms and (b) are true'?
a ^ Logical means roughly the same as “rational” or reasonable

soundCreasonmr4er “ “ h'8h'y teCh"‘Cal SUbj'C' “ '° "* pnnCiPles °f

Premises infer conclusions.
People infer conclusions.
Some statements have implications.
Contradictions are statements made up of meantngless words
8 togicai™ “ °f C°nlradict‘ons and tnconsistencies is a mam goal of being
THE FOUNDATIONS OF ARGUMENT ANALYSIS 49

h It's important to avoid inconsistency if you want to have knowledge or be able

to communicate information to others.
i Criticism of arguments does not, whereas creation of arguments does, require
imagination.
j Creativity or originality can be distinguished from random novelty or mere
variation only by the use of critical standards,
k A logically sound argument might have false premises.
I A factually sound argument must be logically sound.
m If you took a course in technical logic, you would probably learn something;
the only problem is whether what you learn would have any relevance to the
abstract theory of argument.

Answer

Most of these statements require no comment; they just follow immediately from
material in the text, to which you should refer again if your answers were wrong. The
alternatives which you should have checked as correct are: a, b, d, e, g, h, j, k. The
only answer that needs any explanation is m, and the best way to explain why it
should have been marked incorrect is to say that it would have been correct if the last
part of it said: “would have any relevance to your capacity for practical reasoning,”
instead of what it in fact says.

5 Apply the steps of argument analysis, as best you can, to the argument we recon¬
structed in the second question in this quiz. (Hint: it’s a really simple example.)

Answer:

Step 1: Clarification of meaning. No problem here with the stated material.

Step 2: Identification of conclusions (stated and unstated). The implied conclu¬
sion is that Lipton’s Tea is the best, and if this is to have any significance for the
reader—as advertisements are supposed to—it must mean “best for you.”
Step 3: Portrayal of structure. This is a simple structure with one premise and
one conclusion, and we can abbreviate the somewhat long-winded presentation of
the premise by stating it as follows:

Professional tea-tasters prefer Lipton’s tea. (Premise)

Lipton’s tea is best for you. (Conclusion)

Step 4: Formulation of (unstated) assumptions. Now that we have the argument

set out in very simple and clear form, it’s obvious that it won’t work, that is, that the
conclusion won’t follow from the premises unless you assume something. The as¬
sumption is, roughly, that the taste of professional tea-tasters is a good guide to your
own taste. Actually, it’s a little stronger than this, because the conclusion is about
what is (supposedly) best for you, not what you’d like best. So there’s a second
assumption—that your taste is a safe guide to what’s good for you. (Or you might
50 CHAPTER 3

combine them if it’s assumed that the professional’s preferences are a good indication
of what’s best for you.) The structure, if we keep the assumptions separate, now looks
like this:

® + © + ©
s-v-'

©
Step 5: Criticism of the premises and the inferences. The premises now include
the “missing premises,” that is, the unstated assumptions that we’ve discovered are
needed in order to make the argument work. Since we’ve constructed these so that
they do make the argument work, we won’t have anything to criticize under the
heading of inferences, so let’s concentrate on our criticism of the premises.
The first premise is the one that was stated, and we really aren’t in any position
to deny it. Maybe it’s true, maybe it’s not; but, from the logical point of view, there
isn’t anything to be said about it. The main question is whether, assuming for the
moment that it’s true, the (implied) conclusion follows from it.
Well, it will follow if the assumptions on which it depends are correct. If it’s
true that the taste (the preference) of professional tea-tasters is a reliable guide to
what’s good for you, then the argument is just fine.
There are two problems about these assumptions. The first is that there’s a
distinction between the taste of professional tasters and that of ordinary tea-drinkers.
Very often, those who are professionally required to drink huge quantities of a sub¬
stance develop a quite different taste from those who drink it only occasionally; for
example, they will sometimes demand a tea (or a whiskey, a cheese, or a wine) that
is stronger, more distinctive, subtler, or smoother than the usual run-of-the-mill ver¬
sion. Whether you know this to be a fact or not, you can easily see that it’s a
possibility, which is all that you need in order to see that the assumption isn’t some¬
thing one can wholly rely on.
Second, people’s tastes are not necessarily a very good guide to what’s good for
them. For example, for a great many years Camel cigarettes were the most popular
brand, not only among the general smokers, but also, it is said, among the profes¬
sional tobacco buyers and blenders. But Camels were, in fact, not very good for you,
and indeed, appear to have been among the worst for whoever smoked them. Once
you’ve seen that the taste of the professionals is not necessarily a good guide to your
own taste, and also that taste (of anybody) is not necessarily a good guide to merit or
benefit, you have seen that there are serious weaknesses in the two missing premises.
And you have expressed some convincing criticisms of the argument. They may not
be fatal criticisms, because evidence might be produced to refute them, that is, to
establish the truth of the assumptions; but it’s not going to be easy, and, as far as I
know, it isn t available at the moment. So you have shown that the argument is a
very weak one at the moment. (And probably forever, since it’s unlikely that tannic
acid does your stomach much good.)
One might have gone about this criticism in a slightly different way. Instead of
adding missing premises, and thereby making the inference satisfactory, one might
have left them out and criticized the inference instead of the assumptions. Either
THE FOUNDATIONS OF ARGUMENT ANALYSIS 51

method is exactly the same in its net effect. It turns out to be a little more effective,
practically speaking, to try to formulate explicitly the assumptions that would be
required in order to make the argument good, than just to accept the stated premises
as all the premises there are and then try to criticize the inference. For “criticizing the
inference” is shooting at a rather hazy target; stating the assumptions forces you to
clarify the criticism considerably, and this clarification often turns up some new
loopholes (extra assumptions).
Step 6: Introduction of other relevant arguments. It’s possible that one might
want to introduce questions of cost, which is relevant to the conclusion of what’s best
for you, at least in a pragmatic sense. If we were talking about wines or spirits rather
than tea, we would surely have to consider cost, since it’s a substantial factor there.
However, this isn’t the kind of topic on which there are a large number of entries for
Step 6—unlike an argument about a topic like the best schools for you or your
children. (The comment about the probable general effects of tannic acid is a second
relevant, though speculative, consideration.)
Step 7: Overall evaluation. Here we have to employ some judgments. How
serious are these difficulties that we’ve raised with the missing premises? Are they
really just nit-picking points, or do they represent serious possibilities that make the
jump from the given premise to the implied conclusion a pretty speculative one? In
this particular case, one really has to conclude that they raise pretty serious possibili¬
ties and that the obviously intended implications of the quoted passage are not well
supported by it.

QUIZ 3-B

1 Literary prose, as distinct from practical prose:

a Can always be assessed in terms of style, originality, and other literary stan¬
dards
b Can sometimes be assessed in terms of the standards of reason—but not al¬
ways
c Can always be assessed in terms of the standards of reason
d Other
2 The power behind the throne of reasoning is said here to be the requirement of
consistency, which:
a Means the avoidance of contradiction
b Has little force for people that aren’t hung up on a highly scientific, rational¬
istic approach to life
c Can always be ignored, at some price
d May be recognizable only if one knows something about the facts
e Is one step less severe a constraint than the avoidance of improbability
f Obliges one to accept the conclusion of a sound argument, with all its assump¬
tions stated, if one accepts the premises

3 If you’re criticizing an argument, rather than giving one, the appeal to inconsis¬
tency often comes in because you are:
a Showing that it’s inconsistent to accept the premises and reject the conclusion
52 CHAPTER 3

b Showing that the conclusions (or the premises) are intrinsically inconsistent
c Showing that it is not inconsistent to accept the premises if you accept the
conclusion
d Other

4 Logic or reasoning:
a Does not generate new knowledge
b Is the only route to new ideas
c Requires imagination in order to criticize certain types of argument
d Often requires you to think up new explanations of particular facts, that is, to
be an inventor, to be original

5 Argument analysis involves:

a Either criticizing the premises of an argument or (alternatively) criticizing its
conclusion
b An early effort to find the conclusion(s) among the assertions in the argument
c Accepting whatever meaning the arguer gives to the terms being used
d Classifying every assertion in the argument as a premise or a conclusion
e Sometimes dividing sentences up into several assertions in order to reveal the
structure

QUIZ 3-C

1 A.E. van Vogt wrote a series of science-fiction novels about what he called the
“World of Null-A,” centering on heroic figures who had achieved the capacity to
think in “Null-A” (non-Aristotelian) terms. He defined these terms to mean
thinking which abandons the assumption that every claim has to be either true or
false (the Aristotelian Law of the Excluded Middle).
a Can you give some examples of claims that are neither true nor false? (Don’t
confuse this with the question of whether they are now known to be true or
false.)
b In general, what do you think can be meant by “alternative logics”? Are there,
could there be, totally different “ways of thinking”? Give an extract of a few
sentences from an imaginary tape-recorded dialog between your New Logic
hero and a supporter of the Old Logic (our present one). Or can’t this be done,
and if not, what is the significance of that?

2 Isn’t there some sense in which creativity does not involve critical skills? Make
the question specific; think of a particular type of creative person (writer, painter,
musician, scientist, or designer) and start defining what they do that establishes
them as creative. Does that same process or sequence involve critical ability?
Could there be instinctively original creative artists? If so, would their existence
settle the question?

3 How would you settle the question of whether studying technical logic helps one’s
practical reasoning? Suppose a study showed people who had taken the technical
THE FOUNDATIONS OF ARGUMENT ANALYSIS 53

course to be better, on the average, at practical reasoning than people who had
not. Would this finding settle the matter? Would a positive answer justify teach¬
ing technical logic?

4 Find a short example of reasoning in today’s paper, and run it through the Seven
Steps.
 Chapter 4

Developing the Concepts

of Argument Analysis

By now you have a general picture of the Seven Steps approach to argument
analysis. What remains is to become so familiar with it that it becomes an
instinctive approach; and to develop some refinements in it; and to learn
how to apply it to some special cases. There’s obviously nothing very techni¬
cal about it, it’s very much a matter of common sense. But acquiring real
skill with it is a matter of a great deal of practice and a much deeper
understanding of the steps that we have so far provided. If you do feel,
however, that you have a good grasp on what it’s all about, remember that
you can skip through to the end of this chapter and try yourself out on the
A quiz to see if there’s any need for you to work through the chapter in
detail. (If you do well on it, perhaps you can turn in your answers to the B
quiz right away, and skip some lectures while doing some extra reading or
working on a C quiz question for a term paper, depending on what arrange¬
ments for acceleration your instructor has made.)
You won’t ever be required to memorize the Seven Steps. In fact, you’ll
come to remember them without making any effort, because there aren’t
very many and you’ll get plenty of practice. But there’s no reason to feel bad
54
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 55

about going back and looking them up, even if you have to keep on doing
this indefinitely. Remember that some of the most sophisticated, highly
trained, skilled people, for example, airline pilots, use checklists as a basic
part of their operating practices, checklists which they never count on being
able to memorize and always have available. Whenever an airplane lands or
takes off, its pilot is required to go down the appropriate checklist, item by
item, just to be absolutely sure that nothing has been forgotten on this
occasion. There, the checklist is a matter of life and death. In argument
analysis, it's usually not quite that important, though when the arguments
are legal ones about capital crimes, or hearings about the adequacy of field
tests of a new drug, it is also a matter of life and death. And argument
analysis is very frequently crucial for decisions of great importance, involv¬
ing property and the welfare of others as well as yourself, and hence, it does
no harm to make yourself go back and check the list step by step, however
many times you’ve worked with it. It’s extremely important to realize at this
point that making yourself do argument analysis in a step-by-step, system¬
atic way, is exactly what leads to doing it better than if you did it in just a
casual way. The business of formulating the assumptions, for example, will
time and again make you realize that the argument depended on an assump¬
tion which you hadn’t noticed until you made yourself try to state it exactly.
In the same way, the procedure for “structuring” the argument, which forces
you to decide exactly which premise is related to which conclusion, and
which sentences are essentially irrelevant, will time and again lead you to
realize something of the greatest importance about the argument which es¬
caped your notice at first. It may, for example, lead you to realize that
there’s a secondary argument buried in the primary one, an argument which
throws a red herring across the track of the main argument and must be
separated from it in order to be able to cope with either effectively.
In this chapter, we’ll expand on a number of the concepts and proce¬
dures involved in the steps, with illustrations, and we’ll also introduce some
further refinements. In particular, we’ll have to work very hard on the basic
concept of the soundness of an inference and on how it is related to the truth
of premises and conclusions. Later chapters will focus on particular steps in
the argument analysis procedure and on particular types of argument, in still
further detail.

4-1

Let’s begin by reviewing the relationship between argument and inference.

The simplest possible argument consists of a single premise, which is as¬
serted as true, and a single conclusion, which is asserted as following from
the premise, and hence also to be true. The function of the argument is to
persuade you that since the premise is true, you must also accept the conclu-
56 CHAPTER 4

sion. This persuasion will be powerful if it is clear that the inference from the
premise to the conclusion is sound, that the premise does in fact imply the
conclusion. (Other ways to put this are to say that you can legitimately infer
the conclusion from the premise, or that the conclusion is in fact a conse¬
quence of the premise.) Since that argument rests for its strength entirely on
the strength of the premises and the strength of the inferences, these are
naturally the two points to attack, if one is operating in the critical mode; or
the points to reinforce, if one is operating in the expository mode (that is,
setting out an argument instead of criticizing it). Once you have finally
formulated an argument, these two lines of attack are entirely independent.
(Until then, you can always “convert” inferences into parts of complex as¬
sumptions, plus simpler inferences.) To say these lines of attack are indepen¬
dent is to say that the premises can be true or false and the inference still be
sound; and, on the other hand, the inference can be sound or unsound
whether the premises are true or false. For example, from the premise that
the moon is made of green cheese, it does in fact follow that the moon has a
high protein content. The premise isn’t true, but the inference is just fine. On
the other hand, from the fact that the moon’s surface consists largely of
rocks and dust, which is true enough, it does not follow (that is, an argument
which drew this conclusion would be invalid, unsound, no good) that the
moon is uninhabitable. Nor does it follow that the moon is inhabitable. The
question of whether it’s habitable or not (by our species) depends primarily
on its atmosphere, not on its physical composition.
So any combination of the premise being “good” or “bad” (that is, true
or false) with the inference being good or bad is possible. But, if we bring in
the question of the truth or falsity of the conclusion, then, although we can
still have various possible combinations, one combination is ruled out. The
whole point of arguments, you remember, depends on the fact that if you go
from a true premise by means of a good inference, you have to arrive at a
true conclusion. (We can modify this a bit to say that if the premise is just
probable, then the conclusion will be just probable.) Flence the one combi¬
nation that is ruled out is the combination of true premise(s) with a good
inference and false conclusion(s).
Becoming clear about this point is really the crucial step in becoming
clear about what reasoning is as opposed to what truth is, which is really the
distinction on which logic rests. It won’t hurt to go through this step once
more. If you have a true premise and a good (sound) inference, then—by
definition—they will lead to a true conclusion. This follows by definition
because what we mean by “good inference” is exactly “one that will lead us
from true premises to true conclusions.”
Hence it follows that the one combination we can never encounter is
that of having true premises, a sound inference, and a false conclusion.
The big trap, when one begins to think about this, is that of believing
that you can tell that an inference is no good by noticing that it leads to a
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 57

false conclusion. Suppose I make a number of claims about the results of

some field trials on beer-drinkers. On the basis of these measurements about
the amount of beer drunk, the length of time that digestion takes, the rela¬
tive rates of absorption of alcohol, sugars, and the other substances con¬
tained in beer, the rate at which the fluid passes through the system, and so
on, I draw the conclusion (as was done in a famous published article) that it
is impossible to get intoxicated by drinking beer. Now you may know from
your own experience or observations that the conclusion is false. You cannot
conclude from this that the inference I made was unsound. It might instead
have been that some of the premises were false—that is, it may be that a
number of the observations were incorrectly made or recorded. Remember,
then, that you can’t tell the source of the trouble in an argument from the
discovery that the conclusion of the argument is false; it may have been due
to a mistake in inference, or to a correct inference and erroneous premises,
or to both.
Nor can you tell from the fact that an inference is no good that the
conclusion will be true or false. Bad inferences can lead to true conclusions
or to false conclusions, since bad inferences can lead anywhere. That is, if
you stick any wholly irrelevant conclusion in at the end of an argument, and
say that it is the conclusion of the argument, the inference to it will be
unsound; and of course, it might have been a true or false irrelevant state¬
ment.
Nor, of course, can you tell from the truth or falsity of the premises
alone either whether the inference in the argument is good or whether the
conclusion is true or false.
Let’s sum it up, then, by reminding ourselves of the whole point of logic
and reasoning. Reasoning is the process whereby we get from old truths to
new truths, from the known to the unknown, from the accepted to the debat¬
able (in order to make it less debatable), and so on. If the reasoning starts on
firm ground, and if it is itself sound, then it will lead to a conclusion which
we must accept, though previously, perhaps, we had not thought we should.
And those are the conditions that a good argument must meet; true premises
and a good inference. If either of those conditions is not met, you can’t say
whether you’ve got a true conclusion or not.
The other side of this coin is worth noting. If you know the conclusion
is false, then you know that something went wrong: either the premises were
false or the inference was unsound. Hence, if you know that the conclusion
is false and that the inference was sound, the premises must have been false;
or, if you know that the conclusion is false and that the premises are true, it
must be the case that the inference was unsound. This fact gives us one of
our best procedures for checking on the soundness of inferences.
Let’s see if we can sum up these points in a table.
First let’s set out all the possible combinations and what we can say
about them. Then we’ll condense them into a table.
58 CHAPTER 4

1 If the premises of an argument are T(rue) and the inference (the

reasoning) is S(ound), then the conclusion must be T(rue).
(This is the main justification for improving one’s reasoning skills.)
2 If the premises are, on the other hand, F(alse), and the inference is
still S(ound), the conclusion may be T or it may be F—no way to tell.
3 If the premises are F and the inference, on the other hand, is
unsound, the conclusion may be T or F.
4 If the premises are, on the other hand, T while the inference is still
U(nsound), the conclusion may be either T or F.
Now suppose that we happen to know something about the truth of the
conclusion. Can we—or when can we—say anything about the truth of the
premises?
5 If the conclusion is F, and the reasoning is S(ound), then we know
the premises must have been F (because, if they had been true, the conclu¬
sion would have had to be true, from point 1 above).
6 On the other hand, if the conclusion is T, and the reasoning
S(ound), we can’t say whether the premises are T or F. (This one is a bit
surprising to most people.) How could one possibly get a T conclusion by
sound reasoning from an F premise? Surely, one feels, any falsity in the
premise would carry through into the conclusion? Not so; the premise may
be false because it claims more than is true: Vienna is 1,000 miles from
Milan, from which there certainly follows any more modest claim; for exam¬
ple, “Vienna is more than 100 miles from Milan” certainly follows from (is
implied by) the assertion that it’s 1,000 miles from Milan; and the more
modest claim may be true (as in this case). So false premises may imply true
conclusions, and so, of course, may true ones. Hence, the opening statement
of this paragraph is correct.
7 If the conclusion is T, and the reasoning U(nsound), the premises
may be either T or F. (This is obvious enough, since any statement whatso¬
ever, whether T or F, can be the premise of an unsound argument with a T
conclusion).
8 If the conclusion is F, and the reasoning is U(nsound), the premises
may be either T or F (for the same reason as in point 7).
Finally, let’s suppose that we have information about the truth or falsity
of the premises and of the conclusion. What, then, can we say about the
soundness of the inference? As before, there is only one case in which we can
say something definite.
9 If the premises are T and the conclusion is F, then the reasoning
must have been U(nsound) (from the definition of sound inference). Here is
a key way to check out the reasoning in an argurment, if we happen to be in
possession of the required knowledge about the truth or falsity of the prem¬
ises and conclusion. Of course, since an argument is usually employed to
help us decide whether the conclusion is T, we usually do not know in
advance that it is false, and we have to use other standards to tell us whether
the reasoning (and the inference) was sound. We’ll get to them later.
 DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 59

10 If the premises are F and the conclusion is F, the reasoning might

have been S(ound) or U(nsound). (Take any two unrelated false statements
you like, call one the premise and the other the conclusion, and you have an
example of unsound reasoning that fits this pattern. But the inference from 5
= 6to5 + l= 6+ lis Sound, though both premises and conclusions are
F; so either option is open.)
11 If the premises are F and the conclusion is T, you can’t say
whether the reasoning was S(ound) or U(nsound). (Look at points 3 and 2
above and make up your own example.)
12 If the premises are F and the conclusion is T, you still can’t tell if
the inference was S(ound). (For example, the premises and the conclusion
might be about entirely different topics, in which case the reasoning would
be unsound; or they might be components of a perfect argument.)

The table follows; the facts you know at first are underscored.

Premise Inference Conclusion

* 1 T S T
2 F S ?

3 F u ?

4 T u ?

* 5 F s F
6 ? s T
7 ? u T
8 ? u F
* 9 T u F
10 F ? F
11 F ? T
12 T ? T

The asterisks in the table indicate the only cases

where we can make a definite decision. If you look at them
carefully, you’ll see they are really just different ways of say¬
ing the same thing, or expressing its consequences; that a
good inference from a true premise leads to a true conclu¬
sion. (As we've mentioned before, you have to qualify this
slightly if you want to talk about probably true statements,
but the general pattern remains the same.)

So far, we’ve mainly talked of “basic arguments,” that is, arguments

with a single premise and a single conclusion. What we’ve said applies
equally well to complex arguments, those with more than one premise and/
or conclusion, provided that you remember one thing. If there’s more than
60 CHAPTER 4

one premise, for example, they are all being asserted to be true jointly;
hence, even if only one of them is false, the joint assertion of all of them is no
longer true. So, to take an easy illustration, if we have an argument with
several premises of unknown truth which, by means of sound reasoning, lead
us to a conclusion which is false, we know that there’s something wrong with
the premises. If there was only one premise, we’d know that it had to be false
(see the above table). If there are several premises, all we can conclude is
that at least one of them must be false. We can’t conclude that all of them
are false, nor that only one of them is false. You can read the table in just the
same way, as long as you remember that where it says that the premise (or
the conclusion) must be false, with a set of several premises or several con¬
clusions, it means that there is something wrong with the set as a whole; that
is, at least one of the component statements is false.
Another type of more complex argument has already been illustrated in
the last chapter. This is what we can call a “chain argument,” that is, an
argument where the conclusion from one argument becomes the premise for
another. The simplest possible form of this is an argument with a single
premise which leads to a single conclusion, which in turn leads to another
conclusion. The only thing you have to remember about this kind of argu¬
ment is that the conclusion of the first subargument is a premise in the
second subargument. Hence, you should treat it as consisting of two separa¬
ble arguments. These will be covered by different lines in the table above,
depending on which argument you’re looking at in detail. Notice that the
simplest possible chain argument is quite different in structure from an argu¬
ment with two premises and one conclusion, essentially because it involves
two inferences instead of one, and hence is open to attack on the soundness
of its inferences at two points instead of just one.

4-2

If you have a firm grasp on the contents of the preceding section, you will
have no difficulty in mastering the following distinction, which is quite im¬
portant when you want to speak precisely about arguments. Suppose that we
want to talk about the inference in an argument, as opposed to the premises
in it, or the argument as a whole. Suppose that we let the letter p stand for
the premise(s) and the letter q stand for the conclusion. Then the argument
as a whole involves at least the assertion of the premise p and the assertion
(explicit or implicit) that the premise implies q. It may also involve the
assertion of q, or it may just leave this as an unstated implication of the two
preceding claims. The point of the argument is, of course, to establish q. We
do it by marshalling, in its support, the premise p and the claim that p
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 61

implies q. There’s an admirably brief way to express an argument of this

basic form, which is, “p, so q” (or “q, because p”). In this very concise
expression, the part that is implicit is the inference itself.
Now notice the crucial difference between “p, so q” and “If p, then q.”
The first asserts that p is the case, and it asserts that q is the case. The second
does not assert that either p or q is the case. The second only makes a
hypothetical claim; on the assumption (hypothesis) that p is true, then q
would have to be true. The first claim, the most concise possible expression
of an argument (“p, so q”), actually involves the second as one of its compo¬
nents—in this case, as an unstated component. But it goes much further; it
also asserts the truth of the premise(s) and the truth of the conclusion(s).
It follows from all this that you have to be very careful not to attack the
statement “If p, then q” on the grounds that p isn’t the case, or that q isn’t
the case. The hypothetical statement in quotes (sometimes also called a
conditional statement), asserts only that there is a connection between the
truth of p and the truth of q; this is the sort of connection whose meaning is
illustrated in the table above. It says only that if p is true, then q has to be
true; but it doesn't say that p is true, or that q is true.
Conditional or hypothetical statements are a way of expressing the in¬
ference on which an argument depends. They involve a much weaker claim
than the claim involved in an argument, because they assert only that a
connection exists between two possibilities and not that either of them is an
actuality.
Now a conditional statement (“If p, then q”) may express a connection
which is true as a result of the meaning of the terms involved in p and q,
which would make it a “logical (or definitional) truth”; or the connection
may be one which depends for its truth upon laws of nature or particular
facts about the world, in which case it is a scientific or empirical truth. And,
as we’ve previously mentioned, the line between these kinds of truth isn’t a
sharp one; they are just conveniently distinguished.

4-3

We’re now in a position to clarify a pair of terms that are common enough
in the language, extremely useful in assessing arguments, definitions, scien¬
tific relations, and so on. These are the terms “necessary condition” and
“sufficient condition.” To say that p is a sufficient condition for q is to say
that the truth of p suffices to guarantee the truth of q, or that “If p, then q.”
As a simple example, let’s say that p might be the claim that a supertanker
has broken up in a storm in the western Pacific, and q might be the claim
that the rainfall west of the Rockies in the United States will go down by 50
62 CHAPTER 4

percent in the ensuing nine-month period. It has recently been suggested

that one of the really big supertankers—the ones displacing more than a half
million tons—could spill a big enough oil slick to cover some millions of
square miles, and that this spill would sufficiently affect the evaporation rate
so that the water intake into the cloud system would be very seriously af¬
fected. That cloud system, originating in that particular area of the Pacific, is
what feeds the West Coast with moisture. Hence, it is being asserted that a
supertanker breakup in that area would cause the loss of a great deal of
rainfall (and probably also nearly all the crops, including those dependent
on irrigation, since the irrigation comes from the snow pack which also
depends on precipitation). Here’s a case of an (alleged) sufficient condition.
A certain state of affairs is, or is said to be, sufficient to produce another
state of affairs.
The p here would be the assertion that a state of affairs occurred since
we have arranged to use p as an abbreviation for an assertion. But we
normally say that the state of affairs to which p refers is a sufficient condi¬
tion for that to which q refers. These are called, respectively, the antecedent
and the consequent conditions or circumstances. The “assertion” p is of
course only a hypothetical assertion here, and is sometimes called the hy¬
pothesis', the assertion q is said to be a consequence of that hypothesis. Of
course, this is an empirical sufficient condition and not a purely logical one
(that is, one that depends purely on the meaning of words). Even though it’s
a disputed question whether there are really any “purely logical” sufficient
conditions, it’s clear that there are some that come so close that it’s a useful
approximation to treat them as being purely logical; for example, it would
ordinarily be said that having a daughter is a sufficient condition for being a
parent, in virtue of the very meanings of the words involved.
A necessary condition, on the other hand, is a condition which must
obtain (is necessary) in order that another condition obtain. Someone might
say that it’s necessary (or essential) to provide typhoid patients with enor¬
mous quantities of fluid, in order that they will survive. One can express this
by saying that a necessary condition for survival of a typhoid patient is the
intake of very large quantities of liquid. (Thousands of deaths resulted from
the failure to recognize that it was the dehydration resulting from the dys¬
entery associated with typhoid fever that was causing the deaths of many
victims.)
Necessary and sufficient conditions are not, of course, the same. Provid¬
ing a great deal of water for typhoid patients doesn’t guarantee, that is, isn’t
a sufficient condition for, their survival, but it is a necessary condition for it.
The breakup of a supertanker in the western Pacific isn’t a necessary condi¬
tion for a catastrophic reduction of rainfall in the western United States,
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 63

since other things, such as a shift in the jetstream, could also cause it, but it
is (allegedly) a sufficient condition for it. Banning the sale of spray cans
containing Freon may not be a sufficient condition to prevent the destruc¬
tion of the ozone layer in the atmosphere but (it is often argued) it is a
necessary condition. That is, we have to do that, it’s necessary to do that, in
order to prevent that disaster; it may not be enough, but we must do it or
else we can’t avoid that result.
All these assertions about necessary and sufficient conditions can be
expressed using the terminology of “implies,” “consequences,” “can be in¬
ferred from,” and the like, but they are useful abbreviations for what is often
a complicated circumlocution expressed in those other terms. Wherever it’s
true that q follows from p, it’s true that |p {and the state of affairs referred to in
p) is a sufficient condition for q {and the state of affairs referred to in q).
Which leads us to an interesting discovery. If p implies q, which means
that p is a sufficient condition for q, then p guarantees the truth of q, that is,
it makes the truth of a necessary. Hence, p can’t be true unless q is true, so
q is a necessary condition for p. So, whenever p is a sufficient condition for
q, q is a necessary condition for p. Reread that sentence, because it shows
the way in which these concepts are completely complementary. Notice that
it does not say that if p is a sufficient condition for q, then p is a necessary
condition for q. It says that if p is a sufficient condition for q, the q is a
necessary condition for p. The terms are opposites in the same way that
“taller than” and “shorter than” are. If Jones is taller than Smith, then
Smith is shorter than Jones. If p is a sufficient condition for q, then q is a
necessary condition for p. (The technical term for these is “converse rela¬
tionships.”)
Referring back to our previous examples, we can see that, while the
supertanker spill is not a necessary condition for the reduction of rain, the
reduction of rain is (allegedly) a necessary condition for the supertanker
spill. That is, according to the claim being made here, the reduction of rain
is a necessary consequence of the spill, that it must necessarily occur if the
spill does occur; that is, it is a necessary condition for the spill. This sounds
a little odd, in this particular case, because the reduction in rainfall will
occur after the spill, and we usually employ the necessary and sufficient
condition vocabulary to refer to conditions which occur before or concur¬
rently with whatever they are said to be necessary and sufficient conditions
for. But that’s not an essential part of their meaning, it’s just usually so. The
exact way to formulate the condition is to say that the reduction in rainfall
over a certain period is a necessary condition for the spill to have occurred
earlier. In the same way, when we’ve said that the ingestion of a very large
quantity of water is a necessary condition for recovery from typhoid, it
follows automatically that recovery from typhoid is a sufficient condition
64 CHAPTER 4

from which we can infer that there was at an earlier stage a substantial
intake of water.
When we turn to examples of logically necessary and sufficient condi¬
tions, we don’t have to worry about these somewhat awkward sounding
changes of tense. If we assume for the moment that arithmetic, and mathe¬
matics in general, is a rigorous, logical development from the definitions of
the basic concepts of mathematics (the usual view), we can give plenty of
examples of necessary and sufficient conditions from arithmetic. For exam¬
ple, it’s a sufficient condition for being divisible by 3 that a number should
be divisible by 9; but of course it isn’t a necessary condition. Since it’s a
sufficient condition, however, the converse must be a necessary condition;
let’s see if this is true. Is it a necessary condition for being divisible by 9 that
a number be divisible by 3? Of course it is.
There are many cases where a condition is both necessary and suffi¬
cient. For example, it is both a necessary and a sufficient condition for being
divisible by 9 that a number should be divisible by 3, leaving a dividend that
is itself divisible by 3. If either of these conditions is met, then the other is
met. Each is a necessary and sufficient condition for the other. They are
sometimes said to be “equivalent conditions.” Conditions that are both nec¬
essary and sufficient for each other occur in the empirical realm as well,
though they are somewhat less common there. In the field of medicine, a
pathognomonic symptom is one which is a unique identifier of the particular
disease or condition. If the condition is present, then that symptom is pre¬
sent; if that symptom is present, then the disease is present. Each is a neces¬
sary and sufficient condition for the other. (Most symptoms are neither suf¬
ficient nor necessary conditions for the disease of which they are symptoms,
but the collection of symptoms, known as a syndrome, which definitely iden¬
tifies the standard form of the disease, is a sufficient condition and some¬
times is a necessary condition.)
The most useful employment of terms like “necessary” (and “suffi¬
cient”) conditions in the empirical realm occurs when we wish to make a
distinction between the relationships of certain conditions to an effect. It is
often very important to make it quite clear that, whereas the provision of a
certain type of medication is essential (necessary) in order that the patient
should recover from a certain disease, it by no means guarantees (is suffi¬
cient for) recovery. Similarly, in talking about the analysis of circuits or
economic conditions, we often need to be very careful to distinguish suffi¬
cient conditions for a certain state of affairs from necessary conditions for it.
We started off this discussion of sufficient and necessary conditions by
using the letters p and q to stand for statements or assertions. Along the
way, we occasionally used them to stand for states of affairs, and even for
properties. There are certain circumstances in which it’s really important to
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 65

distinguish statements, for example, from properties, but this doesn’t happen
to be one of them. When talking about sufficient and necessary conditions,
one can always convert the possession of a property into an equivalent
statement to the effect that that property is possessed by the object in ques¬
tion, and vice versa. So we can say either that certain states of affairs whose
occurrence is expressed by statements, or that certain properties, possession
of which is a state of affairs, are necessary or sufficient conditions for other
states of affairs or properties. The language is equally useful in either case.

4-4

You are now in possession of a fairly precise vocabulary for the major part
of argument analysis. It should be stressed again that some of the distinc¬
tions that we will be using, and have already introduced, are not absolutely
sharp ones. We have pointed out that the distinction between factual incon¬
sistency and logical inconsistency isn’t a sharp one. We’ve talked about the
way in which some “If p, then q” statements (conditional statements) are
statements of implications, whereas others are statements of causal connec¬
tion and others are not even clearly that. Those distinctions aren’t sharp,
either. Nevertheless, you can see that a conditional statement is sometimes
telling you about an inference that you can make, and on other occasions it
seems to be telling you something about the connection between possibili¬
ties. In the context of a complete argument or a conversation in a particular
setting, you will rarely find any difficulty in making the distinction, so don’t
worry that it isn’t possible to give an abstract definition of the distinction.
But you need to be on your guard against another (related) word or two
that can mislead you seriously in the analysis of practical prose. Just as the
form of words “If p, then q” is often, but not always, an indication that
somebody is telling you something about what can be inferred from what, so
the form of words “q, because p” is often an indication that q is the conclu¬
sion for which p is the reason—but not always. Sometimes it’s just telling
you what the cause of a particular event was. For example, “I got wet
because the top on the car is leaking” isn’t telling you about a premise from
which you can infer a conclusion; it’s telling you about a physical cause for
a physical effect. “He picked up venereal disease because he would insist on
joining every swinging group in range” cannot be reconstructed in the form
of an argument with a statement about the man’s activities as the premise
and his contracting venereal disease as the conclusion; the first didn’t “lend
support to” the second, or even make it very likely it produced it or caused
it. (It only increased the possibility of it, not necessarily to a very high level.)
The very word “because” obviously has an origin that is connected with the
origin of the term “causation”; but it has become one of the principal words
66 CHAPTER 4

that we use in order to indicate the connection between reasons and conclu¬
sions. And there are plenty of intermediate cases where it’s very hard to say
whether “because” is being used to indicate a cause or a reason, but again
that doesn’t prove to be very bothersome in the full context of an argument.
And you can always cover yourself in argument analysis by considering both
alternatives explicitly. The main point to remember is that you have to keep
your eyes open to see whether we’re looking at a case which is clearly one
where a mere causal connection is being asserted, or a case where we clearly
have a reason that is being put forth for a certain conclusion.
The final irony is that even the word “reason” is sometimes used to
indicate a cause. “The reason why the cooking stove in the boat won’t work
is that somebody put kerosene into it instead of alcohol.” This is really a
factual claim, a causal claim; it’s just like “He died of a heart attack.” It
doesn’t give you reasons for you to accept the conclusion, it gives you an
explanation of an event. Most heart attacks are not fatal, but this one was.
Compare that with “I think this stove was made in Switzerland, because
it has the word “Berne” stamped on the base, though the rest of the stamp
marking is pretty well impossible to read.” Here, somebody is giving you a
reason for drawing a certain conclusion, the reason being that the name of a
Swiss city is stamped on the stove, and one of the missing premises is, of
course, that it’s quite common for manufacturers to stamp the place of
manufacture on their products. (What’s another missing premise in this ar¬
gument? Quiz 4-B, 5.)
In general, you won’t have too much difficulty with identifying real
arguments for argument analysis, as opposed to complicated descriptions
and case histories, as long as you constantly bear in mind the importance of
getting a general sense of the purpose or function of the whole passage
before you start the analysis. Don’t lose the woods for the trees, don’t be
misled by the words like “because” or “reason” or “since” or “if . . . ,
then. . . .” None of these will guarantee that you’re looking at an argument.
They should make you prick up your ears because they are often used in
arguments, but they don’t guarantee the presence of an argument; they’re
not a sufficient condition or a necessary condition for the presence of an
argument.

4-5

Let’s look at this question of identifying arguments in a slightly broader

framework. One of the quiz questions at the end of Chapter 3 required you
to go out and find an argument for criticism. Some of you, if you did that,
will undoubtedly have brought back material that looked like an argument
to you, but you couldn’t sell the instructor on it. Let’s look at a few of the
usual mistakes that people make in identifying arguments.
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 67

First, they tend to pick on something that is argumentative, that is,

controversial or emotional. But often people who are talking about a dispu¬
tatious subject do so by means of a long harangue, which is merely emo¬
tional discourse “expanding on the theme” of their point of view, but which
doesn’t provide arguments (not even bad ones) for the conclusion which they
are espousing. Patriotic speeches, lengthy condemnations of the evils of our
times, often look like very good material to the eye of the inexperienced
argument analyzer. But they frequently don’t pass the test of providing rea¬
sons for a conclusion, reasons that can be separated from the conclusions.
Sometimes, though, you can work out an implicit conclusion and reconstruct
them as arguments. Keep asking yourself whether what you are about to
identify as the premises of the argument could be established independently
of the conclusion. If they can’t be established independently, then they’re
really just another way of expressing the same facts, and all you have is a
kind of verbal translation of one sentence into another. Here’s a letter in a
national magazine, commenting on an article it had published on aviation
safety. The writer says that the article is:

The most distorted, ill-conceived and damaging piece of sensationalist journal¬

ism I have ever read. As an air-traffic-control specialist as well as a commercial
pilot, I feel qualified to comment. Through the efforts of many knowledgeable
and dedicated professionals within the aviation industry, a rapidly increasing
number of people and organizations are recognizing air transportation as the
safest, most efficient and most economical means of moving people and cargo.

The writer is obviously in an argumentative mood. But what is he presenting

in the way of an argument? He is saying that he personally thinks the article
was very bad, and he does know something about the subject. Now you
can’t write controversial articles without somebody disagreeing; that’s what
it means to call them controversial. And they would hardly qualify as con¬
troversial if the only people who disagreed were people with absolutely no
knowledge or experience in the areas. So the fact that this person, with
considerable experience in the area, disagrees doesn’t give us a reason for
rejecting the arguments in the article. The next sentence of the letter is really
just another plug for aviation, and all it does, again, is to say that a large
number of people think air transportation is great. Did the original article
say that it wasn’t great, or just that it had a lot of serious things wrong with
it which ought to be corrected? We don’t know and, in the absence of any
specific quote from the letter writer, what he says here doesn’t appear to
count against anything that you can tell was in the article; hence, it isn’t an
argument against the article.
As a matter of interest, the letter, analyzed as straight prose, makes a
number of extremely dubious statements; for example, the claim that air
68 CHAPTER 4

transportation is the safest means of moving people. Even the Air Transport
Association’s own statistics place trains ahead of planes as safe carriers.
The letter illustrates an important point to watch when you’re trying to
develop criticism. Don’t confuse it with letting off steam, with merely an¬
nouncing that you and perhaps many others disagree. Nobody’s going to be
very surprised by that. What the writer should have done, of course, is to
show that some claim or passage or general implication of the article is in
fact wrong by appealing to independently gathered or faultlessly objective
evidence. Indeed, if you think carefully about the piece, you’ll realize that
his very appeal to experience contaminates him as an independent judge. It
turns out that he is an air traffic control specialist, that is, somebody who
earns a living from the success of the airlines. However conscientious people
may be, it’s hard to be sure that they will be unbiased in assessing claims
about the industry which supports them. Similarly, the fact that the letter-
writer is a commercial pilot also shows that he has considerable investment
in the continuing success of aviation. The fact that he disagrees with an
attack on its safety may well show only that he feels severely threatened by
the attack, not that his experience objectively shows the attack to have been
wrong. As we’ll see when we get to a more detailed discussion of bias, one
can’t act as if everybody involved in or dependent on a field is automatically
biased about it, but one also can’t act as if the connection doesn’t introduce
a serious possibility of bias, and hence, a serious weakness in this kind of
“argument,” which merely consists in giving one’s own opinion.
Reflecting further on the piece, we see that, apart from the letter-writ¬
er’s certainly false claim about the safety of air transportation, there is an
implied claim that an increasing number of people are using it. But the
statistics show that the airlines are in trouble, apparently because of a de¬
crease in use rates per 1,000 head of population (which is the appropriate
index, incidentally, rather than absolute numbers).
Thus, even though he doesn’t present anything that really qualifies as
an argument, it’s possible for us to do some criticism of the ways in which he
fails to present an argument, so the example is not without some potentiality
as an exercise. But it certainly isn’t an example of a good, or even good-
looking but fundamentally fallacious, argument.
The Acting Administrator of the Federal Aviation Administration, the
key agency in this area, also wrote in to complain about this particular
article. And, for a change, he quotes a passage from the article; he then
produces facts to refute it, the right approach. It turns out that he doesn’t do
very much better, although his approach is a very much better one. Here’s
what he says;

I take particular exception to Gonzales’ gratuitous remark that generally “the

F.A.A. won’t issue an air-worthiness directive (which has the force of the law)
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 69

until something terrible happens. As a matter of fact, the F.A.A. issues or

revises around 200 air-worthiness directives (A.D.’s) every year as a result of its
regular surveillance program, . . .

That looks like a good, straightforward, honest refutation of a careless

generalization, doesn’t it? But as you look at it really carefully, you notice a
couple of words which weaken its impact a good deal. Before you read any
further, look at the passage and see if you can identify the two key qualify¬
ing terms that greatly weaken this attempt at refutation. Both key words
appear outside the passage that is being quoted from the original article. The
first of them is “generally,” from which we can infer that Gonzales did not
categorically state that the F.A.A. won’t issue an A.D. “until something
terrible happens,” but that he only said that this was generally or usually the
case, presumably meaning that this was more often than not the case. Now
what has the Acting Administrator got to do to refute this? He has to show
that most of the A.D.’s that are issued by the F.A.A. are based on something
other than a terrible occurrence. What does he in fact show? Assuming for
the moment that the facts are exactly as he describes them, “the F.A.A.
issues or revises around 200 A.D.’s every year.” So the second term that
weakens the apparent refutation is the term which I’ve italicized in this
quotation, the word “revises.” If Gonzales was talking about issuing A.D.’s
and if the Acting Administrator is talking about issuing or revising, the one
claim hardly refutes the other. They are about categories of A.D.’s that
overlap to an unknown extent, but might overlap only to a very small extent.
That is, most of the 200 A.D.’s that change in some way every year may be
revisions. The case against Gonzales is looking weaker.
Nor is that the end of the trouble. Remember that Gonzales’s basic
claim, as far as we can infer from these letters which are intended to be read
by people who don’t have the original article in front of them, is that it takes
“something terrible” to get the F.A.A. to produce an air-worthiness direc¬
tive. The reply from the Acting Administrator would be a relevant counter¬
argument if there were many fewer than 200 terrible events in the aviation
industry every year. It looks as though the Acting Administrator is operating
on that assumption because if there are, for example, 400 air disasters each
year, the fact that 200 A.D.’s are issued hardly shows that it doesn’t take
something terrible to produce one. But there are more than a few hundred
crashes, near-crashes, near-misses, or structural failures in the course of a
year in the aviation industry, so that figure of 200 doesn’t cut much ice as a
counterargument. The best interpretation one could make in favor of the
Acting Adminstrator is to say that the key point here, although it’s not well
brought out, is that these 200 A.D.’s come from the F.A.A. “as a result of its
regular surveillance program,” and that a “regular surveillance program”
has nothing to do with the kind of event that we would refer to as “some-
 70 CHAPTER 4

thing terrible. That may be so, but it’s not too obvious, since one woul4
think that the regular surveillance program would involve, amongst other
things, investigation of wrecks and structural failures which could be re¬
garded as something terrible.
So here are a couple of examples of careful reasoning at work, the first
devoted to showing that what looked like an argument really was hardly
worthy of the name, whereas the second one showed that what looked at
first sight like a good tight criticism was in fact very weak (which is not to
say that further investigation might not show that it could be improved).
Our verdict must be, on the evidence available, that the first letter really
makes no significant contribution to the discussion at all, and that the sec¬
ond one raises a point which it certainly fails to show is a conclusive argu¬
ment against the original article.
You 11 have further opportunities to try out your skill in distinguishing
arguments that are worth examining from those that aren’t, and in particu¬
lar from those that are mere rhetoric, in the quizzes at the end of this chap¬
ter. One impression that you should not pick up from this discussion con¬
cerns the relevance of emotional and “loaded” language to logic. It’s often
thought that logic is somehow a cool and precise discipline, in which emo¬
tion has no place. On the contrary, there are many occasions where reason
forces one to the conclusion that an outrageous crime has been committed
and should be condemned as such, or to one that fully justifies enthusiastic
and affectionate support of some person or resolution. In the same way, it’s
a complete mistake to automatically criticize arguments on the ground that
they use emotional or highly loaded language. That language serves a func¬
tion, it has a meaning, and whatever that function or meaning, it may be
appropriate or inappropriate in certain circumstances. The task of the rea-
soner is to determine whether, and when, that language is appropriate to the
evidence or the situation, and to criticize it only if it’s not appropriate. Most
of the really interesting uses of reasoning center on highly controversial and
emotional issues; they’re interesting not just because it’s often harder to
handle this kind of argument but also because it’s often more important,
since issues where people get upset and deeply involved are likely to be more
crucial.
Another point to remember is that the fact that a passage may consist
intrinsically (that is, taken just by itself) of a rhetorical tirade does not justify
supposing that it shouldn’t be reconstructed as an argument; after all, it’s
quite likely that it is, in that context, meant to support a particular conclu¬
sion. If you can see what that conclusion is, then state it, and begin to ask
yourself whether the rhetoric supports it or merely reaffirms it.
Perhaps you are now beginning to see just how much importance there
is to the first two steps of argument analysis (“Clarify the meaning” and
Identify the conclusion”). It’s often not easy to do these, and the whole
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 71

apparatus of your analysis depends entirely on getting these steps right,

although not necessarily at the first try. The crucial point to remember is
that both meaning and conclusion must be worked out by consideration of
the context of the passage, not just by looking narrowly at the words in it.
You are an extremely sophisticated language user, and you have no trouble
at all, in practical terms, in sensing that somebody who tells you about the
experts’ opinion of Lipton's Tea is trying to sell you the stuff. What is a little
harder is the business of formulating exactly what the implicit message is.
We re in the business of judgment here, and there will be times when what I
have to say in this text will strike you as not the most plausible interpreta¬
tion of the passages that I am quoting. You may well be right about this; ask
somebody else to give you an opinion. Write it up from the point of view
that seems best to you. Remember that in many actual situations, you’d be
able to confirm which interpretation was correct by talking to the person
who had spoken or written the passage, or by reading or listening further.
Additional disagreements between us about the relatively short passages
that we deal with in a text would likely diminish with this further exposure.
But, even if we disagree now and again, perhaps on important points, it
doesn't mean that we won't agree about most of the important points in
analyzing arguments. Don’t carry away the picture that a single error in
argument analysis is always fatal. It is much more likely to be fatal if it
occurs at the point where you’re trying to get a general sense of the meaning
of the passage. But even then, you may pick up a number of valuable criti¬
cisms further down the Seven Steps.

4-6

Now it’s time to introduce you to what we might call the ethics of argument
analysis. The dominant principle here is what we can call the Principle of
Charity. The Principle of Charity requires that we try to make the best,
rather than the worst, possible interpretation of the material we’re studying.
That is, even if, as a matter of strict grammar, we could shoot the writer
down for having said something that doesn’t follow or isn’t strictly true, it
may be more charitable to reinterpret the passage slightly in order to make
more “sense” out of it, that is, to make it mean something that a sensible
person would be more likely to have really meant. We’ll do this all the time.
It doesn’t mean letting people off the hook entirely by assuming they
couldn’t possibly have meant something just because it turns out to be un¬
sound or untrue; most of us make such mistakes quite often. What the
Principle of Charity does mean is that “taking cheap shots” is something we
shouldn’t waste much time doing. Other words that come in from ordinary
language about this point are “nit-picking” and “attacking (or setting up) a
straw man.” These terms all refer to poor argument analysis, either to mak-
 72 CHAPTER 4

ing irrelevant criticisms or to making criticisms that are not relevant to the
main thrust of the argument or that are unfair in some other way.
The Principle of Charity is more than a mere ethical principle, but it is
at least that. It requires you to be fair or just in your criticisms. They can be
expressed in heated terms, if that is appropriate; they may involve conclu¬
sions about the competence, intellectual level, or conscientiousness of the
person putting forward the argument, all of which may well be justified in
certain cases. But your criticisms shouldn’t be unfair; they shouldn’t take
advantage of a mere slip of the tongue or make a big point out of some
irrelevant point that wasn’t put quite right. So the Principle of Charity does
coincide with good practical advice about powerful and efficient argument
analysis. It tells you that you want to interpret the argument’s meaning in
whatever way makes the most sense and force out of it, because otherwise, it
can easily be reformulated slightly in order to meet your objections. That’s why
it is sound practical advice. “Setting up a straw man,” means setting up_for
the purpose of attack—an argument which is not the real one to be discussed
but a poor imitation of it, which is easier to criticize, and which is sloppy
argument analysis because it can so easily be rebutted. There will no doubt
be political occasions where you can get away with this kind of thing, but it’s
a very cheap victory when you do, because it’s not only unethical but so
highly vulnerable to counterattack that it s a dubious strategy even on selfish
grounds.
We 11 make a good deal in this text of the difference between “strong”
and “weak” criticisms of arguments. Roughly speaking, the difference is that
a weak criticism is one that can be refuted by rather modest changes in the
argument, or that depends upon a rather dubious interpretation or on what
is, at best, a possibility. By contrast, a strong criticism is one that gets at the
heart of the argument, hits it in a way that can’t be evaded without giving up
the main thrust of the argument, rests on indubitable interpretations and
highly likely claims, and so on. You’re on your way to making strong criti¬
cisms rather than weak ones when you take account of the Principle of
Charity in these early steps where you’re trying to get both a sense of the
general meaning of the argument and everything in it and a sense of what its
main point is, that is, what conclusion it’s trying to establish.
Suppose somebody says, “It s a good thing that President Ford wasn’t
assassinated in Sacramento in July 1975, because then Rockefeller would
have become President, which would have been a disaster.” An example of
a cheap shot would be to say something like: “You mean that you don’t care
at all about the life of a man who’s given many years of devoted service to
his country, that all you care about is the political situation after he is
infamously killed? And you don’t care about the fact that this would have
been a vicious and pointless act of violence—you’re more interested in the
balance of power in the Republican Party?” That kind of response is very
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 73

common, and it can’t be justified on the basis of the passage quoted. The
speaker who is worried about President Ford’s successor may very well have
deep concern for these other matters, but simply be focusing on one of the
several dimensions of the effect of such an assassination, when making this
comment. It's easier to criticize the speaker by bringing in all these other
points because they all are very relevant to the conclusion and important. If
you can successfully accuse an arguer of having overlooked relevant and
important points—for example, if this speaker had claimed to be giving an
overview of all the reasons why the assassination would have been unfortu¬
nate—of course you have a strike against him. But you have no real basis for
that line of criticism, so we regard it as a cheap shot, and certainly as
violating the Principle of Charity.
The example quoted serves as a pretty good illustration of some of the
difficulties that people have with Step 2, Identifying the Conclusions. In one
of the trial classes using these materials, several students identified the claim
that Rockefeller would have been a disastrous President as the conclusion.
One of the reasons they gave for making this identification was the impor¬
tance and the controversial nature of that claim. But the force of a claim
doesn’t show it’s the conclusion of an argument. It’s true that we often try to
establish (that is, make into a conclusion) the most controversial claim in an
argument because it’s the one that most needs support, but sometimes, too,
we use controversial claims as premises. In this case, it is because the arguers
believed that Rockefeller would have made a bad President that they con¬
cluded that it was fortunate that Ford wasn’t assassinated in Sacramento.

4-7

What we’re doing in this chapter is exploring the general procedure for
argument analysis in enough depth to give you a good understanding of the
basic logic behind it, and some examples of its application. We’re still talk¬
ing about fairly general logical issues that come into argument analysis; in
later chapters, we’re going to get into some of the detailed procedures that
fall under each of these headings, but first we need to get the basic concep¬
tual framework set up solidly.
One of the “basic framework” questions that seriously bothers begin¬
ners concerns drawing the line between factual criticism and logical criti¬
cism. You will have noticed in some of the examples we’ve just been discuss¬
ing, about aviation safety, for example, that I occasionally make some
reference to facts that bear on the argument under discussion which you
may not have known. While it’s a good general principle that the main
function of argument analysis is to concentrate on errors of logic rather than
on errors of fact, and while it’s certainly true that you, the student, are not
going to be held responsible for knowing all sorts of details about the opera-
 74 CHAPTER 4

tions of the Federal Aviation Administration and a million other subject

matters of examples that you’ll come across, there are two points to bear in
mind. First, even if you are not certain about the exact facts in a particular
case, it may be a matter of common knowledge or belief that there are some
exceptions to a generalization which is being quoted as a premise to an
argument you are scrutinizing; it’s perfectly appropriate for you to raise
these possibilities of error. You’re not saying that the premise is definitely
wrong, but that it would need to be checked out carefully in certain respects,
which you are pointing to. So don’t underestimate the extent to which you
know a little bit about a lot of things and the extent to which this knowledge
can be useful in pinpointing possible weaknesses in an argument. Argument
analysis doesn’t always score big by mentioning such possibilities, but it has
done a useful task because it has indicated where the argument needs further
support or clarification. If you’re the person who is putting the argument
forward, this indication is helpful, just as it’s helpful if you’re investigating
the argument to see whether it ultimately stands up. And there is one impor¬
tant type of argument where discovering a possibility is completely fatal to
the argument. What is that kind of argument? (Quiz 4-B, 6.)
The second reason why I sometimes bring in a few extra comments of a
factual kind that I wouldn’t expect every reader to know about is simply that
they are interesting and worth knowing for purposes of assessing other argu¬
ments, or even the particular ones under discussion, for their impact on your
own behavior and beliefs. Of course, it’s reasonable to suppose that I’ll be
wrong about some of these so-called facts that I produce, and you should be
on your guard against accepting them from me any more than from the
authors of other textbooks, or from arguments that you encounter. Regard
them as interesting and relevant claims, and if the matter has considerable
importance to you, check them out further before relying on them.
What you are expected to know about is the meaning of terms, because
you not only speak the language but you certainly have access to a dictio¬
nary which you can use if the term either has a technical flavor or is a
relatively uncommon one. With arguments that you are going to consider
seriously, for a written assignment perhaps, it’s strongly recommended that
you turn to an encyclopedia and not just a dictionary, because you’ll often
find that facts and possibilities will emerge from the articles in there that will
sharply affect your appraisal of the arguments, including even your ap¬
praisal of the inferences and not just their (factual) premises. Certainly you
are in a better position to criticize arguments about a technical subject if you
know a good deal about the subject matter; but you may still be able to
criticize them on purely logical grounds although you know nothing about
the topic. One often reads discussions or sees television shows in which the
justifiability of a certain type of surgical operation is being debated, and
without knowing anything at all about the specific details of the medical
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 75

risk, you may easily see ways of handling the problem of whether and how
much to inform the patient. Or, accepting the figures produced by the admit¬
tedly biased advocates of the surgical procedure, you may still see analogies
with other risks in our society, such as the risk of death in traffic, which can
lead to a very different viewpoint on the advisability of the operation in
question. Or you may see loopholes in the experiment that is being quoted as
supporting a position. Errors of reasoning occur even in the heart of science;
and indeed, they are surprisingly common there, particularly errors of rea¬
soning about the practice and the funding of that particular branch of sci¬
ence, rather than errors of reasoning within the science itself. Scientists are
very prone to think that their care, skill, training, and frequently high intel¬
ligence will automatically carry over, for example, to the administration of
science, the funding of science, or the application of science to human af¬
fairs. It’s an open question whether these experts are as good at arguing
about such matters as an intelligent lay person, and on the track record, we
have to conclude that they are very often highly unreliable there, just as is
the lay person.
The question of how to draw the line between factual criticism and
logical criticism, which we’ve been discussing in this section, naturally leads
to the question of exactly how to distinguish between the premises and
inferences in an argument. Much of what we were discussing in the opening
sections of this chapter bears on this point, and as we are now about ready
to consider the steps of portraying structure (Step 3) and identifying the
missing premises (unstated assumptions) (Step 4), it’s appropriate for us to
turn to this point.

4-8

Here we are, then, facing a great big slab of argument—perhaps a para¬

graph or a page from a book, or even an entire article. Let’s review the
preliminary assessment procedures. We’ve read the piece through very care¬
fully, trying to get a sense of what it’s about. We’ve decided that there are
several arguments involved in it. That is, there are several conclusions put
forward as following from certain other statements which are obviously the
premises (since they’re not themselves in turn supported). Some of these
conclusions turn out to be used as premises in a later argument, so we’re
looking at a passage of argument which involves some chains of argument.
Studying the material still more carefully, we see that some of the sentences
merely involve reviews of things that have been said or have been proved
earlier and do not need to be taken into account as fresh material for pur¬
poses of analyzing the whole passage. We also notice that some parts of the
argument involve historical allusions to earlier thinkers or politicians, and
that they are not intended to add new evidence, but only to be of general or
 76 CHAPTER 4

passing interest. So we drop them from our serious analysis, along with the
repetitions. A sentence or two is obviously intended to be a rhetorical flour¬
ish, perhaps a rephrasing of one of the conclusions that has just been estab¬
lished, in more flowery or stirring language. Careful consideration shows
that nothing is being added here. There are a couple of terms in the argu¬
ment which we re not too familiar with, and we’ve taken the opportunity to
jot down at the top of our page the relevant dictionary definitions of them.
Looking carefully at the passage in the light of these dictionary definitions,
we re pretty confident that there hasn’t been any shift in the meaning of
these terms all the way through the argument, from one dictionary definition
to another, or from the dictionary definition to some illicit but attractive
alternative. If in any doubt about this, we’ve translated the clauses in which
these terms occur into other words, to see whether a switch has occurred. It’s
also clear that all the conclusions that are being "aimed at” by this passage
of argument are explicitly stated within the argument. If they aren’t there,
we’ll state them as carefully as we can and write them in the margin at the
appropriate place on our rough working copy, attaching to them a number
corresponding to the place where they occur in the argument, all of whose
functioning assertions we have now numbered in sequence. If you’re having
any real difficulty in sorting out what the arguments are trying to establish
or how they go about doing it, now is the time to write down some criticisms
of unclarity of expression or structure. Later on, you’ll come back to look at
whether these are serious criticisms that should be given high priority in the
final report, or whether they’re relatively minor and deserve just passing
mention toward the end of your report. Never expect to be able to write the
final report straight off. Write a first draft after you’ve filled out the working-
sheet entries, and then try for a final draft.
Now that your general picture of the passage is beginning to take shape,
go back and read it again and see whether you’re missing something. Per¬
haps some of those sentences that you dropped out do add a bit to the
argument—perhaps they add a little extra information that should be taken
account of as a premise, even if you have to rewrite them into a very con¬
densed form in order to express this little bit of information clearly. Since it
saves a lot of rewriting, one can usually immediately number the assertions
as they occur in the printed passage that you’re considering. If you have to
rewrite one of the sentences, cross it out in the passage but leave the number
there; then, on your rough working sheet, write out the condensed form with
the appropriate number next to it. That way, it will still be easy for you to
read down the passage, come to a certain number that’s attached to a
crossed-out clause or sentence, and know just where to find the condensed
version of it. Be sure that wherever there’s a vague sentence, whose meaning
can be understood only after you have read the whole passage (and have
taken into account the Principle of Charity), you cross it out (leaving a
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 77

number in its place) and rewrite it in a more precise and perhaps more
charitable form.
Remember that the crucial difference between premises and conclu¬
sions, when you’re trying to distinguish them, is that the premises are sup¬
posed to supply reasons for believing or accepting the conclusions. The basic
premises of the whole argument will not be supported within that whole
argument. Later on, you may find assertions that have been previously sup¬
ported, that is, that were conclusions of an earlier phase of the argument,
but which are now being put forward as reasons for yet a further conclusion.
These will of course be “intermediate conclusions” in a “chain argument,”
and in the tree diagram of the argument they will have lines leading down to
them as well as lines that lead down from them.
While you’re going through this process, you will of course be begin¬
ning to notice places where the argument is vulnerable to criticism. And
we’ll get on to that in more detail in the next section. First, let’s see if we can
apply what we’ve been saying to an argument of a length that’s slightly
greater than the one about Lipton’s Tea, but not yet as complicated as the
kind that we will eventually handle.

We can be proud that America has turned the corner on the depression of the
last few years. At last the many indexes of recovery are showing optimistic
readings. The rate of inflation has slowed, unemployment has more or less
stabilized, inventories are beginning to drop, advance orders are starting to pick
up, and—the best news of all—the average income figures are showing a gain.
The doomsayers have been discomfited, and the free enterprise system once
more vindicated.

Here’s a pretty straightforward passage. First of all, we have to see if

there’s anything there that we don’t understand. Perhaps you’re not certain
about what “inventories” are; they’re the supplies of raw materials or com¬
ponents that are stored by manufacturing companies from which they draw
the materials for production, or at the retail end, they are the supplies of
products that are warehoused in anticipation of sales. It’s a sign of a depres¬
sion in your situation if you can’t move your inventory, and if it’s true in
general, the economy is probably in a depression. When you’re having trou¬
ble selling goods, your inventories remain stable. They should be dropping
so that you can start placing further orders for manufactured goods or raw
materials and hence, during a depression, a drop in inventories is usually
taken to be a sign that recovery has started up. As your inventories drop,
you begin to place orders for delivery at some future time; these are called
advance orders.
So the general meaning of the passage is clear, but now let’s get down to
a little more detail. Exactly what is the main conclusion, or what are the
 78 CHAPTER 4

several conclusions? This question is not quite so easy to answer and there
are three candidates. First of all, to avoid having to repeat long sections of
the passage, let’s label the assertions in it. It’s helpful to use a different color
of ink for this, so that the contrast stands out; and you may want to use
colored parentheses (to distinguish them from parentheses that are part of
the passage) to indicate the beginning and end of each of the numbered
assertions. The assertions are really the pieces of the jigsaw puzzle, and your
picture of the structure of the argument is going to relate them, showing
which are conclusions and which are premises. You separate the ones which
you are likely to comment on separately, even if they all serve as premises
for the same conclusion and therefore could technically be given the same
number. Even with these instructions, a few judgments have to be made in
allocating the numbers, so let’s look at one way to do it.
1
We can be proud that (America has turned the corner on the depression of the
2
last few years.) At last (the many indexes of recovery are showing optimistic
3 4
readings.) (The rate of inflation has slowed), (unemployment has more or less
... 5 6
stabilized,) (inventories are beginning to drop,) (advance orders are starting to
7
pick up,) and the best news of all—(the average income figures are showing a
8 9
gam.) (The doomsayers have been discomfited,) and (the free enterprise system

once more vindicated.)

Let’s walk through this, step by step. The opening words, “We can be
proud that are just a rhetorical flourish or an underlining of the significance
of the argument. They aren’t “what it’s all about” and we can skip them
right here. The first main assertion is the one that we’ve labeled number 1.
Now we come to the second sentence. I could have just skipped this one on
the ground that it merely summarizes what is going to be said with care in
the following sentence. However, it does go just a little further than the
summary. It says “the many indexes” and then it gives a few of them; hence
it isn’t adequately supported by the next sentence, which, for example, says
nothing about a very popular index, the gross national product (GNP). It
also doesn t say anything about retail and wholesale prices, about farm
prices and foreign orders, about new housing starts and other indicators
trom special industries, such as automobile manufacture, that are bellweth¬
ers for the economy as a whole. So we scent a slight overgeneralization here,
and we need to give this sentence a number in order to make clear that it
represents a conclusion that is open to question when we analyze the argu¬
ment. 6
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 79

In the next sentence, I break out the various claims separately, because
the facts about each of them are rather different and we need to consider
them separately.
I skip over one little comment in the middle of the sentence, where the
writer says that “the best news of all” is that assertion 7 is true. To say that
income gain is the best news of all is not just to wave a flag about it but
actually to make a claim about it which could be challenged. It would be
entirely appropriate to have labeled the phrase and picked up any com¬
plaints, about its justification in the analysis of the argument, but one
doesn’t want to get into too many of the details, and I’m inclined to think of
this as being very much like the “we can be proud” kind of flourish. But
nobody’s going to fault you for having separated it and said something
about it. You need to be clear right from the beginning of this game that,
except for the very simplest arguments, there’s nearly always some room for
alternative ways of analyzing them. The way you structure them is only a
means to an end, after all. The big question is whether you're going to come
out with the really serious criticisms at the end, and not some trivial ones, or
some completely unfair ones.
A similar general problem arises about the assertions that are marked 8
and 9. One could argue that these are more rhetorical flourishes, embellish¬
ments on the fact expressed in 1. But it’s a little more precise to treat them
as further conclusions, because they do go a little further than 1. To be
specific, 8 draws a conclusion about the prophets of doom, and, after all, no
evidence was given earlier in the argument as to exactly what they said other
than the rather general claim that we were heading for trouble; and 9 does
generalize to vindicating the free enterprise system, which might take a little
more than showing that we pulled out of one slump.
Now what is the main conclusion, if there is one? The key conclusion is
assertion 1. Several pieces of evidence are being marshaled to support it, and
they are, of course, assertions 3, 4, 5, 6, and 7. They are summed up in
assertion 2. Then, from this key conclusion (assertion 1), we’re drawing two
further assertions, namely, assertions 8 and 9. So the total structure of the
passage looks like this:
 80 CHAPTER 4

Notice that 9 doesn’t follow from 8 or vice versa; they are both conclusions
from 1 (along with some other assumptions about what the doomsayers said
and what free enterprise amounts to).
Notice, too, that we draw separate lines from each of the various prem¬
ises 3 through 7 to conclusion 2 and hence 1, not because any one of them is
said to be a sufficient condition for 2 or for 1 (a situation which would also
diagram the same way), but because each of them is said to support 1 at
least to some degree, and we need to look at each such claim separately. We
might have used the kind of diagram, illustrated earlier, where a whole set of
premises is tied together with plus signs and a horizontal bracket, but it is
used primarily in situations where each of the considerations has weight only
in conjunction with the others; whereas here, each is supposed to have some
weight by itself. Again, the difference is a matter of judgment as to the way
the arguer feels about the force of these premises. To call them “indexes of
recovery” suggests that each of them has some significance in its own right,
so to speak.
Notice that although 8 and 9 are the “bottom line” conclusions, they
are not the key conclusions. They are, if they can be established, politically
and economically more significant than the single fact that the American
recovery has begun on this particular occasion, and hence, in some sense,
they are more important than what I’ve called the key conclusion. But that’s
in global terms; assertion 1 is far more important for the analysis of this
argument. If it can t be established, then 8 and 9 are nowhere.
The decision to represent the inferences from 3 through 7 to 1 (with an
intermediate stop at 2) by separate lines involves a certain risk which we
should now state, to avoid your falling into a particular kind of trap later. It
might turn out that the argument from each of these indicators is fairly
weak, and we might point that out in the course of our analysis. It does not
follow that the argument from all five of them can be dismissed. (That’s one
reason for using the form of diagram, shown earlier, which reminds you to
hold them all together and to look for their overall strength.) True, we’ll pick
this point up as we get further down the Seven Steps anyway, but it’s well to
be forewarned.
This is a relatively simple argument, compared with many that you’ll
run into, and one of the ways in which it’s simple is that only one person’s
argument is involved. Often, in reading a newspaper report, you’ll see that
the report itself is interpreting some facts in a particular way (this usually
shows up in the heading), but the people who are being reported on are not
arguing in the same way as the newspaper. If you turn back to the last
exercise in the B quiz in Chapter 2, a story from the San Francisco Chronicle,
you’ll see that the headline says, “U.S. Survey Finds Bias in Home Loans,”
but the second paragraph says, “The Federal Home Loan Bank Board,
which drew no conclusions from the survey of lending practices in five
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 81

cities, . . So you’d have to separate two lines of argument in doing your

analysis of that passage, one corresponding to the line of argument of the
newspaper and the other corresponding to the line of argument of the fed¬
eral agency. There are various ways you might do this, and we’ll look at
examples like that later, but just remember that you can’t assume that all the
arguments involved in a passage of argument should be treated as part of a
single continuous stream.

4-9

Now we come to what is probably the hardest aspect of argument analysis,

conceptually speaking. It isn’t usually hard at the practical level, but there
will be times at the practical level when you need to understand the concep¬
tual issue in order to be able to proceed, and this need is what we are going
to take up. We are now moving to Steps 4 and 5 in the Seven Steps; the steps
that require you to identify unstated assumptions (Step 4) and criticize infer¬
ences and premises (Step 5). Here’s where you begin to be able to use your
schematic diagram, your “tree structure” which lays out the form of the
argument. It identifies the targets for your attack. Every line on the diagram
represents an inference; every number represents a premise. Some of those
premises are supported and some aren’t. The ones that are supported are the
conclusions of an earlier argument and hence vulnerable as a result of criti¬
cism of that argument. Your basic targets are set up; now, how do you go
about deciding whether they are immune to criticism, or what kind of criti¬
cism that is appropriate for them? We’ve already done a little bit of recon¬
struction; that is, we’ve filled in any conclusions that are implicit in the
argument or in its context. We now face the question of whether to go
further with the reconstruction and fill in what are called “missing prem¬
ises,” namely, the further assumptions that are required, in many cases, to
make an inference satisfactory. We call them unstated assumptions because
some assumptions are stated within an argument and identified as assump¬
tions. It is not these we are talking about in Step 4, but the missing ones.
Let’s begin with a simple example. Suppose some woman says, “She’s a
redhead, so she’s probably quick-tempered.” Ask yourself right away, what
is the missing premise? Before you read any further, try to write it down on
a piece of scrap paper. What is the speaker assuming?
A popular answer to this is that the woman is assuming, “All redheads
are quick-tempered.” That’s incorrect; it’s much too strong an assumption to
saddle the speaker with. By making this the assumption, you’re setting up a
“straw man,” and it seems easy to refute the argument by saying that it
depends on an assumption to which you know a number of exceptions.
At this point, you may notice that the conclusion is “she’s probably
quick-tempered.” It occurs to you that to get to this conclusion from the
82 CHAPTER 4

premise, you need only a much weaker generalization, namely, “Most red¬
heads are quick-tempered.” The fact that you know some redheads who are
even-tempered doesn’t count at all against this more modest (“merely statis¬
tical”) generalization. So the argument stands up against what at first
seemed to be strong contrary evidence.
But you still don’t have the assumption correct. There’s a feature of the
argument, simple though it is, which tells you that you’re still saddling the
arguer with too strong a generalization. It might be the case that most
redheads aren’t quick-tempered, but the argument might still be perfectly
sound. Why?
The reason is that the argument is about a woman, and to support it, all
you need is the assumption that most redheaded women are quick-tempered.
Do you really have evidence that the statement is false? Whether it’s true or
false, you’re at least in a much more difficult position to find a weakness in
the argument than if you had presented it as depending on the original
alleged assumption.
Now let’s look at the argument as we’ve reconstructed it. We’ve con¬
structed an extra premise (“Most redheaded women are quick-tempered”),
which we can add to the stated premise (“She’s a redhead”) in order to draw
the conclusion that she’s probably quick-tempered. In terms of diagraming
the argument, you should use letters instead of numbers to refer to the
missing premises that you have constructed and add them to the diagram at
the appropriate points. So, whereas this was originally the simplest possible
form of tree diagram, with a single premise and a single conclusion, the
diagram shown here is now one stage more complicated, with two premises
and one conclusion, the second premise being labeled a.

+ 0,
—y-'

©
Stand back and look at what we’ve done. When we first looked at the
argument, it had a single premise and a single conclusion. The inference
from the premise to the conclusion is not exactly an obvious one. It hinges
upon the truth or falsity of what we’ve called the missing premise. But if we
add the missing premise to the given premises, as we have in the recon¬
structed form, what can we now say about the soundness of the recon¬
structed argument?
Clearly, the reconstructed argument involves sound reasoning. Previ¬
ously, it was an argument in which the inference was questionable; now it’s
an argument in which the inference is perfectly sound. Surely, it’s absurd to
reconstruct arguments in a way that completely changes the status of the
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 83

reasoning in them from unsatisfactory to perfectly sound? That doesn’t seem

like an appropriate reconstruction of the original argument.
It’s appropriate enough, though, because it has simply shifted the target
from one place to another. The weakness will now be in that extra premise,
whereas, previously, it was in the inference. And this illustrates a very gen¬
eral principle about argument analysis. You can either leave the argument
the way you find it, in which case a good deal of your criticism will be
criticism of the inferences in it; or you can patch it up by adding some
assumptions on which it is obviously depending, in which case the inferences
will be pretty satisfactory, but the assumption will now come under fire.
There is no essential difference between the inferences in an incomplete
argument and the missing premises in a complete argument. They both serve
as the link between the original given premises and the conclusions of the
argument. For this reason they are often said to be “inference-licenses.”
Laws of nature are a very common type of inference-license. If somebody
says that Jones can’t be in Honolulu at noon if he was in Capetown one
microsecond before noon, the soundness of this inference depends upon the
law of nature which states that nothing can travel faster than the velocity of
light. We are using that law of nature as a missing premise in our argument,
as a license for inferring from Jones’s presence in Capetown at a microsec¬
ond before noon to the conclusion that he can’t be several thousand miles
away at noon. If we were to infer from the fact that he was in Capetown a
minute before noon to the conclusion that he couldn’t be in Hawaii at noon,
we wouldn’t be relying on a law of nature, but upon a technological truth of
the moment, which is that we have no means of transportation that can
travel at the velocity required to make this jump. In that case, the missing
premise is an assumption about contemporary technology, and its truth
serves as a license for the soundness of the inference.
So we have the option of criticizing the inference step as it stands or of
converting the principle that underlies it into an extra premise and thereby
ensuring that the inference in the reconstructed argument will be sound,
although the argument will still be open to criticism because its underlying
assumption, now expressed as an extra premise, is debatable. Which is the
best way to approach argument analysis?
In cases where the inference steps are already very simple and clear, it’s
best to treat them as they stand. They may involve mistakes; for example,
somebody may make an inference in doing a mathematical problem which
involves a mistake because of a failure to put the decimal point in the right
place or to transform from one set of units to another correctly. In such
cases, one can merely point out that the inference is at fault, with the argu¬
ment in its original form. Arithmetical errors and other computing errors are
typical of these cases. If Jim Black has multiplied both sides of an equation
by 7 and erred in multiplying the right-hand side of the equation, it’s pretty
84 CHAPTER 4

pointless to say that this step in his argument “rests on the assumption that
six 7s are 43.” He didn’t make that assumption, he just made an error in the
mathematical transformation, i.e., the inference.
Another kind of case where it isn’t worth reconstructing an argument
by adding missing premises is when the missing premise which you are
thinking of adding simply restates the whole argument. Suppose Mr. Lyons
argues that it was entirely appropriate to discharge a certain patient from a
psychiatric ward because her depression had shown marked signs of lifting;
it’s obvious that he is arguing from a single premise about the lifting of
depression to a single conclusion about the appropriateness of discharging
the patient. Now you can add a missing premise which says, “Whenever this
particular patient’s depression is lifting, it’s appropriate to discharge her,”
and of course the argument will now, in its reconstructed form, be logically
bulletproof. On the other hand, it is somewhat superfluous to have added the
missing premise, since it tells you only that the original argument was satis¬
factory. We haven’t obtained any illumination by adding this premise. All
arguments depend upon the “assumption” that you can get from their spe¬
cific premises to their specific conclusions, and adding this claim as a missing
premise gets us nowhere. (In the example about the redhead, as a contrast,
the process of formulating the missing premise led us to see exactly what was
being assumed, and it wasn’t just “from the premise that she’s a redhead,
you can draw the conclusion that she’s probably quick-tempered.”) So here’s
another sort of case where you don’t want to waste your time adding missing
premises: where the missing premise simply states that the argument as it
originally stood was satisfactory.
In short, where there was obviously an error in calculation, a slip in the
logic, or some other kind of error of mechanical reasoning, it isn’t worth
formulating it as a missing premise, largely because nobody would ever have
assumed that. And where the required assumption is formulated in terms of
the exact same premises and conclusion as the original argument, it is obvi¬
ous from the fact that the original argument was put forward that this
assumption was being made, and no illumination is achieved by pointing it
out explicitly. Identifying missing premises has utility when they are neither
trivial claims nor obvious errors in routine reasoning. It is in the large re¬
maining body of cases where the procedure of trying to formulate the as¬
sumptions exactly pays off. The discipline of doing so leads one to realize, in
a way that no other procedure does, precisely what the assumptions are on
which the argument rests. And it sets them up in a form where they can be
carefully assessed for essential loopholes.
Of course, once one has added the missing premises, one should not
then find any remaining errors of inference in the reconstructed argument.
For you have formulated the assumptions just so that they will make the
reconstructed argument sound, as far as inferences are concerned. The focus
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 85

of attention now shifts to the premises alone (as they relate to the recon¬
structed part of the argument).
Exactly how does one correctly formulate the missing premises of an
argument? Here again, we find that imagination and originality are often
required in this basic part of the critical process. You can see from the
redhead example that a process of careful trial and error will often work.
But there are times when the situation is more complicated and it is consid¬
erably more difficult to get the missing premises right. Moreover, there are
often alternative ways of formulating them when more than one is involved.
That’s perfectly all right as long as you can produce a set of premises which
together convey the essential content of the assumptions underlying the ar¬
gument. To do this, you have to juggle three quite different considerations.
First, the assumptions have to be strong enough to make the argument
sound. Second, they should be no stronger than they have to be, since they
might then be too strong to be true, and you would then have constructed a
“straw-man” version of the argument, which you would be able to criticize
even though the original argument was immune to your criticism. Third, a
considerably more subtle point: You also want to try to relate the assump¬
tions as you formulate them to what the arguer would be likely to know or
would believe to be true. The principal function of argument analysis is not
that of reconstructing the state of mind or body of beliefs of the arguer, but
this may nevertheless be a relevant consideration in some cases. You need to
decide whether you are arguing against the arguer or the argument. In a
dialog exchange, of course, you'll be arguing against a person, and then your
best guess at what that person really believes (provided it meets the other
conditions for a satisfactory missing premise) will be very relevant. In other
cases, you may just have to settle for what you think makes the most sense.
Suppose that the argument about the redhead had gone like this: “She’s
a left-handed, cross-eyed, 6-foot redhead—you can bet she’s a quick-tem¬
pered witch.” Now, strictly speaking, all that the arguer is stuck with as a
missing premise is a very restricted kind of a generalization. You could say
that the argument assumes only that most left-handed, cross-eyed, 6-foot
female redheads are quick-tempered (assuming that “witch” is a figure of
speech). If you look back to the outline of the Seven Steps, you’ll see that
what we are talking about here is the “minimal assumption of the argu¬
ment” (Step 4). We don’t have any other evidence about the arguer in per¬
son, so we can’t come to much of a conclusion about the arguer’s assump¬
tions on the basis of any specific evidence. However, in general terms, it’s
likely that the best (the optimal) assumption for the sake of analysis is that
redheaded females tend to be quick-tempered, or perhaps that redheaded,
cross-eyed females tend to be quick-tempered. We don’t need to narrow it
down any more by bringing in the 6-foot height and other features because
it is extremely unlikely that the person giving this argument has any evi-
 86 CHAPTER 4

dence on the effects of any extra conditions on the probability of a quick

temper. Probably, in fact, the evidence that is available bears on the ques¬
tion of whether redheads in general, whether male or female, tend to be
more irascible than people with other skin and hair colorations. (That’s if
there is any evidence at all.) It’s possible that there is some evidence on this
based on sex differences, so, to be charitable, we restrict the missing premise
to females. But it’s most unlikely that it is available in any more refined form
than that, and it’s only a kind of hunch that suggests that people who are
cross-eyed might tend to be more irascible than those who aren’t. So, with a
dash of charity, we might put that in. But the real point is that assumptions
ought to bear some relationship to what the available evidence can support,
if they are to be the kind of thing that can help the argument along. (And, of
course, they have to be logically adequate, so that they will make the infer¬
ence part of the argument sound.) Hence, we look for the kind of factual
claim that is likely to have evidential support, as well as being logically
adequate for the argument. This usually means that we go a little stronger
than the minimal assumption that would do the job, and thus finish up with
what was earlier called the “optimal assumption.”
Let’s take another example and put it through the hoops. This one will
illustrate the way in which assumptions which are put in negative form are
sometimes the most illuminating ones. Suppose that we are listening to
somebody arguing in favor of decontrolling the price of domestic oil and
hence gasoline prices. He says, “The main reason for this is, of course, that
we have just got to provide enough incentive to the oil companies to explore
new fields, given the fact that the cost of explorations in the marginal areas
has now gone up considerably.” The structure is pretty simple. The conclu¬
sion is that we ought to decontrol oil, and the premises are that doing so will
provide enough incentive to explore new fields, and that the profit margin
included in the present controlled price isn’t enough to cover the cost of that
kind of exploration.
Now, what is the speaker assuming? Note first that it’s trivial to say that
he is “assuming” that from these two premises you can draw this conclusion.
Of course he is, and it’s a waste of time, space, and words to put it down.
The question is, exactly what assumption is he making that he thinks is
strong enough to bridge the gap between those premises and that conclu¬
sion? This question isn't settled by asking him directly because he may not
be very good at detecting his own unconscious assumptions—he might sad¬
dle himself with a stronger assumption than he needs. One of the most
important functions of argument analysis is to force one to make one’s own
assumptions explicit, as a result of which one frequently finds that one is
assuming more than, or other than, the things one had supposed.
Putting the key question of argument analysis in terms of a vocabulary
which we discussed earlier in this chapter, what we’re asking is whether the
 DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 87

premises as they stand in this argument provide a sufficient condition for the
conclusion, or, if not, what we have to add to them to get that.
In the first place, of course, the arguer in the present example is assum¬
ing that if domestic oil prices are decontrolled, the price will go up. That’s a
very sensible assumption in view of the market situation at the moment, but
it certainly is an assumption of the argument, and given the fact that over¬
seas producers are having real problems selling their present production, it
may not be true for long. Secondly, he is assuming that the prices will go up
enough to provide the required new incentives. It’s not entirely clear
whether that is going to be true. It’s a question of whether the demand will
hold up at these higher prices. So you might say he is assuming the demand
will hold up well enough to get the prices up to the point where they create
a profit margin from which exploration in the marginal areas could be fi¬
nanced. But, of course, we haven’t come to the end of the assumptions.
There is a particularly important one still buried in that argument. Can you
see what it is? Take a moment to think about it before reading further. Go
back up and reread the argument carefully, and think about what you’ve
heard in recent discussions of this issue.
The further assumption is that the oil companies will in fact put the
money into further exploration, if they do succeed in increasing the prices to
the level where exploration would be financially possible. Putting it nega¬
tively, the assumption is that the oil companies won’t simply pocket the
profits or divert them into other investments besides exploration. Of course,
this really is a big assumption. As a matter of interest, on the track record of
the past year or so, it’s an unlikely bet. The oil companies have been show¬
ing enormously greater profits, and they have also been diversifying (that is,
buying into different industries, such as hotel chains) with the profits instead
of putting them into exploration. Moreover, they have been doing that with
the profits they were making under controlled prices on native crude oil.
This suggests they were making enough to do at least a certain amount of
further exploration, but not doing it. Well, there may be all sorts of deeper
pros and cons about the issue, but a little exploration in argument analysis
has laid bare a series of key assumptions here which will direct our further
research, and which raise questions in our minds which will substantially
impede our willingness to accept this argument immediately.
Notice that we avoided talking about assumptions like “Belief in pri¬
vate enterprise requires decontrol.” We restricted the assumptions to the oil
companies because the arguer isn’t sticking his neck any further out than
that, and because there are certain special features of the oil industry that
don’t generalize across the entire private sector. We haven't destroyed the
argument; it’s quite possible that the reason oil companies have been diver¬
sifying or distributing their profits in recent years is that going into further
oil production was not as profitable as these other enterprises, whereas it
88 CHAPTER 4

would be under decontrolled prices. On the other hand, it is by no means

obvious that, given an enormously greater quantity of profit, they may not
be in a position to make a business choice in favor of other completely new
enterprises not related to oil production. Argument analysis doesn’t always
lead you to the conclusion that an argument is not good; what it may give
you is a much better perception of what its strengths and weaknesses are,
and that is just what it does in this case.
If you are asked to go beyond criticism to take a position on the conclu¬
sion itself, of course you’ll move to Step 6 (“Consider Other Relevant Argu¬
ments”). In the present example, you would look at the question of what the
effects on the domestic consumer will be of substantially increased costs of
gasoline. Such costs will not only have a direct impact on the running costs
of their cars, but indirect effects on the costs of all their purchases involving
plastics manufactured from petroleum residues, on their home heating costs,
and the cost of other goods and services whose production or provision
involves fuel costs or oil costs. Senator Kennedy estimated in late 1975 that
the annual effect of decontrol would be between $800 and $900 per Ameri¬
can family; the administration’s estimates have been nearer to 10 percent of
that estimate. If Kennedy is right, decontrol might produce a new or worse
recession. Picking up another dimension of the issue, there is the crucial
problem of the nation’s dependence on foreign oil and the possibility of a
boycott by foreign oil suppliers as a device to impose a particular political
position upon the United States. President Ford certainly believed that the
nation should be willing to pay quite heavily for independence from foreign
sources, which means that we have to develop more domestic sources. In his
view, this line of argument might concede the truth of the earlier ones but
move in a new direction to avoid decontrol by requiring a great emphasis on
using more efficient methods of transportation, heating, and manufacturing.
The administration is using a strategy of this kind as a backup strategy, not
as a main strategy. It would certainly involve considerable changes in the
life-style of most Americans, changes which would be likely to prove politi¬
cally unpopular even if they were in the long run beneficial.
Having looked at all these considerations (and there are several others
that are relevant), one should then be in a position to take Step Seven
(Overall Evaluation) and come to a decision as to whether the oil prices
should be decontrolled.
Remember that you have to do this kind of thing piecemeal. First you
pick up particular arguments and subject them to careful argument analysis.
Then you gradually try to assemble a comprehensive set of the main argu¬
ments that are relevant to a particular issue—something which you can do
from a study of the public discussions and the technical literature connected
with the topic. Finally, you are in a position to do the integration or perform
a synthesis of all these considerations.
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 89

4-10

Let s go back to the example which we spent some time structuring earlier,
the one about turning the corner on the depression of the last few years.
Let s apply to this some of the discussion of assumption formulation and
criticism that we've been getting into in the intervening sections. This is a
really important type of argument, because it has persuaded many members
of Congress and the media that recovery is just around the corner, or that we
have just turned the corner to it, or that we have just begun it, and, as a
result Congress is prepared to pass legislation that it otherwise wouldn’t
pass. If Congress and the media are wrong, we may be in for a much worse
period ahead. Even if they are right, they may have been right because they
were lucky and not because this argument is any good (which should affect
our assessment of their judgment). So it deserves careful scrutiny. And it’s
also an important type of argument because it’s just on the borderline be¬
tween the technical and the commonsensical; technical enough to frighten
off a good many citizens who, in fact, could perfectly well understand it with
a slight effort and thus be in a better position to judge the rationality of their
elected representatives.
Remember that the argument essentially has the form of quoting five
indicators that have changed recently, concluding from them that “the many
indexes of recovery are showing optimistic readings,” concluding from that
statement that we have turned the corner on the depression, and then going
on to flights of higher political and economic fancy with conclusions 8 and 9,
to which we will turn later.
Let’s use the Seven Steps, although, as always, we may be able to pass
very rapidly over some of them as soon as we have satisfied ourselves that
they don’t contribute usefully on this particular occasion.
Step 1: Clarify Meaning. We have already said something about the
semitechnical terms “inventories” and “advance orders,” but it’s clear that
there’s a rather vague phrase at the core of this whole argument. The phrase
is “turned the corner” as it occurs in “America has turned the corner on the
depression of the last few years.” What does “turning the corner” amount
to? I suggest that the commonsense meaning of this phrase is the idea that
things are getting better. As you will see, a good deal of the argument would
look more plausible if all it had to show was that things were getting worse
at a somewhat slower rate. But it’s no good calling that “turning the cor¬
ner”; it merely means that we have slowed down our rate of decline, and
that we have not even reached the corner. That distinction, simple though it
is, is obviously not clear to most headline writers in the newspapers and
magazines of the seventies. One constantly sees headlines that suggest that
the depression has—to use another phrase—“bottomed out,” whereas the
fine print shows only that the rate at which it’s getting worse has slowed
90 CHAPTER 4

down a bit. The rate at which things get worse is of very little interest if it
continues to get worse. Suppose that it simply stopped getting any worse.
We still wouldn’t have turned the corner; we would have stabilized in a state
of depression. People all too easily switch from talking about states to rates.
And that’s exactly the fallacy in the argument from the first of the specific
indicators that are quoted. “The rate of inflation has slowed” means merely
that prices aren’t increasing as fast as they used to; it doesn’t mean that
inflation isn’t continuing. Hence, it in no way shows that we have turned the
corner. We may just as well expect inflation to accelerate again as to reverse
its direction—there’s no way of knowing. Meanwhile, it continues, to our
cost.
The second indicator, that of assertion 4, states, “Unemployment has
more or less stabhlized.” The level at which it has stabilized is high enough
to produce incredible hardship, including an income below the official sub¬
sistence level, for millions of families. If it remains constant at this level, we
will not have left the depression at all. Even if it has stabilized at this level,
that’s no sign that it’s going to get better. This may be the permanent level
for it, given the present management of the economy. Changes such as de¬
controlling the price of oil would, in Senator Kennedy’s view, start unem¬
ployment on the rise again, and there are many other significant national
policy decisions which might have the same effect. On the other hand. Pres¬
ident Ford may be right, and we may go upward. This argument certainly
doesn’t show that we have turned to corner yet. Remember, too, that official
unemployment figures are highly suspect: they tend not to include people
who give up on finding a job after many months of trying.
The third premise is, “Inventories are beginning to drop,” which shows
only that people are doing some buying, and thus overcoming the excessive
ordering that led to overlarge inventories based upon the rate of purchase of
a prior and healthier year. The drop in inventories means only that one
unhealthy sign has become less unhealthy; that doesn’t make it healthy. If
the inventory level has dropped to what is now an appropriate point to
balance sales, we can’t tell whether we are in the middle of a depression or
not. Excessive inventories are a bad sign for the business person because so
much capital is invested in them, frequently capital that is on a loan at high
rates of interest. Again, it is very important not to have larger inventories
than you need in order to supply orders as soon as they come in. In a
depression, a small inventory will be appropriate, and hence the reduction of
inventories is what you expect as you head into a depression. It’s no sign of
one’s getting out of it. It is a sign that somebody is buying something and,
nationwide, it’s a sign that a significant amount of purchasing is going on.
But that occurs even in the heart of a depression. So the reduction of inven¬
tories, although welcome news to those who were holding excessively large
ones, is of negligible significance for depression-end predictions. Carried too
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 91

far, reduction of inventories is a sign that a company is going out of busi¬

ness, and indeed part of the present reduction of inventories is due to that, as
the increased number bankruptcies makes clear. When you go bankrupt you
completely liquidate your inventory, often at very disastrous prices. So re¬
duction in inventory isn’t a very good indicator that we are turning the
corner; in fact, it doesn’t indicate that at all.
“Advance orders are starting to pick up.” This appears to be a slightly
better indicator than the preceding one, because it does tell us that some
people have got their inventories down to the point where they feel they
have to replenish them. So not everybody is going out of business, and
indeed some people are betting that they are going to continue to stay in
business. But even at the worst depth of the worst depression, there will be
advance orders. Moreover, such orders will have picked up quite a bit since
the period before we hit the bottom of the depression, because up to that
time we were living on the fat of inventories that were too big. Once again,
the only inference that can be made legitimately from the fact that advance
orders are beginning to pick up is that some sectors of the economy look as
if they are not going to get any worse. That’s a long way from saying that
they are going to get better. It should be a sign that they are going to get
better if the advance orders rose to a point above that of the previous
healthy year. So far, we are nowhere near that point. So we certainly haven’t
any evidence from this indicator that we have turned the corner, that is,
have begun to climb back up.
Finally, the indicator which is referred to in the argument as the “best
news of all”: “the average income figures are showing a gain.” It’s likely that
the writer thinks of this fact as the best news because it refers to the eventual
effect on the consumer. That opinion shows that he (or she) has a heart but
it doesn’t say much for the head. What can you tell from the fact that
average income is increasing? The easiest way to see the loopholes in this
inference is by describing a couple of situations where the average income
will show an increase but where one would not be at all inclined to say that
the trough of the depression is behind us (the “counterexampling” tech¬
nique). In the first place, average income will increase at a time of rapid
inflation as a result of modest wage increases that only partially compensate
for the rate of inflation. This certainly describes the present situation. The
number of unemployed may remain the same, or perhaps increase slightly,
and those who are employed may be getting less real income than ever
before, although it looks more in terms of dollars. If we could be sure that
the “average income figures” here were corrected for inflation, we could rule
out this criticism; but we have no evidence to suggest that corrected dollars
are being used here. The lack of sophistication of the general line of argu¬
ment suggests that they are not, or the fact would have been mentioned.
Secondly, the average income figures may be going up simply because the
92 CHAPTER 4

top 10 percent of the population has been receiving very substantial in¬
creases in its income. The fact that the sales figures on extremely expensive
luxury cars, bookings at the most expensive resorts, and so on, have not been
significantly affected over the past year or so, and indeed have in many cases
shown a gain, is a weak indication that we may actually be in a situation in
which the wealthiest people in the society are, on the average, doing even
better than before. That, of course, pulls the average up, even if the remain¬
ing 90 percent are worse off. To mention a third possibility, recent increases
in the period during which unemployment benefits may be collected, and in
the size of the checks for unemployment and for retirement, will tend to
improve the average income figures, in uncorrected dollars, without really
producing a durable gain in welfare. So nearly everybody may be eating less,
having less to spend, living in worse substandard housing, and still this index
might show a gain. It’s a sign of extreme insensitivity to the full range of
implications of the word “depression” that somebody should think this to be
a significant indicator of improvement.
Notice that we haven’t made much effort to formulate the assumptions
that would make each of these subinferences sound, because they fall into
the category of being plain errors of inference, best pointed out as such.
Now we come to another inference, where it’s more useful to formulate the
missing premise. The inference to “America has turned the corner on a
depression” is clearly unsound, even if none of the preceding criticisms had
any merit. For it is supposed to follow from the claim that “the many in¬
dexes of recovery are showing optimistic readings.” And that conclusion
assumes that the indexes quoted are either all the indexes of recovery, or at
least a representative sample of them. Once we have formulated that as¬
sumption, it’s pretty clear that there are troubles with it. Certainly this isn’t
the whole set of indexes, or even the whole set that is usually used; for
example, the gross national product is not mentioned, nor are new housing
starts, the cost-of-living index, the rate of investment in capital improve¬
ments and machine tools, families below poverty level, and similar indica¬
tors. Hence, we can say only that evidence has not been supplied here that
supports this conclusion; it may be true, but as far as this argument goes, it
hasn’t been shown to be true.
Now we turn to the last two conclusions which are said to follow from
America’s recovery. The first says that “the doomsayers have been discom¬
fited.” Since the preceding argument doesn’t show that doom has either been
averted or is showing signs of being averted, the doomsayers have scarcely
been discomfited. At least, not if they have any capacity for argument analy¬
sis. But there isn’t anything new that’s wrong with the inference to that
conclusion.
The last conclusion of all reaches out a good deal further: “The free
enterprise system has been (once more) vindicated.” There are a number of
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 93

assumptions required in order to get this conclusion out of the preceding

argument, beginning with the assumption that the preceding argument is
sound. Since it isn’t, a link in the chain of inference is snapped right away.
But strong criticism always requires that you go on and look at other possi¬
ble criticisms as well, and there are other assumptions involved here. It’s not
too clear what “the free enterprise system’’ is, exactly speaking, or what it’s
supposed to do. Is it supposed to guarantee that there won’t be any depres¬
sions? An extreme laissez faire view wouldn’t guarantee the absence of de¬
pressions. Is recovery from depressions enough to “vindicate” free enterprise
systems? Shouldn’t there be some reference to the welfare of human beings
in the system? Is a system “vindicated” when it has 10 percent of the popu¬
lation suffering from malnutrition? All these questions need to be answered,
and they all bear on the assumption that the free enterprise system has been
vindicated if “we have turned the corner on the depression.” Perhaps vindi¬
cation requires that they not last very long—remember that the last Great
Depression is usually thought not to have ended until World War II began,
so that the free enterprise system, which was somewhat freer then, according
to most definitions, didn’t pull us out of it even after twelve years. And
finally, this last inference certainly involves the assumption that what we
have in America is the free enterprise system. Whereas many of the ques¬
tions that have just been raised are of the kind that would be raised by
people of the left-wing persuasion in politics, people of the right-wing per¬
suasion would be likely to argue that we do not have a free enterprise system
here. And indeed, they might well argue, it is because of the extent of gov¬
ernment control that the depression of the 1970s occurred and continued as
long as it did. So conclusion 9 is certainly the shakiest of all, depending as it
does on all the preceding assumptions as well as these extra ones.
We’ve built into our discussion of the argument so far many of the
other relevant arguments, so there’s no need to do so as a separate step. We
can sum up by saying that this argument is an extraordinarily weak one,
essentially worthless. And that goes for its utility in establishing all the
conclusions that it is supposed to establish.
Does that mean that the country isn’t coming out of the depression?
Not at all. As things stand, further decline is something about which nobody
can be confident. By the time you read this, we may have come out of the
depression or fallen further into it. Good arguments to show which predic¬
tion was correct are not available to us at this stage.

4-11

This completes the general review of the procedures for argument analysis at
a fairly serious level. It represents an adequate foundation for handling the
great majority of everyday arguments about important matters. The rest of
94 CHAPTER 4

the book will concern itself with refinements, elaborations, and special cases,
but by the time you have come this far and have done the quizzes on this
chapter, you should be in possession of substantially improved skills in argu¬
ment analysis, and, in general, reasoning skills that are much more efficient
than those with which you began.

QUIZ 4-A

1 Without looking back in the text, which of the following are correct?
a If the premise(s) of an argument is/are false, the conclusion(s) will be false,
b If the reasoning in an argument is unsound, the conclusion(s) will be false,
c If the premise is true and the reasoning unsound, the conclusion(s) will be
false.
d If a conclusion is false, either the reasoning is unsound or (one of) the prem¬
ise^) is false.
e If the premises are false and the reasoning is unsound, the conclusion(s) will
be false.

Answer

Just don’t read this until you’ve tested yourself thoroughly by evaluating each of
these statements. If you marked any of them true except the next-to-last one, you
were wrong. For examples and explanations, refer back to Section 4-1.

2 The reason(s) for getting clear about all the permutations and combinations in
the previous question (and Section 4-1 in the text) is so that you:
a Will be able to distinguish the question of whether the reasoning in an argu¬
ment is sound from the question of whether the premises or the conclusions
are true
b Won t jump to the conclusion that there’s something wrong with the reasoning
in an argument just because you happen to know the conclusion is false
c Will understand that your task in argument analysis is not over just because
you think you’ve discovered a false premise
d Will be able to see ways to test the truth of a hypothesis when you’re not sure
whether it’s true or not

Answer

It’s easy enough to say that the two possible targets in an argument are the premises
and the inferences (the reasoning). But students find it exceedingly difficult to keep
them distinct when it comes down to cases; they will frequently fall into traps such as
thinking that the reasoning in an argument must be sound since they know the
premises were true and that the conclusion is true. But in such a case, you have no
grounds for this judgment. And when you are assessing the reasoning skill of some¬
one, perhaps an authority you wish to refer to again or a potential employee you’ll
need to rely on, it would be a serious error to infer that their reasoning is good just
because they generated a true conclusion from a true premise; they may have made
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 95

two errors in the reasoning which happen to be in opposite directions and to cancel
each other out, no thanks to the reasoner. Not only people, but also scientific theo¬
ries, generate consequences; and the standard procedure for testing a theory is to see
whether its consequences (i.e., what can legitimately be inferred from it) are true. If
they are not true, and if you can be sure that your reasoning was sound, you can
conclude that the theory is wrong. For all these reasons, you have to be absolutely
confident that you understand the difference between sound reasoning and true
premises or conclusions. That’s why we make such an issue of this difference at this
stage: To understand this is to understand what the study of reasoning is all about
and why it's important. All four of the responses given here are true.

3 Circle the letter S for sufficient condition, N for necessary condition, and B for
both, in the following cases:
S N B a Being born is a-condition for someone’s being alive now.
S N B b Being born is a-condition for your dying someday.
S N B c Being born is a _ condition for your being a great poet.
S N B d A reduction in the rate of population growth is a_condi¬
tion for the avoidance of eventual overcrowding and wholesale
starvation.
S N B e Discovering an extremely lightweight and cheap storage battery
(say 1 percent of the weight of lead-acid batteries) is probably a
-condition for stabilizing our national gasoline consump¬
tion.

Answer
a Necessary but not sufficient, since a great many people were born who aren’t
alive now: we all also needed food and care, and we don’t live indefinitely,
b Necessary and sufficient, since, empirically speaking (which is clearly what we are
doing here), once you’re born, it’s guaranteed that you’ll die, and for you to die it’s
necessary that you were once born. If we were talking about what’s logically neces¬
sary and sufficient, that is, guaranteed by the very meaning of the words, the answer
would be “neither necessary nor sufficient,” since it’s logically possible that people
might be created full-grown by the whim of a wizard, and that they might be immor¬
tal. So birth would neither guarantee eventual death nor have to precede adulthood,
c Necessary but not sufficient; it takes a little something extra,
d Unless an infinitely expandable food supply or an infinite improvement in the
efficiency of its use emerges, the mathematics guarantee this result. Since there are no
real infinities, this is pretty close to being a logically necessary condition,
e S (probably) but not N—there are other ways to achieve this result, hence this
isn’t an N condition for it. It’s only probably sufficient because weight and cost aren’t
the only relevant factors; by-products or risks from explosion or leakage rate are all
possible sources of trouble. Probably we should just say that this is possibly a suffi¬
cient condition; it’s a start toward one.
4 Which of the following is an argument or part of an argument; that is, which is
the assertion of something as a reason which is supposed to justify a conclusion,
a judgment, or an action? Which of them is informative in some other way? And
 96 CHAPTER 4

which is or could be both? Circle A for argument, O for other, and B for both.

A O B a The sailplane with a small auxiliary motor is increasingly popular

partly because it is economical with fuel.

A O B b Hang gliders quite often crash for reasons that are never deter¬
mined.

A O B c Private planes most frequently crash because they weren't tanked

up or their pilot was.

A O B d If we don’t loosen up the penalties for cocaine use and sale, we’re
going to repeat the savagery and alienation caused by the pot laws.

A O B e If we don’t crack down on drug-running, we’re going to see a

tremendous increase in the power and profits of the underworld.

Answer

In the type of behavior which we describe as rational people do whatever they do

because there are good reasons for doing it, and those reasons are what produces or
causes their behavior. Thus, reasons are, in such cases, also causes. Take statement a
above; in part, this property—economy—of the motorized sailplane has caused the
increase in sales and it also provides a good reason for purchasing such a machine.
So one might set this out as an argument, with one premise about the economy and
a missing premise about the increasing shortage and cost of fuel, and we can reason¬
ably infer from those conclusions that these powered sailplanes are sensible pur¬
chases. We can’t infer that they will become, or are becoming, increasingly popular.
That inference would require another premise about people’s being rational, one
about costs being reasonable, and so on. Whenever, as here, a reason serves as a
cause, there will also be an argument which exhibits the fact that it is a reason for the
conclusion which the people it has influenced have accepted. But the best choice here
is O; for, as it stands, the quotation does not try to argue for the conclusion that sales
are increasing, it tries to explain that situation. If you asked a purchaser—say. Bill
Bottom—to explain his action, he’d probably mention the same factor, and it would
still be an explanation, not an argument. But if you asked him to justify his purchase,
he’d produce the argument, appealing to the same factor but leading to the conclu¬
sion that he (or one ’) should make this purchase. The appearance of the term
because thus does not always justify both a causal and a reasoning interpretation.
Look at statement b. The claim here is just a causal claim. This statement is
informative, and quite significant since there were about fifty deaths of hang-glider
pilots in the past twelve months, and there is a move afoot to have the Federal
Aviation Agency send in its crash-investigation teams to find out more about the
causes ot fatalities. Its generally believed that a number of manufacturers are put¬
ting out extremely dangerous designs. Be that as it may, there’s no latent argument in
the statement we have quoted here, and the answer is O. Statement c is also correctly
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 97

identified as an O. Of course, you can construct an argument by adding various extra

premises, which will enable you to infer from lack of gas or excess of alcohol to the
probability of a crash. But that argument is just using c as a premise which expresses
a causal connection. The question here is whether c can also be interpreted as an
argument of the form "q because p”; and it can’t, because you don’t justify crashes
by stating that the pilot was intoxicated, you only explain them that way. In a, by
contrast, you can both explain and justify the purchase of powered sailplanes by
appeal to economy, so its classification depends on the exact form of the conclusion.
Let's look at statement d. Is this just an empirical conditional statement like, “If
you turn that switch to the left, the fanspeed will go down” which just expresses a
causal connection. Or does it also (or instead) give you a reason for a conclusion like,
“If these fingerprints match, then we'll have to conclude that Sandovan was the
thief”? The answer is B (both)—we’re providing explanations that are also potential
justifications or premises for this conclusion. The same is true of e, though here the
conclusion is a prediction, not a moral judgment. (Note: these are difficult distinc¬
tions to make, as hard as anything in this text. Most actual examples are pretty easy
to classify.)

QUIZ 4-B

Answer T(rue) or F(alse) throughout questions 1 and 2.

1 a If the conclusion of an argument is false, the premises must be false.

b If the conclusion is false, the reasoning (inference) must have been unsound,
c If the premises are false and the reasoning is unsound, the conclusion will be
false.
d If the premises are true but the reasoning is unsound, the conclusion will be
false.
e If the conclusion is false and the premises were true, the reasoning must have
been unsound.

2 a “p, therefore q” is an argument.

b “If p were true, then q would have to be true” is an argument,
c “q follows from p” is a conditional (or hypothetical) claim,
d “p, so q” is a conditional claim.
e “p implies q” tells us that p is a sufficient condition for q.
f “p implies q” tells us that q is a necessary condition for p.
g “p is a necessary condition for q” tells us that we never can have p without q.
h “q is a necessary condition for p” tells us that we never can have p without q.
i “r is a sufficient condition for s” tells us that we never can have r without
having s.
j “t is a sufficient condition for r” tells us that r is a necessary condition for t.

3 If you were deciding how to structure (analyze) a passage of prose, in which of

the following cases would you break the sentence up and show it as involving two
components, one of them a reason for the other? (In the remaining cases, you
would leave the sentence as a single causal claim.)
98 CHAPTER 4

a Freud disregarded the danger to his health from his heavy cigar smoking
because he was incapable of giving it up.
b Freud should have given up smoking because it clearly was killing him.
c We know that Freud was a compulsive smoker because he tried to quit on
many occasions, knew it would cost his life to keep on, and was not suicidal,
d Since 80 percent of California motorists ignore the 55 mph limit despite the
California Highway Patrol’s tripling its speeding citations and now spending
50 percent of its time on enforcing this one law, it’s futile to continue to
enforce it.
e Since you went away, it’s been blues every day.

4 In this passage, the assertions have been numbered for you. Answer the state¬
ments a through k with T(rue) or F(alse).

The Ford policy on releasing price control on domestic oil and gas is simply a
way to give big business bigger profits at the expense of the average citizen (1).
Of course, the alleged reasons are to encourage further development of our own
resources (2), and to decrease total consumption (3) and hence the need for
imported oil with its attendant dependence on overseas politics (4). But there’s
no need for exploration and development to be funded out of profits; loans are
available for that (5). And the profits are in fact not going into development
(6)—Mobil just bought a hotel chain, for example (7)—because there are more
profitable places for them (8). No, it’s just another rip-off of the general public
(9)—a tax of over $ 1,000/year per family (Kennedy’s estimate) (10) to support
Big Oil (11) and Ford’s desire for power that can’t be fettered by the sheikhs
(12).

©
a Mostly emotional polemic, not a rational argument,
b Involves some emotional language, but a good deal of argument,
c It could perfectly well be entirely couched in emotional language and still be
an entirely rational argument,
d The main conclusion is assertion 1.
e The main conclusion is assertion 2.
f The main conclusion is assertion 3.
g The main conclusion is assertion 4.
h There’s a subargument here, which looks like this: Unstated conclusion: Price
control on domestic oil should be abolished,
i This subargument is said to be sound,
j One of the premises of this subargument is said to be false,
k The overall structure consists in refuting the subargument and offering an¬
other explanation for decontrol of prices.
DEVELOPING THE CONCEPTS OF ARGUMENT ANALYSIS 99

5 What’s another missing premise in the arguments (a) in the next-to-last para¬
graph of Section 4-4 (p. 66), and (b) on pp. 86-87?

6 For what kind of argument is it a fatal defect if it can be shown that there is even
a possibility of a certain type of situation occurring? (See the end of the second
paragraph of Section 4-7, p. 74.)

7 “It's worth using some cumbersome or even ungrammatical language to avoid

sexism, just as it would be to avoid racist or religious slurs. The language is
committed to sexism far more than it is to other forms of prejudice, so corrections
are more intrusive in this case. But surely no less important! Grammatical stan¬
dards are not as important as the avoidance of sex-role stereotyping.”
Is this argument or diatribe? Analyze it to answer, and if it’s an argument, go on
through the Seven Steps.

QUIZ 4-C

This quiz is simply a one-step-more-difficult version of Quiz 4-B; it covers the same
ground, instead of getting into more philosophical issues as we sometimes do with
the C quizzes.

1 Without referring to the text, fill in the gaps in the table below when it is possible
to give one definite answer, using the letters F for false, T for true, and S or U for
sound or unsound reasoning. Put a question mark where it isn’t possible to be
definite.

Premises Reasoning Conclusion

a F

b S F

c U F

d T s
e s T

2 The differences between an argument (let us say, from a premise p to a conclu¬

sion q) and the inference (from p to q) on which it depends are that (mark T or
F):
a The argument but not the inference attempts to establish the truth of q.
b The argument but not the inference involves the claim that p is true,
c The inference but not the argument is unsound if p is true but q is false,
d The inference but not the argument expresses a hypothetical or conditional
connection between p and q, that is, it asserts only that if p were true, q would
be true.

3 To the best of our knowledge, which of the following is a sufficient condition and
which is a necessary, condition for whatever follows the blank space? Circle S, N,
or B (for both) in the margin if and as appropriate.
 CHAPTER 4
100

S N B a Overeating and underexercising—premature death.

S N B b Heavy cigarette smoking, defined as three packs or more a day-

lung cancer.

S N B c The number x is divisible by 2—The number x is divisible by 4.

S N B d Being a strong supporter of one side in a political issue—being

biased.

S N B e X occurred before Y—X caused Y.

4 Do a complete analysis and assessment of the argument in question 4 of Quiz

4-B.

5 Suppose two educational experts had the following argument:

“A text which is supposed to teach skill in the interpretation and use of lan¬
guage,” one expert said, “should at least be well written, and laid out and printed in
exemplary fashion.”
“Not if it means a 25 percent increase in cost,” the other retorted, “and/or a
year’s delay in publication, and/or the impossibility of updating the contents at
intervals of less than several years, and/or censoring many examples, perhaps not
coming out at all—as long as the skills are taught.”
“Some skills might still be taught, but the bad example of such writing might
unteach other, delicate, skills—the skills of writing with style and sensitivity, for
example.”
“If they can be unlearned so easily, they won’t survive long in this world any¬
way. Of course, it’s much better if an author does combine the two—literary and
analytic skills—but in an imperfect world, some things may be worthwhile that do
just one job well. Or even fairly well, if the job is done better than elsewhere.”
“Justifying sloppiness sounds like sophistry.”
“Valuing refinements over reasoning competence sounds effete.”
Who won, if anyone? Apply your conclusions to this text.
 Chapter 5

Clarifying Meaning

You now have some basic fishing tackle in the box; the time has come to
add extra reels and lines and rods for a more specialized and powerful
treatment of particular cases. We’ll do this first by expanding on the main
steps in the argument analysis procedure, and then we’ll pick up a series of
special cases that bring in several parts of argument analysis—for example,
special techniques for handling long arguments.

5-1 An Illustration

This chapter on the clarification of meaning has, of course, a great range of

application in reasoning outside argument analysis. In all functional and
much literary prose, the techniques described here will often turn out to be
useful. It must be stressed, however, that we are trying to develop much
more than a list of procedures that you can look up in the way you might
look up the instructions in a home repairs guide when the plumbing gets

101
102 CHAPTER 5

stopped up. We are trying to develop a habit of thought, an attitude, a set of

instinctive reactions that will become part of your normal thinking patterns.
When you hear a commercial that says a particular brand of ballpoint will
write “up to 40 percent longer or more!” your response should not be to
suppose that the manufacturer is claiming savings for you of more or less 40
percent. You should immediately recognize this as an essentially empty
claim. Most of you probably do realize this intuitively, but let’s look at
exactly how your reasoning would appear if we had to “unpack” the claim.
Doing so will help us with less obvious examples.
The first key move is to ask “longer than what?” Longer than you
sometimes probably get from your present brand? Individual pens of the
same brand often (as you know from experience) vary by 100 or 200 percent
in the length of writing life, so even if you keep using the same brand it will
be able to make a better claim than the ad does; its advertising can say “up
to 100 percent more.” Longer than the most widely used ballpoint? Than the
worst of those commonly sold? But any brand, even the worst, could make
this claim, since some samples of it will certainly write “up to 40 percent
longer than others. So the first loophole in the claim is the lack of specificity
about the comparison. Let’s shut this loophole up and see what others there
are, if any. In closing this loophole, we’ll learn something about how to
express claims like this better when we ourselves want to make them.
Let’s add to the claim, after the word “longer,” “than the average for
the most commonly sold brand.” And, since we should be comparing aver¬
ages with averages, we’ll assume that the claim is to be taken to mean that
the average performance of the new brand is (so much) longer than the
average of the most commonly used brand (Bic, at the moment).
This tidying up of the claim has been done for our benefit, to make it
easier for us to see if there are any other sources of trouble (that is, unclarity)
in the claim. The fact that we’ve tidied it up doesn’t in the least mean that
there wasn’t something wrong with it as it stood. We have identified the
“something wrong,” namely, the failure to specify what the comparison is
with, and that loophole has now been identified, charged against the claim
and put on one side for the time being. (It’s charged against the claim, to
repeat, simply because the claim is true of any brand as it stands; hence it in
no way supports the superiority of the advertised brand, which we know
from the context is what it is supposed to do.) Now let’s look at the modified
claim. The claim, as rather generously refined by us, reads that Brand X (the
one being advertised) will write “up to 40 percent longer, or more, than a
Bic,” where Brand X’s performance is to be defined as the average perform¬
ance of a large number of samples of that brand. The remaining problem
concerns the meaning of “up to 40 percent or more.” Five percent is obvi¬
ously within the “up to 40 percent” range. So is 1 percent; so indeed is 0
percent. It looks as if the real meaning of “up to 40 percent” is “not more
CLARIFYING MEANING 103

than 40 percent. Since that includes 0 percent, it’s a really empty claim.
Any brand of anything in the world is up to 40 percent better than any other
brand, because it’s at least no percent better.
But that's being a little too persnickety. When people say, “up to 40
percent,” they certainly are committing themselves to the claim that at least
sometimes there’s a gam of 40 percent. When we add to this the phrase “or
more,” we certainly mean that the gain is at least very occasionally more
than 40 percent. Now, what gain? The trouble is that when we compare
averages, as we’ve agreed to do to make some sense out of the rest of the
claim, there isn’t going to be any “sometimes” feature about the results. The
avefage performance of Brand X will be exactly 2 or 25 or 43 percent more
(or less) than the average performance of the Bic. What has really happened
here is that we have so vague a claim that any attempt to clarify one part of
it makes the other part senseless, and vice versa. Or are we being unfair?
Was our particular way of tightening up too restrictive?
Suppose we hadn’t tightened up the implicit comparison by specifying
what it was a comparison with and how the comparison was to be made.
Then the phrase “up to 40 percent longer or more” would make sense,
because particular samples of Brand X will certainly produce results that are
anything up to 100 percent better than particular samples of any other
bfand (indeed, of Brand X itself!). But the claim is now empty just because
all it tells us is that particular samples of any brand vary a great deal, which
hardly shows any superiority of Brand X over any other brand.
Now that’s a fair amount of unpacking of an apparently simple phrase,
but it’s a phrase that is very commonly used in advertising, and the exact
reasons why it is a misleading phrase do take a little sorting out. It’s mislead¬
ing because it suggests that it’s a tangible, indeed a quantitative, claim about
the superiority of Brand X, but when we get down to unpacking it, we
discover that it really doesn’t contain any promise of actual superiority at
all.
Let’s reflect for a moment on what sort of procedure we used in the
course of this analysis. We considered various specific and clearly meaning¬
ful interpretations of the claim, and then matched them against it to see if
we were either missing or oversimplifying some part of its meaning. Then,
having identified one particular source of trouble, we tidied the claim up so
that it wouldn’t confuse us while we were searching for other sources of
trouble (in this case, “trouble” meant lack of clarity). We’ll find that these
procedures are frequently useful as we go further into this chapter. Making
something more tangible, making it more precise, more specific, is a most
important way of trying to find out if you understand what it means. Con¬
versely, when you want to make something clear yourself, then you try to
make it as tangible and specific as possible.
104 CHAPTER 5

5-2 Real Precision and Fake Precision

The preceding example is one where an exact percentage is quoted, but in a

context which takes away from it any real meaning. That’s a case of fake
precision. Recently we’ve seen examples of TV commercials showing a small
imported car traveling from Los Angeles all the way up through and past
San Francisco, with an accompanying patter which shows that this is a
specific instance of what it means to have a car that will do “540 miles on a
tankful of gas a day (Environmental Protection Agency figures).” That’s a
slightly different example of fake precision. Since the EPA figures are ob¬
tained on a dynamometer and make limited allowance for wind resistance,
and the ones being used refer only to highway driving at a speed well below
the normal highway driving speed, and since traveling over that stretch of
road involves a great deal of city as well as highway driving, there is no way
in which any ordinary driver could expect to obtain anything like 540 miles
to a tankful. The 540 is indeed a very precise figure; the trouble is that it
doesn’t apply to the picture. The picture is one of a fantasy performance, not
a real one; the 540 miles to a tankful is so specific that it looks like a figure
obtained from an actual performance. You are therefore being persuaded
that you are seeing photographs of a car actually getting this precise mile¬
age, whereas what you’re really seeing is a photograph of a car doing an
actual trip, combined with an exact figure that is obtained from calculations
that have very little to do with that trip. This is theoretical precision mas¬
querading (or being misrepresented) as practical precision.
Another type of fake precision is involved when somebody quotes a
person’s height at 5 feet 8.62314 inches. Measurements of a person’s height
are not meaningful beyond about a tenth of an inch (and perhaps not even
that far) because slight posture variations or the length of time that the
person has been standing up before the height measurement is taken, or the
length of time since the person has had a night’s sleep, or the extent of
fatigue from other causes, can well produce variations of that amount sev¬
eral times during the course of a day.
There are many other ways in which numbers are used in order to
exploit the impression of precision that they give, but in a context which
makes clear that the precision is inappropriate. Some of these are quite
subtle. In the example of the fake range on a tankful of gas that we’ve just
discussed, the way in which you’re persuaded to think of this as realistic is
by the combination of photographs of a real car traveling on a real highway
with the superimposed mileage figure. Nobody actually said in the commer¬
cial that you would really get this mileage from your specimen of this car;
indeed, the advertisement even says (in small print) that the figure is based
on an EPA estimate for highway mileage. But few of the viewers realize the
extent to which that undermines the figure completely. It says 540 miles; the
CLARIFYING MEANING 105

real figure isn't 539 or 538 or 530 or 500; the real figure for that trip, run at
normal speeds, would probably be around 375. (We happen to have that car
in the family, and it has yet to produce as much as 400 miles on a tankful of
gas even on steady freeway driving at or below the speed limit, and dealers
as well as other owners agree that this is normal.)
Another style of TV commercial that is currently popular involves the
photographed interview with somebody who has been asked to identify
which pile of laundry has been washed with the detergent that’s being adver¬
tised—let’s call it Bright. A sense of veracity, of accuracy, of completeness of
reporting is provided by the fact that you’re obviously watching a film or
video tape of an actual interview. (Let’s put aside for the moment the ques¬
tion of whether actresses are used to fake that aspect of the matter and
assume that, as appears to be the case, it really is an interview with a house¬
wife.) It turns out that the housewife identifies, in this “blind” test, the
whitest clothes as being the ones that were washed with Bright. (In a similar
commercial, a housewife identifies the cracker that was spread with Imperial
margarine as “the one that was really spread with butter.” Should you, or
should you not, give some weight to these commercials as supporting the
view that there is an obvious benefit from using the advertised product?
Our answer depends entirely on how typical this filmed event is of what
happens in reality. That it happened in reality we can conceive. But is this
all that ever happens in reality? To be blunt about it, did the ad agency have
to film fifty of these scenes before it came across one where the housewife
thought that the Imperial was the butter, or the Bright pile was the cleanest
pile? If there were fifty filmings and one case in which things went the way
the advertiser hoped they would, then we have an extremely misleading
commercial, even though what it portrays as happening is exactly what
happened. So “the camera cannot lie” is far from the truth, at least is far
from the truth that we’re interested in. The camera isn’t lying about what
happened; but it is being used to lie about the significance of what hap¬
pened. You can’t blame the camera, but you can blame yourself if you make
the step from photography to interpretation too hastily. So accuracy or pre¬
cision is a function of context as well as of the use of quantitative measures
or photographic reproductions. In short, there’s often an element of interpre¬
tation involved that undoes much of the apparent accuracy or precision that
the use of figures or photographs may convey.

5-3 Precision versus Imprecision or Vagueness

Of course, the natural contrast with precision is imprecision, the difference

between giving somebody’s height in terms of feet and inches, and perhaps
the nearest quarter inch, and saying that “the person is quite tall.” Just as
precision is phony when it represents more accuracy than can in fact be
obtained, so it is unnecessary when it represents more accuracy than can
106 CHAPTER 5

reasonably be used. It is possible to weigh diamonds to the millionth part of

a gram, but there’s no real point in doing so, and the process would be
immensely complicated and expensive. For example, one couldn’t use even
the most refined top-loading balance because the kind of air currents that
result from changes of temperature in a room in the course of a day, or
because of the movement of people in the room, or because of leakage of air
around the windows, would deflect the balance by many times a millionth of
a gram. How much more absurd it would be to talk about weighing pack¬
ages of sugar to the millionth of a gram, although it is technically feasible.
Precision is, like many other devices in reasoning and in science, a tool to be
used appropriately, not a quality to be worshiped blindly. Hence, it is a
mistake to criticize an argument or an assertion on the ground that it con¬
tains some vague language—perhaps it describes some people as being
“pretty tall” or some sacks of sugar as being “heavy” instead of giving a
quantitative measure of their height or weight—unless the lack of precision
hurts the argument.
When people begin to work on reasoning skills in a careful way, when
they begin for the first time to become analytical and careful about the
analysis of meaning, the three most common mistakes they make all fall
under the heading of this chapter. They tend to complain inappropriately
about the use of vague terms, about the use of emotional language, and
about the failure to define tricky terms. As those of you who have taken
courses in science and mathematics will be well aware, they make a good
deal of the idea of “the number of significant digits” in spelling out a quan¬
tity. You were taught carefully not to give more digits than represent phys¬
ical reality, or the limits of measurement, or are mathematically appropriate.
For example, if you are determining the average weight of thirteen samples
of a chemical, and you can measure each of them accurately to a tenth of a
gram, you should not quote their average weight to a ten-millionth of a gram
just because, when you come to divide by 13, you can nearly always generate
as long a string of decimals as you like by just continuing the calculation.
We’re making the same kind of point here, but in the general context of
reasoning. Don’t complain about axes because they’re not scalpels, don’t
complain about useful approximations because they’re not exact.
On the other hand, there are situations where it’s really crucial to have
a certain amount of precision in order to justify the conclusions that you
need. If you want to determine whether drinking Coca-Cola causes cancer,
you’ll have to be extremely accurate in your measures of the age of the
patients that you put into the experimental and control groups, and as accu¬
rate as possible about the length of time that they have been drinking Coca-
Cola, and the amount of it they drink. Otherwise, you may well pass over
what is actually a very significant result. Or you may be led to conclusions
which can t be justified by the degree of accuracy provided in the data. So
CLARIFYING MEANING 107

here we have a case of the exercise of judgment; you have to learn how to
judge what an appropriate degree of accuracy or precision is for certain
types of conclusions. We’ll come back to this point from time to time, with
various examples. It’s only a matter of refining your common sense some¬
what.
First, let’s look at the way in which increasing precision can be useful
(assuming it can be justified by the appropriate measurements). Suppose
that we described somebody as “tall.” Let us suppose that the complete
sentence says “Jones is tall.” What does this tell you? Certainly it’s vague,
imprecise, but still it tells you something. We might say that it certainly tells
you that Jones is more than 5 feet 6 inches in height. But wait a minute. A
man that is said to be 5 feet 6 inches high wouldn’t be called tall, but a
woman of that height might be said to be tall. Certainly a Japanese woman
of that height would be considered to be tall. Even more certainly, a man
from one of the pygmy tribes in Africa would be said to be tall if he was 5
feet 6 inches high. So if you try to translate “Jones is tall” into “Jones is
more than 5 feet 6 inches high,” you’re immediately making some assump¬
tions. You’re assuming that “Jones” is the name of a Western European or
an American man. Now, Jones is indeed a very common name for people
from English-speaking countries, but, though less commonly, it is also the
name of people of all races. Certainly one is not entitled to the conclusion
that Jones has to be the name of a man, since women are increasingly
referred to by their family names in a way that was once upon a time
reserved for men. What we see here is a typical example of a contextual
inference, and a rather shaky one at that. Of course, the fact that this hap¬
pens to be a shaky example in no way justifies the conclusion that contextual
inferences in general are unreliable. Since a very large part of our inference
involves the use of contextual cues, and since a great many of these are
extremely reliable, it would be foolhardy to draw a generally skeptical con¬
clusion about the inference.
Let’s pursue this example a little further. Can we at least conclude that
“Jones is tall” means “Jones is at least 5 feet high,” if the sentence is encoun¬
tered in some context which shows that it refers to an individual in, let us
say, the North American continent. This takes care of other races whose
average stature is considerably less than that of the usual American or Cana¬
dian; but it still won’t do. It’s quite common in certain schools to refer to the
children by their family names. And a nine or ten-year-old, or even a twelve-
year-old, might well be said to be tall (meaning “tall for his or her age”) even
though the child was only 5 feet high. So “Jones is tall,” even with the help
of some contextual assumptions, won’t do you much good unless you can
definitely count on Jones being an adult. If the context is such that we can
tell the “Jones” being referred to is an adult, it is reasonable to conclude that
the sentence implies that Jones is more than 5 feet tall (or have I overlooked
108 CHAPTER 5

another possible exception?). And that’s quite a significant piece of informa¬

tion. Probably it’s correct to say that the sentence contains rather more
information than this. For example, one might say that it implies that if
Jones is an adult man (and a resident of North America), then Jones is more
than 5 feet 6 inches tall; and if Jones is a woman, then Jones is probably
more than 5 feet 3 inches tall. It’s less and less likely that one would describe
a woman as tall as her height gets nearer to 5 feet; hence, the meaning of the
assertion makes it virtually certain that the individual is more than 5 feet 0
inches high, pretty certain his or her height is more than 5 feet 2 inches, very
likely that it’s more than 5 feet 3 inches, quite likely that it’s more than 5 feet
4 inches, and likely that it’s more than 5 feet 6 inches. Notice that in order to
understand the real import of this sentence, you need to know quite a lot
about the distribution of heights in the relevant population, so the “mean¬
ing” of a sentence will often depend upon the possession of a good deal of
factual information. In a very strict sense, the meaning is just what follows
from the dictionary definition, which supposedly doesn’t depend upon mere
facts. But, to be realistic, the meaning of “tall” would shift pretty quickly if
the population of North America lost several inches of height overnight. For
“tall” is a comparative term, as we’ve come to see in considering cases of
pygmies and children, men and women, Japanese and American. Couldn’t
we say that the true meaning of “tall” is “greater in height than the aver¬
age”? We could, but what this comes down to, in terms of feet and inches,
will of course vary with the average. A strict logician might wish to continue
to argue that “tall” doesn’t have as part of its meaning any “feet and inches”
content at all. But, in the practical language of everyday life, it does have
that content because the description of a person as tall always occurs in a
context where it’s possible to work out the race, age, and sex of the person,
and a contextual inference thus makes possible a conclusion (an interpreta¬
tion) in approximate measurement terms. If you had to give an “absolute”
definition of “tall,” one that would work in all contexts, you’d have to say
that it refers to the average, whatever that happens to be: that is, it would
have to be a relativistic or comparative definition, with no specific content at
all. But if you were giving an explanation of what it means in practice in our
context, when applied to people and not coastal redwoods, you could do
better than the comparative definition.
Thus we see that a vague term (1) has some meaning, (2) often has a
good deal more meaning when we have a context from which we can make
further inferences, (3) can be given an abstract definition by referring it to
another abstract term (“average height”), and (4) can be given a somewhat
more precise meaning if we restrict that meaning to uses of the term in
particular contexts. All these are important messages about the clarification
of meaning, its limitations, and its possibilities, and we’ll rely on them on
many future occasions.
CLARIFYING MEANING 109

5-4 Information Content

Why is precision ever valuable? How can we ever legitimately complain

about vagueness? It may be helpful to think about these concepts in terms of
a notion which we can call “information content.” What we shall try to
show is that the information content of a sentence, or of a description, or of
a descriptive term, goes up with increasing precision, provided that the pre¬
cision is real, not phony, precision. This will lead us to a general perspective
on the meaning of terms and of sentences that we’ll call the “contrast theory
of meaning.” This perspective isn’t anything very elaborate or abstract, but
it’s a useful point of view. You can see what’s meant by the contrast theory
of meaning if you look back at what we were doing when we were trying to
work out the meaning of the assertion “Jones is tall.” What we asked our¬
selves was, What does the assertion exclude? That is, what can it legitimately
be contrasted with? The contrast theory of meaning maintains that the best
way to understand the meaning of a claim (or a term, etc.) is to work out
what it is to be contrasted with. This can be done at various levels; for
example “tall” is of course to be contrasted with “short,” and also with
“average” (in height). That is, it marks one end of a spectrum of relative or
comparative heights. Its meaning is to be gathered from what it is contrasted
with on this spectrum. When we move to a more specific level where we try
to find some kind of feet-and-inches equivalent of “tall,” we see that, given
a good deal of help from the context, we can (sometimes) contrast it with “5
feet high.” So, on this level, the meaning of “tall” is “more than 5 feet high”
(and probably more than 5 feet 3 inches, fairly probably more than 5 feet 5
inches, and so on).
Let’s try to unpack this notion of the “information content” of a claim
a little further. Suppose we’re talking about somebody’s height. Suppose the
context makes it clear that it’s an adult Anglo that we’re talking about (and
not one who’s a basketball player or otherwise athletic). From your back¬
ground knowledge of adult male Anglos, you know that the scale of feasible
height, that is, physically possible height, runs from somewhere around 2
feet 6, for dwarfs, up through somewhere around 7 feet 6 for giants, with
perhaps a very tiny number of individuals going beyond these extremes. So
your background information, without anybody’s telling you anything specific
about Jones, tells you that his height is somewhere in this range, and in fact
tells you that it’s more likely to be in the region around 5 feet 9 or 10 than
out toward the extremes. We can actually represent this in terms of a very
rough kind of curve, showing the probability that Jones has any height from
2 feet 6 to 7 feet 6, and this curve would obviously start with a very low
probability down near the 2-foot-6 end and finish up with a very low prob¬
ability at around the 7-foot-6 end, and bulge upward pretty strongly above
the 5-foot-6 area. (See the accompanying diagram.) This curve is technically
110 CHAPTER 5

described as the distribution of the probabilities that a randomly chosen

individual from the population that we’re talking about (adult male white
Anglos) will have the height that is represented on the line along the bottom:
it is shown in a solid line below.

Here’s a graphical representation, then, of your background knowledge

about Jones’s height, with your knowing nothing specific about Jones, but
only these various contextual facts about the sort of person he is, in very
general terms.
Now consider what happens when somebody says, “Jones is tally We
can immediately replace the curve that we have from our background
knowledge with a new curve that looks something like the dotted one on the
diagram. It shows the distribution of height amongst men that might reason¬
ably be called “tall” by someone familiar with both the English language
and the facts about the height of North American Anglo males. The infor¬
mation content of the statement “Jones is tall’ might, for certain purposes,
be said to be represented by this second picture. But it might also be said to
be represented by the difference between the first and second pictures. What
has happened is that you’ve been able to exclude, with varying degrees of
probability, a whole range of possibilities that previously existed for you.
These two senses of information, the first sense referring to what we might
call the “intrinsic information content” of a claim, and the second to the
“new information content” of a claim, are both frequently used in our lan¬
guage. When somebody hears a political commentator tracking down the
sources of the large contributions to a particular political candidate’s reelec¬
tion campaign, and comparing these sources with the candidate’s voting
record on certain key issues, one might respond to these disclosures by say¬
ing “very informative,” meaning the information that one has learned a
good deal that one didn’t know before, and that one regards the information
as valuable. Somebody else, more of a political specialist in this particular
area, might comment that he hadn t found it so. He’s telling vou that it
didn’t give him any new information. To say that something is very informa¬
tive is typically to say that it s telling you something new. The individual
who didn’t learn very much from the political commentator’s remarks isn’t
CLARIFYING MEANING 111

for a moment claiming that these remarks have no information content at

all. They do assert certain facts, which as far as he knows are true; and there
are a great many of these facts, so that, in one sense, a good deal of informa¬
tion was being produced. But it wasn't very informative for this person; that
is, it wasn’t new for him. So “information” sometimes means “facts” and
sometimes it means “new facts.” Notice that the content will also not be said
to be information if it is false. “Very informative, if true,” the cynic may say.
Information consists of facts or knowledge, both of which terms imply truth;
and sometimes it is restricted to novel facts or knowledge, perhaps even to
new and relevant or interesting facts and knowledge.
So the information content of “Jones is tali' can be thought of as repre¬
sented either by the second curve above, or by the crossing out of the part of
the first diagram that is excluded by the second curve. That’s where we get
the “contrast theory of meaning” perspective: the meaning of a claim is
what it rules out. Just as there are two senses of information, corresponding
to the absolute or intrinsic information content and the new information
content, so there are two senses of “meaning,” corresponding to the “ab¬
stract or purely logical meaning” and to the “practical or significant” mean¬
ing. “Jones is tall” has a certain absolute meaning in that it excludes “Jones
is of average height or short”; in a particular context, however, it may have
an imprecise but substantial feet-and-inches meaning; and in a very specific
context, it may have even more “meaning” than that. The great detective is
informed that Jones is tall; he says, “Aha!,” and Dr. Watson asks him what
is so exciting about this rather vague and scarcely exciting description.
“Why, don’t you see what this means? It cannot have been Jones who was
seen by the stage door attendant at the theater; and that means that Jones
could not have been the murderer.” Logicians have often tended to suppose
that using the term “meaning” in this extended sense where it covers the
implications of a claim in a particular context is a rather sloppy use of the
term; that its proper sense is its acontextual or context-free sense. But recent
work in logic, and more attention to the subtleties of common language,
suggest that the detective’s use is an entirely proper use of the term, and a
much closer relative of the logician’s sense than was previously supposed.
The meaning of a claim, the information it contains, what we learn from it,
its significance, what it tells us—all these phrases have the double function
in our language of referring either to the changes in our knowledge state that
result from adding the new information, or to the strict out-of-context con¬
tent of the claim. Thus, switching from one of these interpretations of
“meaning” to the other, the same claim may be informative and not infor¬
mative, it may tell us something or not tell us something; “Jones is tall” of
course tells us that Jones is tall, but if we already knew this, then we might
equally well say that it tells us nothing. Sometimes we use the phrase “tells
us nothing that we don’t know already” in order to make this distinction;
112 CHAPTER 5

but sometimes we just say “He told me nothing at all,” although of course he
told you a number of things which you believed to be true. It was just that
you knew them already.
We’ve seen how even a vague statement like “Jones is tall” can have
considerable information content in either of the two senses that we’ve dis¬
tinguished. Consider now the statement “Jones is extremely tall.” We’d rep¬
resent that in the same way, with a curve which pushes the probabilities
considerably further up to the right-hand end of the spectrum of heights.
You might say that it narrows down the range of possibilities for Jones’s
height considerably. Notice that it’s becoming increasingly precise as a de¬
scription, and that as it becomes more precise, it rules out more and more of
the range of previous possibilities. That’s the general feature which lies be¬
hind the value of precision. Now consider “Jones is the tallest man alive.”
What you see is a continued shrinking of the range of possibilities until
finally it contracts to a dot. The description becomes increasingly precise,
has more and more information content for you just because it narrows
down the range of possibilities. Thus it excludes more and more of the initial
possibilities. The informativeness or information content of a claim is di¬
rectly proportional to the range of possibilities that it excludes. The more it
excludes, the more informative it is.
Now look at the situation if we start talking about quantitative mea¬
sures. Suppose we say, “Jones is more than 5 feet 6 inches tall.” We can
represent this on the diagram by blocking out all the heights below 5 feet 6.
Suppose that the description is now changed to “Jones is between 5 feet 6
and 5 feet 8.” Now we can block out all the heights above 5 feet 8, and we’ve
narrowed down the range of possibilities to a 2-inch range. Clearly the infor¬
mation content, the informativeness, of our description has increased dra¬
matically. Now if we say that Jones is in fact 5 feet 7 inches high, we’ve
narrowed it down even further. And, where the precision isn’t phony, as we
increase the accuracy of the measurement, we exclude more and more possi¬
bilities and hence increase the information content of our claim. (Notice that
this is true in either of the two senses of information that we distinguished
previously.)
This method of conceptualizing precision, information content, and
meaning helps one considerably in the reasoning process where one is trying
to clarify the meaning of the assertions in an argument that are possible
premises or conclusions for the argument; and it also helps one in under¬
standing and appraising merely descriptive prose.
Now let’s apply this approach to a general view of the world. Consider
the “Conservation of Luck” claim “Good luck in this world is always bal¬
anced off by bad luck. It sounds like a philosophical equivalent of one of
the conservation or symmetry laws in physics: for every force there is always
an equal and opposite force and so on. But what does it really mean? Ask
CLARIFYING MEANING 113

yourself what it excludes, what contrast it draws. For openers, does it mean
that good luck is always balanced off by bad luck for each individual per¬
son? Or does it mean that the balance is struck across the whole population
of the world? (One might get some pretty lucky people, but there would also
be some other people somewhere else who were unlucky.) Until we get some
clarification on this point, the Conservation of Luck principle is so vague
that it really doesn't tell you what to expect at all. Does it rule out people’s
being steadily lucky throughout their lives? Does it rule out, for instance, a
man's being stupendously lucky—winning the Irish Sweepstakes, for exam¬
ple, and thereby making over a million dollars which he invests in a com¬
pany which is wildly successful, the only occasion in which he has ever
invested in anything, enabling him to retire in his midtwenties in good
health which he retain during the rest of a long and happy life? Of course, if
it rules out this kind of thing, then it’s simply wrong, because there are
plenty of examples of both these types of persons. Or so it seems. Perhaps
the Conservation of Luck proponent thinks that there’s always some hidden
sorrow under the surface in these good-luck stories. Does he or she have any
evidence for that? Or is this just an article of faith? Of course, such questions
concern the issue of whether the claim is true, rather than what it means.
But, since people who put forward views like this want them to be true, they
often will interpret them in such a way as to make them more likely to be
true. And one way to do that is to make it more difficult to falsify them. And
one way to do that is to make them so vague that it’s almost impossible to
see what would falsify them.
So a Conservation of Luck supporter might well interpret the claim to
mean that there’s always some bad luck in everyone’s life, even though it
may not be quite as much as there is good luck in some cases (and vice versa
in other cases). So the Conservation of Luck claim might be meant to apply
to each individual, but it may not be the claim that the amount of good luck
is exactly the same as the amount of bad luck. Here we’d have the balancing
of good luck by bad luck, but only in terms of the number of good luck
events and the number of bad luck events, not the quantity of good luck and
the quantity of bad luck.
So there are two completely different questions that have to be settled
before we can tidy up the vagueness in this claim. First, is the Conservation
of Luck principle meant to apply to the life of each individual? Second, is it
meant to apply to the quantity of luck and not just to the occurrence of luck
of both kinds?
Notice that if we take it in the “weakest” of the interpretations that
we’ve been considering, the claim becomes almost empty. That is, it be¬
comes the claim that someone somewhere gets some bad luck sometimes to
compensate for the good luck that other people, elsewhere, get at some other
time. In other words, there is both good and bad luck in the world. That
 114 CHAPTER 5

doesn’t turn out to be a very interesting claim, since we all know it’s true,
and it isn’t at all what we had in mind when we first heard the Conservation
of Luck claim. The phrase “balancing off’ in the claim suggested a kind of
precision that turned out, under pressure, not to be definitely there.
Notice that, in order to understand what people might mean by this
claim, we’re asking them. If the claim were a simple one, expressed in Eng¬
lish, we’d know more or less exactly how to take it, and we wouldn’t have to
ask the speaker. But a phrase like “balanced off” is a metaphorical phrase in
this context; we know exactly what it means when we’re talking about
weighing one thing against another on a balance, but we don’t know exactly
what it means when we’re talking about weighing luck, especially good and
bad luck.
That example is fairly typical of a large number of claims about the
world that are expressed in terms sufficiently vague so that one really can’t
tell whether they are interesting or not. If we took the claim in its most
precise form, interpreting it to mean that within the life of any single indi¬
vidual, the amount of good and bad luck exactly balance, then, although it
isn’t too easy to say how a piece of good luck balances against a piece of bad
luck, we could certainly say that this is an interesting claim and perhaps
investigate it. All we’d have to do is to find a few truly fortunate individuals
whose good luck obviously was far greater than their bad luck, in order to
show that it was false. The claim that there are no such individuals does
seem pretty implausible; but it’s certainly a definite claim, and one that we
can investigate systematically, and one that, if true, would be pretty infor¬
mative. Notice the way in which the reduction of vagueness again increases
the amount of information content.
Let’s take another example. One constantly runs into remarks in low-
grade “Buyer’s Guides” columns in magazines to the effect that “you get
what you pay for.” This axiom is usually offered as a piece of homespun
wisdom subsequent to the discovery that certain feature or performance
characteristics of some product are fairly expensive. By and large, however,
it’s a sensationally stupid remark and, into the bargain, fairly vicious. Of
course it could be made so vague that it’s impossible to falsify it; suppose,
for example, that it is interpreted as meaning that what you get when you
pay more is something more expensive. Then it reduces to “you get whatever
it is that you pay for.” Interpreted in this way, it would be true but trivial.
Interpreted in the way in which it is usually meant, which can be translated
as “when you pay more you get more quality; when you pay less you get less
quality, it s not trivial and it s also not true. Any issue of Consumer Reports
contains a listing of products—usually half a dozen listings—which show
that inexpensive products often outperform, and give you better quality
(e.g., durability) than, more expensive products. In fact, anyone familiar
with the marketplace knows that there are always some products around
CLARIFYING MEANING 115

which are marked up simply in order to be expensive, in the hope that

people who are stupid enough to believe that “you get what you pay for”
will buy them in the belief that they’re better, or on the ground that one
achieves status through being able to purchase extremely expensive prod¬
ucts, regardless of whether they’re worth it. What one should achieve is a
reputation for being unintelligent, of course; but we’re a fairly badly edu¬
cated bunch ot consumers although we're overcoming that gradually, and
we still have people running around acting as if cost is a good indicator of
quality.
Look at the sources of error in that assumption. First, it completely
ignores any reductions of cost that are possible through large-scale produc¬
tions. In the automobile field, for example, a new Rolls-Royce Camargue
will run you over $70,000, whereas the Mercedes 690 SEL costs half that.
But, because the Rolls-Royce is made in extremely limited numbers, its
manufacturers are unable to capitalize on the economies of scale that are
available to the Mercedes factory, which can afford considerably larger runs
of components and accessories from its subcontractors, and hence pays con¬
siderably smaller unit costs. In terms of quality of workmanship, it simply
isn’t possible to argue for any systematic difference between the two. There
are a few little goodies on the Rolls-Royce, like the use of solid wood inlays,
that will not be found on the Mercedes; but there are a dozen engineering
goodies on the Mercedes that cannot be found on the Rolls-Royce, which
add up to the fact that the Mercedes 690 SEL handles far better, is a consid¬
erably safer car, a far more economical car, and a better-riding car, as well
as a better-performing car than the Rolls-Royce. Of course, it isn’t a Rolls-
Royce. So, in the trivial sense of “you get what you pay for” it’s true that
you get something for the $70,000 that you don’t get for the $35,000 or so.
But it’s also true that you get not just something but a hell of a lot of things
for $35,000 that you don’t get for the $70,000. (For those who are interested,
these include air suspension, automatic compensation of height as the load
of passengers and luggage is increased, the capacity to swerve to avoid an
accident and brake to stop before hitting a stalled car, with a huge margin of
safety in situations where the Rolls would add to the wreck, substantially
better safety-car engineering if it does come to a crash, and many other
advantages.)
To give an example from the more humdrum end of the consumer
purchasing spectrum. Consumer Reports did a nice study of rug cleaners
some time ago, in which it compared all the major rug cleaners on the
market by doing very careful comparisons of their ability in cleaning rugs
made with various kinds of fiber and soiled with various types of dirt. Just as
a matter of interest, they also included in the study the use of a dilute
solution of Tide detergent. Since rug cleaners are sold in a rather small
market, and are sold in a way that resembles cosmetics rather than soaps,
116 CHAPTER 5

they have very high advertising budgets, and also relatively high manufac¬
turing costs, per unit sold. Hence the cost difference between the rug clean¬
ers and the dilute solution of Tide was astronomical—10 or 20 to 1 in favor
of Tide. Well, it turned out that Tide beat the hell out of the rug cleaners in
terms of performance. What is one going to say about this if one really
believes that “you get what you pay for”? Presumably what you’re paying
for here where the prestige factors that might enter your purchase of a Rolls-
Royce are hardly relevant, is rug-cleaning power. And you’re not getting
more by paying more. There are dozens of examples of this. In the perfume
field, it’s well known that Jean Patou’s perfume “Joy” is inflated in price just
so that the manufacturer can use the advertising slogan “the costliest per¬
fume in the world.” It’s also widely believed that the profit margin on Cad¬
illacs, although this is not publicly released by General Motors, is from 4 to
20 times that on a Chevrolet; that is, a great deal of what you’re paying for
when you buy a Cadillac is General Motors’ profit, rather than quality in the
product. This is possible because of the extraordinary prestige position that
the Cadillac occupies in the automobile market. It may well be the case that
there are certain gains in quality, but they are certainly not present to the
extent that you pay for them, i.e., to the same extent as the difference in
price.
Once again, we see that if the general claim is made more precise, it
becomes less likely to be true; and if it’s made more vague, its truth can be
preserved but its information content evaporates. This again illustrates the
general principle that the informativeness of a statement goes up as its pre¬
cision increases, but at the same time, the chances that it will be true are
reduced just because it has, so to speak, made itself more vulnerable. Thus,
the clearer and stronger the contrast between a term or a statement and
what it rules out, the more informative it can be (and will be. if it’s true). The
vaguer the term or statement is, the less it rules out. and hence the less it tells
you about the world when a claim is made involving it.
An example you might want to reflect on—you can treat it as a kind of
advance exercise from Quiz 5-C—is one that is of great importance to us in
our political thinking: “Democracy is the best form of government.”

5-5 Exploiting Vagueness

Vagueness is sometimes undesirable, but then it’s sometimes necessary.

When you don’t have enough information to be able to make a precise
statement, or the time to express it, you may be able to express the informa¬
tion you do have in a vague statement and be able to convey something very
valuable by doing so. Maybe you can’t be very precise in locating the rea¬
sons why you found somebody an unsatisfactory employee, but even if you
say something rather vague like “Somehow she just didn’t seem able to
CLARIFYING MEANING 117

accept any responsibility for the work she was supposed to be doing,” you
may be comveying something very important to a prospective employer just
because he needs somebody who is strong in this dimension.
In the preceding section, we discussed a number of cases where rather
fancy-sounding general claims are made, whose truth is preserved only by
interpreting them in an extremely vague way. These are examples of exploit¬
ing vagueness, that is, of using vagueness to preserve truth at the expense of
becoming trivial. The crucial problem in the search for truth in this world is
to combine novelty with accuracy. It’s never hard to make true statements;
you can always say that “either it will rain tomorrow or it won’t.” The
problem is to find a statement which is both true and informative, that is,
not already known to be true. The old criticism of a piece of bad research or
writing is an ever-threatening one—“what’s new isn’t true and what’s true
isn’t new.”
There are other ways of exploiting vagueness which deserve to be con¬
sidered, and which are often employed, sometimes unconsciously, in bad
arguments. There’s a whole family of them that are closely related and that
have acquired special names because they are so common. The underlying
fallacy—the name for a seductive error of argument—is pretty close to being
the same in each case; you can probably tell what it is from looking at a list
of the names for it. It’s called the “bald man argument,” “the slippery
slope,” “the black-and-white fallacy,” the “thin edge of the wedge”; and
sometimes less printable terms are used.
Let’s take the “bald man argument” as fairly typical of the species of
fallacy that we’re talking about. The argument goes as follows: Take a man
with not a single hair on his head. He would certainly be called bald. Sup¬
pose that he had one hair on his head. We certainly wouldn’t say that he had
ceased to be bald. Suppose that he had two hairs. That wouldn’t persuade us
either. In fact, the addition of a single hair is never going to make the
difference between being bald and not being bald. But the difference be¬
tween being bald and not being bald is simply the difference that we achieve
by adding one hair at a time. Since adding one hair never makes the differ¬
ence between people’s being bald and not being bald, there really isn’t any
difference between being bald and not being bald. Thus, the argument ap¬
parently reaches an obviously unsupportable conclusion from apparently
plausible premises, by apparently legitimate inference. In the form we’ve just
expressed it, it’s pretty clear that you’re not going to buy it. But, in more
subtle forms, it becomes a very insidious type of argument. Often it occurs
when rejecting an appeal by someone, say a prisoner on an indeterminate
sentence who’s up for parole. The parole board may argue that, while there
are a number of points to be made in favor of the prisoner, these points are
rather small, whereas the difference between a truly reformed and faultless
record in prison and the kind of record that would be instantly turned down
118 CHAPTER 5

by the parole board is a very large difference. This may be, in particular
cases, a version of the “bald man” argument. After all, the difference be¬
tween cases where, although perfection has not been achieved, enough merit
is evident to justify granting parole, and cases where there isn’t enough
evidence, is only a difference of degree. Hence, it’s inappropriate to turn
somebody down on the ground that all that has been offered are consider¬
ations that amount only to a difference in degree. It would be perfectly all
right to say that there just aren’t quite enough of these considerations. But it
won’t do to say that there are “only” differences of degree.
In the same way, there is the obvious fact that one can move extremely
gradually from white to black through shades of gray so close to one another
that one can’t distinguish any two of them; this fact can be used to support
the conclusion that “really” there isn’t any difference between black and
white. Of course, you don’t accept that version of the argument, which is
why we use the term “black-and-white fallacy” for it. But you may easily be
persuaded that the difference between a war of aggression and a war of self-
defense isn’t really a true difference, because you are led through a series of
cases where a country has had to react against the buildup of forces by an
enemy preparing to invade it, prior to the invasion, or face certain defeat as
soon as the invasion begins. You can, of course, see that these are cases
where the idea of self-defense is predominant, so that it can hardly be called
a war of aggression. You can also see that one can push this kind of consid¬
eration further and further, so that eventually we have a war which is
mounted because of hostile “enemy” propaganda which is interpreted as a
forerunner of an eventual buildup of arms, and is used as a justification of
invasion even though no actual arms buildup has occurred.
Or you may be persuaded that torture is justified in the interrogation of
suspected criminals or prisoners of war by somebody who points out that the
line between physical and mental torture is not a sharp one, and that we
obviously believe that imprisonment and interrogation are in some circum¬
stances justified, so that we really are drawing an arbitrary line when we say
that physical torture is not to be allowed; in addition to being arbitrary, this
line handicaps the forces of right in their quest to stamp out evil. So we
finish up by agreeing that using totally brutal methods of physical torture
while interrogating (or standing by while our allies do it) is legitimate. Part
of the fallacy here is a “bald man” argument.
Another example concerns the difference between brainwashing, propa¬
ganda, and education. It is very hard to draw the line between them, and
that fact has often been used to legitimate types of classroom activity which
are mindless repetition of dogmatic truths about our country or our value
system, treated as if they were of the greatest importance although not sup¬
ported by any kind of rational argument. They are simply brainwashing or
propaganda, not education; and they have no place in the classroom, since
 CLARIFYING MEANING
119

the\ fail to recognize the rights ot the individual to autonomous decision

making, the right which underlies democracy. This example provides one
with a specimen procedure for counterattacking against a “bald man” (or
black-and-white) argument. The most obvious counterargument is simply to
use the phrases which identify this fallacy and say that the line of argument
that s being used would show that black is white or that a bald man really is
not distinguishable from someone with a full head of hair. But another type
of counterargument is a little more powerful, if you can bring it off. Treating
the pledge of allegiance or saluting the flag or reciting a religious code as
something that is (1) obligatory, (2) forced upon children who have not been
able to make up their own minds yet about whether they wish to take such
pledges, (3) important, (4) something the legitimacy of which is not dis¬
cussed seriously, critically, analytically, and at length is, it might be said,
exactly the kind of behavior which a democratic system of government op¬
poses. So, in rebutting the slippery slope argument that tries to show that
such activity is really not very different and hence not “essentially” different
from educational activity, one turns the tables in order to show that, if this
isn t an important difference, the very difference which is extolled in all these
propaganda activities is itself not an important difference; that is, it is not
the difference between us and our “enemies.”
In the same way, in dealing with the slippery slope arguments which are
used to legitimate torture by the police or by army interrogators, one should
consider the possibility that it is precisely the avoidance of such torture, and
similar recognitions of the right of the individual to the integrity of his or her
person and freedoms, that constitute the difference between us and our en¬
emies. So, if the slippery slope argument is to be accepted, we must also
accept the conclusion that we have no reason to be fighting the enemy (or, in
the case of the police, to be fighting organized crime).
Let me run through that once more. This is really a case of turning the
slippery slope argument on itself, and in common with a number of other
cases of reasoning where one can do this, it’s an uncommonly powerful
procedure. (Another example is illustrated by the obvious criticism of “noth¬
ing is really certain,” namely, that criticism applies to itself and hence can’t
be treated as reliable.) The argument justifying brutal interrogation of a
military prisoner has two strands, one of which is the slippery slope argu¬
ment and the other of which is the moral argument. The slippery slope
argument involves saying that we already accept very long and exhausting
interrogation, prison camp conditions, shouting at prisoners in the course of
interrogation and probably shining strong lights in their eyes, and so on. It’s
only one step from these to a little cuffing around, or sticking the point of a
knife into somebody, and only a step from that to chopping off a finger or
shooting off an ear or tying a wire around the throat or applying an electric
prod to the genitals. And, on the other hand (the moral argument), our cause
120 CHAPTER 5

is just and the information we extract may be vital in saving lives of our own
soldiers who are in the moral right, and we’re using these techniques on
somebody who represents the enemy, who is morally wrong. The tough line
of counterargument that we’re considering now runs as follows: First, we
say to the person who has put forward the preceding argument, you’re argu¬
ing that there’s only a difference of degree between what we would anyway
accept and what is being questioned here (the use of physical torture); and
then you say that although this is a difference, it is only a difference of
degree and that it is overwhelmed by the moral differences that are involved.
But the problem is that there’s only a difference of degree between the moral
position of the enemy and that of ourselves, and hence your very same
argument would show that that moral difference isn’t a real difference or a
really important difference, and hence that it can't be used to justify the
difference between yelling, shouting and shining lights, and shooting off an
ear or electroshocking the prisoner, which is an absolutely unmistakable,
clear, visible, and painful difference.
So you have attempted to argue out of existence the one difference that
is really obvious by appealing to another difference which is only a matter of
degree, and the way you’ve chosen to do this is by arguing that matters of
degree are not very important. But the moral difference is most certainly a
matter of degree, and if such differences are not really important, then they
most certainly cannot be used to justify differences of treatment which are
quite clearly not just a matter of degree at the sensory level of the sufferer,
that is, at the most obvious level of direct sense perception. (In case you
think that the difference between, say, the communist and the anticommu¬
nist sides is not just a difference of degree, then it would be worthwhile
trying to spell out just what features of life under the anticommunist j-egime
in South Vietnam could be said to be more democratic or more beneficial to
the people than corresponding conditions in the communist North at the
height of the war in Vietnam. The fact is that the difference here may have
been not only a difference of degree, but it may have been a difference in the
opposite direction from the way the arguer is assuming it goes, which of
course opens up another line of counterargument.)
In the case of using torture on a suspected criminal, the problem is that
“suspected” is not a category of criminality, but just an indication of the way
that the evidence points prior to careful examination by a judge, jury, and
the adversary system. To apply torture to somebody in such a situation is to
treat the person as somebody who has been proven guilty of criminality
beyond any reasonable doubt; that is, it is to treat “being a suspect” as one
of the most serious crimes in the book (since very few jurisdictions allow
torture for anything except the most serious crimes, even if then). So the
attempt to justify torturing suspects, in order to get either a confession or
information, by use of the slippery slope argument is open to the “fast
 CLARIFYING MEANING
121

switch" counterargument that if there isn’t any real difference between bru¬
tality and interrogation of the more verbal kind, then there isn’t any real
difference between the behavior of the interrogator and the behavior of the
criminal, it indeed the suspect is a criminal. Hence one slippery slope counts
against the other and leaves us with a standoff situation, rather than a justi¬
fication of torture. Incidentally, this kind of switching around and using an
argument against itself is called sophistry if you don’t understand it, or don’t
believe it, and clever, or devastating, if you do.
The important thing at this stage is to get a firm grip on some of these
examples by discussing them carefully with your fellow students, instructor,
or triends. They represent a very common type of argument with respect to
many of the most important issues that face us. Think about the distinction
between an abortion and infanticide. Think about the distinctions between
tobacco, alcohol, marijuana, LSD, cocaine, and heroin. How often is the
slippery slope argument used here? Think about issues where you’ve heard a
regulatory board or an appeals board say, “If we made an exception here,
we'd be making them all the time so we really have to draw the line.” Many
times not every time—this is just a way of saying that the board members
get nervous about being drawn anywhere into the gray area, and in order to
make things easy on themselves, they will continue to make dinstinctions
only between black and white, even though there are a serious cost and
injustice that could be remedied if they were prepared to say that there is a
difference in treatment.

5-6 Ambiguity and Its Exploitation

Vague or imprecise terms may be considered as terms which refer to a wide

spectrum of possible interpretations. An extreme case of this is the ambigu¬
ous term, one which refers to two or more quite different meanings. The line
between the two—that is, between vagueness and ambiguity—is not a sharp
one, and you will often hear complaints made about terms like “democracy”
in terms of their ambiguity, when in fact one might more precisely say that
the problem with them is their vagueness.
But let’s look at this in a little more detail. One might define “democ¬
racy” as a form of government in which the rulers are elected by public
voting of all adults. Or one might define it as a form of government in which
the government is subject to recall by a vote of the people. The term, if
you’re not clear which one of the two definitions it refers to, is certainly
ambiguous. One could easily have an elected government which isn’t subject
to recall. The two concepts are logically distinct. However, even though
many people have their own favorite definitions of democracy, and even
though these definitions are logically distinct and not part of a continuous
122 CHAPTER 5

spectrum as most vague or imprecise notions are, the general truth is that
the term “democracy” is a vague term rather than an ambiguous one, be¬
cause its common meaning shifts along a complex spectrum, even though
certain individuals have particular meanings for it that are conceptually
distinct. The upshot of all this is that you don’t want to be too rigid about
whether the term is ambiguous or vague, until you’ve decided whether
you’re talking about its usage in a particular context or about its common
meaning. In a really serious scholarly work on the nature and justification of
democracy, for example, in C.B. Macpherson’s Democratic Theory, one can
indeed argue that it’s used ambiguously, sometimes to refer to “one man
(woman?), one vote” and sometimes to refer to “the maximization of per¬
sonal freedom.” But even that isn’t a particularly serious criticism because
Macpherson explains the ambiguity and operates with one or the other sense
of the term at a time, fairly carefully. But in a general discussion, rather than
in the context of a carefully structured book, the term is much more likely to
be ambiguous in a bad sense (people disagreeing about it mainly because
they mean different things by it) or just vague (people disagreeing about it
because they haven’t thought about where they’d draw the line between a
democratic and an undemocratic society).
With these preliminary cautions, however, one can go ahead and talk
briefly about the relatively rare case of plain ambiguity and the effects that
it can have on argumentation.
The most important case of the exploitation or abuse of the ambiguity
of a term occurs in arguments where there is a shifting of the meaning of a
term during the course of the argument. The technical name of this error of
reasoning or fallacy is the “fallacy of equivocation,” a label which you do
not have to remember. We’ll normally refer to it as simply a shift of mean¬
ing. A typical case would be an argument which opens with one premise in
which a term, let us say “criminally insane,” is used in the premises with a
certain meaning, in order to develop intermediate conclusions. Then, in rea¬
soning from these intermediate conclusions, together with some extra facts,
toward the final conclusion, the term is used in a slightly different way, and
the argument wouldn’t work if the same meaning had been used throughout.
Either the first premises wouldn’t have been true, or the inference from the
intermediate conclusion to the final conclusion wouldn’t have worked.
That’s a typical example of the “shift of meaning” fallacy, or the fallacy of
equivocation. It’s a fairly crude kind of affair, and you only have to keep
your eyes and ears half open in order to pick it up. In most cases! But, like
any logical fallacies, once in a while one comes across a case where it’s
pretty hard to spot the cause of trouble. These cases are nearly always very,
very long ones—for example, a whole chapter, or a whole book, where the
shift occurs gradually in midstream—and so they’re not very easy to quote.
CLARIFYING MEANING 123

The ones that we can quote look pretty implausible, and you imagine that
you couldn’t possibly be taken in by anything so crude; but then once in a
while you are. A crude example, involving the term “criminally insane,”
would look something like this:

The crucial pragmatic significance of criminal insanity is that it absolves indi¬

viduals so described from legal responsibility for their actions. The extent of
true criminal insanity is probably something like 10 percent of the cases that are
thus identified under the present inefficient system. However, there are certain
clear cases of insanity, and we shouldn’t allow our legitimate worries about
overextension of the term to prevent us from treating these unfortunate crea¬
tures with justice. It would be ridiculous to act as if a paranoid schizophrenic
was in full possession of his or her faculties and someone who should be held up
for the full rigors of the maximum punishment afforded by the law. Yet this is
exactly what has happened in the Manson case. A classical case of paranoid
schizophrenia, diagnosed by several extremely competent psychiatrists, was
treated as if it was a standard example of heinous crime. We have to be very
careful not to let our nervousness about the excesses of psychiatry destroy our
basic sense of justice.

What’s happened in this argument is that we’ve shifted from the opening
definition of criminal insanity, which ties it definitionally to the absence of
responsibility, to a later definition which identifies paranoid schizophrenics
in the psychiatric sense as being criminally insane, which isn’t obvious at all.
In fact, it’s exactly what the debate is about in many court cases. The para¬
noid schizophrenic, operating within an elaborate delusional schema, may in
fact, with respect to the act in question, be able to choose rationally, recog¬
nize right from wrong, and otherwise meet the standards for legal sanity.
(Some more detailed discussion of particular cases will be found in The
Meaning of Criminal Insanity by Herbert Fingarette, University of California
Press, 1972.)
How should you tackle this type of fallacy when doing argument analy¬
sis? Look back if you don’t remember.

5-7 Dictionary Definitions—Real and Ideal

We’ve been touring through the messy territory of troubles with meaning,
and it’s now time we started looking at ways of clarifying meaning. Tradi¬
tionally, people think of the dictionary as the source of concise definitions.
And so it is; but the kind of definition that you find there is not by any
means the ideal kind that is so often discussed in logic texts. An ideal kind of
definition would look something like this:

Triangle = three-sided geometrical figure

124 CHAPTER 5

That’s one of the kinds of definitions that you’ll find in a dictionary; or,
more likely, one piece of a definition that you’ll find in a dictionary. It is a
valuable ideal, and we want to begin by discussing its properties so that we
have this ideal firmly fixed in mind before we start moving to the reality
levels of the more frequently encountered definitions and to alternative pro¬
cedures for specifying meaning. Notice that we sometimes call the whole
“equation” the definition, whereas sometimes we call the right-hand side of
it the “definition” of the left side, which is the term being defined. The key
features of the ideal type of definition are as follows:

a The meaning on each side of the equals sign is exactly the same:
that is, one could replace either the term “triangle” or the above
definition of it by the other, in (almost) any context, without loss of
meaning. (The most obvious context where you could not do this is
in the definition itself, since the definition clearly tells you some¬
thing which could be quite informative, whereas, if you substitute
the word “triangle” for the right-hand side of it, you’d get

Triangle = triangle

which is trivial and not informative at all.)

b As a practical consequence of feature a, the consequence that is

crucial for testing purposes, anything which is correctly described by
the left-hand side of the equation must be something which is cor¬
rectly described by the right-hand side, and vice versa. (That’s just
an unpacking of the phrase “has the same meaning.”)
c Another unpacking of the definition, which is useful in special cases,
tells us that it would be a contradiction to describe anything by the
term on the left-hand side and by the negation of the term on the
right-hand side (or vice versa). That is, you couldn’t say something
was a triangle and didn’t have three sides, or that it was a triangle
and had four sides (which implies that it didn’t have three sides).

Notice right away that the term “triangle” is also the name of a musical
instrument, and it’s also the name of a drawing device used in drafting, and
it’s also a term used to refer to the course that racing yachts sail, and so on.
All these other meanings show us that the term, out of context, is ambigu¬
ous. Not that it’s vague, which it might or might not be as well, but that it’s
ambiguous. It might be ambiguous and some of its meanings might be
vague, or all of them might be vague. Or it might be ambiguous and precise
in each of its meanings. What we have been looking at is a definition of one
sense of the term “triangle.’ What you find in a dictionary is usually a listing
of all the senses, and hence a whole series of definitions of a term which is.
CLARIFYING MEANING 125

out of context, ambiguous. Terms that are ambiguous out of context are,
however, rarely ambiguous in a given context. That’s how we are able to
understand the English language, nearly all of whose most common nouns
are given multiple definitions in the dictionary. We understand which of the
meanings is intended by contextual cues. This is perhaps the single most
important feature of the real, as opposed to the ideal, language.
Now if somebody proposes a personal definition (of a particular sense)
of the term, you may use any of the conditions a, b, or c above to see
whether the definition is, strictly speaking, correct. Example: If someone
proposes to define “good teacher” as “someone who scores well above the
average on student evaluation forms,” you might go about scrutinizing this
proposed definition by applying condition b above. Are there any cases of
good teachers, in the popular sense of the term, who would not score high on
student evaluation forms? And are there any cases of people who score high
on student evaluation forms who are not good teachers? Before reading
further, ask yourself if you can see any difficulties with, or answers to, these
questions.
As we begin to reflect on them, we should immediately see that we’re
facing a problem of imprecision. There are many student evaluation forms,
some of them hopelessly unreliable. If one were to use a student evaluation
form which is essentially a rating of the extent to which the student likes the
teacher (and there are such forms), one might easily find cases of rather
strict, cold, and severe teachers who would score badly on these forms but
who would nevertheless be very successful in teaching the students whatever
they had enrolled in the class to learn.
You will notice that we have made two crucial moves as we get down
to discussing this example in detail. First, we have started to talk about how
there might be a good teacher who did badly on some such form, as if this
was relevant to the accuracy of the definition. And second, we have started
implicitly generating a more satisfactory or more basic definition of “good
teacher” (in this case, as someone who succeeds in teaching whatever it is
that the course is supposed to assist the student in learning). Each of these
steps deserves a word of explanation, as it is extremely important.
First, it is crucial to understand that in order to criticize a definition,
you need to think only of possible exceptions, not of actual ones. The reason
for this is that the claim involved in the definition, and expressed as claim a
above, is the claim that the two sides of the equation are synonymous, that is
exactly the same in meaning. Hence, wherever either one could be applied
correctly, the other one would have to be applied correctly; and that doesn’t
mean the same as “wherever one has already been or is now applied cor¬
rectly, the other is now applied correctly.” When we’re talking about mean¬
ing we’re talking about all possible (proper) applications of the term, not
merely its present applications. You can see that this is crucial for a lan-
126 CHAPTER 5

guage because a language has to be usable to describe new phenomena, and

of course it couldn’t describe something new if the rules for its use tied it to
only those cases that have existed until now. One couldn’t write imaginative
literature if one could not describe imaginary scenes. We can visualize and
describe triangles which are tiger-striped, leopard-spotted, and polka-dotted,
even though we may never have seen such a thing and may not even be at all
confident that there is or will be such a thing. The meaning of the word
“triangle” is just the same whether it’s describing A or a triangle of immense
dimensions which you have just visualized as part of a description of a
spatial survey performed by a planetary exploration spaceship. Hence, if the
definition above does correctly represent the exact meaning of the term, it
will correctly translate the term in any reference to an imaginary object just
as well as it will when referring to a real object. So, in order to criticize a
definition, all you have to be able to do is to imagine a fair case of something
which would be correctly described by one side of the definition, but would
not be correctly described by the other side; such a case shows that the two
sides do not have the same meaning. That’s quite different from the issue of
whether, as a matter of fact, all triangles happen to have three sides. For
example, as a matter of fact, it might happen that the only featherless,
tailless bipeds on the earth’s surface were ourselves; but that wouldn’t show
that the term “human being” means the same as “featherless, tailless biped,”
since if we then encountered a plucked chicken running around, or a kanga¬
roo with its tail amputated, we most certainly wouldn’t refer to it as a
member of the human race. This shows that we have in mind a meaning of
“human being” that is not fully exhausted by “featherless, tailless biped.” So
there’s no question that “featherless, tailless biped” does not cover the mean-
ing of “human being” even though it may happen in fact to refer to the same
group of individual entities at a particular moment in time.
In sum, then, the logical scrutiny of definitions must not be restricted to
the mere question of whether the entities described by the left-hand side of
the definition happen at the moment to be the same set of entities as those
described by the right-hand side. The definition must go further, and answer
the question of whether all entities that could be properly described by one
side would have to be properly described by the other side. Hence, there is
an exercise in imagination involved in the criticism of definitions. You have
to try to imagine cases where it’s perfectly clear that people who understand
the language would immediately apply one side of the definition but not the
other. When we apply this test to the above definition of “triangle” in the
sense where it refers to a geometrical figure, we find that there really aren’t
any exceptions of this sort. Hence, we have succeeded in expressing the
meaning of one sense of the term “triangle.” But when we look at the defini¬
tion of “good teacher” above, we can see that there could easily be good
teachers who weren't rated well on a particular student evaluation form. We
CLARIFYING MEANING 127

can also see that there might be teachers who are rated well on such a form
who are not good teachers at all, but who are just popular, friendly, giving
high grades, or saying something that seems to be true and interesting but is
really false.
It might be a little unfair to be criticizing that definition simply by
appealing to a case where a bad student evaluation form is used, which
identifies good teaching with likeability, for example. It’s true that the origi¬
nal definition is vulnerable to this type of criticism, but it is a relatively weak
criticism, since the definition could be reformulated. Suppose we devise the
best possible student evaluation form. It will still reflect the opinion of the
students, and one distinguishing feature of students is that they are not so
well informed about the subject matter as the most distinguished experts in
that subject matter. It is therefore possible, though in practice not very
likely, that the students might be greatly impressed by a teacher who had all
the appropriate pedagogical skills and who appeared to be laying out for
them a subject matter that made a great deal of sense, whereas, in fact, the
information that was being offered was out of date and currently known to
be incorrect. Hence, doing well on even the best student evaluation form
cannot encapsulate the whole meaning of “good teacher,” since that term
certainly involves teaching what is true.
The second feature of this line of discussion is that, clearly, we're bring¬
ing in certain considerations that we implicitly think are part of the concept
of “good teacher” but that aren’t covered in the proposed definition. The
two that we’ve mentioned so far are skill in producing learning on the part
of the student in the appropriate area, and teaching what is in fact true or
correct. These are, to use a terminology we have discussed earlier, necessary
conditions for being a good teacher, but they’re not sufficient conditions. An
ideal dictionary definition of the type we’ve been talking about is supposed
to provide conditions that are sufficient and necessary for the term that it
defines. Since either side of the equation is supposed to be the same as the
other in meaning, it follows that if either side can be applied, then the other
can be applied; consequently, each side is a sufficient condition for the
other; and hence (if you think about it), each is a necessary condition for the
other. So the ideal dictionary definition is a classic example of a sufficient
and necessary condition for the term defined. Here we run into another
difference between the ideal and reality; for dictionary definitions in practice
frequently provide us only with conditions that are usually or generally or
probably sufficient and necessary for the use of the term, but not totally or
logically or strictly or “definitionally” sufficient and necessary conditions for
it.
So, as we criticize definitions, we do so by appealing to examples which
people would instantly recognize and describe by the terms of either the left
side or the right side of the equation, but not both; and we sometimes do so
128 CHAPTER 5

with some implicit reference to at least a part of a better definition; and, as

we’ve seen from the preceding section, we may also do something that’s very
closely related to these, namely, look for contrasts that the term implies and
see whether these contrasts are preserved by the proposed definition. Be
systematic in criticizing definitions; be sure to cover all these points in your
approach. One of the three may work if the others don’t.
Notice that a proposed definition might provide a necessary condition
for the use of the term being defined that isn’t also a sufficient condition; or
it might provide a sufficient condition that isn’t also a necessary condition.
Either would be a fault, and they are faults in different directions. But these
faults are independent: a proposed definition can commit both of them si¬
multaneously. There is another language for expressing this kind of defi¬
ciency in a proposed definition, and it may be a language that is more
graphic for most readers. Suppose we define “triangle” as “any geometrical
figure.” This would obviously be too broad a definition; it includes too
much; we might say it’s too inclusive a definition. That’s another way of
saying that it states a necessary condition but not a sufficient condition.
Something can’t be a triangle without being a geometrical figure; hence it’s
a necessary condition for something to be a triangle that it be a geometrical
figure. But something could be a geometrical figure without being a triangle
(for example, by being a quadrilateral), and hence the proposed definition is
not a sufficient condition for the term defined.
On the other hand, we might define a triangle as a geometrical figure
with three sides, of which two are of the same length. That’s an excellent
definition of an isosceles triangle, which is a particular species of triangle.
But it’s too narrow a definition for the term “triangle” in general. You might
also say it’s too restrictive; this is the same as saying that it provides a
sufficient condition for being a triangle but not a necessary condition.
It’s quite easy to depict the relationship between the two sides of the
proposed definition so as to reflect these various kinds of error. You can
think of the territory to which each of the sides legitimately applies as repre¬
sented by a circle. In a correct definition (Figure 1 below), the two circles
coincide. In a definition which is too narrow, the circle represented by the
definition (Y) falls entirely within the circle represented by the defined term
(X) (Figure 2). In a definition which is too broad, the reverse is the case
(Figure 3). In a definition which is both too broad and too narrow, or too
inclusive and too restrictive, the corresponding picture would be of the two
circles overlapping—some parts of each would fall within the other and
some parts would fall outside (Figure 4). In such a case, we would say that
the proposed definition is neither necessary nor sufficient for the use of the
defined term: or that it is both too inclusive and too restrictive: or that it is
both too broad and too narrow.
CLARIFYING MEANING 129

(Here the proposed definition is X = Y)

There you have the ingredients for the basic procedures for criticizing
definitions. Once again, be systematic. Always look for the possibility that
the proposed definition is too broad as well as too narrow; don’t settle for
finding just one kind of defect in it. Indeed, don’t settle for finding just one
kind of example to prove, say, that it’s too broad; go ahead and find two or
three if you can, since then you have several independent grounds for criti¬
cism in that direction, and will only have to renounce one of them if it turns
out that you’ve made some slight error of interpretation. As practice, criti¬
cize this proposed definition: “college” = where you go to get educated
after you leave high school (Quiz 5-A, question 1).
We’ve already indicated a number of ways in which real definitions
tend to be less exact than the ideal ones we’ve been talking about. First of
all, notice that the short formula we’ve used to represent a definition (X =
Y) is in reality too simple. The Y part actually consists of several properties,
Y,, Yv Yv etc. So the definition usually looks more like this: X = Yx and Y2
and Yy . . . The most extreme deviation from the ideal definition occurs
when what you find in the dictionary is almost totally different from the
ideal definition in that nothing in it (none of the Ys) is a necessary condi¬
tion, although the total set is a sufficient condition for the defined term. In
the ideal definition, each of the properties Yx, Yr Yv is a necessary condi¬
tion, and they are, taken together, a sufficient condition for the proper use of
the defined term X. (And of course, taken together, they are also necessary,
since if several things are necessary, putting “and” between them just makes
the whole set of them necessary.)
Take a definition of a simple word like “apple.” Using your imagina¬
tion, we can readily think of counterexamples to the definitions that you find
in Webster’s Second, Webster’s Third, the Oxford English Dictionary, the
Random House Dictionary, and the other usual standards. That is, we can
easily imagine growing apples in laboratory flasks, so that they’re not the
fruit of any tree; growing them from a skin cell from another apple, so that
they are not grown from seeds; or that they do not have seeds from which
something else could be grown. We can imagine them tasting quite unlike
present apples but being so similar in every other respect that they would
still naturally be called apples, and so on and so on. Every one of the
“defining properties” mentioned in the dictionary can be shown not to be
necessary; hence the sum of them all isn’t necessary. And it is quite common
130 CHAPTER 5

for these everyday terms to be such that we can’t specify any necessary
conditions for them at all. All we can do is to give a set of conditions which
is, as we say, jointly sufficient (that is, fj plus Y2 plus Y3 implies X, but not
vice versa). Or perhaps we can give several sets of conditions, each of which
is jointly sufficient (that is, we can also find a set of Z’s such that Z, plus Z2
plus Z3 also implies A—but not vice versa). Hence we learn that there are
certain cases where the term can be properly applied (e.g., a case where Y,
and Y2 and Y3 are all present); and we learn that there’s a family of cases
which are related in some not too precisely definable way, to all of which the
term applies. (Yj plus Y2 plus Y3 is one case; Z, plus Z2 plus Z3 is another;
and maybe there’s a Wl plus W2 plus W3 case too.) But our insight doesn’t
go beyond that: there isn’t a buried secret formula which encapsulates the
meaning in a more precise way. A famous example is the term “game.” It’s
easy to list sets of sufficient conditions which amount to descriptions of
football, hide-and-seek, pole vaulting, solitaire, each of which is therefore a
description of a game and hence a sufficient condition for the use of the term
“game,” but it is hard to give necessary conditions. Try it! In situations like
this, we sometimes refer to the indicators that are mentioned, the relevant
components of any of these sets of sufficient conditions (in the case of, for
example, “apple,” things like being the fruit of a certain tree, being green in
color at one stage of their development, or being crisp and white-fleshed) as
being criteria for the use of the term, by contrast with necessary conditions for
the use of the term. They are all we have, all that we can appeal to in
explaining the meaning, but they’re not necessary conditions. In a later
section we’ll have a look at the constructive use of this kind of fact in
explaining meaning. What it really tells us is that when we can’t find neces¬
sary conditions to make up a nice, ideal definition, we can still provide
examples to illustrate and thus convey the meaning of the term, even if we’re
not providing a translation. Before we get to that, we’ll look at some other
very significant weaknesses of the “dictionary definition” approach to clari¬
fying meaning. And before doing that, it may be useful to sum up what we
have just been saying.

5-8 Necessary Conditions and Criteria

An ideal dictionary definition defines the term X by referring to a series of

properties, Y,, Y2, and so on, in a formula like this:

X = Y, plus Y, plus Y3 . . .

To qualify as an X (a triangle, for example), something must then have all

the properties Yv Yv etc. (having three sides, being a geometrical figure, for
example). Having each of the defining properties is therefore necessary; they
CLARIFYING MEANING 131

are necessary conditions for being an X. Having all of them is sufficient. The
y1 s are therefore described as individually necessary and jointly sufficient.
Now what we find in practice, in reality, is that such ideal definitions
are usually possible only in mathematics—or when someone invents a new
term and stipulates what it is to mean. Even in math, and soon after a new
term has been introduced (if it “catches on”), we find the situation becoming
cloudier, the connections looser than in the ideal definition. We find that we
can still identify cases which seem to be clearly cases of X, but that the range
of cases of X seems to expand as the concept X develops a life of its own,
and we become unable to say that all such cases must contain certain com¬
mon elements, the necessary conditions (T) that were referred to in the
original definition. This “degenerate” situation, it turns out, is the typical
situation with essentially all the nouns, adjectives, verbs, and adverbs of the
common language and of most of science. We can give rather imprecise
general definitions of these terms (“apple,” “estuary,” “force”), or, some¬
times, exact but abstract definitions which only pass the buck of comprehen¬
sion along to another equally unclear term (“a force is whatever produces or
could produce an acceleration”; “moral virtue” equals “the opposite of mor¬
al vice,” and so on). But we cannot give adequate translations of these terms
into a set of more specific, and hence usually (which is the point) more easily
understandable, terms, each of which refers to a necessary condition for the
application of the term. On the other hand, we can easily identify clear cases
of X, which means that, if we can adequately describe such cases, we can
give a number of sufficient conditions for the use of the term. For example,
we can describe a Gravenstein apple from the tree in our own garden in
sufficient detail so that one may be perfectly confident that anything which
meets that description must be called an apple by anyone who understands
the meaning of that term. That is, we have a case where having the proper¬
ties Wv W2, Wy etc. is a sufficient condition for being an apple, but no one
of these conditions is necessary. We refer to properties like these as criteria
for being an apple; that is, they are relevant components of a sufficient
condition. They would be recognized as relevant by a fluent native speaker
of the language (who would rule out the fact that a fly was sitting on this
particular apple as irrelevant), and the whole set would be recognized as
sufficient for the application of the term, but, unlike the criteria for “trian¬
gle,” they aren’t necessary.
There’s no sharp line between criteria and properties which just happen
to apply generally in practice. For example, being tailless just happens to be
true of humans, but a native speaker would be unlikely to cite it as a crite¬
rion for being human, since someone whose coccyx happened to extend into
a stubby caudal appendage (the anatomical jargon for a tail) would not for
one moment be thought to be nonhuman. In looking for criteria, that is,
while we know we can’t expect them to be strictly necessary, we do look for
132 CHAPTER 5

properties whose absence would be felt to count a little toward weakening

the claim of humanity. If an entity is neither a fruit, nor contains seeds, nor
is crisp, nor is green when unripe, nor has the range of flavors of current
varieties, nor falls into the size range at maturity from a golfball to a grape¬
fruit, then one could hardly call it an apple, even though any one of these
(perhaps any two) could be dropped and we could still be persuaded. No
specific one is necessary, but the presence of some is necessary. It’s necessary
to have “representatives” of this group, but no particular member.
Pictorially, in the search for an ideal definition, we are looking for a set
of properties Y, each of which is a necessary condition for X, that is, each of
which is a broader or more inclusive concept than X:

but such that, when you take them all together, their only common ground
is exactly X:

Y\

What we’ve been saying about real definitions of most terms is that the
outlines of each property aren’t nearly as sharp. You can think of them as
fuzzy circles. So we finish up with a series of fuzzy circles that more or less
enclose X:

Here it’s fuzzily clear that there are some points that fall within X but
outside the range of any given Y, so no Y is a necessary condition for X. Yet
the fact remains that the Fs are all that there is to the meaning of X: If we
try to explain the meaning of X to a foreigner or to someone who had not
previously encountered it, the Vs are what we’d mention and they’re all that
we could mention. (You have to employ some members of this club, but you
don’t have to employ all of them, or any particular ones.)
CLARIFYING MEANING 133

So, in the real world, there aren’t many ideal definitions where the Fs
are all necessary conditions for X. Mostly, the Fs that you’d use to explain
the meaning of X (that is, to define X in the usual way) are only probably
properties of a given X. They are not necessary conditions (nor probably
necessary conditions) but just criteria for X which means (1) they are usually
associated with X\ (2) they are the properties a native speaker would men¬
tion or recognize as relevant to deciding the question of whether something
is really an A or not. The one ghostly remnant of the necessary condition
feature that still applies to these “critical definitions,” namely that even
though one can t claim that each of the criteria is necessary, one could claim
that at least one of them (perhaps several) has to be present, can be ex¬
pressed in terms of the abbreviations, as follows. Although it isn’t true that:

F, is a necessary condition for X

or that

F, is a necessary condition for X

or that

F3 is a necessary condition for X

it is true that: either F, or F2 or F3 is a necessary condition for X

or perhaps even that:

either F, and Yv or F, and F3, or F2 and F3 is a necessary condition for X.

(That is, it might be true that you have to have two of them.)

Now go to your dictionary. How about its definition of “apple”? (Quiz

5-A, 2) Does it have any necessary conditions in it? Can you prove it?
What’s the explanation of this “failure” of definitions? It’s very simple.
The “game” of language is not a game in the literal sense at all. It does have
“rules” but not strict and fixed rules like chess. It is a pragmatic game like
running or hunting, where the goal is success in a certain task—travel, or
escape; a trophy, or food—and the “player” does whatever is most appropri¬
ate for that task. In a very complex world such as ours, the most useful
functions of language, like, for example, communication of information or
instructions, or requests for them, will be most efficiently served by using
language to refer to the most commonly observed clusters of properties.
When we refer to a “table” in our speech, we expect people to understand us
to mean the kind of thing we are used to in the furnishings of our homes,
offices, and institutions. Nobody could care less if there are a few bizarre
objects around somewhere, or on the drawing boards somewhere, or con-
134 CHAPTER 5

ceivable, which do not look like these but which we’d still call tables. In
explaining the meaning of the term, we wouldn’t refer to them. They’re not
pragmatically important. BUT. . .
But they are sometimes extremely important in scientific theorizing; a
precise analysis of what was meant by saying that two remote events are
simultaneous led Einstein to the theory of relativity. And they are important
in showing you something about the nature of meaning and definition,
namely that the meaning of most terms involves a reference to a set of
properties which cluster together even though a few cases of them scatter
around elsewhere. The properties which normally cluster together are the
criteria.

5-9 Fact and Fancy: Definitional Truth and Physical Truth

In these little diagrams that we’ve been drawing, with circles marking out
the domain of X, Yv and so on, we need to be clear what the territory is on
which we are setting out the boundaries. It is the realm of proper use (or the
things to which that use refers), which of course includes actual use but also
includes potential use. We’ve already stressed the fact that knowing the
English language involves knowing how to apply it to many (though not all)
new cases, situations, or entities that have never been seen before. You
would have no problem in recognizing or describing something as a green
rose, though in fact no such entity has ever existed. Hence, when we draw a
circle to represent the range of proper use of the term “rose,” it will cover
many cases that we have not so far encountered in the real world. In a
previous chapter, we have talked about necessary and sufficient conditions
in the real world. Here we have been talking about necessary and sufficient
conditions in all possible worlds. (There would be some which are unde-
scribable in any language, and some which would require substantial addi¬
tions to our language to describe at all thoroughly.)
If a condition is necessary in our world—as, for example, it’s necessary
for an adult human to have passed through a period of childhood—we say
that it’s physically necessary, practically or, more generally, factually or em¬
pirically necessary. If it’s necessary in all the worlds to which our language
can apply, as, for example, it’s necessary for a parent to have had a child, we
say it’s definitionally necessary or logically necessary. Indeed, we usually
make further distinctions under the first of these headings—and remember
that none of them is a special technical term, that they’re part of the logical
vocabulary that already exists in, or is made up from, standard combina¬
tions of terms that already exist in English. These distinctions all exist just
because they’re very useful in argument or meaning analysis.
Under empirical or factual or practical necessity, we distinguish be¬
tween cases where a law of nature “underwrites” the necessity—for example,
 CLARIFYING MEANING 135

it’s physically necessary that solid iron sinks in water because of Archime¬
des’ Law—and cases where we have arranged that something is necessary,
but where it is within the power of people to arrange it otherwise. For
example, it s necessary for you to pass a test in order to get a driving license
in California. One might call these two subcases physical necessity (because
physical law is the source of the necessity) and legal or planned necessity (for
obvious reasons). Where the necessity is neither the result of natural law nor
of planning, but has just grown up as a convention or habit-pattern, we talk
of social or cultural necessity. And finally, where the necessity has ethical
backing, e.g., is an obligation, something that we “have” to do in the sense
of “ought,” we call it a moral necessity. Remember that none of these dis¬
tinctions is absolutely sharp, not even the distinction between logical and
factual necessity. But they’re extremely useful because most cases fall clearly
into one category or the other.
Now what is the contrast dimension of the meaning of these terms,
what’s excluded by calling a relationship between two properties necessary?
Necessary as opposed to what?
In the case of factual necessity, the contrast is with what happens to be
true by accident. The classic example is that it may just happen to be true
that all coins in my pocket are silver—but it’s not necessarily true, either by
virtue of a law of nature (physical necessity) or by deliberate plan or design.
Incidentally, both “accidental” truths and physical necessities are to be con¬
trasted with false claims, as are all necessary truths; but they are also to be
contrasted with each other, since they are true for different reasons. Both,
however, are factually true.
In the case of definitional or logical necessity, the contrasts are with
factual truths and with definitionally false claims. So the set of distinctions
looks like this:

True claims 1 Definitional or logical truths

2 Physically necessary truths
Factual
3 Planned or “legal or 6social or 6moral
truths
7 The rest—"accidental” truths or "mere matters of fact”

False claims 1 Definitional or logical falsehoods

2 Physical impossibilities
Factual
3-6 Planned or legal or similar impossibilities
falsehoods
7 "Mere” factual falsehoods

As an exercise right now, classify the following examples as of types 1, 2, 3,

4, 5, 6, or 7, and as T or F. (See Quiz 5-A.) Add comments where neces¬
sary—and perhaps more than one label if necessary.
136 CHAPTER 5

a You can’t pass this course without reading this text,

b Sons have to have (or to have had) a father,
c Sons have to have a parent.
d It’s impossible for a black woman to be elected President before
1984.
e Democracy cannot exist when there is no freedom of speech,
fThe interior-angle sum of a triangle is 180°.
g You have to go through the hallway to get to the living room in
most house designs.
h You absolutely have to repay that debt this week.

There is an important “pecking order” in these lists. The numerical

order reflects a sequence of strength in the sense of the extent of the domain
over which the truths hold sway.

1 Definitional truths hold in all worlds to which our language can be

applied.
2 Physically necessary truths apply to all possible occurrences in the
past, present, and future of our world. (If they didn’t, the underlying law
wouldn’t be a law of nature.')
3 Social, legal, and planned necessities apply as long as the present
social system applies (perhaps within a particular subculture).
4 Moral necessities are of the same strength as either 2 or 3 above,
depending on whether morality is a kind of natural law or just a kind of
social convention.
5 Accidental truths apply only until some circumstances which affect
them change.

It follows that when you are criticizing a claim—for example, the prem¬
ise of an argument—you are giving the strongest possible criticism if you
show it’s definitionally false, the next strongest if you show it contravenes a
law of nature, the next if it contravenes a law of mankind or the arrange¬
ments of some subgroup thereof, and the weakest if you simply show it
happens not to be true. (But even the weakest is strong enough to destroy
the claim.)
This is just one of several aspects of the distinction between strong and
weak criticism which is of such importance in argument analysis. But it’s
important elsewhere in the domain of reasoning, from the “propaganda
analysis” end to the highest reaches of abstract reasoning. When an advertis¬
er claims that a “properly tuned” engine using the water-injection apparatus
being advertised can show gains of 50 percent in gas mileage, one should

1 This is the usual convention, but, theoretically at least, physical laws might be location-
or time-dependent.
CLARIFYING MEANING 137

immediately wonder whether this claim is true only when a special meaning
is given to “properly tuned,” or whether it’s true as a simple, honest, factual
claim in the ordinary meaning of “properly tuned.” If “properly tuned” is
being used to mean the tuning is done just so as to minimize gas consump¬
tion, the claim may well be true, but it won’t be of much interest to the
average user, because if you tune a car simply for maximum gas mileage, it
becomes virtually impossible to start or accelerate. The claim has been made
a planned or nearly definitional necessity at the expense of being a useful
matter of fact. What “properly tuned” normally means is “tuned so as to
best perform the usual functions of a car”; that’s what you take your car in
for when you ask for a tune-up. Now it really would be important if water-
injection could give a 50 percent gain in mileage without loss of general
drivability, power, and other desired capabilities. But it’s totally uninterest¬
ing to hear such a claim if “properly tuned” is to mean “specially tuned for
high mileage,” since any car, without using a water-injector at all, can easily
be tuned so as to give that much gain (e.g., by using a much smaller carbu¬
retor jet or advancing and jamming the ignition timing), if you don’t mind
loss of power, illegal emissions, and having to run it down a hill to start it.
The trained ear responds instantly to such slippery terms as “properly”
or even “normally” or “generally,” and begins to probe their meaning by
asking questions or examining context. Is the claim going to be converted
into a trival one, or even, in the limit, a merely definitional one, by using
these terms in a special sense?
In the scientific domain, exactly similar problems of this type occur.
The general principle of evolution is said to be “the survival of the fittest” or,
in a fancier version, “the differential reproduction of the best adapted.”
Whether that’s a trivial claim or an important one depends entirely on
whether “the fittest” turns out to be defined as “those that survive,” or in
some independent way. If it is defined as “those that survive,” the claim then
reduces to “the survival of those that survive,” which is totally trivial. There
is a long history of attempts, beginning with those of Darwin and Wallace,
to give some independent meaning to “the fittest” and hence make the prin¬
ciple of evolution enlightening. Exactly similar problems occur with respect
to the gas laws in physics (which are said to apply to “an ideal gas”) and the
laws of elasticity (which apply to a “perfectly elastic substance”). What we
discover is that the cash value of these laws lies beneath the surface—in the
extent to which they approximate the behavior of real gases or substances,
not in what they tell us about ideal or perfect substances, since such sub¬
stances do not exist in the real world. The same is true in ethics: “Killing is
wrong”—but not always. There is self-defense, just war, execution: the cash
value is in the fine print, the application of the principle in practice.
Notice that we are here regarding it as grounds for complaint that such
claims are “reduced to the status of definitions.” That’s not because there is
138 CHAPTER 5

anything absolutely inferior about definitions. Indeed, there is something

absolutely superior about them, namely that they are necessarily true in all
conceivable, or at least in all describable, worlds. But their truth is obtained
at a price, namely that they cease to tell us about this particular world and
start telling us about the meaning of words instead. When we want to know
something about the particular world we inhabit, this switch is not useful to
us, and it may be fraudulent. So definitional truths, despite their truth, may
be “trivial” or “uninformative” about what we need to be informed about,
though they’re just what we need to explain the meaning of a term, and
anything less would be unsatisfactory there.
And in mathematics or in the law or even in ordinary argument, it is
sometimes a great triumph to be able to demonstrate that a particular claim
is, or can be, reduced to a definitional truth or truths. For these may well be
highly informative where they express the kind of information we are seek¬
ing; the fact that they are, when true at all, true with virtually complete
certainty, becomes a matter to rejoice over, rather than to scorn because
such truth was achieved via logic rather than observation or experiment. At
the simplest end of the spectrum, the definition may be informative to some¬
one who didn’t know the meaning of the term defined. At the most complex
end, a theorem of Euclidean geometry—say, Pythagoras’s theorem—is infor¬
mative to any student of geometry, even though it can be reduced (by sev¬
eral exceedingly ingenious constructions and proofs) to the definitional
truths that constitute the Euclidean axioms.
So, whereas we will sometimes rightly scorn a claim when we show it to
be definitionally true, we will in other situations be overjoyed to be able to
prove it so. It’s all a matter of what kind of truth or information we are
seeking in the particular context. Earlier, we showed how increasing the
precision of a claim increases its informativeness. But, we said (in fact, it’s
the other side of the same coin) making it more precise means that it will be
harder to show that it’s true (for example, more careful measurement will be
required), and in the absence of evidence, it’s less likely to be true. Hence,
for both these reasons, the more precise a claim is, the less chance it has of
being true. (Consider the following claims or bets in roulette; the next num¬
ber to come up will be even, will be in the top third, will be one of four
adjacent ones, will be 28.) Since information has to be both new (not previ¬
ously known) and true, there’s a crucial tension in the search for it. We can
increase the precision of our claims, which makes it increasingly likely that
people didn’t know them to be true before—but it also makes them increas¬
ingly hard to prove. Or we can make them vaguer and vaguer, less and less
precise, which makes them more and more likely to be true (“The weather
tomorrow will be warmer or cooler”), but less and less likely to have been
unknown beforehand. The ultimate reduction to triviality makes the claim
 CLARIFYING MEANING 139

definitionally true, and obviously so, in which case it’s worth nothing to
those who already know the language; on the other hand, you surely have
made it true, which on other occasions is the big payoff. So the value (utility)
of a claim varies with context, just as the meaning of a term does.
An extreme view, due to a school of philosophers called the positivists,
held that when a claim was impossible2 to disprove it became meaningless.
What we’ve been saying is that it would be meaningless if it could not be
contrasted with anything; and hence, that if one couldn’t rule out (disprove)
the contrasting state of affairs, one couldn’t prove the claim to be true. It’s a
mistake to compress these two fairly plausible assertions into the positivists’
slogan. For example, what are we to say about “There is no life after
death’’? It’s pretty well impossible to prove or disprove it by any direct
experiment that we can do here on earth, partly because if it’s true, you’ll
never know it by direct experience, and it won’t be testable at all if, as is
quite possible (conceptually), the spirits of the dead cannot communicate
with mortals in any way. But it surely isn’t meaningless; for example, one
knows quite well what it would be like to find out it’s wrong, after dying. In
short, it has contrast meaning though it may lack testability.

Here’s an example that you should spend a little time on. An astrologer
reads your horoscope and announces (amongst other things): “December
11th, in the afternoon, is a very risky time for you to make major decisions.”
Is this provable? Disprovable? How? Is it meaningless? (If not, what does it
mean?)
These distinctions between types of truth can be applied to logic itself.
Using p and q to stand for assertions, we can set out various “principles of
logic” quite easily. As you look at them, you’ll realize that they are them¬
selves definitional truths, just truths about the meaning of the words that
constitute the structure (rather than the content) of an argument. A really
simple example would be: “It is false that p is false = p is true” (where we
once again use the equals sign to convey “means (just about) the same as”).
Another, less simple example (which even has a grand title, being one of
DeMorgan’s laws) is: It is false that p and q are both true = p is false or q
is false. The other DeMorgan law is much the same, interchanging “and”
with “or”: It is false that either p or q is true = p is false and q is false.
If we use an arrow to stand for “implies,” meaning here logically or
definitionally implies, then we can (roughly) translate p = q as p —> q and q
—> p. The equals sign is thus (at this simple level) about the same as a
double-ended arrow.

2 Meaning “impossible in principle”; that is, of such a kind that no conceivable evidence
could count against it.
140 CHAPTER 5

You can see that we’re beginning to develop an elementary part of

what’s called symbolic or formal or mathematical logic. What we just said
about the arrows and the equals sign can be put in these symbols as

(p = q) = (P -> q) and (q -* p)

Remember that “p is a sufficient condition for q” is p —> q; “p is a necessary

condition for q” is q -» p. Notice that there’s no great confusion resulting
from using the -» to refer to logically or definitionally necessary (or suffi¬
cient) conditions here, whereas, in the last chapter, we were using it to refer
also to empirically or factually necessary or sufficient conditions. That’s just
one more example of how one symbol can be (1) ambiguous out of context,
(2) perfectly clear in context.
A slightly more complex logical principle—the one that expresses a
basic form of argument—is:

(p + (p -> q) q

The principle used in complex arguments with intermediate conclusions,

sometimes called the “chaining” principle, is:

(p -» q) and (q -» r) and (r s) (p -» s)

An extremely important principle which can be used to found much of

elementary logic and which should remind you of all the material about the
relationship between the truth of premises and conclusions and the soundness
of the inference from one to another is:

(p —> q) and (q is false) —> (p is false)

(This is the so-called principle of the antilogism.)

A typical fallacy, or false argument, is this one:

(p —> q) and (p is false) —» (q is false)

(That’s called the fallacy of denying the antecedent.)

You will remember this form of argument being rejected in the last chapter
by our pointing out that a false premise can easily imply a true conclusion
(whereas the principle just quoted says that if a premise is false, anything it
implies must also be false). To illustrate: If New York and Los Angeles are
CLARIFYING MEANING 141

10,000 miles apart (p), then they’re more than 1.000 miles apart (q). Now,
they are not 10,000 miles apart, so (p is false). Hence, if this form of argu¬
ment was correct, you could conclude that they are not more than 1,000
miles apart: (q is false). But they are more than 1,000 miles apart. So this
form of argument is fallacious.
There s an associated fallacy, called “affirming the consequent,” which
you'll also recognize.

(p q) and (q is true) (p is true)

Give an example which shows this a fallacy (Quiz 5-A, 6).

You can see how we could easily work up a book full of various theo¬
rems and fallacies expressed in these or more elaborate symbolisms. That’s
what the introductory logic course often does. The trouble is that it can
become a complex game of its own and the messy business of dealing with
the rhetoric and arguments of real life gets left behind. We just use the arrow
symbol as an abbreviation for the word or words we could as well use
instead, and we do not require you to remember the names of any of the
fallacies we’ve just mentioned, although their two millennia of use by edu¬
cated people has made them a part of the vocabulary of many thousands of
literate individuals who never took a logic course. Your effort should go,
rather, into using the points made and in practicing them on real arguments.
When you’re trying to clarify the structure or the defects in an argument, it
may be useful to have one or two of these symbols up your sleeve because
they can make the structure—or the fallacy—stand out more clearly than if
you have to keep repeating whole sentences. So, in this chapter on the clari¬
fication of meaning, we’ve spent a little time on a procedure for clarifying
the meaning of an argument. But now it’s time to wrap the chapter up by
looking at some other dimensions of the meaning of terms and at procedures
for clarifying them.

5-10 The Other Dimensions of Meaning

A dictionary definition, especially the ideal version, focuses on what we

sometimes call the literal meaning of a term. But that’s only part of the
story. We’ve already stressed the contextual considerations which enable one
to identify which of the several possible meanings, or several ranges of
meaning, the same set of letters or sounds can have. (“A very small elephant
is larger than a very large mouse.”) One way to put that point is to say that
the context is one dimension of the meaning of a term, because it is a term-
m-a-context that we normally have to interpret. But there are other dimen¬
sions, too.
142 CHAPTER 5

The whole approach we’ve been taking so far is oriented toward the
usual “referring expressions” like common nouns and verbs. It doesn’t work
at all easily with other terms in the language such as proper nouns, connec¬
tives (and, if, then . . . , but, etc.) and participles. These can hardly be
analyzed into sets of jointly sufficient and separately necessary properties.
It’s better to approach them by analyzing their function. (Look up “the” or
“and” in the dictionary: does it offer a translation? What?) You can explain
the function of many words that you can’t translate. Even with some com¬
mon nouns, such as the names of colors, you can’t give real translations.
What you do normally is to give or describe examples. And that’s very like
what you do with a proper name—you point to or describe what or whoever
it labels or refers to. That description is not a translation of the name, but
just a description of what happens to be the present location of the name’s
bearer. “Jacob Astor” does not mean, though it may happen to be the name
of, the third man from the left in the old yachting photo on page 33 of The
Proper Yacht. “Yellow” doesn’t mean, though it happens to be, the color of
ripe lemons. These aren’t translations because, for example, a disease might
strike lemons overnight which causes every lemon on earth to turn purple
when ripe, but that wouldn’t have the least effect on the meaning of the term
“yellow.”
Just as we sometimes exhibit examples to show or convey meaning,
rather than offer translations into other words, so we may exhibit an exam¬
ple of the use of a term to achieve the same effect. We may explain the
meaning of the word “not” by saying “not doing X” means the same as
“doing something other than, or different from, X.” In this case, we have
actually offered a translation of one particular use of “not,” namely, its use
in the phrase, “not doing. . . .” But we may in fact have achieved our much
more ambitious purpose, which is to convey the whole meaning of “not”; or
it may take a few more examples like this to achieve that general purpose,
depending on the linguistic insight of the listener. (This approach is some¬
times called “implicit” definition.)
So there is a spectrum of procedures for clarifying meaning, from point¬
ing to a particular representative of the class to which the term refers,
through drawing or describing that class, through providing a translation of
one particular use of the term, all the way to providing a completely general
translation or synonymous expression, as the ideal dictionary definition
does. And there are supplementary procedures such as describing the lin¬
guistic function of the term (“It’s the preposition used to refer to . . .”), or
describing the contrasts that the term is used to draw (“The nadir is the
opposite of the zenith”).
The message of this section, then, is that there are many ways of clari¬
fying meaning and that they should be judged, not by the extent to which
 CLARIFYING MEANING 143

they conform to the standard of the idealized version of a dictionary defini¬

tion which many logic texts treat as if it were the only proper model, but
simply in terms of the efficiency with which they succeed in making the listener
able to make sense out of proper usages of the term.
In particular, one must not regard the demand for a definition in the
sense of a translation as always legitimate. With respect to a new or unusual
term, by all means. With respect to an ambiguous term with several clear
and distinct meanings, perhaps. But with respect to an abstract term in
common language, a term like “justice" or “ethics,” a term whose meaning
is as hard to formulate as any theory about the nature of the abstraction
itself, then the demand for an explicit translation is often a red herring in an
argument, about as useful as calling for someone to prove everything she
says all at once. Such terms can be clarified, if necessary, by illustrations, by
giving an example of their use in the sense intended, or by explaining the
contrast intended. But no language can be learned entirely from a diction¬
ary, and no dispute over justice can be settled by appeal to one.
One last note must suffice. We have said little about the subtler flavors
and dimensions of meaning, about what is sometimes called the connotations
or the associations that a term may have. In the first place, these are not
sharply distinct from the core or literal meaning, though it is worth distin¬
guishing them for many purposes. Poetry and rhetoric exploit these rich
fringes of meaning to good and bad effect. We know from metaphor and
simile how, for example, color terms may come to be associated with certain
moods or types of behavior (blue or white with coldness, red with anger and
violence), and in some contexts a full analysis of meaning must go into such
dimensions. We have already done this in examples of practical argumenta¬
tion and will do it again.
In the second place, these connotations or associations cannot be re¬
garded as illicit baggage to be cut out whenever we wish to speak rationally
or argue logically. Such an excision is beyond our powers with respect to
natural languages, but it is also beyond our mission or our rights. In reason¬
ing, we must recognize such components of meaning, we must learn how to
exhibit them, clarify them and work with them, not fight against them or
seek to do without them. Emotional or poetic language is not just a part of
language, but an important part of language. There are times when calling
someone a bloody murderer is not a lapse into the language of the streets, an
abandonment of the standards of impartial reason, but no more and no less
than the truth, and the same with poetry.

QUIZ 5-A

The first few questions in Quiz 5-A are embedded in the text of the current
144 CHAPTER 5

chapter, since, as we get into heavier material, it seems important to give you
a bit of a workout as you go along.
Answer
First you ask yourself whether it’s too inclusive, that is, does “it” (the right-hand side
of the definition) include some things that aren’t included by the term “college”? Of
course, it’s obvious that you get educated, after leaving high school, in other places
besides college; for example, the army, real life, or business school. Still, it might be
said that although these are places where you do get educated, they aren’t places
where you go to get educated; that is, the principal reason for attending such institu¬
tions is not to acquire education but to do something else. This is a rather weak
defense—at least some people join the armed services in order to get free training as
a technician, or to get educated about the world (since the navy guarantees you
traveling). But there are stronger counterexamples; to name one, people join trainee-
ship and apprenticeship programs in order to get educated in particular crafts, skills,
or professions, and these programs are not college. So the definition is too inclusive.
Now you ask whether the definition is too restrictive, that is, whether it excludes
some things that are properly called “college.” This is a little tougher. On the inter¬
pretation previously suggested, which takes “go to get educated” to mean “go with
the intention of getting educated,” one might well argue that a great many people
don’t go to college to get educated but to have a good time, avoid starting real work,
meet a mate, or enjoy other side benefits. Hence, at least some colleges would be
excluded because of the reasons for attendance of their enrollees. However, this is a
pretty weak kind of counterexample, since as long as some people go there to get an
education, and it would be hard to deny that some such people are on every campus,
the place would qualify as a college under this definition, since the word “you” in it
presumably refers to anybody, and not just to the person who happens to be reading
this at this moment. The same kind of defense would apply to attempted counterex¬
amples based on the fact that some people go to college without going through high
school; either they have had correspondence or tutorial instruction, or they have
educated themselves and take one of the “challenge” entry tests. One might say that
the relative imprecision of this definition immunizes it against that type of example.
However, there is a family of considerably more serious counterexamples. College is
not just a place that one goes to, after finishing high school, in order to get an
education, but it’s also a place you go to after finishing college. So the definition is
certainly too restrictive because it excludes graduate schools, and a good many col¬
leges have only graduate schools.
So the definition is both too inclusive and too restrictive, with fairly clear and
not nit-picking counterexamples to support both these defects. We may conclude
that the definition is pretty unsatisfactory. There isn’t any other type of error it could
involve; but it could of course be even further from the truth if the counterexamples
were even more extensive than the ones we’ve mentioned. Obviously, if you look at
it, the definition does cover a good big slice of colleges; it isn’t as far off the mark as
it might be. (Compare it with “college equals a private high school”; in that case
there’d be practically no overlap between the real range of reference of the term
“college” and the range of reference of the proposed definition for it. The circles
representing these would overlap only a little bit.)
CLARIFYING MEANING 145

2 When I looked up “evaluate, the only relevant definition of the common sense
of the term that occurs in the Random House Dictionary of the English Language
is: "to determine or set the value or amount of; appraise . . To define “evalu¬
ate in terms of determining the value of something does suggest that we’re pretty
close to a circular definition. Let's turn to the definition of “value” and see if it
gets us out of trouble. The first definition is: “relative worth, merit, or importance
. . . That would be all right if we could get a definition of worth, merit, and
importance that doesn't involve the term “value” itself. Let’s have a look. Turn¬
ing to the word “worth," we find that it is defined in terms of either “goodness
and importance" or “having a value,” so one of those tracks is completely circu¬
lar, and the other leaves us with only the word “good” as a possible way out of
the circle. The word “good” is given fifty separate definitions. It would take us a
long time to track down the results of looking up every term referred to in these
definitions. Suffice it to say that a great many of them stray back into “value,”
e.g„ through terms like “quality.” It’s fairly common to find that, in the dictio¬
nary, relatively abstract terms are defined in one of these “extended circles.” It’s
not necessarily bad. since somewhere around the circle you may encounter a term
that you already understand on independent grounds, and in that case the defini¬
tion will serve the purpose of explaining the meaning of the word to you. It isn’t
of very much use to define “good” as “the property possessed by things which
exhibit goodness,” but if the circle gets a little more complicated, like the one
we’ve just followed around, then it may well reach out into some other part of
your understanding of the language and make a connection that will give mean¬
ing to the word that you’re puzzled about.
It isn’t often that you find crude errors of excessive inclusiveness or restrictiveness
in the dictionary, but you do find plenty of cases where what are provided as
defining properties are actually “only” criteria. For example, in the definition of
“apple” (in the Random House Dictionary), one finds “the usually round, red or
yellow, edible fruit of the tree Malus pumila”
Now you know as well as I do that there are a number of species of apple that are
green from first formation through full ripeness. You also know that to call these
fruits round is a little imprecise, especially given the extremely elongated shape of
certain species. And of course, there are certainly inedible applies like some of the
crabapples. What are we to do with the word “usually” in that definition? Upon
statistical grounds, it’s probably true that the dictionary “definition” is correct.
But it really doesn’t serve the function of a definition very well if there are clearly
cases of well-known apples that do not meet all these conditions or that indeed
meet only one, and that one only if it is construed very loosely. You and I clearly
have a better concept of apple in our head than that dictionary definition con¬
veys. We have no difficulty in saying of a rather pear-shaped, green, and inedible
fruit that it’s an apple, albeit it’s an unripe apple, even if it was grown in a test-
tube (we’d use taste, texture, scent, and other characteristics as guides). But the
dictionary wouldn’t enable you to identify it as such. That's one of the limitations
of dictionaries—they have to be concise. The extent of our knowledge of the
meaning of common nouns is so vast that we can handle dozens or hundreds of
cases that would not be covered by any short definition, with considerable agree-
146 CHAPTER 5

ment between language users as to the correctness of our decisions. Dictionaries

are just rough guides to the use of most terms, and very useful they are too; and,
with some terms in mathematics, they can be precise and accurate. With most
terms, however, they have to be either imprecise or inaccurate.

3 This question requires you to classify a series of claims as being true or false and,
under each of these headings, as true (or false) by virtue of definition, physical
necessity, plan or design, social or moral norms (“laws”), or as a mere matter of
fact. (See Types 1-7, listed on p. 135.)
a You can’t pass this course without reading the text.
Not an easy one to classify. Since the skills we’re talking about in this text are
very closely tied to commonsense skills, though we’re trying to add a consider¬
able degree of refinement and improvement, and since we carefully avoid
getting into too much jargon, it is just possible that somebody might pass the
tests and hence get through the course without reading the text. Certainly this
isn’t a case of planned impossibility, because it isn’t one of my goals to write
a text that defines a new subject. The text is, for me, simply a means to
improving a type of skill that can be recognized in the outside world anyway,
and there are a number of other ways, and perhaps courses, which would
make it possible for you to pass the exams in this course and hence pass the
course. So this is either (merely) factually true or factually false, and we really
don’t have much evidence about which it is until a few people try to pass the
course without reading the text. The chances are against them, but not so
strongly that we’d have to say this claim is obviously or provably false at this
point; although it is probably false as a matter of fact. (So its correct classifi¬
cation is F-7 or T-7, with the probabilities favoring F-7.)
b Sons have to have (or have had) fathers.
The way the world is at the moment, the evidence strongly suggests that virgin
birth in the human race is practically or completely nonoccurrent. (Even
Christ had a father.) Since this is not a mere matter of fact, but a result of the
nature of the conception and birth process as we now understand it, this is the
case of a physically necessary truth, i.e., T-2. It isn’t logically (or definitionally)
necessary because of course, the idea of virgin birth makes perfectly good
sense within the language, even if it doesn’t happen to occur in fact,
c Sons have to have (or have had) a parent.
Here we’re up against something that looks more like a logical truth (T-l)
because it’s difficult to imagine what could be meant by referring to any entity
as a “son” if he wasn’t the son of somebody, i.e., of a parent. It might be a
mother (in a case of parthenogenesis), but unless there was an identifiable
person who stood in the parental relationship to an individual, one wouldn’t
concede that it was proper to refer to that individual as a son. And if there is
such a person, then that person is a parent, so this is a definitional truth,
d No black woman could be elected President of the U.S. before 1984.
This is certainly a true fact at the moment, and it isn’t exactly a “mere” matter
of fact. It is something which we have allowed to become the case through our
CLARIFYING MEANING 147

societal development, and it is certainly within the power of people to arrange

it otherwise. That’s what’s meant by social necessity,
e Democracy cannot exist when there is no freedom of speech.
Here the answer will depend on how we define democracy. There are some
definitions of democracy—not too bad ones—which would build freedom of
speech into the definition. In that case, this would be a definitional truth.
There are some strong arguments—for example, those by John Stuart Mill—
that it isn’t possible for a democracy to exist without freedom of speech,
although he didn’t define democracy as involving freedom of speech. This
would make it a matter of social necessity, or perhaps a matter of deep psy¬
chological necessity, in which case it would count as a subspecies of physical
necessity. Or one might argue that the necessity is more a matter of planning
or social mores than natural law, in which instance one would classify this
statement as T-3 or T-4; and finally, if one wasn’t persuaded by any of the
previous arguments, one would probably have to accept the argument that in
the absence of freedom of speech, one does not in fact have a democracy
(which would mean this is a T-7 claim).
Of course, we’re talking in a rather long-term sense about this example; it’s
clear that during wartime there are often curtailments on freedom of speech
just for the duration of the emergency, in countries that call themselves de¬
mocracies. One might be inclined to say that such a country is still a democ¬
racy even though there is no freedom of speech for the moment; hence T-l is
out. If one is very literal in interpreting the claim, one might just have to say
that it is F-7, since there can be short periods of time when there is no
freedom of speech and we still have a democracy. The argument isn’t compel¬
ling, since one might also argue that, for the time being, the society has ceased
to be a democracy, and will revert to being one when such restraints as this are
eliminated.
f The interior-angle sum of a triangle is 180°.
This is a definitional truth only with respect to Euclidean triangles. If you
assume that the context is such that we are clearly talking about Euclidean
triangles, although this isn’t stated, then it’s a definitional truth. If you take
the sentence out of context, then it’s definitionally false, or factually false
depending on the geometry, since it is possible to prove from the definition of
non-Euclidean triangles that they cannot or need not contain 180° of interior
angles.

4 This question concerns the provability of an astrologer’s assertion that “Decem¬

ber 11th, in the afternoon, is a very risky time for you to make major decisions.”
Without going into it at great length, it’s clear that this is a very difficult claim to
prove or disprove. If you do make major decisions and they work out on that
afternoon, it can still be the case that you are very lucky and that it is indeed a
very risky time to make them. If you fail to make major decisions on that after¬
noon, perhaps because you believe the prediction, then you’ll never know
whether they would have worked out if you had made them. Nevertheless, the
statement isn’t meaningless, and what it means is much the same as what we
148 CHAPTER 5

mean when we say that your chances are not very good of rolling a double six
with a pair of dice. This isn’t falsified by the fact that you do roll it, any more
than it’s proved to be correct by the fact that you don’t. What would prove it to
be true would be our running a very large number of trials with the dice and
finding that it very rarely happened that you rolled a double six. Now the prob¬
lem with using that line of proof on the present example is that it refers only to a
single day, indeed a single afternoon, in a single year. Hence, we can’t repeat a
whole series of experiments that will be relevant. Nevertheless, we can have good
general grounds for believing that this is true, and we may be able to get direct
evidence in another way. For example, we might be able to use a very large
number of people instead of a very large number of trials by the same person, and
have those people—all of them under the same astrological sign—make or not
make major decisions on that afternoon, and see whether they turned out badly;
then we would have them make some other comparable decisions on an after¬
noon for which the prediction by the astrologer was favorable and see if we got a
large difference. It’s largely because experiments like these have never been done
that the common remark by scientists that astrology is actually false is rather
unscientific. We do have some significant general reasons for thinking that astrol¬
ogy is unlikely to be true, but they are essentially indirect reasons. Direct proof of
the type I’ve been describing has never been attempted under careful scrutiny.

5 The question requests that an example be given which shows that “affirming the
consequent” is a fallacy. “Affirming the consequent” simply means that you jump
from the realizations that (a) one proposition implies another and (b) that the
second proposition is true, to the conclusion (c) that the first one must be true. To
believe this is really to believe that a false proposition couldn’t have true conse¬
quences. But we know from examples that we’ve discussed before that it’s easy
for a false proposition to have true consequences. Any proposition that is false
because it exaggerates the truth will imply more modest claims that may well be
true. For example, the false proposition that all women are feminists implies the
proposition that some women are feminists. But it doesn’t follow that it’s true
that all women are feminists.

6 Set up and criticize a slippery slope argument aimed at proving that we should
not legalize marijuana.
a “If we legalized marijuana, we could hardly avoid legalizing cocaine and then
heroin, which are only different in degree with respect to severity of addictive¬
ness and results of excess use.”
b (The “fast switch” or “dose of your own medicine” move.) “If these differ¬
ences of degree are not significant, as you argue, the differences between mari¬
juana and the various legal drugs like alcohol or nicotine or Valium are not
significant. Hence, marijuana should also be legal.”
c (The “head-on” attack.) “You can die from an overdose of heroin, but not
from an overdose of marijuana.” Withdrawal from heroin has sometimes
CLARIFYING MEANING 149

proved fatal, and has usually been a major trauma; neither phenomenon is
generally true of marijuana, possibly never true. These are not just differences
of degree, but of kind, so the argument doesn’t even get off the ground.

QUIZ 5-B

1 Women are not so strong, physically, as men. There are various ways in which
this claim could be interpreted.
a Give one interpretation—a plausible one, not an absurd one—that makes it
true. (Do this by giving a “translation” that is more precise.)
b Give a plausible interpretation (that is, one which could well be what a user
of the sentence means by it) which could make it false,
c Since the a and b interpretations are both possible, it follows that the claim
is
(1) False
(2) True
(3) Either true or false, depending on the interpretation
(4) Imprecise
(5) Ambiguous
d In the absence of evidence to the contrary, the Principle of Charity would
require you to adopt
(1) Interpretation a
(2) Interpretation b
(3) Neither
(4) Both

2 Reading Dynamics guarantees that you will triple your reading speed as a
result of taking our course, with no loss of comprehension, or your money
back.
This statement is:
a So vague as to be almost meaningless
b Ambiguous
c Imprecise
d Clear enough to provide a reasonable basis for a legal contract, though you
wouldn’t want to pay the fee until you see how the instructor measures
_(fill in).

3 If the function of the CIA is to further our foreign policy aims, and if the
government in a given country always resorts to the killing of political prison¬
ers, guerrillas, or police victims, it is a little hypocritical to act outraged about
the CIA’s planning to assassinate a single key person in that country in order to
advance what our democratic government judges to be the cause of democracy,
a What type of argument is this? Why?
b State in one sentence the strongest possible counterargument.

4 Read the full-page advertisement quoted on page 150 very carefully:

150 CHAPTER 5

DUNLOP HAS HIGHEST RATING IN

CAR & DRIVER RADIAL TIRE TEST

Car and Driver tested 15 different brands of radial tires for overall track per¬
formance. The results were published in September Car and Driver. All quotes
are from that publication.
When all results were in, when all 15 brands were evaluated . . . the
Dunlop SP-4 Radial was classified “Excellent” . . . the highest rating awarded
in the test.
Tests were conducted at the Lime Rock road race course in Connecticut,
“an uncommonly severe test of tire durability.”
“We focused on dry road performance, which gets you down a race track
(or down a country road) quickest.”
One driver tested all tires. To keep results unbiased, he knew only the code
number, not the brand, of the tires he was evaluating.
“We can’t tell you which tire corners best or stops fastest, but we can tell
you which offers the best combination of both. For track use and most road
applications, this information is more valuable than the abstract results of skid-
pad testing.”
“Dunlop is a very easy tire to drive. It’s quick, forgiving, and by far the
most consistent tire we tested.”
Car and Driver told their readers, “If you are looking for a showroom
stock tire, our findings are as valid as a Supreme Court decision. The same
applies if you are after a good handling street tire for your small sedan or super¬
coupe.”
The Dunlop SP-4. Only one of the many Dunlop radial lines, all built to
the same high standards of quality, performance, and durability. When you
need tires, buy the ones the PROS recommend. DUNLOP.

a State in your own words the most natural, or one natural, interpretation of the
heading at the top of the ad.
b Give a description, in a sentence or two, of how the Car and Driver tests might
have looked, making it consistent with the details in the ad’s fine print, and
which makes the ad’s heading extremely misleading,
c State an interpretation of the heading that is not misleading, if different from
your interpretation in a.
d Is the ad misleading? Yes/no.

3 5 “Chair” means “portable piece of furniture with a back, legs, and a seat for one
person.”

6 “Furniture” means “manufactured objects designed to provide surfaces for

sitting, eating, and storage.”

3 Remember that this book is trying to move you toward realistic examples, so it does not
put detailed instructions before these questions. You wouldn’t find any in real life. We could
say, “Comment as seems appropriate,” but that’s trivial. Should you comment as seems inap¬
propriate?
CLARIFYING MEANING 151

7 “Democracy” means “government of the many by the few, for the benefit of
the many.”

8 A sale ad says, “Save on broadloom carpet—list price $9.99 per sq. yd., sale
$6.99, limited time only.” What impression is conveyed, and what fact(s) might
actually be very different from the impression?

9 Argue (a) for and then (b) against the view that “covert operations” (such as
secret activism in politics through cash support and “dirty tricks”) are a legiti¬
mate function of the FBI (at home) and the CIA (abroad), (c) What is your
view, and why do you prefer it?

10 "Lease a new Cadillac, fully equipped, insurance and maintenance and loan
car included, for only $115.90 a month, equity lease.” What are the “catches”
to look out for here?

QUIZ 5-C

1 Look up the definitions of “college” in your own dictionary and see whether they
are immune to criticism.

2 Think of a concept that’s particularly interesting to you at the moment, perhaps

because of some current issue or course in which you’re involved. Track down the
chain of definitions in your dictionary by looking up the definitions of the terms
used in the definitions of your term and then looking up the terms used in those
definitions; and so on. Did you eventually get a circle? Does it matter? Could you
avoid all circles in dictionary definitions? How? (Possible hares to chase through
the dictionary are “equality,” “pornography,” “oppression,” “power,” “free¬
dom,” “insanity,” “reason.”)

3 a What does “democracy is the best form of government” mean? Use at least
the contrast approach to answer,
b Does that approach help clarify meaning here?
c What else helps?

4 a What does “and” mean?

b What does “true” mean?
c What does “mean” mean?
 Chapter 6

The Finer Points of

Argument Analysis

6-1 IMPLICATIONS AND CONCLUSIONS

At the end of the last chapter we were talking about the subtler dimensions
of meaning, and we argued that these, including the emotional and poetic
dimensions, are sometimes crucial for understanding the meaning of state¬
ments, descriptions, or arguments. It follows that they will sometimes be
crucial in argument analysis, because these dimensions of meaning may
contain the implicit conclusions of an argument—the “implications” of an
argument, as we often call them.
The line between the meaning of a statement and what it implies is not
a sharp one. To properly understand the claim that the United States Air
Force provided air support for a number of ground missions by the South
Vietnamese army may require that you understand that “air support” is a
euphemism for (that is, a sugar-coated way of saying) “bombing.” That’s a
case of understanding the finer points of the meaning of a term that has
crept into the language in a semitechnical way. “Military advisors” eventu¬
ally turned out to “mean” combat troops. But, originally, there may well

152
 THE FINER POINTS OF ARGUMENT ANALYSIS 153

have been a distinction between “air support” and plain old bombing. For
example, the spotting of targets and the relaying of information about their
location to the ground gunnery-control centers was an example of air sup¬
port. The use of planes to provide air cover for the infantry, that is, to
provide protective forces that would respond only if enemy planes attempted
to intervene, was another case of air support that was not bombing. But
when the enemy has no air force and when conditions reach the state that
they did in the last days of the war in Vietnam, where, on the one hand, the
forces of the South were losing, and, on the other hand, the pressure of the
media was becoming increasingly critical and powerful, we find an ideal
situation for the term “air support” to become a polite label for bombing
forces and positions on the ground, including of course, many civilians.
Nevertheless, the Air Force found it attractive to be able to continue to use
the euphemism instead of the proper name. Colonel Opfer of the U.S. Air
Force, speaking at a news conference in Cambodia at that time, said to the
reporters present. You re always saying, it’s bombing, bombing, bombing.
It’s not bombing! It’s air support.” (That remark was awarded the Annual
Doublespeak Award by the Committee on Public Doublespeak, a committee
of the National Council of Teachers of English.) It was once said that the
Vietnamese war “was conducted not in secret but in jargon” (Henry Fairlie,
“The Language of Politics "Atlantic, January 1975). Usually the sequence of
events was that perfectly ordinary language was used in a perfectly open and
proper way until something occurred that people were anxious to conceal or
represent in a favorable but unjustified light. For example, in the early days
of our involvement in Southeast Asia, we did indeed send in advisors to the
agricultural, the educational, and the military arms of the government that
we wished to support. Before long, however, it turned out that the advisors
in the military dimension had suddenly become “advisors,” that is, they
were actually an army. One of the distinctions is, of course, that advisors
don’t actually pull triggers; it wasn’t long before our men were pulling the
triggers. Still, the Department of Defense continued to use the term “advi¬
sors”; and so did the State Department, and so did the White House. And it
took a long time for most of the general public to catch on to this. So one
has to be extremely cautious, in analyzing an argument, to determine the
exact meaning of the terms and claims in it—a meaning which may change
depending upon the month or year in which the argument was uttered or
presented.
The situation on the domestic front is no better. In the huge computer
storage system that records all education research (ERIC), the instructions
to people who are summarizing their research articles include the instruction
that one should not use the term “poverty,” but “economic disadvantage-
ment.” That’s a simple case of a euphemism, with even some kind thoughts
behind it; the same could be said of the use of the term “exceptional chil-
154 CHAPTER 6

dren” to refer to children with physical or intellectual handicaps. But it’s not
clear that the saving is worth the misrepresentation. If you are searching for
literature on the effect of federal intervention on extreme poverty, you’re
going to miss some of the research because the federally funded repository of
all that research will not list the natural and correct term. And if you are
looking for programs for exceptionally bright children, you may waste a lot
of time looking through the programs until you realize that the correct eu¬
phemism is “gifted” rather than “exceptionally able” (or successful, intelli¬
gent, or similar designation).
There are many subtler examples that fall under the heading of “impli¬
cations” of claims in our language. One to which a great deal of attention
has been directed lately is the tendency to use “he” as the pronoun to refer
to legislators, and “she” as the pronoun to refer to whoever is running a
home. While there is a statistical basis for these pronouns in that more men
than women are legislators, and the reverse for homemakers, the implication
of using the single pronoun goes a little beyond mere generalization and into
the realm of stereotyping. Where it is possible to avoid stereotyping by some
relatively simple linguistic procedure, it’s a good idea to do so since the
effects of stereotyping are often morally unfortunate. (That is, stereotyping
tends to make a woman company director or a male nurse into something
weird or improper instead of merely infrequent.)
We’re going through a period of reanalysis of the legitimacy of our
public agencies, and this is an appropriate time to tighten up a little on our
language. It’s worth remembering the occasion on which Ron Ziegler, the
press secretary to President Nixon, declared certain previous statements by
the President on the subject of Watergate to be “inoperative,” a brilliant
example of an original euphemism. “False” is a nice old-fashioned word that
does the same job, and it doesn’t even have the implication that the individ¬
ual who uttered the statement knew it to be false, in which case the correct
word would be . . . what? (Quiz 6-A, 1)
In the past few years, a number of books have come out that focus
directly on the use of language as an instrument of political power. Those of
you who are interested in this dimension might want to follow up by reading
some of them, for example:
The Language of Oppression, by Haig Bosmajian, Public Affairs Press, 1975
Power to Persuade: Mass Media and the News, by Robert Cirino, Bantam Books, 1974
Language and Public Policy, edited by Hugh Rank, National Council of Teachers of
English, 1974
Politics of Lying: Government Deception, Secrecy and Power, Random House, 1973
Of course, the forerunners of these analyses are still worth looking at:
The Hidden Persuaders, by Vance Packard, Pocket Books, 1975
“Politics in the English Language,” by George Orwell, reprinted in A Collection of
Essays by George Orwell, Harcourt Brace, 1970
 THE FINER POINTS OF ARGUMENT ANALYSIS 155

(Many of these references came from a special issue of Looking At, devoted
to public doublespeak, April 1975.)
The other side of all this subtlety is that it’s very easy to overreact to
implications that you think are present in an argument, although it’s not at
all clear they’re there. It’s all right to raise an eyebrow, and sometimes to
make quite a point, about a stereotyping assumption that is revealed by the
use of the pronoun “he.” But there are many contexts in which it will only
count as a weak criticism of an argument which essentially centers on other
issues. Remember to use your judgment about strong and weak criticism.
There may be many implications in the way in which an argument is
phrased, some of them quite objectionable; nevertheless, the argument may
quite easily be tidied up by some rephrasing, without losing its main thrust.
Be sure that the criticisms that you make bear on the main thrust, as well as
on the incidental and objectionable implications.
In Chapter 5, we also looked at cases where the meaning really isn’t
clear, as opposed to cases where the meaning is not obvious on the surface.
Sometimes language can hint at meaning when it has little or none. It’s
important not to assume that because there are some subtle hints or conno¬
tations associated with a certain type of language, it must really have a
substantial meaning. A recent development in this area is what R. D. Rosen
called “psychobabble” (in an article in New Times). Here are a couple of
examples: “I’ve been keeping my head together by getting in touch with my
body”; and “I’m into high energy. You know what I mean.” The point is
well illustrated in an anecdote that Rosen recalls about a patient undergoing
psychoanalysis. The patient eagerly responded, at the beginning of therapy,
to each interpretation his analyst made by saying, “I hear you, I hear you.”

“I’m sorry,” said the doctor, “I didn’t know you were a little deaf.”
“I’m not. I hear you. It means I comprehend.”
“Well, what do you comprehend?”
The patient paused, “Jesus,” he replied, “I don’t know.”
“Psychobabble,” the psychoanalyst said, “is just a way of using candor
in order not to be candid.”

Most of us are still in a state where psychobabble looks pretty good to

us. When people say, “You’ve got to go with your feelings. Let it all happen.
It’s all progress,” one still feels that it’s good advice. What does it mean? Ten
years ago, people might have said “Stay cool.” Ten years before that, they
might have said “Take it easy,” or “Relax.” Well, sometimes that’s good
advice and sometimes it’s not. Using it as a reflex comment on everything
makes it meaningless—or false. Sometimes you have to stop its “all happen¬
ing” and go with your judgment instead of your feelings, get to work instead
of relaxing, and so on. Psychobabble language is different, and for a while it
156 CHAPTER 6

seemed to be saying something more interesting, and perhaps at the very

beginning it was. But before very long, it lost almost all its real content, and
today it is nothing more than an empty jargon, part of the performance of
communication without the substance.
So, in argument analysis, the key is to look for the core, for the content,
for the main meaning; then to look for any extra layers of meaning that may
be important. Of course, in practice, the best way to go is to chase all the
strands of meaning to begin with and then start to put them together and to
work out what the mam strand is and what the secondary strands are.
It’s not hard to see from this section the extent to which judgment is a
crucial part of argument analysis, indeed of understanding utterances in the
English language. They are often rich in meaning or in the appearance of
meaning, in connotations or associations or implications, and sometimes in
resounding emptiness. Indeed, it isn’t possible to condemn or praise much
communication without taking account of the context; what may be subtle
and sensitive in one context may be superficial and uninformative in an¬
other. Judgment enters here, and judgment isn’t something that can be
achieved by following a simple formula. Reflecting on the examples we’ve
been discussing and the ones to come is the best way to develop judgment.
Your judgment may not always finish up at exactly the same place as your
instructor’s or your peers’ or mine, but at least you will be able to judge
more quickly and more accurately once you’ve begun to see more clearly the
types of abuse that are possible. For practical purposes of argument analy¬
sis, the implications (if you’ll pardon the expression) of what we’ve been
saying are that the second step of argument analysis—identifying the con¬
clusions—will emerge from the first step (clarifying the meaning) quite
gradually and subtly, and it is indeed just a part of understanding an argu¬
ment that you’re able to say what its conclusion is, whether explicit or
implicit. The practical procedure you should use here is always the same:
Write it out, get it down, spell it out. Try to formulate the main conclusion
and then, later, the intermediate conclusions of an argument, not just in your
head but on paper. The first try won’t usually be right, but get it written
down, look at it, and then reflect on the argument. Have you really ex¬
pressed the main point that the arguer is trying to establish? If not, tidy it up
or rephrase it completely. There is no better way of reasoning than with a
pencil in one’s hand—I say a pencil because it usually has an eraser on one
end, and you are going to need both ends to get these things right.
The general advice that we’ve mentioned before is nowhere more appli¬
cable than here. Argument analysis is a long, hard business. Don't expect
that you’ll be able to do a good job—certainly not the best job of which you
are capable—the first time you pick an argument to pieces as an assignment.
Figure that the first time through is strictly for ideas, although you can set
the task to yourself as getting the job done right right off. That’ll make you
THE FINER POINTS OF ARGUMENT ANALYSIS 157

work at it a little harder, but usually you won’t get the job done right with
that first try. Even this simple step in argument analysis, getting the conclu¬
sions correctly formulated, will often require several trials before you’re con¬
vinced that you’ve done justice to the argument. And usually some more
trials before somebody else is convinced.

6-2 COMPLEX STRUCTURES AND LONG ARGUMENTS

This business of getting yourself into the habit of writing and rewriting, as
you do the argument analysis, requires that you use a set of abbreviations
for your own convenience. You don’t want to have to write out each sen¬
tence or claim to which you are referring every time you put it down in the
structure diagram or refer to it in your criticism. This is particularly true
with long and complicated arguments. In this section, we’ll mention briefly a
few ways in which one can cope with the “monsters.”
As we’ve previously suggested, you can often mark up the original argu¬
ment (preferably using a pencil, because you probably won’t get the best
breakdown the first time through) to indicate all the crucial claims in it that
you want to link together in the structure diagram and comment on in the
course of the criticism. In some assignments, where it’s no use marking up
the original text, since you’re not planning to hand that in, and you can’t
make a photocopy, you may want to set out immediately, at the beginning of
your answer, a little “dictionary” of the abbreviations that you’re going to
use. This might look something like the following:

1 There is hardly any product that cannot be made more cheaply by

somebody who does not care about quality.
2 There is no getting away from the fact that quality takes more time
and costs more money.
3 People who believe that you can get quality for free are simply dream¬
ers—or they’ve never tried to make something decent themselves.
4 You get what you pay for.

In an argument that has a simple structure that involves using the first
three of the above statements as premises, and then has a few connecting
words which indicate that the arguer proposes to draw from those premises
the conclusion listed above as the number 4, you can easily set out a struc¬
ture diagram as before, using just the numbers. It would look like this:
158 CHAPTER 6

Or, depending on the connecting words, it might be this:

This is a simple way of conveying to the reader, as well as yourself,

exactly what the structure is that you are taking to be the structure of the
argument. You’ve worked out the meaning of the separate parts of the argu¬
ment, which is, in this case, an easy task. You’ve completed the second step
of identifying the conclusion, which is also easy, in this case. Now you’ve
completed the third step, setting out the structure—that is, what premises
are supposed to support which conclusion, with the help of what other prem¬
ises. Again, this step was quite simple in the present case. But notice that
you’ve already introduced a set of abbreviations that made it much quicker
for you to complete the third step than it would have been if you’d had to
rewrite all the sentences again. Of course, you had to set out your little
“dictionary” so that the person who’s reading your answer knows just what
you are referring to. But you are going to save on the time that that took
because, later on as you get down to criticizing the premises and conclusions
or the inferences that connect them, you will be able to use the same num¬
bers again. You’ll be able to refer to the premises or the conclusion by a
number, and to the inferences by an arrow connecting two numbers (e.g.,
l->4).
It’s quite simple to extend this procedure, which we’ve used as a stan¬
dard one up until now, to handle very long arguments. Instead of using a
number for each claim or assertion that’s made, which can give you a pretty
complicated structure after you get to about ten or twelve claims, you start
using a larger “unit” than a single claim. Suppose we are talking about an
argument that covers a number of pages in an article that you are reading.
For example, it might be a long article which is attempting to summarize the
case for and against school busing as it now stands. There are many issues
involved in that by now, and an article that summarizes them is going to
take a good many pages. You have the assignment of evaluating this article.
How can you communicate to your teacher, or to another audience such as
a study group or a parents group, the strengths and weaknesses of the arti¬
cle? If it is well written, a very natural step is to use numbers for the para¬
graphs because each one will be devoted to a separate point. Then your
“index” will tie arabic numbers to the various paragraphs. (Of course, we’re
assuming here that your audience has a copy of the original materials; if
THE FINER POINTS OF ARGUMENT ANALYSIS 159

they don’t, you provide that copy, perhaps as a photostat, and you can
attach the numbers directly to it. Or you make up a “dictionary” as before.)
Now. with the paragraphs numbered, you set out a structure diagram
that shows the way in which they are related. Typically, in a well-set-out
argument, each of the paragraphs will be aimed at establishing a point or
further developing a point that’s already been established. Set the argument
out in terms of paragraphs; this will lead to a manageable and comprehen¬
sible structure. It would be tolerably long and difficult to comprehend if one
had to do it in terms of sentences, and even longer if one had to do it in
terms of claims, since there may be many claims in a particular sentence.
Now your criticisms of the article, if there are any, will certainly focus
on some of these paragraphs and not on all of them to an equal degree. With
respect to those paragraphs which you propose to criticize, you now “put
them under the microscope,” set them out with numbers for each of the
claims that are involved, and carry on with argument analysis from there, as
before.
All this effort goes back to the original mandate which we have stressed
repeatedly: It's not enough that you develop a good instinct for what’s a
good argument or a bad one. You have to be able to persuade others of this.
The function of reasoning is nearly always tied to communicating. Even
when it isn’t, setting out your argument analysis in full is really important in
order to clear your own head about the strengths and weaknesses of the
arguments. If you have to communicate with other people, you must use a
system of abbreviations and labels that they can pick up very, very quickly,
or they won’t stay with you (or they’ll make mistakes in understanding what
you’re trying to say). And even when the effort is purely a selfish one in
order to see whether you have made slips, you yourself need a very simple
and clear procedure for setting out structure and “backtracking” your criti¬
cism. One such structure is the use of numbered paragraphs, as just men¬
tioned.
Another method that’s sometimes important, especially where the ma¬
terial has not been broken down into paragraphs that correspond to points
in an argument, consists in identifying the separate strands in the argument
yourself, and providing a summary in a few words that will be clear enough
to indicate the passage to which you are referring. Just as in writing out
your original “dictionary” or index, when the claims get long and compli¬
cated, all you need do is give the first word or two and put in a series of dots
followed by the concluding words of the claim, so you can summarize a
whole subargument by giving just a few words. For example, suppose that in
the discussion of busing, which goes on for several pages with hardly any
paragraphing, one section covers the whole question of whether busing leads
to academic gains by the previously less well performing students. You
160 CHAPTER 6

could display this in your index (as item 7, say), which you are going to use
in order to set out the structure of the whole argument, as follows:

7 “It’s clear from the Coleman data that integration is beneficial ... no
alternative explanation.”

Here we are assuming that the passage as it stands, even though it isn’t
presented separately as a distinct paragraph, is at least separated into a
distinct set of sentences that follow one another. Sometimes even that won’t
be the case. Then you should use a summary of the line of argument, per¬
haps picked out from several different paragraphs or pages in the original:

7 The argument from the Coleman data that integration benefits the academi¬
cally weakest of the populations that are integrated.

Here you’ve used a description, but you could have used a brief summary
(sometimes called a synopsis or paraphrase or transliteration) of the original
argument. In your structure diagram, you will show this as one of the prem¬
ises from which some conclusion, perhaps only an intermediate conclusion,
is drawn. Then it’ll be easy for your reader, or for yourself when reflecting
on the analysis, to see just what you are talking about. It won’t be quite so
easy to avoid making errors when you do it this way, because you’ve done a
great deal of compression of your own already, and there may be errors in
that. So you have to be particularly careful not to dismiss as satisfactory
something that may conceal a number of buried errors, and not to complain
about a line of argument which is, in fact, quite well presented in the origi¬
nal, although your summary doesn’t do it justice.
There’s no reason why this kind of procedure can’t be extended to cover
an evaluation of a whole book. It’s just an extension of “outlining.” One
may use chapters as the “units” for the structure diagram; or one may use
sections or parts of chapters, indicated either by their beginning and ending
or, where the situation is more complicated, by your own summary of the
substance of the argument that’s being covered. Not all books are devoted to
a series of arguments, but a number of the most important in the English
language have that characteristic. It certainly applies to a great many, per¬
haps the great majority, of articles in professional journals, as well as articles
in serious periodicals. So it’s important to be very flexible about the way in
which you set up the units for your structure analysis. Don’t hesitate to use
a mixture of the preceding devices. Don’t hesitate to invent your own. But
the test is always the same test; can the reader, someone who can’t rely on
telepathy to work out what you mean, tell quickly and precisely just what
your labels refer to, and hence what your structure diagram says?
When we get into arguments which are not given in their entirety, that
is, where we have to identify a whole series of assumptions and then tie them
THE FINER POINTS OF ARGUMENT ANALYSIS 161

into the argument itself, these procedures will stand us in good stead. But
they do need to be reconsidered in the light of that special situation. Should
you perhaps distinguish the labels for assumptions or implications from the
labels for what was present in the original argument? Often this is a good
idea. One way to do it is to put parentheses around the numbers that you
allot to material that you have inferred from the original writing or speech,
and to use unparenthesized numerals for material that was present origi¬
nally, or, as we've done earlier, to use letters for assumptions. Another pro¬
cedure is to run through all the numbers that you need to label what was
there in the first place before you start using numbers for assumptions and
implications, so the higher numbers refer to “reconstructed” material. Yet
another procedure is to use letters for both the assumptions and the implica¬
tions, and arable numbers for the material that was originally present. It’s
usually a good idea to employ and state one of these conventions, so that
your reader or you yourself can more readily understand the structure dia¬
gram. When you are filling in implications and assumptions, you have to
write them out in full in your “index” because every single word in them
represents the results of quite a reasoning procedure on your own part, and
they are very open to criticism as being too strong or too weak, not really
built into the original argument at all, and so on.
Finally, it’s worth commenting on certain peculiar forms of argument,
ones that don’t follow the natural kind of tree diagram we have been using
so far. Sometimes, for example, arguers will produce an argument which
they are not endorsing at all but wish to ridicule. They might be simply
quoting an argument from a political opponent which they are going to
refute, or they might be examining a speculation to see what difficulties it
will run into after they have developed some of its conclusions. You have to
watch how to fit these into a structure very carefully. You cannot do it by
simply putting it in as part of the trunk of the tree in the tree diagram. You
must set it off as a separate little tree on the side, leading to an implied
conclusion which you usually have to add; then you can draw a line back to
the main trunk from that implied conclusion. Here's an example:
One of the most attractive lines of argument that the Democrats have used in
order to justify support for a Democratic candidate for President in 1976 is the
unfortunate affair of Watergate. But what guarantee do we have that such an
event would not have occurred under a Democratic administration? Looking
back over the track record of Democratic administrations of the past, it is easy
to point to example after example of corruption, of political misjudgment, of
impropriety and technical breach of the law. This, like other arguments that
they have produced, can’t really be regarded as having any real significance

What’s involved here is the rejection of an argument, not the direct addition
of a further direct argument for the main conclusion (which is, of course,
that we should vote for a non-Democrat for President in 1976).
162 CHAPTER 6

One way of looking at this rather common kind of argument involves

seeing that it consists in heading off an attack that we would normally take
account of when we turn, in the later stages of argument analysis, to con¬
sider alternative arguments. That’s the point where we stand back and look
to see whether we have failed to consider some argument that points in
exactly the opposite direction, or some argument that is even more powerful
than the one we have rejected and which points in the same direction. The
arguer in this particular case is forestalling such a step by considering in
advance one of the apparent arguments in the other direction.
How can one handle this seemingly contradictory argument? From
what we have just said, it is clear that there are two ways to handle it. You
can either pull it out of the whole stream of argument and analyze the rest,
turning to this only when you get to the point in your argument analysis
when it’s time to turn to other arguments that bear on the same conclusion;
or you can treat it at more or less the time it comes up, but do so by setting
it off to one side. It can have its own little tree structure, alongside the big
tree of the main argument structure. And the conclusion to which the little
tree leads is, “The Democrats are unlikely to be any better with respect to
Watergate-type occurrences.” With the addition of the obviously plausible
assumption that either the Democrats or the Republicans are going to win,
one can conclude, “Voting Republican should not be ruled out because of
Watergate-type considerations,” and that conclusion can be fed directly into
the conclusion of the main argument, that one should vote Republican. In
short, if you are very careful about how to phrase the “missing links,” and
very careful to separate the “apparent counterarguments” and their rebuttal
from the main stream of an argument for a particular conclusion, they can
be handled well enough. But you need to watch for them. Once you are used
to spotting them, you’ll see that what they really amount to is a kind of
switch of roles by the arguer from the role of arguing for something to the
role of arguing against the arguments against it. Both are perfectly legitimate
ways to proceed toward a conclusion, but you must of course note that the
negative approach doesn’t leave you with any positive reasons for the con¬
clusion—the argument will ultimately depend on whether there are indepen¬
dent positive reasons for voting Republican, since the counterargument
shows only that you won't be any better off by voting Democratic (in this
respect, at least).

6-3 SIGNIFICANT AND INSIGNIFICANT ASSUMPTIONS

You already know a good deal about how to identify assumptions. The
crucial juggling act that you have to perform is to find something that’s
strong enough to patch up the holes in the argument, but not any stronger
than it has to be, since otherwise it will be open to objections that the arguer
THE FINER POINTS OF ARGUMENT ANALYSIS 163

isn’t really saddled with. In the traditional kind of logic-book argument,

where we give you an argument from the premise, “All warm-blooded Aus¬
tralian animals are marsupials,” to the conclusion, “All warm-blooded Aus¬
tralian animals carry their young in a pouch,” it’s pretty easy for you to
work out that the missing premise is the claim that “All marsupials carry
their young in a pouch." But when we get to real-life arguments, the whole
problem becomes a great deal trickier. In this section, we want to try to set
out some of the traps to avoid, and to give you enough suggestions by means
of going over several examples, so that you'll begin to get a pretty good start
on refining your skills in “assumption hunting.” It’s helpful to begin by
understanding that even the best analysts, people who are really experienced
in the criticism of arguments—for example, lawyers with many years of
experience, or philosophers or social scientists analyzing published articles—
will often make errors in their identification of the assumptions that lie
behind an argument they’re criticizing. This is so even when they have as
much time as they need in order to work the matter out and check it out.
When they’re under pressure, as in the courtroom or in the course of a
debate at a professional society’s convention, their frequency of error mak¬
ing goes up seriously. But it’s also worth remembering that these people,
when they get together in groups and jointly study an argument, can reach
very good agreement about what the assumptions are. So identifying as¬
sumptions isn't really an arbitrary matter, it’s just a difficult matter. And of
course, it’s a very important matter, since if you make a mistake, you finish
up wasting many of your criticisms, because you’re attacking a “straw
man”; that is, either you are saddling the arguer with an assumption that he
or she did not need and did not make, so can’t be criticized for making it, or,
alternatively, you are accepting an argument because the assumption that
you think it rests on is beyond criticism, whereas in fact, the argument
actually requires a further assumption that is very dubious. So the matter is
one of the three most important steps in argument analysis (the others being
the clarification of meaning and the overall assessment of the merits of the
argument).
The first trap to avoid is that of quoting long lists of absolutely trivial
assumptions. The title of this section is “Significant and Insignificant As¬
sumptions.” You have to learn not to bother to point out insignificant ones,
but to focus only on the significant ones. There’s no simple rule for making
the distinction; once more, it’s a matter of developing good judgment. But
here are some specific types of triviality to avoid: First, it is entirely unhelp¬
ful to point out that a particular argument “assumes” that its premises imply
its conclusion. Since the arguer is offering these very premises as support for
this very conclusion, it’s certainly true he or she believes that one implies the
other: but it’s so obvious that you’re not doing argument analysis by men¬
tioning this, you’re only repeating the argument. Putting it another way,
164 CHAPTER 6

what kind of useful criticism could you possibly get from mentioning this
assumption? Criticizing it is just the same as criticizing the argument itself,
so you haven’t done any clarification by bringing it in.
Another version of this trivial kind of assumption is often produced by
beginners, particularly with respect to statistical arguments. “This argument
assumes,” they say, “that there’s some kind of connection (or correlation)
between the things the arguer is talking about in the premises and the things
that are referred to in the conclusion.” Well, you can certainly bet that the
arguer is making that “assumption,” but to say so does little more than
repeat the fact that these premises were offered to support this conclusion,
and hence obviously in the belief that there was some kind of connection
between the two.
Another type of trivial assumption is the “assumption” that the words
in the English language will be understood by the listener in the same way as
they are understood by the arguer. For example, if someone argues from the
fact that Jones is taller than Smith to the conclusion that Smith is shorter
than Jones, one might point out that the arguer “is assuming that ‘shorter’
and ‘taller’ have their usual meanings,” or, to bring in some jargon, “have
such meanings that they function as converse relations.” Certainly true, but
about as interesting as pointing out that somebody who asks a desk clerk
whether the hotel manager is in, is assuming that the desk clerk will not
think the question means that the building is burning down. Your task in
reasoning, and in particular in argument analysis, is to call attention to, and
work on the analysis of, the important, the debatable, the hazardous, and the
hidden points, not the ones that no one could possibly disagree with. An
arguer, when pushing forward some premises for a conclusion, is usually
assuming a great deal about the listener in terms of linguistic knowledge,
capacity to reason, particular knowledge about the issue, and similar capa¬
bilities. General statements to this effect are of no value in argument analy¬
sis. Nor are particular ones about obviously unchallengeable assumptions.
You can see from this discussion that assumption hunting consists mainly in
looking for assumptions that are either obviously assumed but of debatable
truth, or not obviously assumed and hence perhaps overlooked, and also
debatable. In short, you’re searching for a weak link in the chain, not trying
to point out that there are lots of strong links that nobody actually put into
words. Most communication operates on a carrier wave of unstated assump¬
tions, contextual cues, and other things that it would be quite impossible to
express in full.
Let us stress again that trivial assumptions are not confined to vague
and general assertions about arguments; they can be very specific indeed.
Suppose we have an argument about a particular robbery, the premises of
which consist in a recitation of many details about the evidence, and the
THE FINER POINTS OF ARGUMENT ANALYSIS 165

conclusion of which is the claim that a particular suspect was responsible.

One might say of this argument that it assumes that one can infer from this
particular long list of premises to this particular conclusion. It does indeed,
and there’s nothing very general about that assumption, but it’s just as trivial
as any of the ones we've discussed because it’s just another way of saying
that the arguer thinks this is a satisfactory argument, or at least one that
ought to influence you.
Another trivial type of assumption that goes back to an earlier point in
the argument being analyzed—to the premises instead of the inferences—is
one that can be expressed by saying, “the arguer is assuming that there is
some evidence for the premises being put forward.” True enough, since oth¬
erwise there wouldn’t be any point in offering them as “support” for the
conclusion; but again, too trivial.
In general, there’s not much value, except as a preliminary step in
clarifying your thinking, to identifying assumptions of the form; “The ar¬
guer is assuming that there’s some kind of connection between . . . and. . . .”
The crucial question is, Exactly what kind of connection is being assumed?
There are occasionally exceptions to this. If you can prove that there isn’t
any connection between A and B, it’s enough to be able to show that an
argument depends on the assumption that there is some kind of connection.
For example, an argument may rest on the assumption that there’s some
kind of connection between ethnic classifications and honesty; as far as we
know from the evidence available, there is no such connection. Hence, it was
quite enough to be able to state the assumption in these very general terms,
in order to be able to squash the argument. But usually what’s at stake is the
question of the exact kind of connection. For example, an argument may
start with the premise that incidence of cancer in industrial towns is very
much greater than it is in rural areas and move swiftly to the conclusion that
we should crack down on emissions of waste materials by industrial plants.
Such an argument is assuming that there’s some connection between these
emissions and the cancer, indeed, but what’s more to the point is that it’s
assuming that the relation is a causal connection and not a mere correlation.
For only if it’s a causal connection would manipulating the emissions lead to
(and cause) a change in the rate of cancer. Now the premises of course tell us
only that there is a correlation between cancer rates and density of industrial
emissions. In order to show that the connection is stronger than that—strong
enough to uphold the argument—we have to produce a great many other
kinds of evidence about experimental manipulation or controlled studies,
etc. So the argument rests on the assumption of a causal connection between
these two specific factors; not just on the existence of some relationship
between the general conditions of cancer and industrialization. The latter
claim is true enough, but it’s so weak an assumption that it’s an easy one to
166 CHAPTER 6

support, and it doesn’t support the conclusion; whereas the real assumption
behind the argument is a much “stronger” claim and one that is very diffi¬
cult indeed to support.

The “Straw-Man” Problem—Being Fair to the Arguer Remember that

the other side of the problem of finding assumptions, besides finding connec¬
tions that are strong enough to tie the premises to the conclusions, is that of
avoiding the mistake of making them so strong that the arguer doesn’t need
them. Doing so is a typical form of the “straw-man fallacy.”
If it is argued that a particular male candidate is better than a particu¬
lar female candidate for a given job, one can’t claim that the arguer is
assuming that men are always better than women. Getting a fair or accurate
version of the assumption is quite difficult for beginners, and leads to a great
deal of ill feeling in arguments, as well as considerable wasted time.
Remember the argument discussed earlier,“She’s red-haired, so she’s
probably bad-tempered.” The argument consists of a single premise and a
single conclusion—it could hardly be simpler. What is the missing premise?
Let’s review our discussion of it. The usual first reply from a class is, “Red-
haired people are bad-tempered” or “All red-haired people are bad-tem¬
pered,” which are intended to mean the same. That’s quite incorrect, quite
unfair to the argument; it sets up a “straw man.” If the argument had the
conclusion, “She must be bad-tempered,” it would require something more
like this premise. But the argument is much more cautious, and to draw the
conclusion in which “must” is replaced by “probably,” one only needs a
premise referring to most red-haired people. Yet it is still incorrect to say
that this argument assumes that most red-haired people are bad-tempered
(or that a red-haired person is probably bad-tempered). The conclusion con¬
tains a further piece of information which narrows down the required as¬
sumptions; it is the reference to “she.” The arguer may well believe that
most red-haired people are not bad-tempered, only that most red-haired
women are bad-tempered. Here’s where we finally focus on the minimum
plausible claim that’s necessary to make the argument work, and that’s (usu¬
ally) what the assumption is.
The refinement in assumption hunting that we want to remind you of in
this section is contained in that word “plausible” in the definition of an
assumption in the last sentence. Sometimes the interest in plausibility will
make it more sensible to identify a slightly stronger than necessary claim as
the assumption in a particular argument in order to get the “minimum
necessary” claim to connect premises with conclusions. For example, in the
discussion of the red-headed woman, it might be argued that it’s a little more
plausible to suppose that the arguer believes that red-headed people are bad-
tempered than that he or she believes that red-headed women are bad-
tempered; roughly, because there seems to be no evidence about sex differ-
THE FINER POINTS OF ARGUMENT ANALYSIS 167

entiation on this matter, and it might be thought there must be a gene

combination that produces both a tendency to bad temper and redness of
hair.
Suppose that we were to argue, “She’s a homosexual and so she’s prob¬
ably a bad security risk." One could say that the minimum necessary con¬
nection between the premise and the conclusion in this tiny argument is,
“Female homosexuals tend to be bad security risks.’’ However, it might be
more plausible to say that the missing premise is, “Homosexuals tend to be
poor security risks,” because whatever evidence there is about this matter
almost certainly supports the more general claim. A matter of judgment is
involved here. If you were trying to reconstruct the assumptions that the
arguer probably made, you could bet heavily that they are proposing the
broader claim to be true: that all homosexuals are poor security risks. That’s
the direction in which the argument from this point has always gone in the
past, being based on the idea that homosexuals are liable to blackmail. It
didn't make any difference whether they were male or female homosexuals.
On the other hand, if you're analyzing the argument rather than the arguer,
all that the argument requires in the way of an assumption is that female
homosexuals are (tend to be) bad security risks. And it might just happen to
be the case that, even though male homosexuals are no longer very good
territory for blackmailers, lesbians are in fact still much less accepted and
hence more susceptible to blackmail.
So in looking for significant assumptions, one sometimes, especially
when dealing with the arguer in person, may want to look for the assump¬
tions that were most likely made by the arguer, and these may turn out to be
slightly stronger than the minimum necessary connecting claims. But when
you’re dealing with an argument “in the abstract,” that is, out of context,
you usually don’t make these concessions, since what you gain from doing
so—ease of communicating with the arguer who wishes to make a somewhat
stronger claim than the minimum necessary one—is lost by the possibility
that you will finish up creating a “straw man.”
It’s also worth reviewing the possibility of erring in the opposite direc¬
tion in argument analysis. The “straw-man” fallacy involves ascribing a
claim to somebody who doesn’t in fact make that claim. The “iron-man
fallacy,” as we might call it, involves failing to see that someone is vulner¬
able, having made an assumption, when that is the case. For example, it’s
quite common in reading contributions to a discussion in the form of letters
to the editor, whether in a professional genre or in a newspaper, to see a
letter that ends triumphantly with a conclusion like this:

Mr. Jones is therefore quite wrong when he says that there were no above¬
ground nuclear tests during this period.
168 CHAPTER 6

Here we look carefully at the line of argument, probably consisting of docu¬

mentation from official announcements, or perhaps seismograph records, in
order to see whether they support this conclusion. We need to notice very
carefully that what this evidence will support, if the inferences involved are
sound, is the conclusion that there were above-ground experiments during
the period in question. It does not establish or support, by itself, that Mr.
Jones was wrong about anything. For there is an additional assumption
involved in this argument, namely, that Mr. Jones did indeed assert that
there were no such tests. The iron-man fallacy is involved in omitting that
further assumption when doing an analysis of the argument leading to the
above conclusion. What Mr. Jones may have said was that there were said to
have been no above-ground tests during this period, or he may have asserted
that there were no official above-ground tests during this period, or made
some other claim.
Suppose that the conclusion of the argument, instead of beginning “Mr.
Jones is therefore wrong. . . ,” had begun “If Mr. Jones believes. . . , he is
wrong.” Then we don’t need to drag in the extra assumptions. This conclu¬
sion is only a hypothetical conclusion or a conditional conclusion, and it
does not commit the arguer to the view that Mr. Jones did stick his neck out
on this issue.
Now let’s look at a completely different and very interesting range of
cases that call on our judgment in determining what to call an “assumption”
of an argument.
Suppose somebody needs some information about the properties of
right-angled triangles. One way to provide the information is to quote the
axioms of the geometry in question, normally Euclidean geometry, and say
with a flourish, “These contain all the information that you need to know
about right-angled triangles, and a good deal more.” Of course, there is a
sense in which they do, since all the theorems of geometry can be obtained
from them by deduction, that is, are implied by them. Unfortunately, the
knowledge is not being presented in a very useful way. We have not ex¬
tracted or exhibited the relevent and important consequences of the axioms.
They need to know Pythagoras’s theorem, for example, not the basic as¬
sumptions of all geometry. And so it is with respect to the assumptions on
which a particular argument rests. You have to find an illuminating version
of them, and often this means mentioning facts about things not even men¬
tioned in the argument, just as right-angled triangles aren’t even mentioned
m the axioms of Euclidean geometry, though those axioms nevertheless in¬
volve (imply, depend on) claims about right-angled triangles. Assumptions
are one kind of implication of an argument: they are implied premises, not
implied conclusions. Finding the interesting assumptions of an argument is
thus partly a creative act, as is finding the consequence of the axioms of
geometry. Of course, you can always say that the assumption on which an
THE FINER POINTS OF ARGUMENT ANALYSIS 169

argument rests is that one can get this conclusion from these premises, and
there’s nothing false about that; it’s just that it's supremely unhelpful. The
successful argument analyst is the person who manages to uncover and
present in an illuminating and relevant way the interesting and vulnerable
assumptions that an argument makes. It may take a good deal of shuffling
and rephrasing and a substantial slice of original thinking in order to see just
what these are in a particular case. Take, for example, the argument that a
particular candidate for appointment at a high level and in a politically
sensitive branch of the government should be passed over because he is a
homosexual and therefore overly prone to blackmail, and hence too much of
a security risk for the job. This line of argument has been widely accepted in
security work for a very long time in this country and many others. It could
be formulated somewhat as follows:

He is a homosexual.
Hence, he is prone to blackmail.
Hence, he is a security risk.
And, this job requires security clearance.
So, he should not be appointed.

This is an argument of a standard form, and relatively simple. It is really just

a modest extension of an even simpler form, corresponding to the “She’s
red-haired and so she’s probably bad-tempered” example; here, we could
almost compress it to “He’s a homosexual and so he shouldn’t be appointed
to this politically sensitive post.”
Of course, one can say about this argument that “it assumes that homo¬
sexuals are poor security risks.” But there are two things wrong with that
analysis: In the first place, it’s not quite accurate, and in the second place,
it’s trivial. It’s inaccurate, at least for the reason which came up in the
discussion about the assumptions in the red-haired girl case; that is, you
have data here that clearly limit the assumed claim to male homosexuals, so
you’re not entitled to saddle the argument with a claim that refers to homo¬
sexuals in general. It’s also inaccurate because there’s another assumption
involved, since we have an argument that has really two steps to it, rather
than only one as in the case of the red-haired girl. Perhaps you can already
see what the extra complication is in this case. Can you formulate one good
reason that the argument in its compressed form (quoted in the last sentence
of the previous paragraph) does not exactly depend upon the assumption
that male homosexuals are poor security risks? (Treat this as question 2 in
Quiz 6-A.) But let’s get back to the argument as it was originally expressed
and set out in indented form in the last paragraph. There’s a second defect
with the proposed kind of assumption that we’re hunting at the moment. It’s
at too trivial a level. It does little more than repeat what the full form of the
 170 CHAPTER 6

argument says. And that’s not good enough. If somebody said to you that
this argument assumed a certain geographic distribution of homosexuals in
this country, you would at first sight be quite startled by the suggestion. The
argument doesn’t appear to be saying anything about the geographic distri¬
bution of anything. But remember that the axioms of Euclidean geometry
can quite well be said to be “saying ” something rather interesting about
right-angled triangles, namely, that the square on the hypotenuse is equal in
area to the sum of the two squares on the other two sides. At first sight, there
doesn’t seem to be anything in the axioms about right-angle triangles, but
there are consequences in them for all kinds of triangles including right-
angle triangles; they imply Pythagoras’s theorem.
Similarly, the argument depends on certain rather interesting but
nonobvious assumptions. But first let’s pick up a couple of more obvious
assumptions in order to show the difference between umlluminating and
illuminating assumption finding. It is important to notice, in this example,
the extent to which imagination and creativity come into the process of
uncovering assumptions—that is, how different assumption finding is from
some kind of routine calculation—just as they come into the discovery of
new theorems “contained in” seemingly trivial axioms.
As we start picking at the argument we’ve just set out, we start asking
ourselves exactly how and why the arguer thinks it works, making ourselves
state this supposed thinking in detail, sniffing for the difficulties that are
lurking in its structure and that are being successfully bridged—perhaps!
This is particularly hard if we ourselves are somewhat inclined to think that
the argument works. So first one tries to put oneself in the position of a
person with a vested interest in proving the argument wrong, for example,
the candidate who is to be rejected according to the conclusion of the argu¬
ment. Suppose you are that candidate. What could you say against a line of
argument that supposedly shows that you are unfit for the job?
The first thing to pick at is the connection between homosexuality and
blackmail. How is this supposed to work? The idea is that anyone would be
so afraid of being exposed as a homosexual that he or she would either pay
blackmail or reveal secrets, with the first eventually, if not immediately,
leading to the second. Why should someone be afraid of having his or her
homosexual activities revealed? Presumably, there could be two reasons:
First, that the person feels guilty about them, and would suffer humiliation
or shame if others knew of them; and second, that the disapproval of others
might be so severe that the homosexual would be fired if the matter was
revealed.
One of the terms in the expansion of the argument is the word “reveal.”
The moment one begins thinking about the situation of many homosexuals
in real-life positions, one realizes that, in many cases, there is no conceal-
THE FINER POINTS OF ARGUMENT ANALYSIS 171

ment of homosexual commitment. Hence, “revealing” this would not have

any threatening force.
So the first serious problem with the argument is that it assumes the
individual is not already known to be a homosexual. Isn’t this assumption
obviously true with respect to people in the kind of job category we’re talk¬
ing about? It’s not entirely true, and (a quite different point) to the extent
that it is true, it’s true because we make it true; that is, as long as we refuse
to give these jobs to homosexuals, we make it impossible for an overt homo¬
sexual to have such a job, which in turn increases the risk of blackmail with
a covert homosexual who slips through the security check. So, in a way not
unlike the way in which certain drug laws create crime, the ban on homosex¬
uals on grounds of security risk helps create the security risk for homosexu¬
als. In either case, there might also be some deep, underlying reasons that
necessitate, on moral or other grounds, the prohibition of the behavior in
question. But in the absence of these grounds, or as an offset to them, one
must face the self-supporting nature of the combative procedures. To take
another example, if one continues to refuse to appoint women to supervisory
positions on the ground that men aren’t used to taking, or won’t take, orders
for women on the job, one continues to perpetuate both the men’s prejudice
and the absence of women supervisors. So there is one general moral or
“system-type” consideration that counts against accepting the line of argu¬
ment we’re studying. But to return to the more specific objection: It’s clear
than an acknowledged, known homosexual isn’t subject to the force of this
argument at all.
Nor, to be somewhat creative, should the covert homosexual be ruled
out by it, or at least not completely. For it is clear that the person ought to
be given the opportunity to acknowledge homosexuality openly (to “come
out of the closet”), thereby removing the sting from the blackmail and get¬
ting the job (always assuming there are no other overriding reasons why
homosexuals should not get such jobs). So the argument makes the weak
assumption that everyone would rather pass up a job of this type than
acknowledge homosexuality. And that’s something that wasn’t obviously as¬
sumed.
One practical procedure in argument analysis, of particular importance
in moral contexts, consists in insisting that the argument be conducted in
terms of realistic alternatives instead of on an abstract level of approval or
condemnation. So the crucial question in the present argument is not
whether homosexuals ever have some risk of being blackmailed, which is
undoubtedly true in some circumstances, but whether they always (or nearly
always) have a greater risk than heterosexuals. If not, you can’t hold their
sex preferences to be grounds for disbarring their candidacy. With respect to
their nonsexual life, presumably the situation is comparable. With respect to
their sexual life, it’s clear that heterosexuals run certain special risks, albeit
172 CHAPTER 6

different ones from those of homosexuals. In particular, they run the risk of
being exposed for the commission of adultery, “swinging,” the use of prosti¬
tutes, and similar practices. Once one sees that the key question is whether
this particular homosexual is a worse security risk than the competing het¬
erosexual candidates, one begins to see some other assumptions that lie
behind the argument we are considering. For there are some communities
where exposure as a homosexual would have very little impact, whereas, in
others, it would have considerable impact; and the same is true of exposure
for heterosexual “sins.” Remember that we’re not talking about the question
of whether exposure would destroy somebody’s chances for being elected in
a close political campaign, for there is probably a sufficiently large swinging
vote in most communities to produce that effect. We’re talking about
whether there is such a powerful “straight” lobby and public attitude that a
man who had a secure federal job could be removed from it for this reason.
And that might be impossible even though a significant minority, or even a
majority, of voters felt that he should be. The courts do offer significant
defense in such cases.
As one comes to think in detail about the extent to which, and the way
in which, the exposure of a homosexual constitutes a sufficiently threatening
possibility so that the person would be willing to submit to blackmail rather
than hazard it, one begins to see the sense in which the whole argument
depends upon certain geographic assumptions. For it seems clear that the
proportion of adult homosexuals who find their way to certain urban areas
where there is a greater tolerance of a social life for homosexuals is larger
than the proportion of the corresponding “risk” group of heterosexuals (ac¬
tual or potential), namely, adulterers or those who (do or may) engage in
premarital or purchased sex, who move to cities. So the argument will work,
in general, only if the proportion of homosexuals in communities which
would be powerful enough to have them removed from their jobs is greater
than the proportion of heterosexuals in communities which would react just
as unfavorably to heterosexual “misdemeanors.” And it isn’t at all clear that
the reverse is not the case. So the argument, in its general form, does rest
upon an assumption about the relative geographical concentration of homo¬
sexuals and those heterosexuals who are either engaged in, or likely to be
engaged in, heterosexual “misdemeanors” (which covers a good proportion).
That’s an interesting discovery about the argument and a real weakness in it.
To take a specific case, it’s really absurd to suggest that the possibility of
blackmailing even a covert homosexual in San Francisco is serious, and
certainly it’s much less serious than the threat of blackmail to a heterosexual
having an interest in playing around and residing in an uptight rural com¬
munity. Of course, as long as a federal agency continues to have rules
against employing homosexuals, homosexuality will be a ground for exclu-
THE FINER POINTS OF ARGUMENT ANALYSIS 173

sion; but those rules cannot themselves be justified by the kind of argument
here.
The best way to understand what is happening when one is looking for
assumptions is to ask why premises are provided when you’re putting for¬
ward an argument. After all, in assumption hunting, you’re trying to recon¬
struct the premises that someone would have put forward if stating the
argument in full. The crucial point about an extra premise, one that we’re
going to add to the ones that are already visible, is that it should bring to
bear new, relevant, and convincing evidence. As you look at those consider¬
ations, you can see the basis for the points we’ve been making, and for a
slight extension of them. The requirement that the extra or “missing” prem¬
ise (which is what an assumption is) be new immediately precludes a mere
repetition of the supposed connection between the given premises and the
required conclusion. There’s nothing new about the statement that what is
already being provided is supposed to give support to what is alleged to be
the conclusion of the argument. So, typically, an assumption should be refer¬
ring to something else that hasn’t been directly mentioned in the given prem¬
ises, and connecting it with some important concept that occurs in the con¬
clusion. It may also refer to some of the concepts in the premises, but it can’t
(in general) refer only to those and the ones in the conclusion, or else it
wouldn’t bring in anything new.
Secondly, the assumption obviously must be relevant to the conclusion
or it won’t support it; and that’s where the requirement of making it “strong
enough” to do the job comes in. Finally, an assumption—indeed, any prem¬
ise—ought to be persuasive or at least not objectionable in and of itself.
Otherwise, it won’t add any weight to the conclusion because it doesn’t have
any weight itself. And that is connected with the requirement that the as¬
sumption not be any stronger than necessary, since the stronger you make
the claim, the less likely it is to be true, other things being equal. One might
say that the further it reaches, the more likely it is to overreach itself.
As we’ve mentioned before, there are times when looking for an as¬
sumption that is convincing will lead the analyst to identify an assumption
that is somewhat stronger than is otherwise necessary. For example, an
argument may only require the assumption that women are incapable of
jumping 32 feet in order to cross a crevice; but it would be a little peculiar to
offer that generalization when it’s well known that no one can jump 32 feet.
So, in this case, offering the stronger generalization would make for a more
plausible reconstruction of the particular argument.
You can see once more how the various considerations involve some
tension with one another; the assumptions you identify mustn’t be too
strong or they will be an unfair reconstruction of the argument, since they
will be fairly easily refuted even though the argument might still be perfectly
sound. On the other hand, the assumptions mustn’t be too weak, or they
174 CHAPTER 6

won’t connect the stated premises to the conclusion. (Or they won’t support
an independent part of the conclusion that has so far received no support.)
The assumption mustn’t be a triviality of definition or fact—since it then
isn’t worth mentioning. Nor can it be a mere assertion of the fact that the
arguer thinks this is a sound argument, since that’s not worth mentioning. It
must be something new, but on the other hand, it must still be true and
relevant.
This all makes pretty good sense in the abstract; let’s have a look at
another example and see what it means in practice. Suppose somebody says:

“The conclusive argument against nuclear power stations is simply that they
might blow up. They should definitely be completely banned.”

What is the speaker assuming? You might suggest that the assumption is:

Everything that might blow up should be banned.

(One of my class students suggested that the assumption is, “Everything that
blows up should be banned,” but that’s a little weak. It’s a bit late to be
banning things after they blow up!) The original assumption is too strong—
it’s unfair to the arguer to suppose he or she is committed to it. The way to
see that it’s too strong is by asking yourself what would be a good counterex¬
ample to the assumption suggested, and also whether somebody might not
still be able to put forward the original argument without being committed
to denying the counterexample. A counterexample to the proposed assump¬
tion would be a steam boiler, or a pressure cooker, or a gasoline tanker or
truck. They all might and do blow up, but it’s somewhat implausible to
suppose that they should all be banned. At any rate, it’s clear that a person
who thinks nuclear power plants should be banned is not committed to
thinking that all these things should be banned; hence the arguer can’t be
said to be assuming that everything that might blow up would be banned.
Now sit back and ask youself what reason somebody might have for
thinking that nuclear power plants should be banned, other than the mere
fact that they might explode. It’s almost certain that what’s disturbing the
arguer is the result, not just the fact, of the explosion, so we might try
something like this as the assumption:

Any installation that might blow up and thereby release vast quantities of
long-lasting atmospheric poison, especially if there might also be substantial
loss of life incurred in the accident itself, should be banned.

Here we have an assumption that is considerably narrower than the original

one and hence is considerably more plausible, because it isn’t vulnerable to
THE FINER POINTS OF ARGUMENT ANALYSIS 175

all those cases that constituted counterexamples to the original proposed

assumption.
Although the new assumption is more plausible, it still contains serious
difficulties. Can you see a new set of counterexamples? It comes awfully
close to covering the manufacture of cigarettes, for example, and automo¬
biles and gasoline. To make it more plausible, one should probably insert
another qualifying phrase, something like “wherever there is a reasonable
alternative.“ That makes the assumption more plausible; but it also means
that you have to add another missing premise to the argument, namely, that
there is a reasonable alternative to generating power by the use of atomic
plants. So we made one premise more plausible, but at the expense of intro¬
ducing another which is quite debatable. But we’ve got a much fairer recon¬
struction, a more plausible argument, and the Principle of Charity requires
us to prefer it. Of course, we still have problems with the fact that there are
reasonable alternatives to smoking cigarettes. Can you get around this?
There remains one further crucial source of difficulty for the argument,
though discussing it is of no special relevance to the problem of assumption
hunting. That is the problem of how you get from the danger to prohibition.
It’s not obvious that everything that’s dangerous, even extremely dangerous,
should be banned, even assuming that there is a reasonable alternative. We
operate on a rather general commitment to the view that people should be
able to go to the devil in their own way, and it’s only in some cases that we
intervene in the name of their own or others’ welfare. That will be a topic for
Section 8-2.
The example we’ve just discussed should show you how the tension
between plausibility and overgenerahty, between triviality and utility, works
out in one particular case. But you need to work through many more exam¬
ples than this to acquire the skill thoroughly. Here’s one for you to try right
away (Quiz 6-A, 3): Identify the assumptions in the following argument:

The matter is quite simple. The United States government signed a treaty with
the Apache nation, a treaty which has never been revoked, acknowledging their
right to a huge tract of land in the Southwest known as the Chiricahua Reser¬
vation. They then shipped the surrendering tribals to Florida in cattle cars and
opened half that land to white miners, ranchers and settlers. The Apache are
now claiming what is theirs by right and they should receive the land, com¬
pound interest on its value, and damages for the suffering due to its loss.

6-4 COUNTERARGUMENTS AND COUNTEREXAMPLES

If you refer back to the outline of the procedure for argument analysis, given
earlier in this text, you’ll see that we’ve now covered it in some depth up to
176 CHAPTER 6

Step 5, which involves criticism of the premises (including the missing prem¬
ises or assumptions) and criticism of the inferences. We’ve already done a
great deal of this, because many premises, whether explicit or implicit, are
definitions or quasi-definitions, and we have spent a great deal of time on
teaching you how to criticize them.
One generally doesn’t do much about the large group of factual prem¬
ises unless one has relevant, specific, factual information. Obviously, it isn’t
possible for a book on reasoning to impart all the factual information that
you would need in order to criticize all the factual premises of all the argu¬
ments that you’re going to encounter. We’re focusing on the special kinds of
premises that are open to logical criticism, of which the most obvious kinds
are definitional premises or ones that assert an inference to be sound, and
also on the criticism of inferences themselves, that is, the allegedly logical or
reasonable steps from a premise to an intermediate or final conclusion.
It may be worth your while to reread the first pages where a brief
outline of criticism strategy is given. You have, in the intervening pages,
been provided with many further perspectives on the kind of criticism strat¬
egies that are possible—for example, through talking at length about the
difference between sufficient and necessary conditions, and through using
the technique of diagraming overlapping circles, etc., to illustrate overre-
strictive and overinclusive claims. Here we want simply to pick up this gen¬
eral theme of the criticism procedures used in attacking premises and infer¬
ences and to tie it to Step 6 in the outline—the introduction of other relevant
arguments.
The chief procedure for criticizing the kind of premises and inferences
that we can criticize rests upon realizing that they are very commonly gener¬
alizations of one kind or another, or equivalent to such. The premises might,
for example, say something about the morality of affirmative action quo¬
tas—that is, about the practice of requiring that a certain number of people
from minority groups or women should be hired within a specified time. The
assertion that affirmative action quotas are immoral is of course a general¬
ization, indeed a universal, categorical, exceptionless generalization! That is,
it denies the existence of any justification for such quotas. In order to criti¬
cize such a generalization, you naturally begin to think of possible special
cases where such quotas would be justified—for example, where we’re deal¬
ing with relatively unskilled labor, paid at very high rates, the skills being
such that they could be acquired in a few weeks of on-the-job training, and
we’re looking at a labor union with many thousands of members and per¬
haps half a dozen blacks. In a situation like this, given a job market where
there are a large number of people who are qualified, or could easily become
qualified, for such a job, and who are very anxious to obtain it, one might
well conclude that a good case could be made for requiring that certain
modest quotas be met. It this is so, then one has attacked the universal
THE FINER POINTS OF ARGUMENT ANALYSIS 177

generalization which was the premise under consideration. Notice that this
isn’t a case of attacking a definitional premise. There are a good many
premises like this one which involve analvses or interpretations that can be
attacked on philosophical or logical grounds, even though they’re not simply
definitional. We’re all perfectly familiar with the procedure of criticizing
generalizations by pointing out exceptions. That’s the procedure of “coun¬
terexample." Of course, if the generalization is somewhat more guarded—if,
for example, it refers to quotas being generally or normally or usually or
typically improper—a single counterexample isn’t going to refute it. We’ll
have to look at the question of whether there’s a preponderance of counter¬
examples, or perhaps a very large number of important ones, even if not a
majority. (Technically speaking, each one of these should not be called a
counterexample individually; but they form a group which, if large enough,
constitutes the evidence refuting the generalization, and we will sometimes
refer to this group for convenience as a group of counterexamples.)
Another example of a similar kind of premise, in that it involves a
philosophical analysis rather than a mere definition (which would be a lin¬
guistic analysis) or a purely factual claim, is, “The moment that the cells
have united to form an embryo in the womb, at that moment an entity with
moral rights is created.” That is again a universal generalization, and one
way to attack it is to look for a counterexample. There are other ways. One
might look for arguments that point in the other direction. That procedure is
what we call “counterarguing,” as opposed to counterexampling. The quota¬
tion we’ve just given, which is, of course, a key premise in many of the
arguments of the Right-to-Life advocates, i.e., those who oppose a woman’s
unrestricted right to abortion, is not easily refuted by counterexamples. The
nature of the claim is really one that relates to a whole moral system; and
one normally attempts to refute such claims by developing an alternative
approach. In this particular case, for example, the usual line of opposition
consists in arguing that moral rights are not inherent in biological entities as
such, but only in people; and that the embryo at the time of conception is a
long way from being a person. Indeed, so that argument would continue,
personhood involves such things as having a self-concept, being aware of
pain and pleasure, being capable of reflective action, and so on. It’s clear
that the entities that we normally refer to as people with moral rights do
exhibit all these qualities. Are they defining properties or not? Here we’re
into the questions of the definition of the person and the definition of moral
rights, and we may be able to appeal to counterexamples to refute certain
oversimplified definitions of “person” or of “moral right.” But we may also
have to set out a substantial part of a moral theory to support our criticism.
So the procedure of criticism involves both counterexampling and
counterarguing, and the procedure of criticizing a particular controversial
claim may lead one quickly into both these approaches.
178 CHAPTER 6

Let us look again at the procedure for criticizing inferences. An infer¬

ence is a step from a premise or premises to a conclusion or conclusions. It
is “represented by” the claim that the premises are sufficient conditions for
the conclusion or conclusions. Let’s take the case of an inference from a
single premise to a single conclusion. If we represent the assertion made by
the premise by the letter “p” and the conclusion by the letter “q,” the claim
that q follows from p, or can be concluded from p, is the same as the claim
that p implies q or that q is a legitimate inference from p. All these claims
amount to the claim that in all cases where p is true, q will also have to be
true, or will probably be true. That can be seen as a variety of, or analogous
to, a generalization.
Let’s take an actual example. “If we postpone the presently scheduled
1978 emission requirements on automobiles, we’ll save about 40 million
gallons of gasoline per year.” That’s what we usually call a conditional
statement, which asserts that it’s legitimate to infer from the condition (or
antecedent) stated in the first part of it to the conclusion (or consequent)
stated in the last part. It can equally well be regarded as equivalent to a kind
of generalization, one which asserts that relaxing emission controls will
cause a reduction in gas usage, which is just the kind of generalization you
might run into as a straight premise in an argument. As we’ve seen before,
the inferences in an argument often can be reexpressed in a way that makes
them equivalent to a premise, reducing the remaining inferences to very
simple definitional ones. So inferences can often be criticized just like gener¬
alizations or just like definitions. In the automobile-emissions case, of
course, the generalization that relaxing emission controls will reduce gaso¬
line consumption is true only with respect to certain types of technology for
meeting the emissions requirements. The use of stratified-charge engines,
some types of Stirling closed-cycle engines, and a number of other technol¬
ogies makes it possible for a manufacturer to meet the emissions require¬
ment without any cost in gasoline consumption, and some already have.
After all, one wants to remember the basic fact that burning up any compo¬
nents of a fuel generates more of the heat that you’re trying to get out of the
fuel, and thereby produces more energy from it, while at the same time
leaving less in the way of residue, hence reducing emissions. Of course, it’s a
complex technological question whether you can do this efficiently. At the
moment, many of the technologies use up more energy to burn up the resi¬
dues that they produce in the way of heat while doing the burning; so it
costs us gasoline to operate the emissions control systems. But it doesn’t
have to be that way. There are other weaknesses in the inference, that is, in
the generalization; for example, the assumption is being made that people
will not continue their present trend toward more and more economical cars
in a way that is encouraged by the imposition of federal emission controls.
Again, if the electorate feels that the avoidance of smog and conservation of
THE FINER POINTS OF ARGUMENT ANALYSIS 179

oil are both desirable outcomes, and sees that these can best be facilitated by
gradually shifting to increasingly economical automobiles, the relaxation of
emission controls will not produce the kind of saving that is being mentioned
here. And so on.
Take an inference which is supposed to be a strictly logical one, that is,
one that follows directly from definitional considerations. For example, con¬
sider the inference from “Eureka is north of San Francisco, and Coos Bay is
north of Eureka” to the conclusion “Coos Bay is north of San Francisco.” If
one want to attack an inference like this, one has to attack the definitional
rules which are involved in the concepts of geographical direction. And
that’s simply a matter of attacking a definition, which one does by counter-
exampling in the way discussed in the section on meaning.
So it’s easy to see that the counterexampling and counterarguing proce¬
dures cover just about all the attacks one can make on an argument, whether
on its premises, its inferences, its assumptions, or its conclusions.
We need say only a word or two more about counterarguing in order to
have covered the next-to-last step in argument analysis, which is the intro¬
duction of alternative arguments. This last point to be considered in this sec¬
tion concerns the problem of drawing the line between criticism of a given
argument and the volunteering and weighing of further relevant arguments. It’s
clear that if you’re asked simply to criticize an argument, you can just point
out whatever weaknesses in the premises and inferences that occur to you,
and let it go at that. Supposing that you are able to find a number of such
criticisms, you will certainly have shown that there are no good reasons for
believing the conclusion of the argument is amongst those that have so far
been mentioned. But, of course, there may be other reasons for believing it,
or other reasons for disbelieving it, which you are aware of. Should you
mention them? This is really a question of the situation that you are in, the
role that you are playing. With respect to your own views, naturally you
should take these other points into account. With respect to attempting to
help others to formulate a correct view of the matter, you might or might not
want to mention them. It might seem to be personally intrusive, it might be
pedagogically bad (perhaps it’s better to encourage students to discover
these other arguments themselves), or you might be in a special role, such as
that of a lawyer attempting to defend a client against certain charges. Elere
you are required to provide only the best defense, which usually means the
best criticism of the prosecutor’s argument, and it is not appropriate for you
to volunteer further argument or interpretations (or even facts) which may
tend to support the prosecutor’s conclusions.
But I must confess that I think the limited role of “sniping” at argu¬
ments is one that should be accepted only when there are special reasons for
doing so. In general, it seems to me that the criticism of arguments naturally
180 CHAPTER 6

and properly involves considering whether there are other arguments for
and against the conclusion of the argument under consideration.
As a matter of fact, one can take all this a step further—bring it nearer
to home, if you like—by suggesting that one should always conclude the
evaluation of an argument with a period of reflection on your own position
in the light of the argument. You may feel quite pleased with yourself for
having identified some weaknesses in the argument. But does this lead you,
personally, to reject its conclusion? If not, you are in possession, presumably,
of some other argument for the conclusion. Bring it in, make it explicit; it’s
appropriate to do so as part of argument analysis. For example, the argu¬
ment that you’re criticizing has certain weaknesses, but you find yourself
fully convinced that the conclusion is completely false even though you can
demonstrate only that a particular argument for its truth doesn’t work; then
you probably have some other arguments in mind, consciously or uncon¬
sciously, which count against the conclusion. Pull them out and set them
out, and look at them critically; subject them to the same scrutiny that you
have turned on the argument under consideration initially.
So don’t “let go” of an argument until you have forced yourself to
decide what your own situation is with respect to the conclusion. Your
feelings about that conclusion may indicate that you do have other relevant
arguments in mind, and maybe you should mention them. If the argument
fails and you finish up without anything supporting the conclusion, but, on
the other hand, you have no very good reason for thinking the conclusion is
false, you should end with a perfectly open mind about the conclusion. Are
you truly doing that? Be sure you are really pulling out of your head all the
relevant considerations that lie buried in it. Just as your knowledge of the
meaning of terms is usually much more subtle than any simple formula that
you can pull out to define them, so your knowledge of the probability of
conclusions is often considerably more complicated that can be expressed in
a single argument. Try to support your judgment, perhaps in a series of
arguments, perhaps just by giving a series of generalizations that seem to
you to have some bearing on the conclusion, though you’re not quite certain
how to construct an argument out of them. Reasoning is very largely the
conversion of unconscious judgments, feelings, and knowledge into something
more explicit. When these considerations are made more explicit, it’s very
much easier to analyze them or assess them and to decide on their true
worth. That’s why this process of role playing, of putting yourself in the
position of having to decide whether you really and truly believe or reject (or
are perfectly open-minded about) the conclusion, is worth doing: It helps to
force you to make the implicit explicit, the unconscious conscious, and the
covert overt.
Here’s an example to test these skills on. Somebody puts forth the
argument that giving people bad grades is a form of punishment, indeed
THE FINER POINTS OF ARGUMENT ANALYSIS 181

sometimes quite a severe form of punishment as far as their career or schol¬

arship plans are concerned, and that associating education with punishment
is a surefire way to turn people off it. Therefore, the arguer concludes, grades
should be abolished. Assess this argument. (Quiz 6-A, 4)

6-5 REFORMULATIONS OF CLAIMS AND ARGUMENTS

We’ve been talking about providing counterexamples to generalizations, and

counterarguments to arguments. Now it’s time to look at the defense against
such approaches. Suppose that you give an argument for the evils of using
gas-fired clothes dryers, using as your reason the fact that they consume on
the average more natural resources than electric clothes dryers do. Of
course, one of your assumptions is that it’s desirable to slow down on con¬
sumption of fossil fuels, but then this isn’t (it turns out) an assumption which
your critic is challenging. Instead, your adversary points out that what you
say applies to the average or typical gas dryer, but that it doesn’t apply to
all. In particular, the latest generation of gas dryers does not use a pilot light,
and that it is the pilot light which accounts for the higher total cost in fuel of
using gas dryers. The actual cost in fuel of doing one load of drying is
considerably less with the gas dryer (or so your critic says).
If you have no reason to doubt your opponent’s premises, then your
best defense consists in reformulating your conclusion, and perhaps also
your own premises. Instead of arguing that gas dryers should be banned or
that they are bad, you simply circle around the counterexamples that have
been pointed out and reformulate your conclusion so that it doesn't cover
them. It will now read, “Gas dryers with pilot lights should be banned (or
otherwise curtailed).” Your original generalization was a bit too sweeping,
as the counterargument showed, so you narrow it down.
If you are attracted by the general moral argument that euthanasia is
wrong because it involves the risk of depriving people of many years of life
if a cure is discovered for their condition, somebody might point out that
sometimes the person requesting euthanasia (or for whom euthanasia is re¬
quested) is so old that death would be certain to come very soon from one
cause or another, even if a cure should be discovered for the condition—let
us say cancer—which is the most immediate threat to the person’s life. Of
course, you might meet this argument by saying that even a slight saving of
years of life is worthwhile; but that might be met by the further argument
that some of these individuals are so unwell that life would not be a pleasure
to them even if the cancer could be cured. A natural next move for you is to
fall back to a slightly less strong claim and simply say that you believe
euthanasia is not justified for people who would have a normal life expec¬
tancy of a good many years if the condition which is immediately threaten¬
ing to life could be cured. That’s a case of reformulating a generalization in
182 CHAPTER 6

order to take account of a counterexample to its initial form. The generaliza¬

tion is now a weaker one in the sense that it doesn’t make such a strong
claim; but, of course, it’s a stronger one in the sense that it is now immune
to a certain type of counterexample that was fatal to the original generaliza¬
tion.
In an analogous way to that just described, where we have been refor¬
mulating claims in the light of counterexamples, we can also reformulate
arguments. Of course, reformulating claims means reformulating an argu¬
ment. But there are a great many other ways. For example, one may add
extra premises, delete some premises, get to the conclusion through a differ¬
ent intermediate conclusion, once it has been pointed out that the inference
to the intermediate conclusion that was originally suggested is unsound, and
so on.
In short, the first round of fighting in argument analysis is by no means
the end of the story—at least in many cases. And the moral of that point is
that you need to be very careful about the kind of criticisms you make when
you are criticizing an argument, so that they won’t be easily handled by
some relatively trivial reformulation of a claim or of the argument. When
you’ve made some critical comments on an argument, look ahead to ways in
which they might be met by reformulation. Role-play the arguer. Try to
imagine yourself defending the argument in the light of the critical points
that you have raised. If this can be done, you should certainly take account
of the defensive moves you have just thought up when you’re trying to
decide how to sum up your assessment of the argument.
One of the most important types of reformulation involves clarifying
and sometimes even slightly shifting the meaning of a term. An argument
may begin about whether childbirth at home is as safe as it is in the hospital.
Various considerations are raised, such as the increased probability of pick¬
ing up an infection in the hospital and the increased risk that an emergency
in the home couldn’t be coped with because of the absence of a specialist or
some special drug. At this point, the person who’s arguing for the superiority
of home childbirth says, “Of course, if it wasn't for the absurd prejudice of
doctors against attending child delivery in the home, one could have a spe¬
cialist (instead of a midwife) present, with most of the necessary drugs and
apparatus.” This is an immediate shift in the nature of the argument. It is a
case of reformulating the claim that is at stake. The argument began by
being an argument about the relative safety of home childbirth and hospital
childbirth, but now it’s become an argument about the possible relative
safety of the two approaches if doctors worked in a way different from the
way in which they do currently work. This is a reformulation of the conclu¬
sion, and it might be argued that it really amounts to clarification of what is
meant by the conclusion, rather than an alteration of the claim itself. It isn’t
really worth a big fight to decide which of these is the case: that would be an
 THE FINER POINTS OF ARGUMENT ANALYSIS 183

ego trip. The main thing is for you to be on guard against the possibility that
you are interpreting some term, or some phrase, or even a whole sentence, in
the argument in a way different from the way the arguer now interprets it.
Suppose the argument is about whether it’s sexist for schools to have
different schedules of required exercises in the physical education classes for
girls and boys. The counterargument is produced that girls aren’t as strong
as boys, and it would be unreasonable to set them exercises which demand
equal physical strength. The reply to this is that the only reason why girls
aren t as strong as boys is continuing discrimination, beginning in very early
childhood, of which this discrimination in physical exercise tasks is just one
example. All such discrimination is sexist, it is said, part of the sexist culture,
and hence, it is certainly true that discrimination between the required exer¬
cises is sexist.
Notice what has happened here. The argument has shifted from what
one might call the question of whether any “added sexism” is involved in
setting different exercises to the question of whether setting different exer¬
cises is “part of” the general sexist pattern. There’s been a shift of meaning,
at least as far as the critic of the argument is concerned. It may be that the
critic misunderstood the arguer; or it may be that the arguer, faced with a
certain type of counterargument, reformulated the conclusion in order to
withstand the threat better. The main function of an argument should be to
arrive at the truth, and not so much to score points along the way; so it’s
often best to adjust to the new sense, to treat it as a clarification of the
original position, rather than to claim a partial victory through having
forced a withdrawal from the original position.
It’s pretty clear that arguments about sexism or racism, or the causes of
crime, or artistic brilliance, and so on, are likely to involve some skirmishing
about the exact meaning of these terms, and the skirmishing is likely to
result in at least clarification, but possibly also reformulations. Simply call¬
ing for an explicit definition of the key term is not a particularly helpful
procedure in most cases. It will be useful where the term is a technical one
that admits of a precise definition, but it generally won’t be useful where the
term is a crucial theoretical, philosophical, or ethical term, because its full
meaning is not easily set out in a definition. It will be better to use the
methods of examples and contrasts (discussed earlier) to clarify meaning, in
the course of an exchange between two people. When you yourself are under
attack, it will seem to you that clarification of your original intentions is
taking place; when you are the attacker, it will seem to you that the arguer
is reformulating (evading) rather than clarifying. Whichever it is, the deci¬
sion isn’t very important except perhaps in a debating society. The main
point to remember is that in formulating as well as in reformulating argu¬
ments, a great deal can often be saved even when an important point has
been made against the original form of the argument. One frequently sees
184 CHAPTER 6

people giving up on an argument because somebody has pointed out a very

serious weakness in it, when what they should be doing is making quite sure
that their instincts have nothing more to be said for them. There may per¬
haps be a way to reformulate the argument that salvages an important form
of it. One of the real signs of maturity in people engaging in argument, or in
reasoning in general, is their willingness to offer reformulations to their op¬
ponents when they can see that such reformulations would save an impor¬
tant point for “the other side.”

6-6 STRONG AND WEAK CRITICISM

We have referred throughout this book to the difference between strong and
weak criticism, and to the desirability of ensuring that one learns how to
make that distinction both in the role of defending an argument and in the
role of criticizing it. We have illustrated, on a number of occasions, particu¬
lar ways in which one can distinguish between strong and weak criticisms.
And we have stressed that the idea of a “strong” argument or claim is
confusingly ambiguous. A strong claim may mean one which sticks its neck
out a long way like the claim that white males are more aggressive than
Asian-American males. Or it may mean a claim for which the evidence is
strong, in which sense “strong” could be fairly well translated as “well sup¬
ported.” The stronger a claim is in the first sense, the less likely it is to be
strong in a second sense, other things being equal. You will remember that
one example of this came up in connection with the discussion of precision.
As we make a claim more and more precise, it becomes increasingly infor¬
mative, if true; but it also becomes increasingly difficult to support, requiring
more and more refined methods of measurement. So as it becomes stronger
in one sense, it becomes less likely to be true, other things being equal. That
is, it is less strong in the other sense.
When we talk about strong and weak criticism, the same ambiguity is
present. It’s a strong criticism of an argument to say that it’s “entirely hope¬
less”—much stronger criticism than to say that it “doesn’t seem to be quite
satisfactory.” But this is the sort of strength that refers to the strength of the
claim, rather than the strength of its support. One might equally well say, in
the other sense of “strong,” that a strong criticism of an argument is one that
really turns out to be well supported, and that the second one just quoted is
much more likely to be well supported, since it takes much less to support it.
So it is stronger.
Given the ambiguity in the terms strong and weak, let’s focus on the
type of criticism that is desirable, that is, well supported and incisive criti¬
cism. To use other terms, we want to be careful to produce criticism that
“goes to the heart of the argument” and that succeeds in showing that it
failed there, not in some peripheral way. We’re not particularly interested in
THE FINER POINTS OF ARGUMENT ANALYSIS 185

criticism that is strong only in the sense of merely claiming to utterly destroy
an argument. We need criticism that really undermines it completely. Or, to
be more precise, we need to distinguish carefully between criticism which is
really serious for the argument and criticism which requires only a reformu¬
lation to be taken care of. In short, we want criticism that is as strong as
possible both in the sense of making a strong claim and in the sense of
having strong support.
One of the most common faults in people beginning argument analysis,
or reasoning in general, is a tendency to throw in the weak comments with
the strong ones. Your instructor will frequently remark, perhaps verbally or
perhaps in the margin of your assignment, that a particular criticism or line
of argument is weak, or perhaps that another is strong or very strong. Make
sure that you understand thoroughly the basis for these assessments. They
constitute one of the most valuable kinds of judgment that your instructor
can provide for you because they gradually teach you to discriminate be¬
tween the most useful and the most trivial types of argument, criticism, or
reasoning.

6-7 OVERALL JUDGMENT

The last stage in the procedure that we outlined for argument analysis con¬
sists in tying together everything that has gone before, and summing it up (if
at all possible) with an overall judgment, assessment, or evaluation of the
argument. Just as in grading student work, it’s often much better to point
out that there are certain good features and certain other bad features in a
complicated argument, rather than just assign an overall grade. But this
business of “fractionating” an evaluation can also be used as a cop-out. That
is, a teacher who doesn’t like to be critical will search for “something nice to
say” to a student, and then say that, giving it about as much force as some
rather mildly expressed criticism. The overall impression that this creates in
the student’s mind, when reading (or hearing) these comments, is that he or
she has made some errors but has also done some good things, and that
these more or less balance out. The real situation may be entirely different.
The good point may be a finer point; the weakness may be very serious. It’s
extremely important that instructors call a spade a spade in the evaluation of
reasoning, because the procedure of evaluating reasoning itself exemplifies
reasoning, and if it is done in a misleading way, it commits one of the faults
that it is supposed to correct in the student, a hypocritical situation. Some¬
times one hears that there is no way to compare apples and oranges, and
hence that it’s impossible to give an overall single grade to work which
exemplifies, on the one hand, considerable creativity, but, on the other hand,
significant defects in the validity of the arguments provided. But the process
of overall judgment requires you to weigh apples against oranges. (So does
186 CHAPTER 6

ordinary practical shopping.) Make no mistake about it, that’s exactly the
game that reasoning is obliged to play for excellent practical reasons on
many occasions. When you go out to make a major purchase, perhaps of a
television set, you will have to weigh apples against oranges: you will have
to weigh the fact that a certain set is easier to carry around against the fact
that it has a somewhat less satisfactory picture. You’ll have to weigh the fact
that it’s considerably cheaper against the fact that the place where you’re
buying it is unlikely to provide service if it breaks down. Of course, these are
all apples, oranges, peaches, and pears, incomparable in some sense. But
there’s such a thing as putting them all together in a rational way, rather
than in a stupid or irrational way. If you let price be the absolute determiner
of what you purchase, you will finish up making some very foolish pur¬
chases. If you ignore price entirely, you will finish up being unable to make
many purchases that you would very much like to make solely because you
don’t have any money left, when you could easily have bought what you
wanted and still had money left. Both errors are errors of reasoning.
The process of evaluation is, on the one hand, the most elusive and
difficult of all reasoning activities and, on the other hand, the most impor¬
tant. It is the most important because it is the most closely tied to action.
When we have found out that there are certain weaknesses in a certain
argument, that some of them can be met by reformulation and that some of
them can’t, that some criticisms of some parts are strong and some criticisms
of other parts are weak, we have found out all that we can about the basis
for our final evaluation. But we haven’t made it yet. In order to make it, we
need to “weigh up” all these considerations. That means we have to decide,
and have good reasons for deciding, how much is left after the demolition
work is finished and what it’s worth. Why do we have to make such deci¬
sions? Why can’t we just leave the list of established criticisms and refuted
criticisms as they lie? Simply because we’re going to have to perform actions
on the basis of this argument; risk our money on the basis of this other
argument; risk our reputation on the basis of yet other arguments. When
these things are at stake, you have to decide whether to go with the argu¬
ment on balance, or against it. Merely listing pros and cons doesn’t get you
to “the bottom line,” the real pay off.
The argument may be one about the superiority of a particular life
insurance policy over another, and on your decision hinges the future wel¬
fare and happiness of your children or other relatives. You look carefully
into a number of competing policies and rapidly discover that each of them
has certain advantages and certain disadvantages. You must now go about
the process of weighing these advantages and drawbacks in terms of their
relative importance to you, and then, of course, adding up the “scores” that
each of the policies merit in each of these respects, in order to arrive at a
final conclusion. You go through exactly the same kind of consideration in
THE FINER POINTS OF ARGUMENT ANALYSIS 187

an informal way in making decisions about what to do with a free evening,

or whether to get married or to have kids, or where to go to school or what
courses to take. Making the considerations explicit is a way to make the
whole procedure more reliable, because the apparatus of criticism comes
into play much more easily with respect to explicit argumentation.
You want to buy an automobile? There will be arguments for buying
one of a certain size, type, price, maker and importer, dealer, guarantee, and
performance; and there’ll be arguments for another or for several others,
along these dimensions as well as other dimensions. Then you have to get
down to the business of deciding how important to you the various dimen¬
sions are; then you can combine the factual data with your carefully justified
estimates of importance into an overall judgment. And you can’t avoid com¬
bining them into an overall judgment if you’re finally going to have to pick
one or the other alternative. It's no good complaining that you had to weigh
seating capacity against reliable service and that these aren’t really compara¬
ble qualities.
In the same way, an instructor, when providing an overall grade for a
student, will have to combine a number of different considerations in order
to provide that single letter grade. It isn’t perfect; it doesn’t tell the student
as much, in some ways, as a more detailed description would. But it does
give an overall estimate of quality that is often highly useful, and it tells you
a great deal more than mere fractionated grades would (that is, grades on
each of the dimensions of performance). Similarly, when you evaluate an
argument, in the end you will have to decide whether or not it affects your
degree of belief in the conclusion; and the extent to which it does that is
really the grade that you are giving it. Don’t shirk that obligation, as long as
there is any point to it at all. Make an overall judgment of the worth of an
argument when you finish all your considerations and analysis. Remember
that, even if it doesn’t direct you toward a particular action, it does direct or
redirect you toward a particular belief, and that later actions may depend on
that belief. Reasoning is not just an abstract matter, like analyzing chess
positions; it is the skill that lies behind making good decisions, performing
sensible actions; that is the skill that leads to developing sane and humane
attitudes.

QUIZ 6-A

1 Press Secretary Ziegler once declared certain previous statements by President

Nixon on the subject of Watergate to be “inoperative.” This is a nice example of
an original_“False” is a nice old-fashioned word that does the
same job and doesn’t even have the implication that the individual who uttered
the statements knew them to be false, in which case the correct word would be
188 CHAPTER 6

Answer

Of course the answer here is “he,” and the distinction between lies and false state¬
ments is often a crucial one in arguments. Fundamentalist Christians sometimes
argue that one must believe either that Jesus was a liar, or that He was the Son of
God. This is an example of a false dilemma, that is, a pair of possibilities that do not
exhaust the alternatives. The third possibility is of course that Jesus was mistaken,
but not lying.

2 Consider the argument, “He’s a homosexual so he shouldn’t be appointed to this

politically sensitive post.” Does this argument assume that "homosexuals are
poor security risks”?

Answer

No. Although the arguer might well be making this assumption, you can’t infer that
from the argument. You can infer only that the arguer is making some assumption
about the connection between homosexuality and ineligibility for a politically sensi¬
tive post. And that is not an illuminating assumption. What’s the most likely missing
premise? Perhaps it would be something like, “It would be politically embarrassing
to have it discovered later that he was a homosexual”; together with some other
assumption like, “The degree to which his performance would be superior in this
post would far offset the costs to the party or the country of the embarrassment that
disclosure of his homosexuality would produce.” Probably the next most likely as¬
sumptions would be ones that refer to the possibility of blackmail of homosexuals
and the consequent possibility of a security leak through this individual. If you are
obliged to attempt a serious evaluation of the brief argument quoted out of context
here, you would undertake to create several of these possible lines of defense and to
critique each of them, as we can easily do here. If you can ask the arguer what
assumptions he or she is making, the answer will simplify the analysis considerably.
But, in any case, you must be clear that you cannot make a reliable inference here as
to what the assumptions were in the abstract; the best you can do is to cover some of
the most likely combinations of assumptions.

3 What are the assumptions in the following argument?

The matter is quite simple. The United States government signed a treaty with
the Apache nation which has never been revoked, acknowledging their right to
a huge tract of land in the Southwest known as the Chiricahua reservation.
They then shipped the surrendering tribals to Florida in cattle cars and opened
half that land to white miners, ranchers and settlers. The Apache are now
claiming what is theirs by right and they should receive the land, compound
interest on its value, and damages for the suffering due to its loss.

Answer

The basic point of the argument is simple; a broken contract should be remedied by
(a) restoration of the property which the contract called for turning over, (b) com-
THE FINER POINTS OF ARGUMENT ANALYSIS 189

pound interest on its value for the time that the Apache did not receive benefits from
it, and (c) damages for the suffering due to its loss. The land in question appears to
be half the Chiricahua reservation, rather than all of it. The damages claimed are all
that might be reasonably argued for, since the claims involve no reference to dam¬
ages for inhumane treatment above and beyond deprivation of the reservation. That
part we will put on one side, not even considering it among “other relevant argu¬
ments" since it is not relevant to the conclusions of this argument. (But it is relevant
to action on the general question of damages. If that general question is not going to
be taken up, the damage done by inhumane treatment provides us with another
reason for paying these damages, namely, that they are property compensations for
something else for which no damages are to be paid.)
One of the important facts about this argument is that it’s a highly plausible
one. Indeed, it has achieved a great deal of political support in Native American
circles, so that it must be completed or rebutted in order to avoid political repercus¬
sions that, in some areas particularly, are becoming pretty serious.
It is interesting to note that the law does not accept this argument as it stands,
and the law developed that position without prejudice toward minorities. The law
rejects this kind of approach for the calculation of very old debts due to breach of
contract, and there is some good common sense behind it on this point. (Here we’re
considering other relevant arguments in order to set up a kind of perspective on this
one, and perhaps to help us improve on assessment of the assumptions which we are
about to dig up.) The reason the law doesn’t accept this line of argument is very
simple: The amount of recompense, computed in terms of compound interest, is
certainly disproportionate to the way in which assets appreciate. If a deposit in a
bank, not land, had been stolen from the Apache, and if the thief had been allowing
that deposit to accumulate interest in his or her own name, the court would unhesi¬
tatingly transfer ownership of the bank account to the Apache, when it had judged
the treaty to have been valid and breached invalidly. But the fact is that land cannot
be treated as equivalent to its then cash value in a bank account. It does not appre¬
ciate over long periods at the rate that compound interest increases the size of a bank
deposit.
Now let’s look carefully at the argument for assumptions. There are three com¬
ponents in the conclusion which we should separate. First, there is the claim that the
Apache should receive the land. Second, there is the claim that they should receive
compound interest on its value. And third, there is the claim that they should receive
damages for the suffering due to its loss. Different assumptions are required in order
to reach each of these conclusions. The first assumption is that an initially illegal
transfer should be revoked, no matter what other innocent parties will suffer and no
matter how long it is since the original breach of contract occurred. Notice that we
have introduced a reference to other interested parties, although no such reference
occurs in the original argument. This is an instance of bringing in insights that
illuminate the assumptions. It doesn't require any specialized knowledge of the his¬
tory of the Southwest to see that the argument has already mentioned white miners,
ranchers, and settlers, and that these individuals took up the land by government
authorization; that much is present in the argument. One might adopt a fairly tough
line and argue that they should have seen that the government authorization was
illegal, and that they should have been held responsible as cocriminals. However,
190 CHAPTER 6

whatever one says about them, it is very difficult to argue that their grandchildren,
inheriting established ranches in that area, should be treated as having no right to be
considered because of their criminality. In short, other people who had operated in
what seemed clearly to be a perfectly legal way have developed interest in the land.
One might still argue that they should be recompensed by the government which
originally led them into this situation. But then, of course, that government isn’t
around any more. Should the present government accept all responsibility for the
actions of its predecessors? The argument assumes that it should, or else that the
interests of the present tenants should not be considered. The latter assumption
would be about as cruel as the original deprivation.
Now, should a government be responsible for all the consequences of errors by
its predecessors? One of the reasons that we elect new government officers regularly
is so that each one can make a fresh start. Our government officials are not automati¬
cally entitled to abrogate treaties, but, on the other hand, they’re not automatically
to be saddled with all the sins of their predecessors. Perhaps some compromise is
appropriate, from this point of view.
But, whereas the previous argument is a highly debatable one, there is a serious
assumption present in the passage that we’re analyzing, about which we haven’t
spoken so far. That is the assumption that the Apache have received nothing at all in
the intervening period from the government. But they have of course enjoyed a slice
(though not all) of the so-called advantages of citizenship. They have received pro¬
tection of the defense forces of the nation, some of the protection of the police forces
of the nation (and some of the repression from that source), and, perhaps most
significantly, they have received very substantial support in the dimensions of welfare
and education from the federal government. While it is quite inappropriate to sug¬
gest that these perquisites offset the massive rip-off completely, they do require some
adjustment in the rather simple formula proposed in the argument as compensation
to the tribals for their land.
The assumption that they should receive compound interest on its value has
already been criticized; a more appropriate formula would involve reference to the
present value of the land involved, which would be considerably less. However, the
land as it stands has been substantially improved by fertilization, irrigation, and
other processes, and the value due to those improvements needs to be deducted from
the appreciated value. However, again, the income taken out of the land over a long
period of time, where it was not turned into land improvement, needs to be added
back in, and it may well be appropriate to add interest on that.
The assumption that people who are compensated by transfer of the property
that should originally have been theirs, plus compound interest on its value (even
with the adjustments just suggested), should in addition receive damages for the
suffering due to its loss is a little more complex. Most of the suffering due to its loss
presumably arises from deprivation of income, etc. that would not have existed had
they not been deprived of it. However, that loss has been covered by the preceding
reimbursements. Are there further “damages from the suffering” over and above that
deprivation? One would need to argue rather carefully for this.
The question requests only the identification of assumptions; as usual, only by
looking a little bit ahead of that, to the justification of those assumptions, or their
refutation, are we able to get a fairly clear formulation of the assumptions them-
THE FINER POINTS OF ARGUMENT ANALYSIS 191

selves. In the process of doing this, we have anticipated some but not all of the final
evaluations of premises and the argument as a whole. One special feature deserves
comment: When three conclusions are drawn, as here, it is important to set the
assumptions out so that they represent the fact that the conclusions are not wholly
isolated. You notice that our assumption for the third conclusion (or the third com¬
ponent of the one conclusion) has to be formulated as “It’s appropriate to compen¬
sate people for suffering due to the loss of property through breach of contract, over
and above return of the land and (adjusted) return of the equivalent of the income
from the land." Otherwise, we would have an assumption which would support the
obviously appropriate conclusion that some kind of damages were due, but it would
not reflect the missing link in an argument where other components have already
afforded claims for substantial reparations.
Still interested in the argument as a whole, and where one should finish up on
it? Do question 3 in Quiz 6-C.

4 Giving people grades is a form of punishment, indeed sometimes quite a severe

form of punishment as far as their career or scholarship plans are concerned, and
associating education with punishment is a surefire way to turn people off to it.
Therefore grades should be abolished.

Answer
Let’s do this one briskly in order to show that it isn't always necessary to get into the
more elaborate forms of structure analysis. The meaning of the terms in the argu¬
ment is pretty clear, given that one understands that “punishment” is being used to
cover disappointment. The structure is simple: Bad grades are punishment; punish¬
ing turns people off: hence grades should be abolished.
The first assumption that’s going to cause trouble is the one that bad grades are
the only kind of grades. Since they obviously aren't, the most that can be said for the
argument is that it would support the conclusion that bad grades should be abol¬
ished. Clearly, the current phenomenon of “grade inflation” shows that a great many
instructors have accepted the conclusion that bad grades should be more or less
abolished, without accepting the conclusion that all grades should be abolished. So
the argument is crucially defective in that respect.
The second point of attack is simply to register a complaint about the premise
which asserts that giving people bad grades is a form of punishment. In a technical
sense that identifies punishment with negative reinforcement, that is, with any stimu¬
lus that tends to diminish the frequency of the response to which the reinforcement
was applied, bad grades certainly have that effect. And it is on the basis of that
technical sense that the argument proceeds to infer that people will be turned off as
a result of getting bad grades. But that involves an oversimplifying assumption,
namely, that getting a bad grade in one particular course will (a) turn people off the
whole educational system of which that course is only a small part, and (b) not be
accompanied by other reinforcers, such as positive vocational guidance suggesting a
constructive alternative path to the present career goals or alternative attractive
career goals that will “take the sting out of” the “punishment” of the bad grades.
Since the influence of one bad grade obviously is quite easily overcome by getting
good grades in a number of other courses, to the point where people are not turned
192 CHAPTER 6

off the whole system, the first part of this assumption is unsound, and the second—
though less obviously—is certainly unsound in particular cases. For example, some¬
body may get bad grades in a course which she doesn’t like and wanted to drop, and
be able to persuade her counselor or parents that dropping plans for a medical career
is now necessary; she may therefore see the bad grade as a blessing in disguise, and
not as a punishment at all.
Hence the premises and the assumptions lead to some very serious criticisms of
the argument. Moreover, there is another argument that bears on the same conclu¬
sion which has to be considered. That is the argument that the function of grading is
to inform people, not to provide them with enjoyment, and that education can hardly
produce socially and individually good results if it doesn’t combine an adequate
system of information about the success of students on various subject matters with
the provision of good training in the subject matter. Without the discriminatory
procedure of differential grading, students have no basis for making a discrimination
between more or less promising areas for concentrating their time. Hence, a system
of education which does not provide bad grades where bad grades are appropriate
will not serve the individual or society well. Providing enjoyment does involve elimi¬
nating or reducing punishment; but providing enjoyment is not the principal task of
the system of education. Hence the line of argument presented rests on a further
assumption, which we hadn’t previously noticed, namely, the assumption that the
maximum yields of a system of education will be achieved by abandoning honest and
skilled feedback to the students as to their performance, if that feedback will prove to
be disappointing. That assumption is obviously unsound, and the other relevant
argument that we’re now considering shows it to be unsound; so we have to abandon
the line of argument presented entirely.

QUIZ 6-B

1 “The mere fact that anything exists at all would be inexplicable unless there were
a Creator. No more compelling proof of the existence of God is required than
this, the hallowed cosmological argument.”
Criticize this argument, paying particular attention to Steps 6 and 7 of the proce¬
dure.

2 “No one should blame the ‘addict’ for his or her addictive behavior. For everyone
alive has bad habits—eats too much, watches the Late Show too often, is a
fingernail biter, avoids exercise—and these are just as irrational and as persistent
as those labeled “addictions,” for example, drinking or smoking too much. In
fact, the relatively recent identification of gambling addicts and “workaholics”
just shows how arbitrary the distinction is. These behavior rigidities are just part
of the human condition, not “sins.” They require understanding, not sitting in
judgment.”

3 a Is there an afterlife? List some facts that seem to support your answer to this
question. Can you construct a strong argument for your position?
b Construct the best argument you can for the (or a) position you do not agree
with as to the existence of an afterlife.
THE FINER POINTS OF ARGUMENT ANALYSIS 193

c Give one strong criticism of the argument you have just constructed,
d Reformulate the argument to handle the criticism, conceding as little as possi¬
ble.

4 Why do you think the questions in this quiz, near the end of the text, are not
multiple choice?

5 At the end of each description of a tour in a current American Airlines travel ad,
there’s a sentence much like this: “For further information, call your travel agent
(or us) and ask for details on tour 637AA721WAD28704.”
a In terms of what we have said in the scope of this book, can any criticism be
made of the quoted sentence?
b What criticism?
c How can it be justified?
d Is it an example of strong or weak criticism?
e Is your answer to d a strong or a weak evaluation?

6 Give the best arguments you can for getting a college education and for not doing
so. Evaluate them and justify your evaluation. Are you biased on this issue?
Might you be? Should that affect your evaluation of the arguments? How?

7 CHARLES McCABE
HIMSELF
RETHINKING RAPE

The thing to do about rape, in my opinion, is to take the sex out of it.
I am not being frivolous. The crime that we call rape—having sexual
intercourse with a woman forcibly and without her consent—would be much
better handled by society if it was treated, not as a sex offense, but as what it
truly is—a heinous and aggravated form of personal assault.
As such, the punishment should be properly grave for the convicted, like
maybe 15 years in the cooler.
Thanks to the feminist movement, rape is a big thing these days. The
recent trial of Inez Garcia brought out some pretty weird feelings about the
whole subject.
Rape is said to be the country’s fastest growing violent crime. Convictions
for it are rising about 15 per cent each year, due largely to the fact that women
are more willing to report the crime than formerly, and to take the police
hassling that often goes with it.
Police understandably don’t like rape cases. One of the reasons is the
woman scorned. She will, and far too often does, report a rape when the truth
is that she was more than willing. She is advancing the charge as a kind of
revenge.
Then, too, the rape case too often resolves itself into a case of one man’s
word against one woman’s word, and is thus ultimately unresolvable.
194 CHAPTER 6

But what is really wrong with our rape laws, from the point of view of the
victim as well as that of law enforcement, is that they are all tied up with sex,
and our curious way of handling sex crimes under the criminal sanction.
Morals and shame and all kinds of irrelevant nonsense are contained in
the word rape. If a raped woman were to come to the police with a complaint
for aggravated personal assault, the matter could be seen as straight police
work. It would be handled like other cases of personal assault which are also
usually not easy to prosecute.
If, however, the rape should be followed by an unwanted pregnancy, and
perhaps an equally unwanted abortion, the crime would be treated with all the
more gravity.
Inez Garcia went to trial for shooting to death a 300-pound man who held
her down while, she charged, another raped her. Her defense was based on the
fact that a woman should have the same right as a man to resort to violence to
protect her honor.
She blew her case in mid-trial when she rushed to the bench and shouted
to the judge, “I killed that mother f——, so why don’t you just find me guilty
and put me in jail, you lousy pig.” She was later sentenced to life imprisonment
for murder.
Charles Garry, Mrs. Garcia’s lawyer, later told columnist Shana Alex¬
ander: “I’ve never seen a woman turn on her own case like that. She scared that
jury silly. At first I didn’t understand why she wouldn’t tell the court she was
sorry, since she’d already told the prosecution psychiatrist God would punish
her. Now I realize she was playing to the gallery. She didn’t want to let her
supporters down.”
Rape is too important a thing to be left to the extreme feminists. For them
it is just another weapon in the war between the sexes which is their stock in
trade.
If instead of taking to her gun Mrs. Garcia had complained of grievous
personal assault, and if, and this is the really big if, police attitudes on the
subject had been de-sexualized, justice might have been served, and served
fully, without the murder for which she was convicted.
I realize that this is rather Utopian thinking. But if you think the unthink¬
able long enough it often becomes thinkable, like the near-impeachment of Mr.
Nixon. Our rape laws just don’t work, like nearly all of our sex laws. This is one
case where we might solve a problem, merely by changing its name.

QUIZ 6-C

1 Is there a “reasonable alternative” to smoking, and what is it?

a For adolescents
b For smokers
What is your definition of “reasonable alternative” as revealed in your an¬
swers to a and b? Does it imply that you should give up driving a car? Having
sex? Why not?
THE FINER POINTS OF ARGUMENT ANALYSIS 195

2 “The very notion that it’s appropriate to apply one single letter, a grade, on the
end of a 30-page term paper, representing perhaps 120 hours of work, is absurd.
And that’s what grading is, absurd. ” Comment.

3 What reparations are due the Apache? (Complete the analysis of question 6-A,
3.)
 Chapter 7

Special Types of
Argument

We have now covered, first briefly and then in considerable detail, the basic
procedures of argument analysis and their ramifications into the criticism of
functional prose and, in general, all prose when the perspective of appraisal
is other than aesthetic. We conclude the book with two chapters in which we
indicate some of the ways in which the basic material applies to some special
cases (Chapter 7) and ties into certain other subjects that should be thought
of as natural continuations of one’s thinking about the processes and appli¬
cations of reasoning (Chapter 8).

7-1 SAMPLES AND GENERALIZATIONS: SCIENCE

Reasoning in science is in fact of a great many kinds. If you read through a

textbook on theoretical physics, you will find most of the reasoning to be of
the same kind as that in any mathematics text; that is, it exemplifies what is
technically called deductive argument, argument in which each step follows
from the previous one by the application either of definitions or rules, or of

196
SPECIAL TYPES OF ARGUMENT 197

more complex principles that have themselves been developed from defini¬
tional rules.
Then again, another substantial part of scientific reasoning in almost
any field involves the evaluation of competing hypotheses. This is done by a
complex process of reasoning, one that involves the derivation of conse¬
quences of the hypotheses and the comparison of these consequences with
known facts (or, first, the investigation of reality to discover the relevant
facts); the reformulation of the hypotheses in order to generate more nearly
exact consequences (if this is necessary); the rejection of the hypotheses if
the consequences are too grossly incompatible with the facts; attempts to
refine, by simplification and by relating to other areas of scientific knowl¬
edge, the basic tenets of the hypotheses, and so on and so on. Certain aspects
of this process have been called “hypothetico-deductive” reasoning. But the
process as a whole is a good deal more complicated than that particular view
of it suggests. Notice that it includes the “reformulation in order to bypass
counterexamples” step that we’ve already covered; and the identification of
implications and assumptions also comes in. Scientific reasoning involves
many of the same elements that come into argument analysis. There is,
however, another extremely important type of inference which does charac¬
terize science, the details of which have been developed in the present cen¬
tury to a degree of refinement never previously attained, one which is of
particular importance in the social sciences, which are themselves essentially
a product of the twentieth century. This is the process of reasoning to a
generalization from observations of a limited number of particular cases.
The semitechmcal phrase for it is reasoning from “characteristics of a sam¬
ple” to the “characteristics of the population” from which the sample is
drawn. Some of the governing principles of this type of inference are covered
in a branch of statistics known as inferential statistics. (The contrasting
branch of statistics is descriptive statistics, which concerns the representation
of facts by means of the use of such types of “summarizing statistic” as the
mean, median, and standard deviation.) We’re concerned here, not with the
mathematical discipline of statistics, but with the basic reasoning steps that
underlie it. Even with respect to these, we will be covering a very limited
part of them in this section, but it is close to being the heart of inferential
statistics.
The fundamental type of inference here is well exemplified in the stan¬
dard opinion poll. When the Gallup poll, for example, is to determine the
opinion of the country about a particular candidate for President, it does so
by interviewing between 1,000 and 2,000 citizens. From this sample of the
population, it is possible to draw extremely reliable and rather precise con¬
clusions about the opinion of the population as a whole. It is possible only
because the sample is very carefully chosen. About 200 million Americans
198 CHAPTER 7

are not interviewed at all, yet we can conclude reliably what their opinions
are from the few interviews that we do make. This is a remarkable achieve¬
ment of applied social science, and it is one that, though open to abuse, can
also keep lawmakers more closely in touch with public opinion in a way that
makes possible the better functioning of a democracy. But of course the
opinion poll has immense significance outside the political domain. It can
tell us about the probable success of a new product, when it is used in
market research; it can tell us about the extent of knowledge of basic facts
needed to survive in a modern society, as when it is used in large-scale
educational testing; and so on.
The sample is itself still quite a substantial group of individuals. When
we have finished the interviewing of the sample, we have to boil down the
results so as to obtain a useful description of the opinion of the sample. This
process is part of descriptive statistics; it involves only deductive reasoning.
For example, we may discover that 54.6 percent of those interviewed favor
candidate A over candidate B. Reaching this conclusion is no more than a
merely arithmetical step, though it is part of the process of reasoning. But
now we come to the procedure of generalizing to the population. There is no
trick to this if certain conditions are met about the typicality of the sample.
One simply asserts that approximately 54.6 percent of the population as a
whole prefers candidate A. All the skill here lies in the process of picking the
right sample in order to get the desired degree of approximation. And that is
what we need to look at briefly.
It is convenient to think about the process of identifying a sample which
will enable one to generalize to the population as a whole as involving two
basic requirements. The first requirement is that the sample be representative.
The second requirement is that sample be adequate, i.e., large enough.
The requirement of having an adequately large sample is obvious
enough. A sample of one, in a political opinion poll, would clearly not yield
a basis from which one could generalize to the voting behavior of the popu¬
lation of the United States as a whole, since a single person would presum¬
ably be voting either for candidate A or for candidate B, and you can be
sure that 100 percent of the population will not be voting for one of these
candidates. A sample of two or three or four is going to be just about as
unsatisfactory. A sample of 100 percent of the population will of course give
us the right answer, but it will destroy the whole value of surveys, namely,
that they cost less than the election itself. Do we need a 50 percent sample?
Without previous knowledge, one would guess that a 50 percent sample
would be pretty safe. Could we get away with a 25 percent sample? Here is
where one’s intuition about the matter begins to weaken. The suggestion
that we could do a good job with a sample of only 1 percent of the popula¬
tion would be pretty implausible if we didn’t have the enormous body of
successful predictions behind us that have resulted from modern survey
SPECIAL TYPES OF ARGUMENT 199

techniques. Still. 1 percent of a population of 200 million is still 2 million

people, and interviewing that many is going to be a highly expensive busi¬
ness. The situation today is that very good, although not the best possible,
results can be obtained with a sample representing Viooo of 1 percent of the
population. It’s obvious enough that the whole trick lies in the procedure
whereby the sample is picked, and we’ll turn to that in a moment. First let’s
get an idea of the general level of accuracy (or approximation) that we’re
talking about. This will depend upon a great many factors, but, roughly
speaking, if we go from a sample of 2,000 to a sample of 4,000, we will not
be more than a very few percentage points more accurate (that is, if we
predict 54.6 percent will vote for candidate A from a sample of 2.000, we
might, although it’s quite unlikely, find that the sample of 4,000, equally
carefully chosen, would lead us to correct this to 56.8 percent). It would be
more typical for the change to be only a few fractions of 1 percent. If we go
to a sample of 6,000 or 8,000 or 10,000, we’ll virtually never pick up an
improvement as large as 1 percent, more usually, it will be 0.1 percent, and
obviously this is very rarely worthwhile. So it’s a measure of the extreme
efficiency of the sample selection procedures that they will get us so close
with such a tiny sample, and we have to bear in mind that even if we took a
sample 10 times as large, we could still be off substantially as far as predict¬
ing an election is concerned because of events that occur between the time of
the polling and the election. Notice that there is a two-stage inference in¬
volved in political polls. First, there’s the inference from the sample to a
conclusion about which candidates people in general prefer at the time the
sample is taken. Then, there’s the prediction that that percentage of prefer¬
ence will hold up until the election. The second prediction is much chancier,
and errors in it are really not something one can blame on the polling
organization when, for example, a late-entering candidate begins to catch up
very fast, or a scandal breaks. In any case, the kind of bet that’s involved
there is not the type of inference that we’re concerned with at the moment;
it’s not an inference from a sample to a population, but an inference from
the population’s preferences at one time to its preferences at another time,
which obviously doesn’t depend on statistical considerations alone but on
political stability and the emergence of new factors.
How do we go about selecting a sample in such a way that a size of
Viooo of 1 percent of the population is adequate? This is where we get to the
question of the representativeness of the sample. There’s no abstract rule that
will answer this question for us. A sample may be representative with re¬
spect to political behavior, but completely unrepresentative with respect to
marketing behavior. This is why George Gallup can command a rather
substantial fee for his services: he knows what principles to use in selecting
a sample for particular types of generalizations.
200 CHAPTER 7

Basically, you have to start off by determining what factors affect politi¬
cal preferences. For example, it’s well known that people with very large
incomes tend to vote Republican rather than Democratic, so it’s clear that
your sample will have to include a proportion of upper-income people as
well as lower-income people, or else it will be subject to what’s called bias
for picking from a group which doesn’t have typical political preferences.
Similarly, it’s known that many more union members will tend to vote
Democratic than nonunion members. So you have to be sure that your
sample contains some of both. How many of each? The same proportion of
each as the population contains. If 35 percent of the population (in this case,
the population of voting adults is the one that you’re concerned with) are
union members, you take care that 35 percent of your sample will contain
union members. Since men and women will quite often have different re¬
sponse patterns, especially but not only if a woman candidate is running, it’s
important to ensure that each 50 percent of your sample represents one or
the other sex. Actually, it would be more precise to ensure that your sample
has about the same proportion of each sex in it as actually vote; this might
turn out to be a larger proportion of men than women in some cases, and
possibly the reverse in other cases. Again, where issues that affect children
are at stake, as is often true, a sample which represents the population’s
distribution between parents and nonparents is likely to be required.
Now you can begin to see that the selection of the sample will be very
severely constrained, if it is to be a representative, yet small, sample. To tell
if it is of adequate size, one will need other information of a purely empirical
kind about the accuracy that can be achieved with samples of a certain size.
If we have to meet twenty or thirty requirements on the sample in order for
it to be representative with respect to the differences among people that
affect their voting behavior, we may well have to go to a sample of 1,000 or
so in order to meet all these conditions. (Obviously, you can’t have a sample
of ten people that will meet even the conditions that we’ve already de¬
scribed.) So the pressure of your knowledge about the number of (indepen¬
dent) factors that influence voting preferences will drive the sample size
upward, and your interest in accuracy will of course also tend in the same
direction. The great compromise that has to be made by the pollster is to
decide on the point of diminishing returns, that is, the point at which further
increases in the sample size will only produce improvements in accuracy that
are insignificant for the purposes of the poll. If the poll is done (as many are)
in secret, to advise a candidate of his or her chances, the results need be
accurate only to within about 5 percent, or perhaps, at the beginning of a
campaign, even 10 percent, since a serious effort in the campaign can swing
the results that far without too much difficulty. If the poll is for national
publication, and if the reputation of the pollster is at stake, it will be neces-
SPECIAL TYPES OF ARGUMENT 201

sary to go to a rather larger sample, since a swing of only 1 percent or so will

quite often decide a major election.
The Nielsen ratings, which determine the fate of most of the shows on
television, at least the shows which represent the most substantial invest¬
ment by the networks, are based upon the reactions of a very small sample
of families; but, once again, we have empirical evidence that using a larger
sample doesn't lead to significant changes in the results. Of course, it’s an¬
other question entirely whether one should let the fate of a television show
be determined by the number of families that are watching it. Since one is
attempting to advise the advertisers who foot the bill for the show, the
crucial relevant information would be the number of people who are led to
purchase products by the advertising that occurs in the commercial breaks
during the show. But we have so little empirical evidence about this that we
fall back on the very crude measure of “total number of viewers,” on the
assumption that any viewer is equally likely to make a purchase of whatever
is being advertised. That assumption is surely incorrect, but we don't know
the direction in which it’s incorrect. Sometimes people make guesses that
certain types of product would be more appropriately sold to some types of
viewer, the ones who are viewing a show of a particular kind, such as the
broadcast of a football game, and they operate on these assumptions. But
these are not really well based empirically. They are simply plausible bets.
The television rating game is much more seldom based on scientific evidence
than is the public opinion poll.
How does one tell whether a sample is representative? Obviously, from
the preceding discussion, one needs to know a great deal about factors that
influence whatever it is that one’s trying to generalize about (which is nor¬
mally called the criterion variable). We all have a great deal of knowledge of
a general kind about what kinds of factors affect what kinds of behavior,
and consequently, if we use our imagination, we’re usually in a position to
apply at least some basic tests to the sample that’s being used in order to see
whether it meets the standard of being representative. It’s more difficult to
judge whether it’s adequate; there are technical statistical considerations that
come into this, as well as a great deal of empirical knowledge. About all one
can do here is to be very suspicious of very small numbers, unless one has
prior “track record” evidence that suggests that the pollster has the required
skill and evidence to make generalizations from very small samples.
An early disaster in the history of political surveys was the Literary
Digest poll. This was a poll done to predict the outcome of a presidential
election, and it was done by taking a substantial random sample of the
names from telephone directories across the country, conducting the inter¬
view by telephone, and summing up the results in the usual way. Before
reading any further, ask yourself whether that sample was a representative
202 CHAPTER 7

sample for the purpose. If you think it was not, write down, as precisely as
you can, your reason for thinking that it wasn’t representative.
The reason that it was not a representative sample is that, particularly
at that time, people who owned telephones were not a representative sample
of the economic strata of the voting population. They tended to be wealthier
than people without telephones, and political voting is significantly affected
by economic classification. As a result, the poll was excessively optimistic
about the fate of the Republican candidate, so much so that its forecast was
entirely wrong, and the failure was so widely known that the Literary Digest
ceased to exist as a magazine.
Let’s take a slightly more subtle case. One of the major polling organi¬
zations determines the sample of people that it wants to interview and then
sends its interviewers out to find these people at their home addresses. If
nobody is home when the pollster calls, the interviewer moves to an alter¬
nate name on the list, the name of somebody drawn from the subpopulation
who has the same general characteristics as the person who’s not at home. A
rival organization has a slightly different approach. If the person selected is
not at home, the interviewer is required to go back twice, at different times,
in order to see if the person may then be home, before moving to an alter¬
nate name. Can you see what the reason might be for adopting the second
procedure, by comparison with the first? If so, write it down before proceed¬
ing any further. (Forcing yourself to write down your reasons here is very
important since it requires you to express them carefully and correctly. It
may well be that you have a vague sense of what’s wrong, but that if you
were to try to express your feeling exactly, you would make a claim that
wasn’t correct. If you had tried to express it, you would notice this; and you
would be forced to think the matter through a little further before settling on
an answer.)
You have to ask yourself, What could explain the fact that somebody is
not at home when the pollster calls? Well, during what hours is the pollster
most likely to call? Interviewers normally work during the daytime, al¬
though they also work during the evenings. People who are most likely to be
away from the house include all the adults in the family who hold jobs.
Although it’s also possible that people are away from the house because
they’re shopping or at a doctor’s or dentist’s office, it is virtually certain that
they will be away if they hold a regular job. If you immediately move to an
alternate choice, you will tend to be picking up an unrepresentatively small
portion of families where both husband and wife hold jobs. Does that mat¬
ter? It does matter, because such families have a stronger tendency to vote
Democratic than other families. You might not know that fact directly, but
you might be able to reason that it’s rather likely, since, for example, the
chances are greater that both members of the family will belong to a union,
and you know that union membership is tied to a tendency to vote Demo-
SPECIAL TYPES OF ARGUMENT 203

cratic. Hence the "return visit” procedure, though more expensive, is likely
to give better results.
Suppose that we want to take a poll on the subject of student opinion
toward having campus security guards carry firearms, and we set ourselves
up to interview people at a convenient campus location, let us say near the
entrance to the campus cafeteria. Now we interview every tenth person who
comes through the door. That looks like a pretty unbiased approach; but is
it? (Treat this as question 7-A, 1.)
In talking about sampling procedures, one will often hear reference to
the desirability of using a “random” sample. We’ll conclude this section by
saying a word or two about the concept of the random sample. If you want
to make a generalization about a whole population, and it might be a popu¬
lation of insects or birds or trees, rather than the population of the United
States, and you don’t know anything about what characteristics of the popu¬
lation affect the property about which you want to generalize, your best bet
is to take a purely random sample. This kind of sample we’ve been talking
about is a hybrid between a purely random sample and an entirely specific,
or nonrandom, sample. Suppose you are going to make a generalization
about the purity of aspirins that are produced by commercial manufacturers.
You are going to pull a sample of the pills from the various production lines
and give these samples a complete chemical analysis. How should you pick
the sample that you pull? A few obvious facts would occur to you; for
example, you might suspect that the first few pills produced when the ma¬
chines start up in the morning may not be very typical, since the chemicals
may tend to concentrate or evaporate overnight, so you’d want to pick your
sample from the main body of the day’s production run (or to pick only a
few from the early part of the morning run). Suppose that you decide, on
statistical grounds, that you ought to look at perhaps one pill in four or five
hundred. Suppose the day’s production is 100,000 pills. That means you
ought to pull a sample of about 200 pills. Should you step in in the middle of
the run and take 200 pills off the moving belt all at once? No, because you
know very little about what affects the purity of the pills. There may be
various factors that depend on the time of day, the operating temperature of
the machinery, the addition of new loads of chemical, shifts in the machine
supervisors and operators, and so on. It will clearly be much better to space
out your sample collection across a day’s production run. It may be better
still to space it out across several days. A good strategy might seem to be to
pick out every five-hundredth pill that comes down the line, since whatever
factors are affecting production, they are unlikely to be precisely timed so
that every five-hundredth pill will have characteristics different from the rest
of the production run. Still, the machinery may make changes every 500 or
5,000 pills. So you will get an even more random sample by using a table of
random numbers to determine time intervals or pill numbers, and then pick
204 CHAPTER 7

whatever pill happens to be passing down the production belt as your timer
or counter indicates the next random number from the table of random
numbers.
If you’re interested in the details of how this might work on a time
basis, you would use a clock that registers in hours, tenths of hours, and
hundredths of hours (instead of the usual hours, minutes, and seconds).
Then you’d pick three digits from your random-number table, say 391, and
you’d pick a sample at hour 3.91—whatever pill was passing by a certain
point on the belt at that moment. Now you pick a second three-digit number
from your random number table, or perhaps just use the next three digits
from the table, and pick a pill from the production line corresponding to
that time, and so on until you have your sample of 200 pills for the day. This
system will certainly give you a sample that is not correlated with any sys¬
tematic defects in the equipment.
If you want to do an experiment on the relative benefits obtained from
the use of different reading programs, you normally have to compare the
achievements of a group of pupils that uses one of these against the achieve¬
ments of a group of pupils that uses another. How do you tell which students
should go into which group? A good procedure is to “randomly assign”
pupils to the two groups. How do you do this? A common procedure is to
alphabetize the names of the pupils and then allocate alternate names to
each of the groups. This assumes, and it’s a very plausible assumption, that
there is unlikely to be any correlation between natural talent for learning
reading and a person’s numerical position on an alphabetized class-order
sheet.
Using a more selective procedure than pure randomization, as done in
the Gallup poll, increases the “efficiency” of the sampling; that is, it means
that you can achieve a generalization with the same degree of reliability
from a smaller sample. The technical name for the procedure of placing
certain constraints on your sample—for example, that it contain about the
same proportion of men and women as the population that you want to
generalize about—is stratification. We talk about stratifying a sample with
reference to sex, socioeconomic status, age, IQ, and other characteristics. In
the experiment on the relative effectiveness of two reading programs, we
would normally stratify the sample by requiring that each group contain an
equal number of boys and girls, roughly an equal number of high, middle,
and low IQ pupils, and so on. The way we would do this in practice would
involve grouping the students under these headings—e.g., group the names
of all the boys together, and the names of all the girls, and then we would
randomly select half the boys to go into the experimental group and half the
girls to go into the same group. We could make this random selection by the
procedure of taking alternate names from the alphabetized list that we’ve
already talked about, or by various other procedures including the use of
SPECIAL TYPES OF ARGUMENT 205

random-number tables. If we carefully “match” the two groups in terms of

these variables (sex, IQ, etc.) that definitely do affect ability to learn from
certain reading materials, we can attach greater reliability to our generaliza¬
tion about the population from which we took the sample; or we can use a
smaller sample and achieve the same degree of reliability in the generaliza¬
tion. This is the use of “stratified random samples” and “control groups,”
two of the most important procedures in scientific method.
We've been talking about inferential statistics; there are also interesting
problems about reasoning with descriptive statistics. Everybody knows that
talking about the “average American” or the “average woman” can conceal
a multitude of misleading conclusions. A more careful study of descriptive
statistics leads us to distinguish between three or four quite different senses
of the term “average,” for example, the types of average that are described
as the mean, the median, and the mode. We won’t go into detail about these
matters here, since most of you will have had some elementary statistics in
one or another social science course, or can achieve a very good understand¬
ing of these points by reading the first chapter in any statistics text. The
discussion of inferential statistics above is somewhat unlike that normally
provided in a statistics text, and it stresses certain similarities to other types
of reasoning that we’ve discussed elsewhere in the book, rather than treat it
all as a mysterious new subject. In particular, we had to set up the discussion
of samples in order that, in Section 7-4, we can talk about causation a little
more adequately. Section 7-4, as well as 7-2 and 7-3, also involves a discus¬
sion of types of reasoning that are very important in science, so we don’t
want to suggest that inference from and to populations is the one key scien¬
tific type of reasoning. But it is a very important type of scientific reasoning.

7-2 INDIVIDUALS AND GENERALIZATIONS:

ETHICS, BIAS, AND PREJUDICE

We have been talking in the previous section about the inference from facts
about individuals to facts about the population of which they are parts. In
this section we’re going to talk about reasoning from facts about a whole
population to conclusions about the individual members of the population.
This type of reasoning has about the same kind of status in ethics as the
previous type of reasoning has in science. That is, it is one important type of
reasoning, and it raises a number of crucial ethical questions; but it isn’t by
any means the only kind of reasoning in ethics, and indeed it represents a
type of reasoning that is very important in science, too.
Suppose you know that women are, in general, less strong physically
than men (taking up again an example we discussed earlier). Suppose that
you are in charge of the selection of drivers for a taxicab company. You
know that there are occasions when some physical strength is required in the
206 CHAPTER 7

performance of the tasks of changing a wheel and moving fairly heavy lug¬
gage for people who can’t lift it themselves, and perhaps, in some cases,
resisting physical assault. Are you entitled to place an advertisement for
applicants in which you require that only men apply? That’s an ethical
question,1 but of course it’s also partly a question about the reverse type of
reasoning from the kind we discussed in the last section. Given the truth of
the generalization about women, what do you really know about the individ¬
uals? Of course, the answer is that you know very little. The ethical point
arises because you really don’t know very much about the individuals. It
stresses the fact that since your action will affect individuals, you must not,
morally or ethically speaking, act in a way that condemns individuals as a
result of their membership in a population which can have certain general
characteristics that they, as individuals, may not share at all. To put the
matter more tangibly, you are certainly not entitled to restrict applicants to
men, since there may well be, and often are, women who are just as capable
in performing the whole range of tasks involved as men, and may indeed be
far more capable than most men. And, more to the point, more capable than
many or all of the male applicants.
It’s important to realize that there are conceivable circumstances where
a generalization about a certain group may be a satisfactory basis for action,
indeed may be a morally sound basis for action. For example, when calling
for volunteers for an especially hazardous mission, the commanding officer
may restrict the volunteers to people with no dependents. Obviously, he’s
working off the generalization that people with no dependents typically have
fewer responsibilities, or fewer serious responsibilities, to other human be¬
ings than those with a family. However, this may not be true in all cases.
The excuse for using it as the basis for action is the lack of time to go into a
detailed history to decide which single people support aged parents or young
siblings, and as a generalization, it’s probably true that people with families
do have slightly more “lives at stake” than single people do. That is the
crucial feature of this situation. First, there is a true and relevant generaliza¬
tion, and it isn’t offset by any other equally probable generalization that
points in some other direction. Second, there isn’t time to get down to the
details of the particular cases.
It’s because the second of these conditions is not met in the usual em¬
ployment situation that restrictive advertising or recruiting is usually not
justified. Of course, one might try to justify it on purely practical grounds.
It’s more likely that a male applicant will pass the minimum tests for phys¬
ical strength that the Detroit police department uses, hence it saves money
to restrict applicants to males. And this is quite certainly true, but it’s not

1 And also, thanks to recent legislation, a legal question.

SPECIAL TYPES OF ARGUMENT 207

enough to offset the very serious general consequences of accepting the argu¬
ment. The general or system consequences of accepting the argument are
that we exclude all women from police forces, even those who have the
necessary physical strength. As a result, a certain point of view on crimes
that affect women—especially rape—and a crucial dimension of representa¬
tiveness, which can strongly affect the rapport with and faith in the police
force, are lost.
These general consequences are often overlooked by employers making
a choice that is, in the short run, advantageous for them. And it is for this
reason that the state argues it should step in and enforce laws about affirma¬
tive action and other fair employment practices.
It’s undoubtedly true that there are a great many white men in the
country who would not enjoy taking orders from a black supervisor or office
manager. And, although the reverse is certainly true in some cases, the
generalization that “blacks will tend to encounter more resistance in getting
their orders carried out” is true. If we use it as a basis for restricting applica¬
tions for senior positions to whites, we will of course be preserving the status
quo, and ensuring that the prejudice that presently exists is not overcome.
These general effects we are very anxious to avoid in the name of justice,
and it is because these effects follow that we argue against the legitimacy of
using the generalization as a justification for action in individual cases—
except for emergency situations.
You can see how this ties into the type of issue mentioned at the end of
the preceding section: descriptive statistics tells us something about the
whole population, but it does it by condensing the information about indi¬
viduals, and usually by representing the population in terms of its average,
its range, or other classification. Such general descriptions of a population
are extraordinarily useful for many decision purposes, for example, they
may help us decide where we should put our effort in changing the school
system. But they must be regarded as items of information to be used with
very great care when we wish to take action on the fate of an individual
person.
The tendency to act improperly from generalizations of this kind is part,
though not the worst part, of what is referred to as racist or sexist behavior.
The worst type of racist or sexist behavior, or prejudiced or biased behavior
in general, occurs when the action is based on a false generalization that
should be known to be false. The generalization may be false because it is
expressed in terms of all members of the given group having a certain char¬
acteristic, whereas only some or perhaps most of the group have that charac¬
teristic. If all members of the group had the characteristic, and if that char¬
acteristic is relevant to appointing, promoting, selecting, and rewarding
employees, it would be appropriate to use the generalization. But very few
generalizations about human beings fall into this category, and to act as if
208 CHAPTER 7

they do, when such action may adversely affect the interests of the people to
whom the generalization refers, is to act from bias or prejudice, which in
particular cases is called sexism or racism.
Just as descriptive statistics provides extremely useful ways to interpret
or understand huge bodies of data that would otherwise be quite unintelligi¬
ble, so the procedure of stereotyping is a very natural and useful type of
human thinking. It amounts to representing groups of things or people by a
single, allegedly typical, example. But the “fallacy of stereotyping” consists
in treating this stereotype as if it represents all the individuals in the group of
which it may be a fairly typical example. So the stereotypes of the campus
radical, or of the hippie, or of the businessperson or of the surgeon or lawyer
or Mexican or professor, are all useful devices, in a rough-and-ready kind of
way, for certain types of thinking. They may represent the general character¬
istics of a population quite well and may be useful for contrasting one popu¬
lation with another. (Think of the “typical” Republican versus the “typical”
Democrat, for example.) But when it comes down to cases, you have to be
willing to face the fact that an individual may, as we say, shatter the stereo¬
type. The extent to which human beings are capable of reacting flexibly,
openly, and rationally to a person who is very unlike the stereotype of the
population from which he or she comes is a measure of their reasonableness
and of their capacity for justice.
It’s unwise to act as if all stereotyping is always fallacious. It is some¬
times used as a shorthand for the “fallacy of stereotyping,” in which case it
does mean something fallacious or morally improper. But, to be precise
about the term, we have to recognize that it’s no more fallacious than talking
about ideal types (the ideal manager or parent) or typical Catholics, atheists,
and Baptists. There is scientific information which entitles one to make
claims about typical people from certain groups; but it in no way legitimates
drawing conclusions about a given individual with a membership in that
group unless absolutely nothing else is known about the individual, and then
only a probable conclusion. It doesn’t justify action with respect to that
individual, unless no other information about that individual can be discov¬
ered, say during a job interview.
Bias and prejudice, then, are to be seen as conclusions or actions based
upon either false generalizations or generalizations improperly applied to an
individual case. Notice that there is an error in either the premises or the
inference of an argument in each of these cases. It’s crucial to realize that
bias and prejudice are forms of error. This seems obvious enough, but if you
listen to the way the words are used today, you’ll notice that we have drifted
toward calling people who have definite views on a particular subject “bi¬
ased with respect to that subject, regardless of whether their views are right
or wrong. This is a particularly unfortunate position, since it implicitly as¬
sumes that people can’t be correctly convinced about any issue on which
SPECIAL TYPES OF ARGUMENT 209

somebody else is convinced to the contrary. If this were true, it would be

prejudiced to regard the earth as being round rather than flat. There are still
some enthusiastic members of Flat-Earth Societies around the globe.
Let’s be clear about what we’re saying. Suppose that we’re going to
appoint a new commission to investigate the legitimacy of the death penalty
in the state of Idaho, where the matter has recently become a controversial
one. A particular man is proposed. Someone says that he’s biased because he
is well known for his strong commitment to the necessity of the death pen¬
alty. It’s of course a general truth that we don't want to entrust a matter of
importance to biased people, because “biased” carries the implication of
falsehood. In this case, however, we’re not entitled to that implication unless
we've already settled the issue of which position is false, but that’s why we’re
setting up the commission. You’re not entitled to call a man biased because
he supports the death penalty vehemently, or because he opposes it vehe¬
mently. You're entitled to call him biased only if his reasons for supporting
or opposing it are mere matters of prejudice—that is, if they involve demon¬
strably incorrect inferences or correct inferences from premises that are de¬
monstrably false. You might think that at least somebody with a very strong
prior commitment on an issue like this is incapable of understanding the
alternative point of view and hence would not be a proper person to appoint
to a group that will have to discuss the issue. But you have no idea whether
that’s true. These are separate dimensions of ability, and, although there’s no
doubt as to the statistical connection between them in that some, perhaps
most, people who have strong views on one side of a matter are incapable of
listening to the other side, the best people in the field are often not thus
prejudiced. They hold a view which, although strong, cannot be faulted in
terms of reasoning except by extremely sophisticated means, which have not
yet been uncovered. They have, in short, an argument which appears to
them, as rational people, to be a conclusive argument for their position.
They recognize that there are arguments on the other side, but they believe
they can identify weaknesses in those arguments. It’s a fallacy and indeed a
slur to identify such people as biased or prejudiced. The only respect in
which they’re like someone who is prejudiced is that they have a strong view
on the subject. Unlike identifying somebody who obviously is biased and
prejudiced, it isn’t possible to show—certainly not easily—that they are rea¬
soning erroneously. Hence, it isn’t possible to tell whether people who main¬
tain a certain position vigorously are biased or prejudiced. It’s a matter of
what their reasons are, and what their knowledge of and capacity for assessing
the reasons of the opposition are, and what their reaction would be to new
evidence. If they are immune to new evidence, if they dismiss the arguments
of their opposition with trivial counterarguments or misrepresent these argu¬
ments, if they draw their own conclusions in a way which suggests that they
are only providing rationalizations rather than good reasons for their oppo-
210 CHAPTER 7

sition, they are biased or prejudiced. They may still be biased or prejudiced
at some unconscious or not easily detectable level, but you can’t hang some¬
body for that kind of possibility. If you could, suspicion would be grounds
for imprisonment.
In the light of this discussion, you should answer one of the following
questions and regard the one you pick as question 7-A, 2. Is a Women’s
Studies department entitled to restrict applicants for its faculty to women? Is
a Black Studies department entitled to restrict its applicants to blacks? Is a
department of religion entitled to regard an atheist as an ineligible applicant
for a position? Is a department of political science entitled to regard a mem¬
ber of the Communist party as ineligible? Is a department of political science
entitled to regard a committed Fascist and racist, for example, a neo-Nazi,
as an ineligible applicant?

7-3 PROPERTIES AND GENERALIZATIONS: ANALOGY

Argument from analogy is a vitally important and extremely tricky type of

argument. On the one hand, in science or ethics or the law, it is often the
only type of argument that can be developed on an important issue. And,
even where other arguments can be developed, argument by analogy is often
the most powerful and compelling type of argument we can use. On the
other hand, it is elusive in the sense that it’s very hard to give its precise
structure or to spell it out in full detail; and it is often double-edged in the
sense that an analogy can be turned against the arguer to point to the
opposite conclusion from the one which was proposed.
One of the reasons why analogies are difficult to handle is that they are
a more abstract type of argument than the two kinds that we have been
considering in the preceding sections. Just how much trickier they are is
perhaps best demonstrated by saying that argument by analogy is identified
as a fallacy in at least one widely used elementary logic textbook. What the
author really means is that it’s a fallacy to suppose that arguments by anal¬
ogy give you as much certainty as deductive arguments—that is, those in
which the steps depend only on the meaning of the terms. But it’s not likely
that anyone has ever really thought that argument by analogy provides one
with that degree of certainty. Analogies supply us with what are sometimes
called “plausibility arguments”: Sometimes they provide us with a kind of
grasp or understanding that persuades us of something, but their strong
point is not the capacity for conferring absolute certainty upon the conclu¬
sions which they are used to support, or even a certainty as high as that of
the premises from which the conclusion is inferred. Analogy is just a more
explicit kind of reasoning than the simile or the metaphor, to which it is a
natural cousin. Similes and metaphors are often used to convey implicitly
some suggestion—some implied conclusion—which the analogy makes more
SPECIAL TYPES OF ARGUMENT 211

explicit. They are legitimate inductive or probabilistic arguments, and are of

central importance in science and law.
“Banning the possession of handguns is like shackling everybody except
criminals.” As expressed, this could well be described as a simile; but, be¬
cause of the implicit implications, it can also be regarded as a condensed
form of an analogy, which is really just an extended or complex simile—a
claim about the similarity of two things. Now a simile can be used to sug¬
gest, as in this statement, that a certain conclusion should be accepted. It
then becomes a compressed argument from analogy. In the example given,
the implied conclusion in most contexts is that it would be wrong to ban the
possession of handguns (since it’s obviously wrong to shackle the innocent
and let the guilty go free). Hence, argument by analogy consists of arguing
from certain similarities between two things that are said to be analogous, to
a conclusion (or conclusions) about one of them which is said to follow since
the corresponding assertion holds about the other. Thus, if the things that
are said to be analogous are X and Y, the analogy is supported by certain
similarities (A.B.C) which may be left unspecified since they are thought to
be so obvious. Our attention is then drawn to the fact that X or Y has the
property p; hence, it is suggested, since X and Y are analogous, the other (Y
or X) has the “corresponding” property q (the analogous property).
Running through the handgun example, we have an analogy between
banning the possession of handguns (X), and shackling the innocent (Y);
this would be supported by pointing to the fact that in both cases, people
who are unwilling to defy the law will be handicapped (this is the similarity
A). Now, since it clearly is morally wrong to shackle the innocent (this is the
property p), it follows by analogy (so it is suggested) that it is morally wrong
to ban the possession of handguns (this is the property q that corresponds in
the case of X to the property p in the case of Y).
The logical structure of this argument is essentially that of reasoning
from the existence of some similarities to the presence of another similarity.
It could be described as reasoning from a sample of properties, which turn
out to be common to two systems, to the conclusion that another property is
also probably common to both systems. If you think back to the type of
inference that’s involved in going from a sample to properties of a popula¬
tion, you can see that this is a case of inferring from samples of the proper¬
ties of two objects to a generalization about most of their properties, and
hence to the conclusion that a particular one of interest will be shared by
both of them.
Notice that it would be quite wrong to say that arguments from anal¬
ogy involve jumping to the conclusion that all properties of the analogous
objects are the same, since if all properties were the same, the objects would
obviously be identical rather than merely analogous. One of the old favorites
in discussing arguments by analogy is the ancient analogy between the state
212 CHAPTER 7

and the human body, with the head of the state (notice the metaphor) repre¬
senting the head (of the body), the police and army and the other branches
of the executive representing the arms and legs, the consumption of goods by
the state being represented by the consumption of food and other things by
the individual, and so on. Now it’s clear that nobody putting forward this
analogy thought that there were no important differences between the hu¬
man body and the state. For example, it’s obvious that the separate limbs
cannot survive independently of the body, whereas the individuals in the
state, or groups of them, can survive, at least for long periods of time, and
often long enough to form a separate state, without connection to the rest of
the state. Hence, argument by analogy cannot be thought to be an argument
to the conclusion that all the properties of the two analogous systems are
held in common, although it would certainly follow from that, that the one
the conclusion refers to is held in common. Such arguments are, by their
very nature, arguments to the probable sharing of a certain property, based
upon the substantial extent to which other properties are shared.
Once we have this general picture of argument by analogy, it’s obvious
how to go about criticizing it, though it’s not at all obvious what constitutes
a definitive refutation of an argument by analogy. One criticizes such an
argument by pointing to what are called dysanalogies between the two sys¬
tems, that is, properties which they do not share. An example is the capacity
for independent survival, just mentioned, in the analogy between the parts of
the body and the parts of the state.
Given that the argument is not intended to involve, as an intermediate
conclusion, the assertion that all the properties are the same, then what one
has to refute is the claim that most of the properties are the same, or that
properties of the type to which the conclusion refers are the same or mostly
the same. It’s much harder to refute these “weak” generalizations involving
terms like ‘most” than it is to refute categorical generalizations referring to
“all” properties or objects. But the line of attack is clear enough; one has to
go about the process of looking at the properties of the two things being
compared, and see whether one can’t make a case that they are mostly
different or that the particular type of property to which the conclusion
refers is, in fact, usually different in these two types of things.
A neat and powerful, but rather different, move consists in using the
very same analogy to argue for the opposite conclusion. In the case of the
analogy about the handgun laws, one can argue that shackling innocent
people makes it harder for them to move into the path of crime. The implied
conclusion is that the handgun law at least prevents, or tries to prevent,
people who are not already gun-wielding criminals from joining that group.
Some restrictions on freedom, it might be added in a conciliatory footnote,
are necessary in order to preserve the freedom of all.
SPECIAL TYPES OF ARGUMENT 213

The direct attack on the analogy, by listing the differences between the
things that are being compared in order to undermine the plausibility of the
suggestion that there’s a crucial similarity, can also be illustrated in the case
we’ve been discussing. Here we would say that banning handguns is not like
manacling the innocents since it puts exactly the same manacles on the
criminals, i.e., their possession of a gun now becomes grounds for prosecu¬
tion, and we’re thereby able to imprison many of them before violence is
done to others. This is not paralleled by any feature of the alleged analogy of
manacling the innocents. Moreover, it might be argued that possession of
handguns results in a great many unintentional, serious, or fatal accidents,
because most people who have guns are not experienced in taking the care
necessary to use them properly and safely. Hence, there’s a bonus for the
innocent in the handgun legislation which doesn’t correspond to any bonus
in shackling the innocent.
In the course of the Watergate hearings before the Senate investigating
committee, the second-in-command of the White House staff John Ehrlich-
man, was asked if he could explain why the White House had thought it
appropriate to hire burglars to break into the office of the psychiatrist who
had been treating Daniel Ellsberg, the Pentagon consultant who released the
Pentagon Papers to the press. It was clear that the White House hoped to get
some “dirt” on Ellsberg, which could be used in a direct blackmailing proce¬
dure, or in a discrediting procedure with the press. Why was this thought to
be an appropriate kind of procedure for the White House to engage in?
Ehrlichman replied with an analogy, and you should treat it as an exercise to
assess it, criticizing it if you feel that that’s appropriate (Quiz 7-A, 3). Ehr¬
lichman said that the situation was like the following: Suppose that you
learned that in a safe deposit box in a bank vault in Washington, D.C., there
was a map showing the location of an atomic bomb that was timed to go off
the next day and blow up a large part of Washington. Breaking into that
bank vault would, he suggested, be like breaking into Ellsberg’s psychi¬
atrist’s office—what else could one reasonably do? It may be relevant to add
a word or two about an exchange that followed immediately between one of
the senators and Ehrlichman. The senator said that it occurred to him that
the appropriate thing to do would be to call up the president of the bank and
ask for the keys, explaining one’s reasons. Ehrlichman replied that they had
done the equivalent of that; they had attempted to bribe one of the nurses in
the doctor’s office to get them the file.
The final comment on analogy is to mention that the next step beyond
it, in the sequence from metaphor through simile to analogy, is the concept
of a model. Models, which are crucial in scientific thinking but also in politi¬
cal and business planning, are really elaborate analogies, sometimes of a
highly formal kind (mathematical models, computer models) and sometimes
 214 CHAPTER 7

of a vague and an almost poetic kind (the free enterprise model). One of the
classic models in the history of science is the so-called planetary model of
atomic structure. This was the view that the atom could be conceived of
rather as a miniature solar system, with a nucleus that was extremely heavy
compared with the “planets” swinging around it (which correspond to the
electrons). These planetary electrons would be in relatively stable orbits, at
distances from the nucleus which were very, very great compared with the
diameter of the nucleus, which was, in turn, very great compared to the
diameter of the planets. Now it turned out that this way of thinking about
the structure of the atom was able to account for a great many of the
phenomena that were known to physics at the time it was most popular.
Notice that nobody was suggesting that all the properties of the solar system
were shared by the atom. For example, there was no suggestion that the
planetary electrons had satellites. A nice feature of models, and of analogies
in general, is that they often suggest further similarities which, on investiga¬
tion, turn out not only to hold but also to explain some puzzling phenom¬
enon. In this particular case, an experiment, known as the Stern-Gerlach
experiment, produced results quite impossible to explain on any known
model of the atom, including the simple planetary model. Then somebody
thought of the fact that in the solar system, the planets not only revolve
around the sun, but they also rotate about their own axes. Suppose that one
was to postulate the same in the atom, that is, suppose that each electron
had some spin as well as a rate of revolution about the central nucleus. This
would clearly give the system dynamic properties different from one in
which the electrons had no significant spin; for example, they would have a
store of energy over and above the kinetic energy due to the rotation in
orbit. Hence, the results of interactions between streams of bombarding
particles—or strong magnetic fields—and atoms would be different. It
turned out that this difference was just about right to explain the results of
the Stern-Gerlach experiment.
This kind of example has often been taken to show that the role of
models in science is essentially psychological, that is, it assists the scientist in
thinking up new hypotheses, but it has no real explanatory power in itself.
The scientist uses the model to generate the new hypotheses, which he or she
tests explicitly and if they turn out to be confirmed, adds them to the body of
knowledge on the subject. While this is partially true, it involves a mistake
about the nature of explanation, and to that topic we will turn in the next
section. But, as a final example for this section, to link it to the next, you
might discuss the widely held idea that scientific theories are analogous to
scientific instruments, i.e., they serve to achieve an important end, but they
do not have any intrinsic, true scientific content. Specifically, we invent
notions like the electron and photon, the relativistic increase of mass with
velocity, the gene, and so on, in order to enable us to generate predictions
SPECIAL TYPES OF ARGUMENT 215

and assertions, but we should not conclude that these “hypothetical entities”
have any real existence, any more than we should think that a surgeon’s
scalpel contains life-generating properties just because it can be used to save
lives. Do you agree? (Quiz 7-C, 1)

7-4 EXPLANATIONS AND JUSTIFICATIONS:

CAUSES AND REASONS

The topic of explanations and their nature is a vast one, and it is extensively
discussed in the philosophy of science, as well as in the philosophy of history
and elsewhere. We will not attempt here to elaborate on it at any length, but
we will try to draw one important contrast, that between explanations and
justifications, and to stress once more the creative-imaginative element that
is involved in explanation hunting and so important in providing and criti¬
cizing powerful arguments.
You will recall our stressing that the process of criticism of argument
characteristically involves the testing of inferences, and that the testing of
inferences characteristically involves trying to think of circumstances in
which the conclusion could be false even though the premises were true.
Thinking of just one such case destroys a deductive argument completely
(although it might be reformulable as a probabilistic argument); thinking of
important families of such cases undermines and can eventually destroy even
a probabilistic (inductive) argument. We stressed that there isn’t any rule for
thinking up such explanations, and one’s ability to do it is improved by two
activities. The first is a steady increase in one’s general knowledge of what
causes what, what sometimes affects what, and so on—knowledge that leads
one to think of alternative explanations, by analogy if not from direct ac¬
quaintance. The second source of improvement lies in the skills gained
through extensive practice in the criticism of arguments and in the setting
forth of good arguments in the face of criticism.
What we call “imagination” is to a large extent the ability to draw on a
very wide range of knowledge—of models, if you like—which illustrate some
of the pitfalls that other proposed models may fall into. For example, in
much of the research on extrasensory perception, the investigators have been
quite inexperienced, and as a result have been all too easily persuaded that
some extraordinary ability, such as telepathy, is being demonstrated, when
in fact there are other possible explanations. If one has had considerable
experience, both in reading the literature of previous investigations and in
performing some oneself, and if one has acquaintances in the illusionist or
“stage mentalist” guilds, one goes into such an investigation armed with a
dozen or more alternative explanations that have to be ruled out before a
persuasive case for telepathy can be made. The fact that there are a number
of experiments where such alternative explanations do appear to have been
216 CHAPTER 7

refuted is what makes the field truly interesting. The fact that the casual
kind of “parapsychological experience” such as a foreboding that turned out
to be correct (or a dozen of them), or a dream which involved an event
happening somewhere else (or a dozen of these), is of no interest to some¬
body who is familiar with the serious investigation of alternative, naturalistic
explanations of these reports. The same applies to water-divining (“dows¬
ing,” “water witching”), fire-walking, psychic surgery, and like feats. One of
the reasons people are so gullible about these areas is that organized science
has reacted to them in a totally unscientific way, for the most part. Naturally
enough, people whose common sense enables them to see that scientists can
be, and frequently are, prejudiced along with everybody else suspect that the
dismissal of such areas is just a sign of that prejudice. Sometimes it is, and
sometimes it isn’t. If one wishes to take these investigations seriously, one
has to commit oneself to a very substantial course of study. It is now more
than thirty years since I founded the first university-based research group for
the investigation of parapsychological phenomena in the Southern Hemi¬
sphere, and the chain of investigations and discussion that has followed
since has been amongst the most intellectually stimulating experiences of my
life; not surprisingly, it has also led to a considerably expanded knowledge
of ingenious ways of faking—consciously or unconsciously—supernatural
phenomena. Let’s see how ingenious you are at providing possible explana¬
tions of some of the phenomena that have recently been exhibited by the
Israeli “psychic,” Uri Geller (quiz 7-A, 3). One of his performances consists
of asking the audience to pass up to him any wristwatches or pocket watches
that have stopped working. After holding them in his hands for some time
and attempting to “infuse them with psychic energy,” they quite often start
running again. What alternative explanations (to the psychic one) can you
think of for this? How would you go about deciding whether the psychic
explanation was correct?
Remember that, in Section 7-1, where we were talking about criticizing
inferences from sample characteristics to “population characteristics,” that
is, to generalizations about the population, we stressed the fact that a major
thrust in the criticism of such inferences consists in trying to think of expla¬
nations of why the sample would exhibit a certain characteristic although
the population as a whole might not. For instance, a sample of names drawn
from the telephone book, however randomly, will not be a politically repre¬
sentative sample of the population as a whole if people who get their names
into telephone books have any characteristic that is politically different from
characteristics of the population as a whole. Obviously, there is an economic
nonrepresentativeness about telephone owners which invalidates the infer¬
ence. Similarly, in Section 7-2, where we were thinking about inferences in
the opposite direction, from characteristics of the population to characteris¬
tics of individuals, clearly we were trying to think of ways in which the
SPECIAL TYPES OF ARGUMENT 217

assertions about the population could be true, but corresponding assertions

about individuals false; that is, we were trying to think of alternatives to the
view that the population had that property because each member or most of
them, had the property, i.e., alternative explanations. If the generalization
about the population is universal and exceptionless and categorical, and only
then, one can proceed immediately to conclusions about the individuals. If
the generalization about the population says something about what it tends
to do, or what it typically does, or what its average member does, one can’t
draw any definite conclusions about the individual, only probabilistic con¬
clusions that will be open to complete refutation by direct investigation of
the particular individual. So the process of attempting to think of counterex¬
planations is central to the whole business of criticizing, and thus to the
business of providing arguments. The very testing of an inference itself in¬
volves trying to think of a counterexplanation of the facts put forward as
premises, other than the claim presented as the conclusion. (Usually; give an
exception for Quiz 7-C, 7).
So much for connecting the discussion of explanations with what has
gone before. Just a word or two about the nature of explanation itself, and
then we can turn to the distinction between explanation and justification.
Explanation is one of the most controversial concepts in logic and the phi¬
losophy of science. There are many people—probably the predominant
school at the moment—who believe that explanations are just psychological
devices of no intrinsic logical or scientific force. There are others who believe
that explanations are both psychological devices and a key, possibly even the
key, logical and scientific device. There are yet others who believe that the
psychological content is merely incidental, but that a purely formal analysis
of the concept can be provided which shows that it has a basic and impor¬
tant logical structure independent of the psychological aspects of it. It is
certainly true in the history of science that particular models of explanation,
notably the animistic and anthropomorphic models of explanation, have
infused our thinking well beyond the range of their legitimate application.
No doubt the reason why we like to explain phenomena in terms of animis¬
tic and anthropomorphic concepts is that they are so easy for us to under¬
stand because of our familiarity with animals and humans. Nevertheless, the
fact that on occasions our psychological preferences and rigidities have
tended to distort our patterns of scientific explanation is no ground for
supposing that we can dispense with the psychological element in explana¬
tions. (That might simply be throwing out the baby with the bathwater.)
Acquiring understanding, which is what one gets as a result of accumulating
explanations, is the highest goal of learning, and it isn’t just an old myth that
suggests that’s so. Understanding involves obtaining a “mental grasp” of a
subject matter or a phenomenon, the capacity to apply it to new problems,
218 CHAPTER 7

for example. Explanations are much more than mere mnemonic devices
(that is, memory-jogging devices, such as rhymes that give you the names of
the American Presidents); they are richer and deeper than such devices. Like
analogies, they generate much more than they start off with. When you are
provided with a thorough explanation of electromagnetism or of the disease
transmission cycle in the case of malaria, you are put in possession of a kind
of framework for your knowledge in that area that spins off practical sugges¬
tions, theoretical conceptions, and marriages between the two.
Just because of this elusive semipsychological component in explana¬
tion, this requirement that the explanation convey understanding and not
the mere ability to recite a few facts that are contained in the words used for
the explanation, it is very difficult to pin down the difference between a
satisfactory and an unsatisfactory explanation. Astrology offers an excellent
example. Many people believe that they obtain real understanding of the
character of somebody when they learn the sign under which that person
was born. It is a belief which has been with us for thousands of years, and it
is probably more popular now in Western cultures than it has been for
hundreds of years. A group of distinguished scientists recently published a
letter in the Humanist denouncing the resurgence of belief in astrology, but
the letter exhibited very little more than prejudice, as far as an independent
observer is concerned. As mentioned before, no careful, critical experiment
has been performed in which astrological predictions have been matched
against predictions made on other grounds. If one starts discussion of astrol¬
ogy at the general level, of course it’s obvious that the configurations of the
heavenly bodies have very marked effects on earthly phenomena, as tides
and crops show. So, one is left to fall back on the fact that these distin¬
guished scientists, none of whom has studied the subject directly, think
there’s nothing to it. That is hardly a persuasive or indeed a proper line of
approach to the subject. One would have to argue that our knowledge of the
general influences that do exist suggests very strongly that they would not
extend to the formation of character or to a direct control of the particular
kind of events on the earth’s surface that are referred to in horoscopes, etc.
However, one would also have to agree that we simply lack serious experi¬
mental investigations of the validity of astrological predictions.
There have been a great many occasions on which folk wisdom has
turned out to be false, but there have also been a very large number of
occasions when folk wisdom has turned out to be true or to contain part of
the truth, after it had been scoffed at by organized science for a long time.
The most obvious examples of this come from folk medicine, but there are
others, including graphology, weather prediction, agriculture, and innumer¬
able examples in the areas of crafts and athletics. An extremely healthy dose
of skepticism about the reliability of science is an absolutely inevitable con-
SPECIAL TYPES OF ARGUMENT 219

sequence of any scientific study of its track record. Not surprisingly, this is
not a popular subject amongst scientists. People don’t like to be shown to be
inadequate according to their own standards, whether the standards are
those of ethics or those of science. One of the most interesting borderline
cases has been that of psychoanalytic theories, the validation or invalidation
of which has been a bone of contention in psychology for more than half a
century. At the other extreme, the Skinnerian version of behaviorism is
equally controversial. In the political domain, the question of whether the
Marxist dialectic provides genuine understanding is of the same kind.
All these “models” (or theories or conceptual schemes) provide the
practitioners with a special language, a jargon of their own, and a large
number of alleged explanations and predictions. How reliable are these ex¬
planations and predictions? How much real understanding, as opposed to
feeling of understanding, do these approaches provide? How much better are
the predictions that they yield than those of an intelligent observer not using
these theories but using all the other background knowledge that we have
about psychological or socioeconomic events? It’s not easy to answer that
because, in the first place, the experts from any one of these groups do not
agree with one another to a very large extent. A rational approach requires
very careful investigation of the theory for its internal clarity and coherence
before one can even decide what “it” implies. Reason then requires that one
compare its predictions with those based on other general knowledge, rather
than with the mere standard of truth and falsity. It’s easy to produce true
predictions. For example, it’ll surely be true that six years from now you’ll
either be married or not; but of course the value of that as a prediction is
totally trivial. When we start looking for the scientific value of the predic¬
tions, retrodictions, and explanations generated by these theories, we have a
very much harder standard to meet than if we’re simply looking for truth.
We have to find new and valuable, as well as true, predictions.
We now turn to the distinction between explanations and justifica¬
tions—between causes and reasons, as it is sometimes put. The search for
explanations, of which causes are a particular case, is the search for a frame¬
work into which one can fit a phenomenon in such a way that one can
answer a whole series of questions about it, relate it to other phenomena in
the same field, and so on. One is seeking to achieve understanding of the
phenomenon. But the search for justifications, for good reasons for some¬
thing, is conducted in a different framework, a framework with some, but
not very much, overlap with the former. A justification has to be a good
reason for a belief or an action; a moral justification has to be a set of
reasons that are morally adequate to support a certain conclusion or action.
The fact that the cause of your crashing into the plate glass door was your
not seeing it in no way constitutes a good reason for crashing into the plate
glass door. The fact that the explanation of your lying to a congressional
 220 CHAPTER 7

investigating committee was that your job was at stake in no way provides a
moral justification for the lying. All this seems obvious enough, but there is a
philosophical position which has attempted to confound the two questions,
the search for explanations and the search for justifications. This position,
which is technically called “hard determinism,” argues that first, on scientific
grounds, we have to believe that there is an explanation of every event even
if we don’t have it at the moment; and second, that if there is an explanation
of an event, the event had to occur in the way in which it did—could not
have occurred in any other way—and hence any discussion of justification,
or praise or blame or evaluation, is as inappropriate as talking about justify¬
ing the path of the planets in their orbits. What cannot be other than it is
can scarcely be blamed for not being other than it is. Once one understands
fully why something occurred, so the hard determinist argues, one can no
longer see any point to the question of whether it should have occurred or
should have been done, or whether it is or was right or wrong. Similarly with
beliefs; since what we believe is determined by our hereditary composition
and the environment to which we’ve been exposed, we cannot believe other
than we do believe, and hence it is pointless to criticize our beliefs as if they
could somehow have been different.
Without getting into the entire argument about free will and determin¬
ism, of which this discussion is part, it is important to understand the diffi¬
culty that the hard determinist is up against. When one says that a justifica¬
tion, like a proof, has to meet certain standards of evidence, or standards of
moral probity if it’s a moral justification, one is saying nothing more or less
controversial than is involved in saying that in order to be considered for the
Olympic trials, a miler has to have a best time that’s under four minutes.
We’re not judging whether somebody should be blamed or praised in some
peculiar sense in setting these standards; we only wish to assert that this
kind of performance is what it takes to be eligible; and we may also wish to
admire people who can run a mile in four minutes, just as we might also
admire people who can wiggle their ears, or dance gracefully, or reach the
rim of a basketball net without leaving the floor. In the same way, we may
admire a sunset or a glider in flight. Arguments have to meet certain stan¬
dards in order to be acceptable, standards that vary depending upon
whether actions that risk life and death are at stake, whether a scientific
reputation is in jeopardy, whether moral assessment is involved, and so on.
If the standards of good argument are met, the argument in question gives a
justification for the conclusion. If they’re not met, it doesn’t. The success of
a particular individual in providing an adequate justification for an action
may or may not be completely determined by his or her environment and
heredity. But even if it is, it is still entirely appropriate to judge the person’s
performance in terms of the standards of merit that identify adequate justi¬
fications, just as it is with the mile runner. An argument that is less sound, or
SPECIAL TYPES OF ARGUMENT 221

a moral justification that is less than adequate, can’t be made to work any
better by pointing out that it was produced by an illiterate. Of course such
considerations may affect the extent to which you praise or blame the person
for producing a substandard argument, but they don’t have any effect at all
on the standards of merit for the argument.
The French have a saying, “Tout comprendre c’est toutpardonner”—“To
understand everything is to forgive everything.’’ The suggestion is that once
one has a full explanation of somebody’s behavior, one must forgive it.
That’s not to be confused with the idea that once you understand some¬
body’s behavior, you have to approve it. Approval requires justification, and
causes don’t produce justifications. Indeed, they don’t produce arguments at
all. In point of fact, it’s not clear that to understand something, even com¬
pletely, should lead one to forgive it. If one understands that it was an act of
deliberate choice, made in self-interest, that led some man, after collecting a
reward, to kill a kidnap victim so that he couldn’t be recognized by the
victim later, one may still feel totally unable to forgive the action and, in¬
deed, one—in my view—should not forgive the action. However, that’s an
issue for the free will and determinism discussion in a Problems of Philoso¬
phy, or an Ethics course, and we’ll pass by it for now. What we have to cling
to firmly is the distinction between causes and reasons, between explanations
and justifications, and the legitimacy of having standards for evaluating ar¬
guments, whether or not one has the theories and data required for explain¬
ing arguers.
It may seem like a rather abstract point. Why is it important at the
practical level? Perhaps the most important practical connection lies
through the concept of bias or prejudice, which we discussed in Section 7-2.
Biases and prejudices are often causes of beliefs. Sometimes people get hung
up on the fact that biases and prejudices are not always easy to detect, and
conclude that perhaps they are always present and they are what always
determine all our beliefs, the “arguments” that we put forward being essen¬
tially smoke screens, or “rationalizations,” as they are sometimes called. One
needs to keep a firm grip on reality here, and not be misled by an abstract
philosophical argument. No secret bias or prejudice is forcing you to the
belief that 2 and 2 are 4 or that siblings are brothers or sisters, or that
triangles have three sides. These are matters of definition, and the whole
realm of deductive argument is founded on inference according to such
definitional rules. Of course, one can sometimes make errors in very compli¬
cated deductive arguments, and bias may lead us to make such errors. We
should take care to scrutinize deductive arguments very carefully when they
bear on an issue where we have strong feelings. But when all is said and
done, and when several independent experts, with different biases and preju¬
dices in the field, have all carefully checked out a fairly straightforward
deductive argument, one is entitled to say that one has extremely strong
222 CHAPTER 7

grounds for supposing it to be a sound argument. The remote possibility

does remain that there’s an error there which is due to bias, but it is simply
a remote possibility. When you see a car crossing the dividing line of the
highway and heading straight toward you, the remote possibility exists that
it’s a hallucination, but you’d better not act on the basis of that remote
possibility. You’d better get off the highway just as fast as you can and as
safely as you can, and if you can’t do all that, get off the highway unsafely.
Remote possibilities are not what you should guide your life by. The same
applies to the hypothesis of universal bias and prejudice as the “real” deter¬
miner of all our conclusions. If I set out to decide whether it’s better to
invest my savings in a savings account or in tax-free bonds, there’s a proce¬
dure for doing this rationally. It involves calculating the extent to which the
difference in taxation balances off against the difference in safety, and if a
few certified public accountants specializing in tax advising all agree that my
conclusions are correctly drawn, then I can forget about the possibility of
bias and prejudice. It’s there, but it’s too small to take seriously in an action
context.
When we get into the area of race relations or ethics or international
politics, of course it’s much easier for bias and prejudice to operate without
being detectable. But that’s no more interesting than saying that theoretical
physics .is much harder to understand than elementary chemistry, and hence
we have to look more carefully at what appear to be good arguments, and be
willing to recognize that a greater degree of skepticism about the conclusions
is appropriate. It doesn’t follow that nothing can ever be established beyond
reasonable doubt in the area, or that bias and prejudice are determining
most, or even a significant part, of the conclusions that are drawn. We have
a long history of these subjects to guide us, and it shows that mistakes do
occasionally occur and are sometimes serious, even when careful tests of the
argument have been made, but not often. So first we look for counterex¬
planations of the data that haven’t been considered in arriving at the conclu¬
sion that is put forward as the underlying explanation. We look for counter¬
arguments. We look for technical flaws in the premises and inferences. We
use subject-matter experts and logical experts of different viewpoints and
backgrounds to do the looking. At that point, we’re in a position to make a
judgment with considerable confidence. We can still be wrong, but that
possibility is no longer very likely. You may get run over when crossing the
street, but the chances are negligible if you look both ways.
Consider one practical application of these considerations to the role of
the teacher of a controversial subject. Some people have argued that a
teacher of philosophy, say, who is about to embark on a discussion of the
problem of the existence of God should begin by laying out his or her
“prejudices” on the table. This is thought to be “honest” and to show a
SPECIAL TYPES OF ARGUMENT 223

willingness to recognize the fallibility of (even) professors. As exercise 7-A,

5. discuss the following reaction to that position.

Far from being honest and helpful, this is a crowd-pleasing confession of in¬
competence. If what is being laid out are simply one’s positions on the issues
about to be discussed, then—since these issues are within one’s field of profes¬
sional competence—one presumably has what one takes to be good arguments
for them. Hence they’re not known to be prejudices and should not be described
as such, since prejudices definitionally involve identifiable errors. If these con¬
clusions are conclusions that can’t be reached by good reasoning, then it’s pro¬
fessionally inappropriate to hold them as conclusions, and it is a concession of
gross incompetence to maintain them as positions at all. If one isn’t quite sure
whether the arguments one has establish one’s conclusions, then the time to
mention that is when you get to laying them out, at which point you can
indicate that you’re not convinced that they really establish the conclusions.
After all, it’s a professional judgment that an argument is not decisive. Again,
there is no justification for the advance announcements. These advance an¬
nouncements may be crowd-pleasing, but they concede the propriety of prejudice
in the expert. A professor should restrict himself or herself to making clear how
frequently the alleged experts turn out to be wrong or in conflict on subjects
about which they’re quite dogmatic, and add to this—if it’s felt to be neces¬
sary—the fact that it’s clear that this will apply to the present speaker or writer
as well as to most other people. But the one thing that can’t be done is for
somebody to identify the prejudices that are unconsciously affecting them—
that’s a contradiction in terms—and the one thing that’s inconsistent with pro¬
fessional standards is to identify prejudices that are consciously recognized as
such, and still wish to maintain the conclusions they affect as sound professional
judgments.

The last topic in this section concerns causation itself. Everyone knows
that one has to distinguish causal connections from mere correlations. The
Bureau of Statistics tells us that there was a steady rise in real family income
through the 1960s at the same time as there was a steady decline in the
birthrate. By applying statistical tests, we can discover whether this correla¬
tion is truly significant or a mere coincidence—like a tricky run of five
successive throws of “doubles” with a pair of dice (where the two dice come
up with the same number on each throw). But that doesn t tell us whether
increased wealth causes reduced childbearing, or whether something else
(e.g., better technology) produces both effects. The early days of investiga¬
tion into the relationship between smoking and lung cancer mostly involved
reports of high positive correlations, but nobody could be sure that there was
a causal connection. It’s generally much harder to establish causation than
correlation. As usual, the trick is mainly that of trying to find alternative
explanations, and then ruling them out by direct experiments and observa-
 224 CHAPTER 7

tions. For example, the smoking/cancer correlation might reflect the fact
that heavy smokers tend to be city dwellers and that either the stress or the
atmosphere of cities tends to produce lung cancer. To rule out this hypoth¬
esis, we’d look deeper into the statistics to see whether we found a big
difference in cancer rates amongst city dwellers, depending on their smoking
level; and the same for country dwellers.
But what about the possibility that there is some genetic predisposition
in some people that tends to lead them to smoke and also tends to give them
lung cancer? If this were true, the smoking wouldn’t be causing the lung
cancer but of course there would still be a high correlation between the two.
Here we can easily design, in the abstract, an experiment that will decide the
issue: We would simply pick a large random sample of children, divide them
(randomly) into two groups, and ensure, by bribes or coercion, that one of
these groups did not smoke, and that the other did smoke heavily.2 Fifty
years later, we’d know the answer. This type of design, the so-called fully-
controlled, fully-experimental study is always the best design for identifying
causes; but, as here, it frequently involves completely immoral manipulation
of human beings. And it’s just a little slow in getting results.
A number of other very clever insights were required to establish be¬
yond reasonable doubt that smoking really does cause lung cancer. For
example, a study of adult converts to Mormonism (which prohibits smoking)
who previously smoked showed that the younger they were when they gave
up smoking, the better their chances of not getting cancer. Making the
rather plausible assumption that conversion to Mormonism is not controlled
by any (hypothetical) genetic factor that produces smoking and cancer, we
have here a sample of people in which circumstances have almost acted like
an experimenter in “manipulating” the amount of smoking people do, and
we find that it does control lung cancer. There were also experiments with
mice which showed that skin cancers could be induced by painting their
skins with the tars and nicotine from cigarettes. This web of indirect evi¬
dence gradually built up the strength of the causal conclusion beyond rea¬
sonable doubt.
Suppose that we discover there’s a very much higher incidence of can¬
cer of the uterus among women who are using birth-control pills than
among those who are not (that is, we have a high correlation). Do we have
good grounds for inferring a causal connection? If not, exactly why not, and
how could we settle the issue? (Quiz 7-A, 6)
The importance of causes is their connection with control. Correlations
as such give us no basis for control, though they may be useful for prediction.
It used to be thought that there was a significant correlation between the
sunspot cycle and business prosperity. If true, it would hardly help control

2 Except for one possibility—what is it? (Quiz 7-C, 8)

SPECIAL TYPES OF ARGUMENT 225

the latter; but since sunspots are semipredictable, we would have had a
useful predictor nonetheless. The early symptoms of a disease are another
example of something which is a useful predictor but useless for manipula¬
tion.
Suppose there are two or three small towns in the United States with a
high degree of fluoridation of the water, and two or three with essentially
zero fluorides in the water supply. If there is a very large difference in the
incidence of dental cavities between the two groups, favoring the first, do
you have persuasive evidence for fluoridating the water of cities with less
than the fluoride level of the low-cavity towns? Spell out your reasons in full.
(Quiz 7-B, 5)
The governor of Connecticut once introduced mandatory suspension of
licenses for speeding. There was a 30 percent (or thereabouts) decline in
highway fatalities in the first year of the mandate. He was very proud of this.
Before you joined in the congratulations, what would you have wanted to
know that you (and he) could know there and then? (7-B, 6)

7-5 DIALOGUES AND DEBATES

Although dialogues and debates do represent special types of argument, and

a little can be said about them under that heading, they also represent
something more crucial to the process of reasoning than that comment sug¬
gests. One can illustrate this by mentioning that one of the most original and
influential philosophers of the twentieth century, Ludwig Wittgenstein, used
to claim that there was really no point at all in philosophical writing, only in
philosophical discussions (dialogues). What he had in mind was something
like this: On very important issues that involve really complex and abstract
ideas, as do philosophical—and ethical and political — issues, there’s no way
in which all the possible lines of support and defense can be set out within
the available period of time, even if somebody could or would read or listen
to them all. Hence, the only sensible procedure is to begin by indicating a
general line of thought in order to get the immediate reactions of those with
whom one is interacting, to focus on the worries that they have (rather than
all possible worries), to produce one’s argument for handling those worries,
and to go on from there. Thus the dialogue proceeds along the lines where
the needs of the parties involved turn out to be greatest, rather than attempt
to cope with all needs of all possible parties to such discussion or disputes.
There is a long tradition behind this approach, beginning in Western history
with the Platonic dialogues, and in Eastern philosophy long before that. It’s
thus not an accident that the dialogue is the chosen method for dealing with
these complicated issues; it is probably the best.
There have been philosophers who even made a philosophy out of the
structure of dialogue: the so-called Hegelian synthesis was achieved as a
226 CHAPTER 7

result of the process of putting forward a position, called the thesis, followed
by the counterpresentation of an “anti-thesis,” the emergence of a “synthe¬
sis” from these, and then of an antithesis to the synthesis, and so on.
So one extremely important feature of dialogue is that it provides one
with an opportunity to get early reactions to a line of argument. Therefore it
is more appropriate in dialogue to begin by presenting the main thread, or
the first steps of an argument, and then encouraging a response. Long
speeches for openers leave too little possibility for adequate interaction.
(These conclusions might well be applied to most meetings and many lec¬
ture-courses, as is done in the “Personalized System of Instruction” or the
Keller plan, now quite widely used at the college level, emphasizing individ¬
ual interactions and eliminating most lecturing.) It is important to distin¬
guish between serious dialogue or discussion and a certain type of debate,
where the “game” is to “score points,” rather than to reach the best solution.
One of the drawbacks about the “debating team” approach as a way of
instructing people in reasoning techniques (which is not its only purpose) is
that it tends to make too much of a fetish out of point-scoring, rather than,
for example, willingness to compromise when one is convinced that the
opponents have a better argument, and so on. In short, its risk is that it may
lead one to trade in truth for competitive superiority. On the other hand, it
has some excellent virtues; for example, it requires that one fight as hard as
possible, dig as deep as possible, for defenses which at first sight one might
not have supposed existed for the position which one is allocated or adopts
at the beginning. This is an excellent exercise in that it teaches you that one’s
first picture of a position is often very misleading as to the extent of its rather
readily available support. Again, debate format encourages you to try to
anticipate the objections that will be made by the opponent. This is excellent
practice for the step in argument analysis of introducing other relevant argu¬
ments (and counterarguments). It makes straw-man arguments suicidal. It is
also excellent practice for trying to distinguish between strong and weak
criticisms of the opponents’ arguments, since the limitations on time, and the
need to appeal to an audience that will judge the issue, make it necessary to
focus on a relatively small proportion of the possible points that one could
make. Again, public speaking skills can be improved enormously with de¬
bate experience. But there is no substitute for systematic written analysis,
repeated until one has a concise and potent presentation. It is the best basis
for debate—and for all serious thinking.
Dialogue and debate stress the importance of communication, as op¬
posed to mere exposition, a stress that has characterized this text through¬
out. Rhetoric does the same; and many of the skills discussed and presum¬
ably communicated in this book are often covered in courses in a
department of rhetoric, or speech, or communication. This fact illustrates
the growing emphasis in recent decades on the communication aspect of
SPECIAL TYPES OF ARGUMENT 227

reasoning. It isn’t a new tradition, since it is to be found not only in Plato

but also in the Sophists of ancient Greece, and again in the Paris schools of
the Middle Ages. But it is very much more practical than the preoccupation
with symbolic logic that has become the domain of the philosophy depart¬
ment in present-day universities. Although we have not brought into this
text many of the technical concepts and historical schools in rhetoric (or in
jurisprudence, for that matter, where there are other equally relevant tradi¬
tions), this omission is mainly from a desire to avoid greater length, not from
a belief in their lack of relevance.
One of the issues that comes up in the course of dialogue and debate
concerns the credibility of authorities, including sometimes the authority of
the person with whom the debate or dialogue is being conducted. There are
no really useful general rules for determining who is to count as a reliable
authority, except by referring to their track record in relevant cases. It’s very
little use to quote physicists on political issues or politicians on issues in
physics. Nevertheless, there is a tendency to quote scientists outside their
specialty fields as if their scientific approach could be relied on to transfer
across the boundary of their specialty. Although there are some outstanding
examples of scientists with this kind of general rational approach, there are
a tremendous number of them who are in absolutely no way superior to any
other functioning adult with respect to topics outside their fields of exper¬
tise: they are just as subject to prejudice, to lack of relevant knowledge, and
to overestimates of their own acuity. (This, by the way, is certainly a self-
referential statement.) One footnote might be useful: It is worth attempting
to distinguish between the concepts of credibility (of an authority, a witness,
or other qualified person), reliability, and consistency. Reliability (sometimes
called validity) refers to the “absolute” question of whether or not the au¬
thority is right. That’s what we might be able to get from an estimate of
prior track record. In the absence of direct evidence of reliability, we tend to
fall back on two rather weaker indicators of it. Consistency is one of them. It
breaks out into (1) internal consistency, and (2) consistency with others. It’s
a sign of probable lack of reliability, if there is substantial internal inconsis¬
tency. (But some of it can be tolerated, though at the price of not being able
to tell what is being said on the specific points on which there is inconsis¬
tency.) It’s also something of a point against a witness or an authority that
there is poor agreement with other alleged authorities. But again, it s not
really a fatal objection. It just means that a third party, an observer or
listener, isn’t able to make a decision: think of all the court arguments
between “expert” psychiatric witnesses as to the sanity of the accused. Fi¬
nally, there is the question of the credibility of the witness, which will likely
be assured by evidence of reliability, but might require something more than
that. Credibility is a function of the extent to which the audience which we
are currently addressing believes the witness; and that may mean that con-
228 CHAPTER 7

siderations of the extent to which the audience can understand certain tech¬
nical language that the witness uses, the extent to which the witness appears
forthright in manner, meets the eyes of the interrogators, lacks any predis¬
posing personal interests in this case, and so on, may come in, even though
they are, strictly speaking, very, very weak and possibly zero indicators of
reliability. That is, when we are talking about credibility, we have to talk
about what the audience thinks indicates reliability, and not just what does
indicate reliability. This is a serious consideration for lawyers planning
which witnesses to call. In the same way, when you are thinking about
giving an argument in front of an audience, you have to think about what
the audience will understand, not just about what will constitute a valid
argument.
The other side of this divergence between reliability and credibility
(which is not an ideal state of affairs) is that a ruthless dialectician or debater
can often make points illicitly against an opponent by exploiting the extent
to which the audience can be persuaded not to believe someone by means
which should not actually impugn their reliability. Into this category fall
most of the so-called ad hominem moves: attacks on the person rather than
the argument. The attempt to disparage testimony by showing that the mor¬
al character of the witness is questionable, is sometimes appropriate, but it is
often used when it is not appropriate. For example, the sexual life of a
plaintiff, including such matters as whether the person is engaged in prosti¬
tution, has traditionally been brought in at rape trials, although this is, one
might well argue, improper, since it has an effect on the jury that is quite
disproportionate to its relevance to the credibility of the witness. Its use is all
too often a case of the exploitation of prejudice, and it has no place in sound
reasoning or justice.

QUIZ 7-A

1 Suppose that we want to take a poll on the subject of student opinion on arming
campus security guards with firearms, and we set ourselves up to interview people
at a convenient campus location, let us say near the entrance to the campus
cafeteria. Now we interview every tenth person who comes through the door.
That looks like a pretty unbiased approach; but is it?

Answer

The population that you wish to generalize about is all students. The crucial question
is whether a representive sample of all students will come through the campus cafe¬
teria during the time (unspecified) that we sit there. In the first place, the answer is
definitely no, since those who bring their own food to campus and eat in offices or
classrooms will not show up; and they may well be different from those who use the
cafeteria with respect to the question that we’re interested in, since they will certainly
SPECIAL TYPES OF ARGUMENT 229

be different with respect to work habits and socioeconomic status, which are quite
likely to correlate with attitudes toward authority. For example, the students who see
themselves as important components of the campus social system tend to do a good
deal of their socializing in the cafeteria building. These students are more likely to be
identified with student government’s opinion on campus than those who are rela¬
tively more socially isolated and hence likely to be those who eat by themselves
without placing the need for human company very high. Similar points may be
relevant about those who frequent more expensive eating places than the cafeteria.
Again, since the time of our interviewing isn’t specified, the extent to which we pick
up what may be a substantial population of evening students cannot be determined.
And it is well known that evening students tend to come from a different age group,
social class, and motivational group from daytime students.

2 Is a Women’s Studies department entitled to restrict applicants for its faculty to

women?

Answer

No, attractive though the compensatory argument seems (that is, the argument that
women have been underrepresented on college faculties for a very long time and
should at least be given the chance to have a preeminent position in the line for jobs
in “their own” department). The most attractive intellectual type of argument (as
opposed to the social-political one just mentioned) is that only women would have
the kind of experience that would form an adequate basis for teaching the kind of
subjects that are covered in a Women’s Studies department. However, that argument
is not satisfactory since there are two powerful considerations that suggest the appro¬
priateness of at least some men on the faculty of any such department that aims to
provide a reasonably comprehensive view of the kind of topics taught in the depart¬
ment. First, the argument can be made that, at least, “the other side” should be
heard from occasionally; that is, it is a characteristic of courses under Women’s
Studies that they present a rather strong profeminist point of view. It’s not unreason¬
able to require that a total program provide at least some place for a discussion of
either the feminist point of view by someone other than a woman, or even a nonfem¬
inist or antifeminist point of view. The second argument is a pedagogical one: At
least part of the audience for courses on women will be men, or, at the very least,
part of the “audience” with which women will have to interact will be men; a
representation of men will therefore have some advantages in making the learning
more effective, particularly for men (since they will have at least one role model
available), but also for women (who also must deal with men).
You may have chosen one of the other questions. Each has its own special
features, and the answer isn’t the same in all cases; or at least it’s not the same to the
same degree. Read the preceding answer carefully for clues to the question which
you tackled. (See question 5 in Quiz 7-C.)

3 Comment on Ehrlichman’s analogy that was presented toward the end of Section
7-3 (p. 213).
230 CHAPTER 7

Answer
The direct attack consists here in saying that there’s a complete difference in the
potential importance of the evidence that would be obtained by the illegal procedure
in the two cases. In one case, the information was to be used to denigrate political
opponents’ criticisms of the government. In the other case, the lives of hundreds of
thousands of people would be saved. It is a particularly significant fact that such a
difference should have been completely overlooked by Ehrlichman.
The analogy can be turned on itself by arguing that if this kind of analogy is
thought to be satisfactory, then doing anything that would further the cause of the
political party in power would be justified, since it is clear that this is being identified
with the welfare of hundreds of thousands of people, perhaps with the nation as a
whole. It would surely then be quite appropriate to imprison vociferous critics of the
government in power on the ground that they were causing serious disruption in the
society. In short, the standard justification of dictators! (The law is treated as a tool
for the party in power, rather than as a protector of all people alike.)

4 How might you explain Uri Geller’s “trick” of starting up watches that have
stopped running by holding them in his hands in order to “infuse them with
psychic energy”—without appealing to the supernatural?

Answer

The most usual reason for a wristwatch or pocketwatch to stop running, and still be
carried, is that the lubricants have become exhausted or clogged up. (If you drop a
watch, you know why it stops running and you don’t carry it around from that point
on. But watches that stop through lubricant failure do so intermittently, and people
commonly carry them around for quite some time in the hope that they’ll restart.)
The warmth of somebody’s hand is often enough to thin the lubricant substantially,
so that some of it will flow back to the points that need lubricating and the watch will
start running again, though usually only for a short time. Whether or not this is the
actual explanation of Geller’s feat is another question. To settle it, one would have a
watchmaker look at the watches prior to their being passed to Uri Geller and thereby
test our hypothesis as to the cause of their stopping; and one would also try giving
half of them to somebody without any psychic powers but with very hot hands, to see
if that person were just as successful.

5 “Professors should not begin a discussion of controversial topics by setting out

their ‘prejudices,’ because this demonstrates professional incompetence” (ex¬
panded at some length in the full quotation in Section 7-4, p. 223).

Answer

Although a good point is made here, the line is perhaps a little severe, since one may
have general grounds for feeling that one’s enthusiasm for certain conclusions
slightly exceeds the strength of the evidence, though one hasn’t been able to identify
a weakness in one’s line of argument. In such a situation, it might be argued that one
SPECIAL TYPES OF ARGUMENT 231

should both recognize that one probably does have a prejudice here and refuse to
concede incompetence, since one believes also that the argument possibly is correct;
thus one is simply erring on the side of safety and honesty, surely not a bad idea.
However, it is hardly a fully scholarly or fully professional response; surely one
should have reached the point of being able to identify an adequate argument by the
time one is lecturing on it; and if so, one should cut back on one’s enthusiasm for a
position where the argument does not seem to support the enthusiasm. So the quoted
argument seems to be slightly too strong, but not very much so.
Note this is a self-referent argument. What follows from this about the answer
I’ve proposed above?

6 Suppose we discover there’s a very much higher incidence of cancer of the uterus
among women who are using birth-control pills than among those who are not
(that is, we have a high correlation). Do we have good grounds for inferring a
causal connection? If not, exactly why not, and how could we settle the issue?

Answer

The mere correlation does not give us grounds for inferring a causal connection in
general, and in this particular case, it would do so only if no alternative explanations
can be thought of. But it’s nearly always possible to think of some, which is why the
general inference from correlations to causal connections is unsound. In this case, for
example, there’s a whole string of possibilities. For instance, women using birth-
control pills are in a considerably higher economic bracket than those not using
them. They will therefore be more likely to be eating certain types of food and
engaging in certain types of occupations (particularly in cities) and in different types
of sexual activities from those who do not use such pills. So, unless we can rule out
all those other alternatives, we cannot begin to make a case for a causal connection.
How could we rule out those other alternatives? The easiest way would of course be
to use a fully controlled experimental design; almost anything else will only provide
weak contributory evidence. Notice that in this case, unlike the smoking case, we can
avoid both moral and practical difficulties in the setting up of such a design; we can
use volunteers who are willing to be assigned to one of the groups in order to serve
the cause of discovering whether these pills do produce cancer.

QUIZ 7-B

1 “Drawing conclusions from information about the mean of a population is just as

speculative, or at least probabilistic, as drawing conclusions from data about
samples. So what’s the point of the distinction between description and inferential
statistics?” Analyze as an argument as well as, or in the course of, answering it as
a question.

2 Is it statistics, experience, or both that tell you whether a sample of 1,000 or

10,000 citizens is necessary in order to get a good “reading” of attitudes toward
the number of children most people want to have?
232 CHAPTER 7

3 The Texas Bureau of Wild Life Management gets its estimates of desert mammal
population by counting the number of carcasses found beside a sample stretch of
the desert highway. Discuss the pros and cons of this procedure.

4 Are you ever justified in advertising a job for which you stipulate that only whites
should apply?

5 and 6 See last paragraphs of Section 7-4, pp. 224-225.

QUIZ 7-C

1 Do electrons really exist, or are they just a “useful fiction”?

2 To what extent and how (if at all) are reasoning skills involved in the creative
part of scientific theorizing?

3 Why is random allocation of subjects to the experimental and control groups

important if one is trying to design an experiment that will identify a cause?

4 What’s a test of the honesty of someone who refuses to employ long-haired hippie
types in a restaurant on the ground that their hair might get into the food?

5 Answer one question you have previously not answered in the last paragraph of
Section 7-2, p. 210.

6 Find an argument from analogy in an advanced textbook; set it out carefully and
assess it.

7 What other way, besides thinking up counterexplanations, can lead to refuting an

argument? (See p. 217.)

8 What’s the remote possibility that makes the experimental design of the second
paragraph of p. 224 not quite watertight? How could you tighten it up? Could
you then be confident of the conclusions?
 Chapter 8

Extensions and
Ramifications

This chapter is just a list of suggestions about directions to go and connec¬

tions of this subject with others.

8-1 SPEED REASONING

Argument analysis is a laborious business. One scarcely ever gets a good

analysis of a serious argument until the third draft. Nevertheless, getting
some good points can often be done quickly, and the more practice one has
at doing thorough analysis and having it criticized, the faster one becomes at
hitting key points swiftly and expressing them clearly and forcefully. For
some years I have taught a course called Speed Reasoning at the University
of California Extension in San Francisco, assisted at various times by Mary
Anne Warren and Bob Ennis. In it, we try to compress a survey of the basic
structure of argument analysis and some skill practice into two full days.
What remains in most participants’ minds from this compression is the out¬
line of the procedure and some ways to do or avoid key moves, plus the
conviction that “it can be done,” i.e., that quick but reasonably incisive
233
234 CHAPTER 8

analysis can be achieved if enough practice is undertaken. Nobody first

watching a film of one of Nadia Comaneci’s gymnastic routines on the bars
would expect to be able to do it without practice, if then, and intellectual/
verbal skills are no less complex to learn. The extent to which either the
physical or the verbal skills depend on some inherent ability is not clear, but
most people who work on reasoning skills, even for just two intensive days,
show very substantial improvement. Getting a systematic approach and a
belief in the possibility of success is half the battle; developing a repertoire
of standard traps and counterexamples and the judgment to distinguish be¬
tween nit-picking and incisive criticism is most of the rest. You have to go
for sound basic procedures before you can go for speed; the speed comes
with practice, especially practice under time pressure.
One way to improve your skills and your speed is to tape-record some
debates or speeches on controversial issues, and then work on them by play¬
ing back a piece at a time. Gradually, you’ll be able to handle each segment
of the argument faster, and then you can move to bigger segments, until you
can give good replies in the time you would actually have had in real life.
Such TV interview shows as “Meet the Press,” “Firing Line,” or “The Advo¬
cates” are excellent for this; some of the radio “call-in” shows are, too.
Another practice hint is to make yourself find one example of practical
prose to criticize (or praise, for statable reasons) each time you read the daily
paper. Even at first, the ads and the letters column will be easy meat. Then
you’ll progress to the editorials and the columnists and then the news analy¬
sis features. Do not “do it in your head”! Cut the passage out and tape it to
a sheet of paper, mark numbers on it, structure the argument—do the whole
thing. Keep a file of these efforts. Look back after a month or so, and you
will be amazed at the quality of improvement and at the speed improve¬
ment.
Doing this with a friend or a group to help is much better. People from
the Speed Reasoning course (the Speed Reasoners) have continued to meet
in order to sharpen up one another’s performance. They started this activity
spontaneously and it has proved both durable and enjoyable, as well as
effective.

8-2 DECISION STRATEGIES AND REASONING

One of the problems when you get to “the bottom line” of an argument
analysis is to decide what action to take. Of course, what risks you take (i.e.,
how strong an argument you need in order to justify an action) depends on
the possible rewards and losses involved. The whole theory of decision mak-
ing (and the "theory of games”) has developed in the past few decades to
assess the kinds of strategies one can adopt in such situations and is a natu¬
ral extension of the study of reasoning, as well as part of the social sciences.
EXTENSIONS AND RAMIFICATIONS 235

Within economics, the subject of welfare economics involves the exploration

of a number of ways to make a rational subject out of the value-laden
private and public policy decisions affecting our economic behavior. Many
important distinctions have been developed among types of decision strategy
for personal and social decision making. You might want to look up the
meaning of “minimax strategy” and “Pareto optimality” in, for example the
International Encyclopedia of the Social Sciences. Related issues about con¬
trolling standards for choices come up in theoretical ethics, and the “cate¬
gorical imperative” (proposed by Kant) is one example of the attempts to
provide general decision strategies there. The Golden Rule is another.
As an exercise, you might see whether you can state the principle of
choice that justifies your crossing a city street, knowing that you might be
killed by a car running a light, when you could avoid the crossing by walk¬
ing half a mile further to a footbridge. Does that principle of choice (e.g.,
“One should ignore very small risks of death entirely”) show that the sup¬
porters of the speed limit or of the ban on atomic power plants are irratio¬
nal? (Quiz 8-B, 1) Is it always irrational to buy lottery (sweepstake) tickets?
If not, why not? That is, what principle of choice justifies it? Is the same
principle inconsistent with the one quoted (in parentheses) in Quiz 8-B, 1?
With the one you gave as your answer to 8-B, 1? With refusing to participate
in a cham-letter scheme? (Quiz 8-B, 2, all of them)

8-3 BEHAVIOR CONTROL—THE REASONABLE PERSON

The preceding section concerns principles of choice, and is thus still fairly
academic compared to practical suggestions, though it’s a step in the practi¬
cal direction since it’s at least concerned with choices rather than beliefs. But
there still remains the problem of translating choices into actions; and in
many cases, it’s a real problem. One may rationally “choose,” that is, “de¬
cide” to give up smoking, but one may still be a long way from being able to
implement the choice. The true addict is often said to be wholly incapable of
implementing that decision.
Just as one can make a study of decision strategies, one can also make a
study of implementation strategies, of how to get the choice or the decision
translated into action. And this study is what leads to a number of the
expensive services that are now available to assist you in improving your
reading or study habits, reducing weight, getting exercise, or giving up
smoking, drinking, drugs, or gambling. These “applied behavior modifica¬
tion” approaches are often somewhat more optimistic in their advertising
than their results justify, but they are also very often effective—far more
often, on the face of the evidence, than most traditional approaches, like
psychoanalysis, which are sometimes used to tackle these “behavior prob¬
lems.” It is part of a complete approach to reasoning to look carefully at this
236 CHAPTER 8

area, since it constitutes the slip between the cup and the lip for many
decisions. Because that slip so often occurs, people are often motivated to
judge reasoning as largely irrelevant to practical action. Of course, it
shouldn’t be, and it needn’t be, provided one follows through all the way to
the action level, at every step applying the standards of rationality.
For example, one can rationally analyze the problems of the addict,
noticing that the loss of the capacity to act in accordance with reason is
extremely localized. Alcoholics are often said to be unable to resist the offer
or opportunity of a drink when under stress. (Even more plausibly, they are
said not to be able to resist the chance of a second drink after taking one.)
But alcoholics have not lost the capacity to pick up a telephone, or any of
their other abilities (except in very advanced cases). Hence, it is perfectly
possible for an alcoholic to call for help when feeling the temptation to drink
is about to become overwhelming. One of the Alcoholics Anonymous strat¬
egies is to provide a hot line to an emergency squad who will rush right over
and provide diversions and support to prevent a drinking episode from be¬
ginning. That’s a rational solution to the problem, not a sign that reason is
irrelevant to practical problems.
We have often spent time in class on “feasible pathfinding” procedures
for coping with irrational behavior of this and other kinds. The first point to
convince oneself of is that the conventional “addict” is not very different
from you and me, except for a matter of degree. There is no real basis for the
idea that addicts are suffering from some physiological hypersensitivity that
makes them either “sick” or victims of misfortune and hence not responsible
for their situation. There’s nothing physiological about gambling, and gam¬
bling addicts have just the same general symptoms as other addicts. There’s
nothing physically special about water or Coke, but there are known cases of
addiction to both; the symptoms are grossly excessive consumption or use
despite serious adverse side-effects and the wish to “cut down,” plus with¬
drawal symptoms if the supply is eliminated. Start looking at your own
behavior: If there’s overeating, staying up too late to watch bad TV shows,
reading garbage-type escapist literature, playing pool when homework
presses, or compulsive buying sprees, you’re an addict. Finding your way
out of such habits (“addictions”) is a matter of finding a behavior pathway
through a sequence of those choice points where you can make and carry
out the “best” decision, which will eventually lead you to safe territory.
Suppose you have to finish a homework assignment; it’s important, it’s fea¬
sible, it’s not even all that dull. But you know from previous experience that
there’s just no way you’re going to get it done by just going home and doing
it. There are too many alluring alternatives, or even unalluring ones that
begin to look alluring. Start looking for ways to commit yourself to some
course of action which (1) you can perfectly well do, and (2) will lead to
doing the assignment. For example, you can easily decide to spend the eve-
 EXTENSIONS AND RAMIFICATIONS 237

ning at the library (public or campus), and that may restrict your goof-off
tendencies enough to get the job done. If that isn’t tough enough, maybe you
have to make a contract with someone in the same situation, with $20 riding
on it and whoever doesn’t do it paying the other. If $20 won’t work, maybe
$50 will. Or can you arrange a study date, the deal being three hours’ work
before going for a late snack? Here your self-respect when under the eyes of
someone you value is the lever you use on yourself. And so on—the proce¬
dure ot pathfinding is analogous to assumption hunting or argument con¬
struction in that you re trying to balance off the feasible with the successful,
and your imagination can suggest to you a series of pathways to be explored.
Very few people, if any, are beyond the possibility of finding their way out of
the bad-habit maze, certainly on important occasions and, if it’s serious
enough, for good. As usual, it helps to make it a joint program with someone
else.
Doing this with respect to bad habits of thought as well as bad “social
habits” is part of what we expect of a “reasonable person.” The phrase
connotes a lack of excesses, of prejudices; a commitment to moderation (not
meaning passivity or stick-in-the-middle-of-the-road), as well as to rational
thought at the preaction level.

8-4 FLOWCHARTS, FORTRAN, FORMALISMS,

AND FORWARDS

Where else can you go from here? Perhaps straight on to some professional
program, perhaps in the humanities or the arts, in pure science, or in teach¬
ing. But there are some little excursions, apart from those already listed, that
are worth considering and that carry on the development of your own skills
in, and understanding of, reasoning. The obvious ones are the statistics and
experimental design courses, which expand on some of the scientific infer¬
ence discussions we’ve started here; and the symbolic logic courses, which
develop the abstract principles of certain types of formal reasoning into
subjects of great interest in their own right. Perhaps less obviously, there are
courses in business schools and computer science departments that also tie
in. Learning a computer programming language, like FORTRAN, APL,
ALGOL, BASIC, or any other, involves an intellectual discipline that de¬
serves more recognition than it has so far received as a branch of practical
reasoning. It would be my guess that you would probably be more likely to
improve your problem analysis skill (at least) if you pick up a programming
language than a formal logic course or two. In any case, it’s worth consider¬
ing, especially because of its tremendous “instrumental utility,” that is, be¬
cause it is so valuable for many other reasons, including research, business,
and applied science.
238 CHAPTER 8

Somewhere or other on a campus, someone is (usually several people

are) teaching flow-charting. It may be, and probably is, part of the computer
science course, at least in one version. But you may also find it showing up
in a management course in the business school, perhaps under the name
PERT-charting. The whole subject of analyzing sequences of decisions and
actions, evaluations and costing, in a way that clarifies responsibilities and
lead-times, has been extensively developed in recent years and, if not used
obsessively, can be very valuable. It is just an extension of the reasoning
approach mto a particular area.
Those of you interested in the fiscal dimension of reasoning—for exam¬
ple, cost analysis in various forms, and its ties to management through var¬
ious management information systems such as PPBS (Program Planning and
Budgeting Systems), will find these covered, relatively superficially, in ac¬
counting courses, and relatively better in management courses with a strong
practical emphasis. The careful working out of the arguments for alternative
approaches to the “discount rate of return,” for example, will prove fascinat¬
ing for those of an analytical and inquiring (and practical) turn of mind,
though far, far removed from the usual academic company of a logic course.
Somewhere on a big campus, perhaps in the School of Public Policy,
perhaps in the School of Education, you’ll find a course on evaluation, which
will pick up and develop the whole procedure of reasoning-to-an-evaluative-
conclusion, something we’ve talked about in connection with the last step of
argument analysis where you are evaluating an argument, and on a number
of other occasions. Evaluation as a discipline is pretty new, or at least has
taken on a new form in recent years; you’ll find it a very natural extension of
what we’ve been doing here.
And there are other such extensions: operations research is one; some
of the courses in applied ethics, in propaganda analysis, and in rhetoric are
others; even some courses in historiography. They turn up in many disci¬
plines, whenever someone begins to look hard at the reasoning process.
If you are within range of a law school, you may find someone teaching
legal reasoning (the former Attorney General of the United States, Edward
Levi, is the author of a brief and highly respected book on that subject).
Nobody ties all these together so that you can major in reasoning, but that’s
an idea worth considering, and, on most campuses today, special interde¬
partmental majors can be arranged.
If you’re not on a campus, it’s easy to convert all that I’ve been saying
into a reading program and a library development program of the greatest
value to you now and later.
This is one subject which, approached sensibly and seriously, will never
be short of credentials at both the academic and the practical end. May it
continue to attract you and infuse your thoughts and actions!
EXTENSIONS AND RAMIFICATIONS 239

QUIZ 8-A

Questions 1 and 2 (several parts) will be found in Section 8-2, paragraph 2.

For Question 3, see if you can select, and justify your selection of, one further
course (or solid text) that you will undertake in order to improve your reasoning
skills. Make it one you would not have chosen if you had not taken this course (or
read this text).
Index
 Index

Ad hominem fallacy, 228 Argument(s):

Affirming the consequent, fallacy of, criticism of (see Argument analysis,
141, 148 Step 5)
“Air support,” as euphemism, 152-153 in debates and dialogues, 225-228
Ambiguity: deductive, 196-197
distinguished from vagueness, distinguished from argumentative
121-122, 124-125 statements, 67
and fallacy of equivocation, 122-123 distinguished from conditional
Analogy, argument from, 210-215 statements, 60-61
difficulty of criticizing, 212-213 distinguished from explanations,
models, use of, 213-215 95-97
Apache nation, argument for distinguished from implications,
reparations to, 175, 188-191 46-47
“Apple,” defining, 129-131, 145-146 distinguished from inferences, 55-56
Argument(s): emotional language in, 70
from analogy, 210-215 factually sound, 38
assessing, basic criterion for, 30-35 function of, 55-56
chain, 60, 77, 140 good, conditions of, 56-57

243
244 INDEX

Argument(s): Argument analysis: Step 4-formulating

identifying, mistakes in, 66-71 unstated assumptions:
logically sound, 38 by converting inferences into
statistical, 197-205 premises, 56, 83-86, 89-93,
(See also Argument analysis; 178-179
Inferences; Reasoning) distinguishing significant from
Argument analysis, 39, 156-157 insignificant assumptions,
abbreviations in, value of, 157-160 162-175
acquiring skill at, 54-55, 71-72 general procedure, 43, 44
of arguments rejecting opposing minimal, 43, 166-167
arguments, 161-162 in negative form, 86-88
of complex arguments, 59-60, optimal, 43
157-161 value of considering other relevant
ethics of, 71-73 arguments, 189
role of judgment in, 156 (See also Premises)
Step 1-clarifying meaning, 39, 75-78, Step 5-criticizing premises and
89-90, 102-103 inferences: of argument from
common mistakes in, 106 analogy, 210-215
definitions and their criticism, factual distinguished from logical
criticism, 73-75
123-134, 144-145
general procedure, 43-44, 75-94,
distinguishing definitional from
155, 215
factual truth, 134-139
strong distinguished from weak
general procedure, 40, 103
criticism, 72, 155, 184-185, 226
importance of, 70-71
Step 5a-criticizing premises, 39,
(See also Definitions; Meaning)
43-44,81
Step 2-identifying conclusions, 39,
with counterexamples, 44, 176-179,
73, 77-79
189-192
of arguments rejecting opposing
in nonexpert areas, 74-75, 89-93
arguments, 161-162
requiring realistic alternatives,
general procedure, 41, 76-77, 156
171-172
implicit, 152
strong distinguished from weak
importance of, 70-71
criticism, 136
(See also Conclusions)
(See also Definitions, criticizing)
Step 3-portraying structure, 39,
Step 5b-criticizing inferences, 39,
79-80, 82
43-44, 81, 82
abbreviating complex structures,
by converting into assumptions, 56,
157-160
83-86, 89-93, 178-179
of arguments rejecting opposing
with counterarguments, 177-180
arguments, 161-162
with counterexamples, 178-179
general procedure, 41-43
by introducing alternative
Step 4-formulating unstated
arguments, 179-180, 188-192
assumptions, 39, 81-82, 92, (See also Inferences)
188-192
Step 6-considering other relevant
of arguer, 43, 85-86 arguments, 39, 88, 93
avoiding straw-man arguments, general procedure, 44-45
163, 166-167 value of, 179-180, 189
 INDEX
245

Argument analysis: Step 7-overall Causes:

evaluation, 39, 45, 88, 93, 180 correlations distinguished from causal
value of, 185-187 connections, 223-225
[See also Argument(s)] distinguished from reasons, 65-66,
Aristotle, 37 95-97, 219-223
Arithmetical errors, 83-84 Certainty, possibility of, 119, 221-222
Assumptions (see Argument analysis. “College,” defining, 144-145
Step 4; Premises) Computer language and reasoning,
Astrology: 237-238
provability of assertions in, 139, Conclusions:
147-148 distinguished from premises, 77
validity of, 218 implicit, 152
Authorities, reliability of 227-228 intermediate, 77
Automobiles: truth values of premises and, possible
emission requirements for, argument combinations of argument
for postponing standards,
soundness with, 56-60, 94-95
178-179
(See also Argument analysis, Step 2)
engine, “properly tuned,” defining,
Conditional statements:
136-137
distinguished from arguments, 60-61
EPA mileage figures for, precision of,
as equivalent to generalizations,
104-105
185-187
Averages, types of. 205
inferences expressible as, 61
Aviation safety, argument regarding,
Conditions:
67-70
antecedent and consequent, 62
equivalent, 64
(See also Criteria; Necessary
Bald-man argument (see Slippery-slope
conditions; Sufficient conditions)
argument)
Conservation of Luck claim, clarifying
Ballpoint pens, analysis of longevity
claim for, 102-103 meaning of, 112-114
Consistency:
“Because,” uses distinguished, 65-66
Behavior guidance and reasoning, and basic criterion of argument
221-222, 235-237 assessment, 30-33
Bias: and knowledge, 35
error and the likelihood of, 221-223 as requirement of communication, 31
reasoned conviction confused with, (See also Inconsistency)
208-210 Consumers Reports, 114
in science, 5, 216, 218-219 Converse relationships, 63
Birth-control pills and cancer, inferring Correlations, causal connections
causal connection from correlation, distinguished from, 223-225, 231
224, 231 Counterarguments, 177-180
Black-and-white fallacy (see Slippery- defending against, 182-184
slope argument) Counterexamples, 44, 91, 129, 144
defending against, 181-182
use in criticizing arguments, 177-179
Cadillac and “getting what you pay Creativity, reasoning and, 35-37, 170,
for,” 116 215
246 INDEX

Criminal insanity, argument Emotional language, not always

equivocating on, 122-123 grounds for argument criticism, 70,
Criteria: 106
and definition, 130-134 Environmental Protection Agency
distinguished from accidental (EPA) mileage figures, advertising
properties, 131-132 claims regarding, 104-105
distinguished from necessary Equivalent conditions, 64
conditions, 130-133 Equivocation, fallacy of, 122-123
Erlichman, John, 213
Ethics, reasoning in: arguments from
analogy, 210-215
Debates, features of arguments in, bias in, 222
225-228 inferences from generalizations to
Decision strategies, 235-236 individuals, 205-210
Deductive arguments, 196-197 Euphemisms, 152-154
Definitions: Evolution, general principle of,
and criteria, 130-134 clarifying meaning of, 137
criticizing, 125-129, 144-145 Exercises, scrambled and uncued, value
demand for, not always legitimate, of, 15-17, 22, 24-25
106, 143 Explanations:
dictionary, extended circle of, by causes, distinguished from
145-146 correlations, 223-225
ideal, 123-125, 130-132 as conveying understanding, 218-219
implicit, 142 creative element in finding, 215-216
and necessary conditions, 127-133 distinguished from arguments, 95-97
practical (real), 125, 129-134 distinguished from justifications,
of specific terms discussed: “apple,” 65-66, 95-97, 219-223
129-131, 145-146 distinguishing satisfactory from
“college,” 144-145 unsatisfactory, 218-219
“good teacher,” 126-127 Extrasensory perception, research in, 5,
“properly tuned engine,” 136-137 215-216
“triangle,” 123-124, 126, 128, 130
“Democracy,” ambiguity of, 121-122
Denying the antecedent, fallacy of, Fairlie, Elenry, “The Language of
140-141 Politics,” 153
Depression, argument regarding Fallacies:
indicators of recovery from, 77-81, ad hominem, 228
89-93 affirming the consequent, 141, 148
Determinism, hard, 220-221 denying the antecedent, 140-141
Dialogues, features of arguments in, equivocation, 122-123
225-228 iron-man, 167-168
slippery-slope, 117-121, 148-149
stereotyping, 208
straw-man, 71-72, 81, 85, 163, 226
Economics, decision strategies and avoiding, 166-167
reasoning in, 235 Federal Aviation Administration
Education, distinguished from (FAA) air worthiness directives,
propaganda, 118-119 argument regarding, 68-70, 74
INDEX 247

Fingarette, Herbert, The Meaning of Inconsistency:

Criminal Insanity, 122 types of, 31-32, 47-48
Flowcharting and reasoning, 237-238 (See also Consistency)
Folk wisdom, 218 Inference-licenses, 83
Ford, Gerald R„ 88, 90 Inferences:
assassination worry, weak criticism contextual, 107
of, 72-73 convertible into assumptions, 56, 61,
Free will, 220 83, 89-93, 178-179
defined and distinguished from
implications, 38-39
Gallup, George, 197, 199 distinguished from arguments, 55-56
Geller, Uri, 216. 230 distinguished from conditional
Generalizations: statements, 60-61
bias in application to individuals of, sound, 56
205-210 soundness of, possible combinations
reformulating, to defend against of premise and conclusion truth
counterexamples, 181-182 values with, 56-60, 94-95
from samples, 197-205 statistical, 197-205
Grades: (See also Analogy, argument from)
bad, as form of punishment, (See also Argument analysis, Step 5)
argument for, 180-181, 191-192 Information content, 109-116
given by peers, value of, 13-15 Instincts, reasoning and, 3-4
inflation of, dealing with, 18-19 Interrogating prisoners, argument
value of, 18-21, 185-187 regarding use of torture in,
118-121
Iron-man fallacy, 167-168
Hegelian synthesis, 225-226
Homosexuals and security risks,
argument regarding, 167, 169-173, Justifications:
188 distinguished from explanations,
Hypothesis (see Conditions, antecedent 95-97, 219-223
and consequent) moral, 219, 220
Hypothetical statements (see Condi¬
tional statements)
Hypothetico-deductive reasoning, 197
Kant, Immanuel, 235

Imagination (see Creativity)

Implementation strategies, 235-237 Language:
Implications: computer, and reasoning, 237-238
distinguished from arguments, 46-47 and political power, books on,
distinguished from inferences, 38-39 154-155
of euphemisms, 152-154 used by government in regard to
of personal pronouns, 154 Vietnam war, 152-153
Inconsistency: Law, reasoning in, 210
avoidance of, as basis of Lipton’s Tea advertisement, analyzed,
communication and argument 46-47, 49-51, 71
assessment, 30-35 Literary Digest poll, 201-202
248 INDEX

Logic: Nielsen ratings, 201

defined, 35, 37-38 Nuclear power stations, argument for
and emotional language, 70 banning as unsafe, 174-175
principles of, 139-141
[See also Argument(s); Argument
analysis; Fallacies; Reasoning] Oil prices, argument for decontrolling,
86-88
Opfer, Colonel, 153
Macpherson, C. B., Democratic Theory,
122
Marijuana:
Parole, argument against, 117-118
arguments for legalizing, 148-149
Plato, 225
experiment on the effects of 45-46
Polls, opinion, choosing samples for,
Marx, Karl, 219
197-205
Meaning:
Precision:
clarifying, to defend against
information content and, 109-116
counterexamples, 182-184
necessary distinguished from
connotations and associations, 143
unnecessary, 105-109
contrast theory of 109-112
real distinguished from fake, 104-105
criteria clusters and, 130-134
value of, 138-139
dimensions of, 141-143
Predictions, basis for, 224-225
distinguished from implications,
Prejudice:
152-156
error and the likelihood of, 221-223
and information content, 109-116
reasoned conviction confused with,
practical distinguished from abstract,
208-210
107-108, 111-112
in science, 5, 216, 218-219
shift: to defend against counter¬
Premises:
examples, 182-184
fallacy of 122-123 distinguished from conclusions, 77
(See also Argument analysis, Step 1; inferences convertible into, 56, 61,
83-86, 89-93, 178-179
Definitions)
Mercedes 690 SEL and “getting what truth values of conclusion and,
you pay for,” 115 possible combination of
Metaphors, 210-211 argument soundness with, 56-60,
“Military advisors,” as euphemism, 94-95
152-153 (See also Argument analysis, Step 4)
Models, 213-215, 219 Principle of Charity, 71-73, 76-77, 175
Psychobabble, 155-156

Necessary conditions:
and definitions, 127-133 Quizzes:
distinguished from criteria, 130-133 general instructions for, 6-7, 9
distinguished from sufficient value of, 13, 20-21
conditions, 61-65, 95
types of, 64, 134-136, 146-147
Necessity, factual distinguished from Racism:
definitional, 146-147 and bias, 222
(See also Necessary conditions) and improper reasoning, 207-208
INDEX
249

Reading, relationship to reasoning, 1-2, Samples, statistical: representative,

12-13 197-205
Reasonableness, 4, 235-237 stratification of, 204-205
Reasoning (this book): Science, reasoning in: analogy,
general instructions for quizzes, arguments from, 210-215
6-7, 9 errors in, 75
helpfulness of, 5-6 hypothesis evaluation, 197
need for student effort, 21-22 models, use of, 213-215
Reasoning Prize competition, 6, 13 prejudice in, 5, 216, 218-219
unusual aims and features of, 12-23 statistical inferences, 197-205
Reasoning: Sexism, 183
from analogy, 210-215 implications of personal pronoun use,
as best guide to truth, 3-5 154
bias in, possibility of, 205-210 and improper reasoning, 207-208
courses which develop, 37-38, Similes, 210-211
237-238 Skinner, B. F., 219
and creativity, 35-37, 170, 215 Slippery-slope arguments, 117-121,
in decision making and behavior 148-149
control, 234-237 Smoking and cancer, proving causal
developing speed in, 233-234 connection, 223-225
as distinctively human, 2-3, 7 Sophists, 227
emotional language and, 70, 106 Statistical inferences, 197-205
ethical (see Ethics, reasoning in) (See also Analogy, argument from)
hypothetico-deductive, 197 Statistics:
and instincts, 3-4 descriptive, 197, 198, 205
nature of, and conditions of good inferential, 197, 205
argument, 30-35, 57, 180 Stereotyping:
relationship of reading to, 1-2, 12-13 fallacy of, 208
scientific (see Science, reasoning in) legitimate, 208
statistical, 197-205 Stern-Gerlach experiment, 214
uses and value of, 3-4, 8, 29-30, Straw-man fallacy, 71-72, 81, 85, 163,
35-37 226
[See also Argument(s); Argument avoiding, 166-167
analysis; Inferences; Logic] Sufficient conditions:
Reasoning Prize competition, 6, 13 and definitions, 127-130
Reasons, distinguished from causes, distinguished from necessary
65-66, 95-97, 219-223 conditions, 61-65, 95
Redheads and quick tempers, argument logically, 64
regarding, 81-83, 85-86
Rhetoric, 3, 226-227
Rolls-Royce Camargue and “getting
what you pay for,” 115 “Tall,” clarifying meaning of, 107-112
Rosen, R. D., 155 Teachers:
argument for announcing
controversial views in advance,
Samples, statistical: adequate, 197-205 222-223, 230-231
random, 203-204, 216 “good,” defining, 126-127
250 INDEX

Thin edge of the wedge (see Slippery- Vietnam war, government use of
slope arguments) language regarding, 152-153
Tide detergent, and “getting what you
pay for,” 115-116
“Triangle,” defining, 123-124, 126, 128, Watergate and the Democrats,
130 argument regarding, 161-162
Truth, definitional distinguished from Wittgenstein, Ludwig, 225
factual, 61, 146-147 Womens Studies departments,
(See also Necessary conditions, types argument for restricting faculty of,
of) 210, 229

Vagueness:
desirable, 105, 116-117, 138-139 “You get what you pay for,” clarifying
distinguished from ambiguity, meaning of, 114-116, 157-158
121-122
fallacies involving, 117-121
van Vogt, A. E„ 52 Ziegler, Ron, 154, 187-188
 Date Due

PSS- 14=4992 rttr

V
1QQQ '$
4
rto^ Oww yet

« Sin ? 4 £ tasjn
-Attrn ~ae*
 BC 177 .S37 010101
Scriven, Michael.
Reasoning / Mi.chae Scriven^

7404 6
TRENT UNIVERSITY

BC177 .S37
Scriven, Michael.
Reasoning.

DATE ISSUED TO