NATIONAL IN STITUTE OF MENTAL HEALTH

Perspectives
on Deterrence

National Clearinghouse for Mental Health Information

 

AI
delinquency
Lh

   
 

 
CRIME AND DELINQUENCY ISSUES:

A Monograph Series

Perspectives on

Deterrence

by
Franklin E.|Zimring
The Law School

University of Chicago

National Institute of Mental Health
Center for Studies of Crime and Delinquency
5454 Wisconsin Avenue

J

Pl Pee BE AD
January 1971
This monograph is one of a series on current issues and di-
rections in the area of crime and delinquency. Thé series is being
sponsored by the Center for Studies of Crime and Delinquency,
National Institute of Mental Health, to encourage the exchange
of views on issues and to promote in-depth analyses and develop-
ment of insights and recommendations pertaining to them.

This monograph was written by a recognized authority in the
subject matter field under contract number NIH-69-1133 from
the National Institute of Mental Health. The opinions expressed
herein are the views of the author and do not necessarily reflect
the official position of the National Institute of Mental Health
or the Department of Health, Education, and Welfare.

-

Public Health Service Publication No. 2056

 

For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 - Price 55 cents
l/ 742]
[T\ 14314

r
= } )

Foreword  ....c
HEALTH
LIBRARY

The prevention of crime and delinquency covers a wide range
of activities. Usually prevention suggests such activities as:
eliminating social conditions closely associated with crime and
delinquency, providing mental health and related services to
children in danger of becoming delinquent, and reducing the
situations in which crimes are most likely to be committed.

Deterrence, perhaps because it is viewed as a process limited
to the criminal law and its administration, is not often viewed
as a prevention activity. Yet deterrence is inextricably involved
in legislative, judicial, and administrative decisions that consider
sanctions or threats of sanction against deviant behavior. And
as such, it touches the lives of people in everyday life, beyond
those who become directly involved in the criminal justice
system.

Though deterrence is a broad concept, distinctions are drawn
among special, general, and marginal deterrence. Special deter-
rence refers to the threat of further punishment of one who has
already been convicted and punished for a crime. General de-
terrence aims to prevent the population at large (including the
criminals in it) from committing criminal acts. In considering
deterrence, a practical issue that usually appears is: would a
more severe penalty attached to a criminal law more effectively
deter criminal behavior than a less severe one? Marginal deter-
rence concerns itself with this issue. It refers to increasing the
effectiveness of deterrence through variations in the conditions
of legal threats. Marginal deterrence issues may arise in relation
to both general and special deterrence.

Historically, there has been considerable argument about the
effectiveness of deterrence. Witness the debate over whether
capital punishment makes a unique contribution to the deter-
rence of homicide, or whether lesser punishments would prove
less effective as deterrents to drug abuse, “white collar” crime,
shoplifting, parking violations, and a host of other deviant
behaviors.

The current growing incidence of crime and delinquency and
the mounting public concern about it makes the careful examina-
tion of deterrence both timely and important. It appears essen-
tial that the concept, process, and actual results of deterrence
be subjected to sound conceptual and scientific assessment. Such

Tn iii
04760
theory building, research, and the planned use of knowledge
must come about if societal efforts at preventing and reducing
criminal acts are to be based on testable and empirical models,
rather than simply on belief systems.

This volume, drawing on a broad and interdisciplinary range
of literature, reviews the concept of deterrence in detail, and
considers its application to various forms of deviant behavior.
Professor Zimring’s effort represents in my judgment an im-
portant contribution to the analysis of the concept of deterrence.
It also points to specific analyses of some basic assumptions
underlying our criminal justice system, and suggests the develop-
ment of more effective legislative and social policy directions.

In order to provide the author complete freedom to express his
views, no substantive changes have been made by NIMH. The
views expressed, therefore, are those of the author.

Saleem A. Shah, Ph.D.
Chief, Center for Studies
of Crime and Delinquency

iv
Contents

Foreword  ______ oon
Acknowledgments _______
Introduction
Part I. Definitions Cl

Part II. Deterrent Motives and Crime Control Policies
Section 1. Foundations for Official Belief ~~. __
Section 2. Some Ethical Issues in Deterrence
Section 8. Determining Costs in Crime Prevention

Part III. The Deterrent Effect—A Survey of Issues
Section 1. General Deterrence = ___ ______
. Differences Among Men = __________ __
Types of Threatened Behavior =.
Threat as Communication = ______
Applicability and Credibility =... ~~ _
Variations in Threatened Consequences
Variations in Severity of Consequences
The Moral Quality of Threatened Behavior:
Conscientious Objection and Deterrence __
. Moderating Individual Propensities Through
Group Pressure

mn eEEDORE

Section 2. Special Deterrence

Concluding Reflections 7 SR

Page
iii
vi

vii

10
20
25

32
32
34
49
56
65
73
83

92

96
97

108
Acknowledgments

The financial and intellectual support responsible for the pro-
duction of this document came from two institutions and many
individuals. The Center for Studies in Criminal Justice at the
University of Chicago has been supporting my work in deter-
rence since 1965. Its Co-Directors, Norval Morris and Hans W.
Mattick, first nagged this document into existence and then
helped provide the critical suggestions needed to patch over some
of its most distressing lacunae. The Center for Studies of Crime
and Delinquency of the National Institute of Mental Health fi-
nanced and encouraged the writing of this manuscript during
the summer of 1969. Saleem Shah and George Weber of that
Center protected their investment with valuable editorial advice.
Gordon Hawkins collaborated with me on many of the working
papers and articles that form the foundation of this presentation.
Mrs. Sylvia Karjala was, as usual, indispensable in her role as
typist-grammarian.

Anyone who chooses to write about deterrence owes a sub-
stantial intellectual debt to Johannes Andenaes of the University
of Oslo, the most distinguished of the modern scholars writing
about deterrence. During 1968 Professor Andenaes spent five
months in Chicago as teacher and critic of those of us involved
in the deterrence project. He also read and criticized an early
draft of this monograph. It is difficult for me to determine how
many of his insights pass in this volume without proper
acknowledgment.

Maurice Zimring of Hilo, Hawaii, is responsible for the English
translation of this effort.

Franklin E. Zimring
The Law School
University of Chicago

vi
Introduction

The disciplines of law and criminology have a long speaking
acquaintance with the theory that threat of punishment acts as
a deterrent to crime. The broad outlines of the deterrence argu-
ment have been articulated and embellished over the course of
two centuries, as deterrent motives have assumed a greater im-
portance in justifying regimes of punishment policy. In recent
years deterrence has received an increasing amount of attention
from scholars in many fields, and the emphasis has shifted from
the question of deterrence as a general theory to its discussion
as a complex of issues about human behavior in a number of
different situations. A number of thoughtful articles begin by
phrasing the traditional question “Does punishment deter
crime?’ but quickly make it clear that the answer to this gen-
eral question is less important than the significant differences
in situation that condition the likelihood and extent of deter-
rence in crime control.

As the focus of discussion about deterrence has shifted from
the general to the specific, the need for empirical research has
received increasing emphasis in the literature, and the volume
of statistical studies of crime control policies has also increased.
Research work in social psychology and other fields relevant to
deterrence has also mushroomed in the period since World War
II. The net effect of increasing attention and study, however, is
something less than a knowledge explosion. Empirical study of
crime control policy holds great promise, but we are in the
prehistory of such studies. The behavioral models discussed in
deterrence are more suggestive than definitive; there are many
unanswered questions about the relevance and reliability of much
of the interesting psychological experimentation; on most ques-
tions, what is known leaves uncomfortably ample room for specu-
lation. And as knowledge increases by small degrees, the move-
ment appears to be more away from than toward a general
theory of sanctions. But if the boomlet in deterrence study has pro-
duced only modest increments in our understanding, it is none-
theless an important step forward in an area of great importance.

This volume is an attempt to draw together some of the em-
pirical and analytical discussions of deterrence in criminology,
law, and various social sciences, and in doing so, to impose a
conceptual organization on issues in deterrence. The following

vii
pages consider three different but related aspects of deterrence
in crime control. Part I offers a few definitions basic to the
discussion of issues in deterrence, in an effort to avoid one of the
most basic of communication gaps—a difference of language
usage between speaker and audience. Part II is concerned with
deterrence as a motive for official action in crime control. Part
IIT is a survey of issues relating to the effectiveness of threats
as a mechanism for inducing compliance to law. The survey of
contributions relevant to deterrence is hopelessly incomplete. It
would be difficult to find an area of social science research without
possible bearing on our understanding of deterrence, and thus
literature of significance in our inquiry expands quickly out of
reach. The conceptual organization in Part III is most chari-
tably viewed as tentative, and much of the analysis is speculation.
It is published in this form in the hope that it will provoke
further research and analysis, thereby contributing in small part
to the momentum toward a better understanding of the potential
and limits of criminal law as a mechanism of social control.

viii
Part I. Definitions

It is essential, if analysis is even in part to replace emotion in
discussions of deterrence, to read some precision into the defini-
tions of a few central terms.

A sign in the park reads

WALKING ON THE GRASS IS PROHIBITED
$100 FINE
VIOLATORS WILL BE PROSECUTED

Smith and Jones pass the sign while taking a lunch-hour stroll
on the park sidewalk. The sign announces a legal threat: the
attempt of the threatening agency (here presumably the city
authorities) to decrease the amount of threatened behavior (walk-
ing on the grass) by announcing that the agency will impose
unpleasant consequences on those individuals who commit the
threatened act. The sign conveys a legal threat because it is
based on a rule of Jaw issued by an agency of governmental
authority. Like most legal threats, this message is issued to
citizens in general rather than specially communicated to par-
ticular individuals. In this situation the threatened consequences
are a fine and prosecution. Smith and Jones are part of the
audience of the threat. The agency probably considers them part
of the threatened audience whether or not they actually read the
sign, because they are two of the people the threat was aimed at
influencing. However, if our interest is confined to how such an-
nouncements affect human behavior, membership in the threat-
ened audience should be restricted to those individuals who are
aware of a threat’s existence.

The threat communicated to Smith and Jones may or may not
affect their behavior. They may stay off the grass because neither
had intended to walk there in any event. One or both may walk
on the grass in spite of the threat. If the sign produces any
change in the behavior of Smith and Jones, such as inducing
them to keep off the grass, tempting them to walk on the grass
in defiance of the order, or persuading them to walk in some other
park during their lunch hour, the threat will have had a chan-
neling effect on the behavior of one or more members of its
audience, for the channeling effect of a threat is the total quality
of behavioral change in a threatened audience attributable to
the existence of a threat. The term deterrence is used more re-

1
strictively, applying only when threats succeed in achieving their
objective by reducing the number of times the threatened be-
havior takes place. The deterrent effect of a particular threat is
the total number of the threatened behaviors it prevents.

There are a number of situations where threats may influence
behavior in the direction desired by the threatening agency, but
the result will be less than fully law-abiding conduct. The po-
tential bank robber may respond to the threatened penalty for
bank robbery by switching to burglary, heeding one threat only
to disregard another. The presence of a police patrol may reduce
the burglary rate in Manhattan at the expense of the Bronx. Or
intensive law enforcement may persuade the mugger to reduce
the level of his activities from five to three crimes a week. In
each of these cases, the mechanism of deterrence is observable,
but the result of the threat is not law-abiding behavior on the
part of the individual threatened. Yet in all of these cases, a
deterrent effect as defined above will exist, at least on the Island
of Manhattan. A more troublesome case is that where the posting
of a 65-mile-per-hour speed limit causes a driver to reduce the
speed from 90 to 70 m.p.h. In this case, which can be called
partial deterrence, no law violation will have been prevented by
operation of the legal threat, but the speed reduction is still an
important contribution toward the goal of the legal threat.

If one is interested in comparing the rates of behavior that
would result in the case of a particular threat against the rates
of behavior that could be expected if the threat were removed,
the total deterrent effect of the threat is relevant. If, on the
other hand, one seeks to compare the effectiveness of one type
of threat against that of a different threat (say, the threat of
Y years in prison against the threat of Y plus 10 years in
prison), the deterrent effect of the larger penalty is not the
central issue; rather, the comparison should seek to establish
whether the larger penalty functions as a marginal deterrent,
reducing the rate of threatened behavior below that experienced
under the lesser penalty.

A further distinction concerns the difference between the spe-
cial and general effects of threats. Special effects® are the re-
actions that a threat produces among those who have been pre-
viously punished and who, for that reason, may react differently
to threats than the rest of the population. The general effect
of a threat is the response it produces among persons who have
not been punished. For this group the threat, and the example
of the punishment of others, must influence behavior independent
of any personal experience with the threatened consequences.

1 See Andenaes, The General Preventive Effects of Punishment, 114 U. Pa. L. Rev. 949, at
951 (1966).

2
The Deterrent Process

The definitions offered above describe deterrence only as a
result—the number of offenses prevented—without providing any
insight into how legal threats might achieve such results. Com-
mentators have suggested a number of different psychological
processes by which the threat of punishment might result in
deterrence. The mechanism most often referred to is direct or
simple deterrence, but a number of more subtle ways in which
threat and punishment might reduce the number of offenses have
also been suggested.

The theory of simple deterrence is that threats can reduce
crime by causing a change of heart, induced by the unpleasant-
ness of the specific consequences threatened. Many individuals
who are tempted by a particular form of threatened behavior
will, according to this construct, refrain from committing the
offense because the pleasure they might obtain is more than offset
bv the risk of great unpleasantness communicated by a legal
threat. In the simple model of deterrence, the reaction is a very
specific one—comparing this crime with this penalty for one
particular moment >—and the results of this episode of weighing
the pros and cons of lawbreaking do not alter the individuals
personality, or sense of right and wrong, or general propensity
to obey the law. If the individual is to be kept law-abiding, the
process of simple deterrence must confront him at every turn—
making each form of forbidden conduct a risk not worth taking.

In some early writing, the image of the potential criminal
reacting to punishment threats was that of a dispassionate cus-
tomer peering at a price list in search of bargains, perfectly in-
formed and completely rational? But modern writers have recog-
nized the plausibility of the notion that people might refrain from
crime specifically to avoid unpleasantness, while discarding the
image of the super-rational potential criminal that used to ac-
company the theory.' Others, convinced that the notion of simple
deterrence is unalterably bound up with the image of man as
the “lightning calculator of pain and pleasure,” have rejected the
notion of simple deterrence completely.®

In some cases of simple deterrence, the emotional effect of the
threat on the threatened audience may be minimal—the man
"2 The most famous exposition of the theory of simple deterrence is found in the works of
Bentham. See Principles of Penal Law, Pt. II, Bk. I, Ch. VI, Measure of Punishment. I Works,
399 (1843), and Principles of Morals and Legislation, Ch. XVI of the Proportion between Pun-
ishment and Offenses, I Works, 86 (1843). [Citations quoted from Hawkins and Zimring, De-
terrence and Survey Rescarch, Working Paper, Center for Studies in Criminal Justice at the
University of Chicago (1968). ]

3 Ibid.
4 See, e.g., Andenaes, General Prevention: Illusion or Reality? 43 J. Crim. L., C., & P.S.

176 (1952) ; Morris, Impediments to Penal Reform, 33 U. Chi. L. Rev. 627 (1966).
5 See, e.g., Barnes and Teeters, New Horizons in Criminology, at 338.
dissuaded from parking his car in a tow-away zone may suffer
only the mildest regret that the threat of unpleasantness has
made this convenience impractical. In other cases, simple de-
terrence will be achieved because the threat of consequences
produces substantial amounts of anxiety or fear in potential devi-
ants. Thus, depending on the nature of the threatened behavior
and the threatened consequences, simple deterrence may be as-
sociated with different levels of emotional arousal. Within even
the simple model of deterrence, then, a number of different proc-
esses may be at work, producing different kinds of emotional
effect that may condition different types of threat response.
Beyond the notion of simple deterrence, a number of less direct
mechanisms through which the threat of punishment induces
compliance have been suggested, and these more subtle processes
may well be more important than simple deterrence in reinforcing
patterns of law-abiding behavior. The search for these more
subtle effects involves a wide horizon. As Professor Packer puts it,

Our hypotheses about the operation of general deter-
rence should be broadened to include also the effect of
punishment—and indeed, of all the institutions of crim-
inal justice—on the totality of conscious and uncon-
scious motivations that govern the behavior of men in
society.¢
With the scope of inquiry thus broadened, the threat and ex-
ample of punishment may play a role in reducing crime as an
aid to moral education, as a habit-building mechanism, as a
method of achieving respect for the law, and as a rationale for
obedience.

Threat and Punishment as a Teacher of Right and Wrong

Simple deterrence is not the primary explanation of wide-
spread patterns of conformity to most of our criminal laws. As
Professor Toby has said, “ . . . the socialization process keeps
most people law abiding, not the police.” ? But the threat and ex-
ample of punishment may play an important role in the teaching
of right and wrong, and thus in crime prevention, in the larger
sense of socialization. The association of forbidden behavior
with bad consequences may lead individuals to view the behavior
itself as bad. Thus, knowledge that people who steal are treated
badly would lead to the association of wrongfulness with stealing,
and ultimately to the conclusion that stealing is wrong.®

8 Packer, The Limits of the Criminal Sanction, at 42.

" Toby, Is Punishment Necessary? 55 J. Crim. L., C., & P.S. 332, at 334 (1964).

8 Any discussion of the moralizing effects of punishment among American lawyers owes
much to the work of Professor Andenaes, supra notes 1 and 4. The present comments are also
based on Andenaes, Some Further Reflections on General Prevention, Working Paper, Center
for Studies in Criminal Justice (1968).

4
Along a somewhat different tack, the threat and example of
punishment by a legal system will communicate to the individual
that the legal system views the threatened behavior as wrong,
and this information will also affect the moral attitudes of the
individual. The internal dialogue would go: “The institutions I
respect view this behavior as wrong, therefore I should consider
this behavior wrong.”

Finally, threat and punishment may aid moral education by
serving as an attention-getting mechanism. The threat of punish-
ment for stealing forces the individual to think about the moral
quality of the act of stealing. Such reflections may lead to the
conclusion that stealing is wrong because it causes other people
to suffer and undermines the security of a system of private
property. In this last process, which Professor Andenaes calls
“punishment as an eye-opener,” * the threat of punishment pro-
vides only the occasion for reflection rather than any substantial
moral ammunition.

Threat and Punishment as Habit-Builders

On his first few trips downtown, the threat of punishment may
cause Smith to refrain from speeding on his way to work. After
a while, however, more than the simple deterrent effect may be
at work, for Smith may have developed the habit of driving at
a certain speed, and this habit provides additional insulation
against future law violation. And just as the habit of obeying
the law in particular situations may emerge over a period of time,
the threat of punishment may produce a number of separate
habits of compliance which in turn result in a more generalized
habit of obeying the law.1®

Threat and Punishment as Mechanisms for Building
Respect for Law

If the solemn commands of a legal system were not reinforced
with the threat of punishment, many individuals would see no
basis for concluding that the legal system really meant what it
said. More important, even those who would continue to obey
legal commands would be demoralized by watching other people
break the law and escape unpunished. The threat of punishment
is evidence that the legal system is serious in its attempt to pro-
hibit criminal behavior: it is the ‘“‘convincer.” The punishment

9 Andenaes, Some Further Reflections on General Prevention, supra note 8. For comments
on the effect of increasing the severity of consequences threatened on the ‘eye-opening’
effects of punishment, see sub-section III C, infra.

10 “Every one of us is confronted daily by situations in which criminal behavior is a possi-
ble alternative. Sometimes the presentation is sufficiently vivid that we think about it and re-
ject the criminal alternative. More frequently and more significantly, we automatically and
without conscious cognition follow a pattern of learned behavior that excludes the criminal
alternative without our even thinking about it.” Packer, supra note 6, at 43.
of wrongdoers is necessary to continue that respect, because
the unpunished criminal is a direct challenge to the authority
behind the law."* How great would be our dissatisfaction if the
villain in the Western movie could escape his just deserts after
a career of evil!

T'hreat of Punishment as a Rationale for Conformity

Temptation to break the law is often a disturbing experience.
Even a man completely convinced that stealing is wrong may
occasionally need help in fending off the temptation to break his
own moral code. A person in this position desires additional
reasons for not stealing as a defense against pressures within
him that are morally distressing, and the threat of punishment
can provide that additional reason when it is needed.

These, then, are some of the less direct ways in which the threat
of punishment may have subtle but significant effects on the rate
and quality of law-breaking in society. They are hypotheses,
rather than established effects on conduct.’ Little is known about
whether, when, and how much these mechanisms affect conduct.
If the threat of punishment plays a role in the development of
morality and respect for law, it is certainly far from the only
force at work in that phase of the socialization process. And
since the threat of punishment and other socialization processes
are so closely interrelated, it is difficult to isolate the effects of
punishment for empirical study.

The Deterrent Process and Policy Research

There are many different facets to the study of the effects of
threats on human behavior, and many different ways in which
they might be studied. The proper focus for deterrence research
depends on the specific question about penal policy that is of
central importance.

Drunk driving may be punished at present with fines and ex-
pressions of social disapproval which are marginal at best. A
legislator, examining this policy, will have a rather specific frame
of reference when asking questions about the effect of threat
and punishment. With the present penalty as a starting point,
he will cast his eye part of the way up and part of the way
down the scale of possible agency responses to drunk driving,
and he will wish to know what kinds of effects particular changes

11 See Andenaes, supra note 1, at 950.

2 Some psychologists hold, for instance, that the threat of mild sanctions will prove more
conducive to moral learning about the threatened behavior than the threat of severe sanc-
tions, because threat of severe sanctions produces avoidance of the threatened behavior with-
out internal conflict, and such conflict is the occasion for much of moral learning. See, e.g.,
Turner and Wright, Effects of Scverity of Threat and Perceived Availability on the Attrac-
tiveness of Objects, 2 J. Personality and Social Psychology, at 128-132 (1965).

6
will have on the drunk-driving rate. He will probably not be
interested in the effects of leaving drunk driving unpunished
because there are reasons, other than deterrence, for its punish-
ment. He will not wish to know about all the consequences of the
imposition of drastic sanctions, because there are reasons, other
than deterrence, which render drastic sanctions for this crime
undesirable and impractical.

Again, when our legislator enters the capital punishment con-
troversy, his interests are frighteningly specific. He wishes to
know what differential effects can be expected when we sub-
stitute one drastic sanction, life imprisonment, for an even more
drastic sanction, the death penalty. To start a dialogue with our
legislator by outlining the effects that threat and punishment
might have could be misleading, because many of the issues re-
lating to the effects of threats have no relationship to questions
pertaining to the shift from one drastic sanction to another.
For example, it is difficult to suppose that a move between two
such drastic sanctions would substantially affect the moralizing
functions of the example of punishment. That kind of issue
seems to bear with much more importance on questions relating
to whether behavior is to be punished or unpunished, and to
situations where rather drastic jumps up and down the spectrum
of agency responses are contemplated. Shall we send traffic
offenders, now subject to light fines, to prison?

Thus, while threat and punishment may operate in ways that
can be loosely categorized as deterrent at many levels, different
types of penal-policy questions call for different operational defi-
nitions of the deterrent effect that evaluative research seeks to in-
vestigate. If the issue is whether the punishment for a specific
major crime should be reduced or increased by a moderate
amount, the question for research should be whether the higher
penalty is a marginal deterrent, and by how much. The range
of options presented by the problem bears no apparent relation
to the educative or moralizing effects of the institution of
punishment. When the treatment of whole categories of be-
havior as criminal is called into question, the deterrence issue
that must concern us involves the full panoply of methods by
which threat and punishment might affect the rate of behavior,
and the scope of analysis must be broadened, as Professor Packer
said in just such a pursuit, to the

totality of conscious and unconscious motivations that
govern the behavior of men in society.’

We cannot be satisfied with the analysis of only those mech-

13 Packer, supra note 6, at 42.
anisms of simple deterrence when our policy decisions may affect
much more, but it would be equally indefensible to agonize over
the effects of punishment as a mechanism to induce respect for
law when considering whether meat should be added to prison
lunches.
Part II. Deterrent Motives and Grime Control Policies

Legislators, judges, police, prosecutors, and prison officials are
in the crime business, and all would agree that lately business has
been much too good.! Because these individuals operate so close
to the problem of crime in America, they are held—and many
hold themselves—responsible for trends in the crime rate. When
crime rates rise, this sense of responsibility impels that some-
thing must be done to stem the tide. Since he is both powerful
and responsible, it is difficult for the official to feel or announce
that the problem is either insoluble or that the solution lies
outside his sphere of influence. Even if the official can persuade
himself that effective solution is beyond his control, he cannot
persuade his constituency, and because the official must per-
sonally do something about crime, he finds little comfort in
calling for an end to slums, drastic improvement in public educa-
tion, or any other millennial measures. Thus, the law enforce-
ment official and society at large find themselves in complete
agreement that actions taken by officials are intimately related
to fluctuations in the rate of crime.?

But the law enforcement official soon finds that he has a very
limited range of crime control options. He can do very little to
make the life of the disadvantaged more enjoyable. He can make
crime physically more difficult in only a limited number of situa-
tions—urging citizens to lock their cars, fostering the use of
automatic locks on bank vaults, raising the height of fire alarm
boxes to secure them from the whims of five-year-old children,
doubling and trebling the number of police on the street to ham-
per the mugger, the purse thief, and the random attacker. But
many prevention strategies are expensive, and the administrator
is the first to feel the brunt of this type of expense: to double
the police force, we must more than double our budget for
police. If we require some of our more influential citizens to

1See, e.g., both the tone and statistics in the Uniform Crime Reports—1969, compiled and
published by the Federal Bureau of Investigation.

2 There is, to be sure, much official ambivalence about the relationship between law enforce-
ment efforts and crime rates. In discussing the nature of homicide, police authorities are quick
to point out that there is little by way of direct intervention that police can do to reduce the
rate of violent killings. * . . . [Clriminal homicide is, to a major extent, a social problem be-
yond police prevention,” Uniform Crime Reports—1967, at 8. At the same time, police and
many other officials have asserted a direct relationship between rising crime rates and con-
stitutional restraints on police procedures imposed by the United States Supreme Court in
recent years. And four of the eleven factors listed by the FBI as “conditions which will affect
the amount and type of crime that occurs from place to place” concern police, prosecution,
and court policies. See Uniform Crime Reports—1967, at vi.
take expensive precautions, they will object; if we do not require
precautions, many will not heed warnings. Beyond this, many of
the most serious of crimes—homicide, aggravated assault, rape,
indoor robbery, larceny, and crimes against trust—are committed
where police cannot prevent them.

A belief in the efficacy of general and special deterrence is.
therefore, attractive—because it supplies crime control measures
where substitutes are unavailable and does so without great ap-
parent cost to influential members of society. It is thus not sur-
prising that deterrence through threat and punishment is amor
the most valued official weapons in the war against crime. And
while most of the discussion of deterrence in criminal law re-
velves around the effects of particular strategies in particular
situations, there is much that can be said about the function of
deterrence as a motive in crime control that is of general appli-
cation. This segment deals with three aspects of deterrence as a
motive of official action: (1) the foundations of official belief
in the efficacy of deterrence, (2) some ethical problems as-
sociated with the use of punishment for deterrent purposes, and
(3) the proper model of costs and crime preventive results that
should serve the official in making decisions about whether extra
punishment for deterrent purposes is desirable.

Section 1. Foundations for Official Belief

There exists in this country no official dogma that could be
considered a general theory of sanctions. Implicit in many pro-
visions of our penal codes is the assumption that higher penalties
will reduce the rate of particular types of crime. A few laws—
so-called habitual-criminal statutes * and the Federal law ordaining
extra penalties for gun crime * will serve as examples—go as far
as assuming that stiffer penalties can make inroads on crime
in general. But the source of those beliefs about deterrence that
influences our penal policy is not to be found in law itself, but
in the shared attitudes of our lawmakers. When issues arise, the
attitudes of those with legislative, judicial, and administrative
power determine the direction of policy. Thus, the nature of the
official’s attitude toward deterrence and any explanation of its
origins are subjects of independent interest to the student of
penal policy.

It is foolish to try to pin one attitude about deterrence on the
thousands of officials responsible for the administration of our
criminal law. Yet different groups of officials appear to share at-
titudes about deterrence to the extent that generalization does
not seem unfair. When confronted with a crime problem, leg-

? See Morris, The Habitual Criminal.
418 U.S.C. 924(c) (1968).

10
islators often agree that the best hope ‘of control lies in “getting
tough” with criminals by increasing penalties.” Police subscribe
to the notion of “getting tough,” but are apt to put more empha-
sis on what is termed “strict law enforcement,” ¢ a concept that
accords the major role to policing. There are significant exceptions
to both of these patterns. Many legislators, but far from a ma-
jority, now doubt the efficacy of the death penalty as a marginal
deterrent, when compared with the threat of protracted im-
prisonment. Many police express less than total faith in the
ability of “strict law enforcement” alone to make a positive
contribution in the control of prostitution.

But the exceptions are few, and people more often seem to
think in a straight line about the deterrent effect of sanctions:
If penalties have a deterrent effect in one situation, they will
have a deterrent effect in all; if some people are deterred by
threats, then all will respond; if doubling a penalty produces an
extra measure of deterrence, trebling the penalty will do still
better. Carried to what may be an unfair extreme, this style
of thinking imagines a world in which armed robbery is similar
to illegal parking, burglars think like district attorneys, and the
threat of punishment produces an orderly process of elimination
which diminishes the crime rate as the penalty scale increases
by degrees from small fines to capital punishment, with each step
upward as effective as its predecessor. Other officials, mostly
those engaged in prison administration, the discouraging end of
deterrence, will take a different but equally unitary view: Since
human behavior is unpredictable and crime is determined by
other causes, deterrence is a myth.

It is a cheap point, to be made without any semblance of self-
congratulation, that deterrence is far too complicated an area
to accord with either of these polar views. The purpose of pro-
posing these caricatures of official attitudes is to explore what
might account for such patterns of thought. Like deterrence itself,
the explanation of these attitudes is multifaceted.

Belief in deterrence is both natural and in the interest of most
law enforcement officials. It is difficult to deny that, as Pro-
fessor Packer puts it, “People who commit crimes appear to
share the prevalent impression that punishment is an unpleasant-
ness that is best avoided.” To threaten with punishment is

5 Crime and Penalties in California, a publication of the California Assembly Office of Re-
search, discusses two episodes of increased-penalty responses to rising rates of marihuana use
and assaults on police during the 1961 term of the California legislature, at 10-13. Other fre-
quent candidates for “get tough’ legislation in the United States include dangerous drug use
and sale, gun robbery, and organized criminal activity.

6 The difference in emphasis between ‘getting tough” and strict law enforcement appears
to be that the former emphasizes the severity of sanctions while the latter emphasizes risk of
apprehension.

7 Packer, The Limits of the Criminal Sanction, at 149 (1968).

11
therefore a very promising strategy of influencing behavior. And
deterrence, as noted above, is a strategy that shows promise of
working in areas of behavior where the official has no other
technique of crime reduction at hand.

But belief in deterrence and the assumption of a monolithic
theory of deterrent efficacy are two different matters. The prin-
cipal signpost of those who hold monolithic attitudes about de-
terrence—whether affirming or denying the effectiveness of legal
threats—is that they allow themselves only one idea about the
nature of deterrence. And once the complex of issues about de-
terrence is transformed into a yes-or-no question, the results
are predictable. If the sum of people’s thinking is to decide
whether they are “fer it” or “agin it,” the “fers” will achieve a
substantial majority as a matter of common sense. But when a
complex series of different issues is bent into the form of a
ves-or-no question, the margin of error obtained from answering
that question correctly approaches fifty percent.

The tendency of officials to have only one idea about deterrence
may be seen as a symptom, rather than a cause, or unitary at-
titudes about deterrence. The task becomes that of explaining
why this “single idea” phenomenon persists. One factor is no
less basic than human nature; we all would prefer to have simple
rather than complicated explanations for the questions that per-
plex us. Complicated explanations evolve from the pressure that
experience exerts on our simple initial constructs, a process that
takes the classic structure of trial and error. Since there is so
little evaluative research in the area of deterrence, however, the
pressure toward rethinking unitary positions is not great.

Thus, the fact that there is so little known about the effects
of sanction policies means that there are fewer inconsistent re-
sults to sensitize officials to the differences in situation which may,
in turn, condition differences in threat effectiveness. Moreover,
what few data are available can be comfortably fitted into the
official’s initial opinion—it can either be accepted at face value
or rejected as inconclusive. But the absence of reliable research
in deterrence does not mean that officials are without any basis
for their opinions. They fall back on personal and administrative
experience to test their views of deterrent effectiveness.

The official’s personal experience is likely to give strong support
to an enthusiastic embrace of deterrence. Having worked hard to
achieve the regard of his fellows, he is more sensitive than most
to the threat of social stigma. He likes to regard himself as a
rational man, and will be anxious to give himself credit for re-
sponding to threats where any other course of conduct would
be irrational. The official is also a law-abiding man and attributes

12
some of his obedience to the threat of sanctions. He remembers
slowing down when passing a police car on the highway, and con-
sidering the likelihood of audit when filling out his income tax
report. He is less likely to recall deviations.

A more important source of information is the official’s ad-
ministrative experience with crime-control policies. In his official
capacity, the legislator “tries out” deterrent threats; the results
of these trials are integrated into his attitudes about deterrence.
In the absence of controlled research, the apparent results of on-
going crime-control policy are the most important data availa-
ble about deterrence, and these results have a profound effect on
official attitudes. Unhappily, as we shall see, the unquestioning
acceptance of the results of experience with crime-control policy
is not a sufficient substitute for careful evaluative research.

The World’s Three Worst Experiments

Experience is a valuable teacher, but she occasionally codes
her messages in ways we fail to understand. In order to em-
phasize some of the common pitfalls that officials encounter in
trying to read the significance of the results of crime-control
policies, it is best to start with three classic cases of common
errors made in inferring more from statistics than the statistics
will support.

Aunt Jane's Cold Remedy
One common inferential error takes its name from the famous
(though apocryphal) ad:

Take Aunt Jane’s Cold Remedy and
Your Cold Will Go Away Within Two Weeks!

All of Aunt Jane’s customers will probably be quite satisfied
with the medicine, because every time they take it, their cold
disappears within the period advertised. The medicine is, of
course, a mixture of whiskey, sugar and water, with no known
effect on the common cold. The cold goes away within two weeks
because most colds go away within two weeks, whether treated
or not—it is the nature of the animal to get well. But if Aunt
Jane’s customers are faithful, they will have no way of knowing
it wasn’t the remedy. And if they have an emotional investment
in believing in the cure, they may staunchly resist the suggestion
that the remedy is useless—even when their colds go away with-
out being treated with the medicine. These untreated recoveries
will be explained away as milder colds.

Law enforcement officials treat crime, not colds, but they
are prone to encounter the same problem. When life is proceeding
normally, no great pressure is put on the legislator to increase
penalties or on the police official to double or triple patrols.

13
Pressure for strong new countermeasures comes when the crime
rate suddenly spurts. Often the scenario then follows this course:
1. spurt in crime rate; 2. countermeasure; 3. return of crime
rate to more normal historical level.

Did the countermeasure reduce crime? If so, by how much?
The headlines read, POLICE DOGS REDUCE GHETTO CRIME
BY 65%, or COMPUTER REDUCES FALSE ALARMS BY
35%, and enforcement officials are reluctant to avoid credit for
the change. But if we remember that an unusual spurt in the
crime rate was responsible for the countermeasure, another pos-
sible explanation of the reduction exists: The crime rate simply
returned to its usual level, which would also have been the case
whether or not the computer or police dogs or new patrols had
been introduced. A well-documented example of law enforcement
techniques taking credit for reductions that were partly the
result of other factors was Connecticut’s famous “crackdown”
on speeders in the mid-1950’s. After the crackdown began, Con-
necticut’s rate of traffic fatalities decreased dramatically, propel-
ling the Governor of that State into the national limelight.
Investigators ¢ later found that the death rate in Connecticut
just before the crackdown was unusually high and that fatality
rates in States adjoining Connecticut, which had not instituted
a speed crackdown, had also declined, though not as much as in
Connecticut. Thus, a large part of the saving attributed to the
program probably had nothing to do with it.

How many situations are like the Connecticut speed crack-
down? And how can observers discriminate between spurious
and genuine preventive effects? In Chicago, the fire department
used to send a fireman out to talk to the students of any school
that had been responsible for a large number of false alarms.
The rate of false alarms from the treated schools usually de-
creased after the talk. However, when a randomly selected group
of schools with recent false alarm problems was left untreated,
the false alarm record of these schools also decreased.?

In New York City, the rate of taxicab robberies climbed
sharply in 1966. The police department authorized policemen
to take after-hours jobs driving taxis, and the robbery rate fell,
although it remained higher than it had been prior to 1966.1°

8 Campbell and Ross, The Connecticut Crackdown and Speeding, 111 Law & Society Review
33 (August 1968).
? Interim Report, The Chicago Fblse Fire Alarm Project, Center for Studies in Criminal Jus-
tice at the University of Chicago Law School (unpublished, 1967).
10 The countermeasure was introduced in mid-1967. Cab robberies for the period August
through December for the years 1964-1967 were as follows:
1964 1965 1966 1967
241 245 535 346
[Cited from letter from Gordon F. Dale, Technical Services Bureau, New York Police De-
partment, March 1968.]

14
Proof positive of deterrence? A final case history comes from
Great Britain, where traffic fatalities fell sharply after the Road
Safety Act of 1967 introduced a new definition of drunk driving
and new means of enforcement.!

In the Chicago study, it was possible to catch Aunt Jane's
Remedy at work because the untreated group of schools served
as a control, providing a number of schools where the peak rate
was given the chance to diminish without treatment. In Con-
necticut, observers were able to test the claims made for the
speed crackdown by comparing the Connecticut rates with those
of other States, a strategy less clear-cut than use of a randomly
selected control group but helpful nonetheless. In New York,
available figures could only show that the rate of taxi robberies
after the “cops in cabs” experience was lower than just before,
but still higher than it had been in previous years; since it was
impossible to tell what the rate would have been if the program
had not been introduced, it is not possible to determine how much
of the reduction can be attributed to natural causes and how
much can be credited to the program. But knowledge that the
program had been introduced in a peak period should at least
make observers sensitive to the distinct possibility that a good
part of the decrease might have been unrelated to the new
program.

In some situations, detailed study of historical trends in the
crime rate, or comparisons with untreated areas, will show that
the decrease was not solely attributable to Aunt Jane. If the
New York cab robbery rate had fallen to levels lower than those
usually experienced in the past, there would be a strong basis
for concluding that the program (or some other change) was
responsible for extra deterrence. In Connecticut, after a careful
study had shown that much of the rate reduction probably had
nothing to do with the speed crackdown, there was still left a
part of the reduction that could not be attributed to forces
at work in States without the crackdown. The program prob-
ably had some effect after all.*?

In Great Britain, government studies demonstrated clear-cut
results. The drop in fatalities was sharp and came immediately
after the law was introduced. Night accident fatalities (more
closely related to drunk driving) fell more dramatically than day
accident fatalities. And statistics on miles driven in Great
Britain showed that there was no reason to expect a natural
decrease in accidents due to less driving.*?

11 See Andenaes, Some Further Reflections on General Prevention, Center for Studies in
Criminal Justice at the University of Chicago Law School (Working Paper, 1968).

12 Campbell and Ross, supra note 8, at 52. The authors are unable to eliminate regression
as an explanation but note the sustained nature of the decrease.

13 Press release, Great Britian Ministry of Transport, February 8, at 11, 1968.

15
By now it should be apparent, however, that if officials desire
to maintain a unitary faith in deterrent countermeasures, it will
be easy for them to do so. Because new programs are normally
tried during periods when the rate of the particular crime is high,
the official can take credit for any decreases in rate that follow the
introduction of the new measures. If the crime rate stays high,
who can say that it would not have climbed higher were it not
for the new program? In fact the rise in crime shows conclusively
that we need more of the new countermeasure!

Subjecting our crime prevention strategies to evaluative re-
search will prove less comforting but, in the long run, more
valuable. On some occasions, close analysis will show that the
crime rate could have been expected to decrease even if no new
treatment had been administered; in other situations, a steady
rise in crime over a long period of time that can be attributed
to factors such as an increase in the population at risk will
indicate that, in the absence of a new treatment, the crime rate
would probably have continued to increase.” In some, but not all
cases, evaluative research may be able to provide rough estimates
of how much the rate would have increased or decreased in the
absence of treatment and thus give us a baseline for testing the
value of new treatments.

Providing a baseline, so that reliable determinations can be
made about the degree of crime reduction attributable to partic-
ular countermeasures, is an absolute necessity in any but the
most wasteful of crime-control policies. Without such a baseline,
it may be assumed that some strategies reduce crime when in
fact they do not. Then, too, the preventive effects of other pro-
grams may be underestimated because we fail to account for
expected increases in crime. Even when countermeasures have
some effect on crime, overestimating that effect by neglecting the
possibility of natural decrease may provoke the use of programs
that are not worth their cost—and, worse still, postpone the
development of new and more effective strategies.

Tiger Prevention

Another barrier to the revision of crime-control strategies is
the official belief in a number of “tiger prevention” programs.
Tiger prevention takes its name from the story of the gentleman
who was running about the streets of mid-Manhattan, snapping

14 According to the California Assembly Office of Research study, supra 5, “In the
City of Los Angeles, the rate of attacks on police went from 2.5 per 100 in 1952 to 8.4 in 1961,
to 16.8 in 1966. . . . In 1961 the first special penalties for attacks on the police were enacted
by the Legislature and such penalties were further increased in . . . 1963 and 1965.”

15 Thus, it is not possible to conclude, as was done in Crime and Penalties in California,
supra note 5, at 11, that increased penalties for assaults on a police officer were of no deter-
rent effectiveness merely because the rate of such crimes continued to increase after penalties
for that offense increased.

16 See Part II, Section 8, infra, for an elaboration of this theme.

16
his fingers and moaning loudly, when he was intercepted by a
police officer. Their conversation follows:

P.O.: What are you doing?

Gtim.: Keeping tigers away.

P.O.: Why, that’s crazy. There isn’t a wild tiger within
five thousand miles of New York City!

Gtlm.: Well then, I must have a pretty effective technique!

Other factors, of course, account for the absence of tigers in
New York City, but as long as our gentlemen friend continues
to snap his fingers he won’t know that. And if the method of
prevention were any less ridiculous, it would be difficult for
observers to conclude for certain that his countermeasure, which
is, after all, designed to prevent tigers, is not the reason why
the number of wild tigers is low.

In crime control, the tiger-prevention problem is subtle and
difficult to resolve. Officials who administer very high penalties
acquire the firm conviction that only those penalties stand between
them and huge increases in the crime rate. Having assumed
that the penalty is the only reason for the absence of a crime
wave, the official has “proved” that high penalties deter crime
more effectively than less severe penalties. But the proof is the
same as that offered by the tiger prevention specialist, except in
one respect: because penalties may well influence crime rates, it
cannot be assumed, as in the tiger example, that the countermeas-
ure bears no relation to the rate of crime. Being unable to assume
that there is no relation between penalty level and crime rate, our
only method of determining whether and how much the two are
related is to vary the penalty. But since that would involve
risk-taking, the status quo and the “proof” of deterrence built
into it persist. The tiger-prevention argument is not refutable
when posed as a barrier to experimental decreases in punish-
ment; it is, however, patently absurd to present high penalties
combined with low crime rates as proof of deterrence.

Studies of different areas with different penalties and studies
focusing on the same jurisdiction before and after a change in
punishment level takes place show rather clearly that level-of-
punishment is not the major reason why crime rates vary.' In
particular areas, such as capital punishment as a marginal deter-
rent to homicide, the studies go farther and suggest no dis-
cernible relationship between the presence of the death penalty
and homicide rates.'®

Although imperfect, these studies are certainly the best method

17 See Rusche and Kirchheimer, Punishment and Social Structure (1939).
18 See the studies collected in Bedau, The Death Penalty in America (1964).

17
available of testing whether more severe sanctions have extra
deterrent force in particular situations. But even when conduct-
ing comparative and retrospective studies, the tiger-prevention
fallacy may crop up in a more sophisticated form. Consider a
study that shows that homicide is both the most severely punished
and least often committed crime in a particular jurisdiction.
Proof of deterrence? Or tiger-prevention? It is quite possible
that the rate of homicide would remain low even if the penalty
for homicide were less severe, because of the strong social feelings
against homicide. Indeed, one reason why the penalty for homi-
cide is so high is because citizens view this crime as so terrible.
Both the low rate and the high penalty may be effects of the
same cause: strong social feelings against homicide. Thus, show-
ing that crimes which are punished more severely are committed
less often may only be one way of showing how accurately a
penalty scale reflects general feelings about the seriousness of
crimes.

This problem can spill over into some forms of comparative
research. Assume we find out that homicide is punished more
severely in Maine than in Georgia, and that the homicide rate
is much lower in Maine than in Georgia.?® Does the higher penalty
cause the lower rate? It may be that the higher penalty shows
that people in Maine have stronger social feelings about homi-
cide and these feelings, rather than the extra penalty, could
explain the difference in homicide rate. In order to test the
effect of penalties alone, areas that are similar to each other in
all respects except penalties should be sought out and compared.
This would be a strategy far more cumbersome than assuming
that harsh criminal sanctions are keeping the tigers away but,
again, far more rewarding.

The Warden’s Survey Research

A final example of drawing faulty inferences in deterrence
is the proof offered by an imaginary prison warden that deter-
rence is a myth. He has interviewed a thousand men on death
row, he says, and none of them had been deterred by the threat
of penalties from committing their crimes. The warden’s survey
is unpersuasive for two reasons: first, it is not clear that his
prisoners would tell him if they had been deterred by penalties
at some time in their lives, since it would not be in their present
interests; but, more important, the warden’s sample of people
to ask about deterrence is hopelessly biased. The warden has
started with a group of men who have not been deterred or they

1 Tittle, Crime Rates and Legal Sanctions, 16 Social Problems 409 (1969).
20 Ibid., at 416.

18
would not be on death row. If the law has had a deterrent effect,
it has done so by influencing the behavior of people every-
where else.

The warden’s survey problem occasionally crops up in sophis-
ticated efforts at deterrence research. A study conducted for the
President’s Commission on Law Enforcement and the Adminis-
tration of Justice ?* asked incarcerated offenders whether increased
penalties or increased policing would decrease crime. A majority
thought not.?2 This was a modified form of the warden’s survey
because (a) it involved a group of people who might have some
personal interest in keeping penalties from increasing, whether
or not they reduce crime, and (b) it was a case of talking about
deterrence with a group of people who were there because they
had not responded to the threat of punishment for crime.?®

Another subtle form of the warden’s survey problem is found
in Professor Chambliss’ treatment of the problem of deterrence
and drug abuse.?* Basing his conclusion on statistics which show
that a high proportion of treated drug addicts resume their
habits after treatment, Chambliss concludes,

The evidence, then, suggests that drug addiction, like
murder, is relatively unaffected by the threat or the
imposition of punishment.

As Professor Andenaes points out, the data support Chambliss’
conclusion that the threat and imposition of punishment do not
deter those who are already drug addicts from further drug
use; however, the drug laws could still affect drug addiction
rates by deterring people from getting involved with drugs.2®
And it cannot be concluded that all potential drug addicts are
unaffected by the criminal sanctions simply because those who
become addicts were unaffected, for those who did in fact become
addicts in spite of the law are a selected group of deterrence
failures. All of the law’s successes, if there are any, will be
found in the non-drug-using segment of the population.

21 Goodman, Miller, and DeForrest, A Study of the Deterrence Value of Crime Prevention
Measures as Perceived by Criminal Offenders (1966). Unpublished paper submitted by the Bu-
reau of Social Science Research, Inc., to the Institute for Defense Analysis.

22 See Ibid., at 47. The study is considered in detail in Hawkins and Zimring, Deterrence and
Survey Research (1968), Working Paper, Center for Studies in Criminal Justice at the Uni-
versity of Chicago Law School.

23 Putting aside the serious issue of whether the prisoners responded candidly and were in
a position to make accurate judgments about deterrence, the study can be justified as an at-
tempt to determine whether those who have not responded to present sanctions would be de-
terred by higher sanctions. From that standpoint, convicted prisoners are a perfect sample
for testing marginal deterrence, even though they are a poor sample for testing the deterrent
effect of present sanctions.

24 Chambliss, Types of Deviance and the Effectiveness of Legal Sanctions, 1967 Wis. L.
Rev. 703.

26 Ibid., at 708.

26 Andenaes, Some Further Reflections on General Prevention (1968), at 5, Working Paper,
Center for Studies in Criminal Justice at the University of Chicago Law School.

19
Our tiger preventer and our warden have made contradictory
assumptions about the relationship between those criminals in
jail and our potential crime problem. The tiger-prevention advo-
cate assumes that large numbers of law-abiding citizens are
held in check by the threat of penalties, indeed, that only severe
penalties can perform this job. The warden seems to assume that
when he interviews prisoners, he is talking to the totality of the
potential crime problem, or at least to a representative sample.
Because these assumptions are unsupported, both the “proof”
and “disproof” of deterrence must fail.

But the fact that these patterns of thinking fail to prove or
disprove deterrence does not mean they are unimportant. Of-
ficials will continue to act on sincere convictions, whether or not
these convictions are well founded. Thus, a significant step to-
ward more research in deterrence, and ultimately toward a more
rational crime-control policy, would be to make officials more
sensitive to new insights and more vulnerable to understanding
the complexities that undermine monolithic attitudes about de-
terrence. For this reason, the careful study of “cold remedies,”
“tiger prevention,” and other such habits of thought are an im-
portant part of a program for progress in crime-control policy.

Section 2. Some Ethical Issues in Deterrence

In order to make the threat of punishment believable, the crimi-
nal law must follow through by punishing those offenders it ap-
prehends. Punishing people in order to deter them (or others)
from committing future offenses raises some questions about the
justice of pain inflicted for deterrent purposes that should be dis-
tinguished from issues that relate only to the efficacy of deterrent
strategies. When concerned with the efficacy of punishment-for-
deterrence, we ask the question, “Will it work?”’; when concern-
ed with the justice of punishment-for-deterrence, we ask, “Is it
morally acceptable to punish for this reason?’ These two issues
are not identical. It is easy to imagine punishments that would be
effective but unjust, such as the random execution of every tenth
parking violator.

Two basic principles relating to the justice of punishment need
detain us only briefly here, since they have been admirably dis-
cussed in the literature.’ First, it seems clear under the moral pre-
cepts of our criminal law that punishment can only be justified
if its subject has committed a forbidden act. Conscientious ob-
jectors to this principle are hard to find. Some anti-utilitarian
writers have tried to hold extreme utilitarian philosophers respon-
sible for implying that punishment without offense would be jus-

27 See, e.g., Hart, Punishment and Responsibility; Packer, The Limits of the Criminal Sanc-
tion; Moberly, The Ethics of Punishment.

20
tified within their system, because it is possible that more good
may come from the act of punishment (deterrence of others)
than harm suffered (pain to the individual).?® Yet those who raise
this point do so by way of showing the limits of exclusively utili-
tarian thinking. Nobody affirms the justice of punishing the inno-
cent.

In a similar vein, it is hard to find writers who would deny the
injustice of aimless punishment. We have developed a large and
diverse vocabulary of motives for punishment: punishment is jus-
tified as a means of expressing society’s retributive feelings, as a
method of isolating high crime risks, as a deterrent, as a mecha-
nism for rehabilitation. Commentators disagree about which pur-
poses are legitimate, and which of the legitimate purposes of pun-
ishment should be accorded priority. But most would agree that
the gratuitous infliction of suffering, even on those rightly clas-
sified as blameworthy, cannot be justified. To say that punishment
must have a purpose to be considered just, differs from the strict
utilitarian position that punishment must achieve more benefit
than the harm it produces, in that our formulation allows, as
legitimate objectives of punishment, goals that some utilitarians
might reject. And so, having said that aimless punishment is un-
just, it would still be possible to justify punishment that was moti-
vated solely by society’s need to express its disapproval of the
punished conduct. The objection might be made that if pure ret-
ribution is a sufficient justification for punishment, then the lim-
iting principle that punishment must have a purpose is accept-
able to writers of different persuasions precisely because it is
empty baggage. Since, as a first principle, we have limited punish-
ment to the blameworthy, only those who can justly be punished
for retributive reasons are left in the discussion. Since retribution
is a legitimate aim of punishment, saying that all punishment
must have a purpose imposes no further restrictions.

While this observation may hold true when the sole question to
be addressed is whether even the smallest measure of punishment
is just, limiting punishment by requiring that it must be imposed
for legitimate reasons may have a profound effect on how much
suffering may be justly inflicted for a particular offense. The need
for retribution justifies the punishment of offenders, but the
amount of punishment justified is limited to the amount of punish-
ment necessary to achieve the retributive effect. Any extra punish-
ment is unjustified unless it serves other purposes. Limiting the
extent of punishment to those measures which serve legitimate
"2 Even Bentham and Hobbs recognize, however, that the existence of an offense is a neces-
sary condition for punishment. See Moberly, supra note 27, at 63.

20 Morris and Zimring, Deterrence and Corrcctions, 38 Annals of the American Academy of
Political and Social Science 1387, January 1969, at 138.

21
purposes will protect against the overkill of criminal sanctions.

The two conditions for just punishment discussed above—the
existence of blameworthy conduct and a legitimate purpose for
punishment—are necessary but not sufficient conditions for de-
termining that a particular punishment is just.?* Punishment may
serve legitimate objectives and still be too severe. Individuals
who have committed criminal acts may have done so under cir-
cumstances that would make harsh punishment shock the con-
science.’ And punishment for deterrent purposes raises a few
special problems, in that it appears to have so little to do with the
particular offender: such punishment is not determined as the re-
sult of the particular degree of reprehensibility of his conduct,
as with retribution, and it cannot be justified as being designed to
benefit the offender, a familiar justification for rehabilitative
measures.” Indeed, because it stems from other considerations,
such deterrent punishment may generate conflict with the law’s
rehabilitative and retributive precepts.

The special problems raised by punishment for deterrence sel-
dom concern administering harsh penalties because of deterrent
motives to people who would otherwise go unpunished; except
for regulatory offenses, deterrent motives condition the degree of
punishment far more often than they represent the single justi-
fication for punishing crime.

In most cases involving serious crime, we have no difficulty clas-
sifying those who commit them as blameworthy. A street robbery
is objectively dangerous behavior, and the offender’s intentions
are seldom in doubt. When we catch him, the quality of his con-
duct makes it clear that he is eligible for punishment. But having
decided that punishment is justified, the question remains: how
much punishment is just? Beyond deciding that the robber’s
punishment is appropriate, legislators have prescribed a range of
from one to ten years in prison, a judge has sentenced our offender
to five years, and the parole board has refused to let him out after
serving three. The particular penalty for robbery in this sit-
uation exists, as we have noted, for many purposes: to phys-
ically isolate robbers and thereby prevent some crime, to assist
in the rehabilitation of the robber, to express society’s retribu-

% See Packer, supra note 7, at 63-70.

31 The story of Jean Valjean’s pursuit and punishment for the theft of a single loaf of
bread in Hugo’s Les Miserables was written around this theme.

32 “[Deterrence] is peculiar in that it ignores the personal quality of the offender,” Temple,
The Ethics of Penal Action, First Clarke Hall Lecture, 1934. See also, G. B. Shaw, The Crime
of Imprisonment, Philosophical Library, 1946, at 32-33: “[Deterrence] necessarily leaves the
interest of the victim wholly out of account. It injures and degrades him; destroys the repu-
tation without which he cannot get employment; and when the punishment is imprisonment
under our system, atrophies his powers of fending for himself in the world . . . . He is, at
the expiration of his sentence, flung out of the prison into the street. . . . [with] no com-
punction as to society; why should he have any ?”’

22
tive feelings toward robbers, and to add potency to the threat of
the law, in order to promote the deterrence of apprehended robbers
and others who might become robbers. If we assume that robbery
would be punished by a two-year, rather than a five-year, prison
sentence were it not for deterrent motives, we encounter a complex
of issues about the justice of punishing for deterrent purposes.
Our robber will argue that he is, in fact, being punished twice:
two years of prison for the offense of robbery, three years of pris-
on for the service of mankind. The legislator might reply that
once an individual is determined to be blameworthy, the
term of punishment is just as long as it is set as a result of proper
balancing of legitimate punishment objectives, and deterrence is
certainly one legitimate objective of punishment.?* The prisoner
might then argue that, while deterrence is a laudable goal of pub-
lic policy, the second sentence he is serving, if it is designed to de-
ter him from future robbery, seems unjust because it is not at all
clear that he would commit robbery again in any event.** On
the other hand, if his extra sentence is motivated by the desire to
prevent others from entering the path of crime, he feels unfairly
put upon. Why should his grief pay for their moral education 7°

Such a dialogue is only possible because it has been assumed that
an extra measure of punishment was assessed for deterrent pur-
poses only. It is not deterrence as an objective that is cause for
concern but the escalation of sanctions for deterrent purposes.
Thus, the moral problems raised by punishment for deterrent pur-
poses arise only when we impose a punishment for deterrent pur-
poses that is more severe than would otherwise be imposed. And
yet, increases in penalty for exclusively deterrent purposes are far
from rare, if the reasons given for legislative and judicial change
in policy are taken at face value.

The robber’s complaint can be understood most forcibly in those
cases where the punishment is harsh and the disparity between
the punishment that would be set in the absence of deterrent mo-
tives and the increases for the sake of deterrence is most extreme.
Sending offenders to prison for parking violations, much as it
might reduce overtime parking, excites our sense of injustice be-

33 “When a man has been proven to have committed a crime, it is expedient that society
should make use of that man for the diminution of crime; he belongs to them for that pur-
pose.” The Reverend Sydney Smith, as quoted by Radzinowicz and Turner, 21 Canadian Bar
Review 92 1943). (Citation from Andenaes, Some Further Reflections on General Prevention,
Working Paper, Center for Studies in Criminal Justice at the University of Chicago Law
School, at 57).

3% A similar complaint could be registered by those confined in preconviction preventive de-
tention and those held in restraint by civil processes because of ‘“‘dangerousness to others.”
The relationship between ethical problems associated with the rationale of constraint and the
need for evaluate research appears to be similar in these areas. See infra, at Section 3, for a
brief discussion of the “research imperative.”

35 Not unexpectedly, prisoners place little value on the deterrent potential of increased sanc-
tions. See Goodman et al., supra note 21.

23
cause the total penalty seems so grossly disproportionate to the
wrongfulness of the conduct. The same sense of gross dispropor-
tion has overcome many critics of drastically escalated penalties
for the possession of marihuana. These huge differences are likely
to be perceived in connection with behavior that is not considered
a serious threat to the community. In these cases, the community’s
sense of “just deserts” creates a limit beyond which punishment
seems unfair. The notion that the punishment should fit the crime,
and not grossly exceed it, is one natural limit on the imposition of
sanctions for deterrent purposes.

Where criminal conduct has been more serious, the community
is less likely to have any sympathy for the offender; and the gap
between deterrent increases and punishment levels which might
be set in the absence of a desire for marginal deterrence will
be smaller, if only because the base penalty is larger. It neverthe-
less remains true that we are taxing the offender in order to in-
fluence the behavior of others. If this is the only way in which we
can reduce the crime rate, the practice seems inevitable, and can
be more easily justified than if alternative methods of crime con-
trol are available.?”® If alternative methods of crime reduction do
exist, we are in the even less enviable position of explaining to the
robber that his extra punishment is our method of saving other
scarce resources.

Two principles emerge from a consideration of the moral prob-
lems of punishing offenders for deterrent purposes. First, the
harm suffered by offenders as a result of the extra measure of
punishment administered for deterrent motives must be recogniz-
ed as a cost, not insubstantial, to the community as a whole. If the
community should rejoice at the prospect of punishment as an
indication of retributive feeling, no joy should come from punish-
ment in excess of that required to fully express collective feelings
of outrage. The offender is a citizen, and the community’s deci-
sion making process exists to protect his welfare as well as that of
others.?” Considering the suffering of offenders as a cost helps to
make clear that aimless punishment is an irrational community
response. More important, if the offender’s suffering is not taken
into account as a cost, the usual result of comparing the cost of
extra offender suffering and alternative crime reduction methods
will be to opt for the extra suffering. The situation is reminiscent
of the stranger who had just been advised that the best method
of increasing his horse’s efficiency was to violently castrate the

38 Cf. Andenaes, supra note 33, at 60: “But in the ordinary run of things, where general
preventive considerations are taken into account in determining the general level of penalties,
the question seems to me to be primarily not one of the principle, but one of degree. How much
human suffering are we willing to accept as the price for a reasonable protection of life,

provertv and other interests ?”’
37 See Temple, supra note 32.

24
animal with two bricks. When the stranger asked, “Doesn’t that
hurt?,” his mentor responded, “Not if you don’t get your thumbs
in the way.”

Of course, the important question is, “Whom does one count
when determining if the action hurts?” When the community as-
serts its right to punish offenders at higher levels for exclusively
deterrent purposes, it must protect them against unjustified pun-
ishment by fully considering the offender’s interest in freedom
from excessive punishment.

A second principle to be drawn from an otherwise unresolved
dialogue is that administrators have a moral duty to the punished
offenders to do research on the deterrent effect their policies pur-
sue. When the law-enforcement official imposes extra punishment
for deterrent purposes, he holds the lives of the least lovable seg-
ment of our society in trust. To base extra punishment on a belief
in deterrence is morally acceptable only as long as it is necessary.
When facilities exist for the evaluation of sanction policies, a fail-
ure to test our policies while continuing to penalize offenders in
the name of deterrent beliefs becomes morally obnoxious.

But if our administrators have an obligation to test their
methods, it is an easy obligation to ignore. Offenders are the
least powerful of pressure groups, and the official is subject to very
real temptations to avoid research—in addition to the cost of
evaluation, any honest effort to test a program is a threat to its
continued existence and a potential shadow on the prior judgment
of its administrators. Criminal law administration thus represents
an almost ideal mixture for evaluative inaction: powerless sub-
jects, economic excuses for resistance to research, and an adminis-
trative incentive to avoid the test of long-held beliefs.

When the genuine obstacles to effective research are added to
this already potent brew, the absence of research is as easy to ex-
plain as it is difficult to justify.

Section 3. Determining Costs in Crime Prevention

To the official, cost is an important matter. To talk of the effects
that particular programs might have, without addressing the is-
sue of cost, tells the administrator only half of what he must know
before he has a rational basis for acting. Estimates of a pro-
gram’s cost are essential in deciding whether particular programs
will create sufficient returns to justify the commitment of re-
sources necessary to bring them into existence. Cost estimates are
also necessary to help the official choose between alternative crime
prevention strategies. In their simplest form, these statements are
self-evident: it would certainly be worth one million dollars to cut
Chicago’s larceny rate in half, but not one billion; it would be

25
folly to spend one million dollars on a program if the same results
could be achieved by a different method for $500,000. But typical
cost problems in crime control are far more complex.

Many different types of cost are incurred by law enforcement
systems. The commitment of resources to publicity campaigns
and to the administration of criminal justice in courts involves
monetary costs and the use of personnel that could be deployed
in other phases of crime prevention. The nonmonetary costs of
publicity campaigns include the environment of fear or tension
that such campaigns might produce and the possibility of a grad-
ually diminishing willingness on the part of the public to pay
attention to official pronouncements. The nonmonetary costs of
a low commitment of resources to courts include delay, injustice,
and the appearance of unfairness to those who come before the
bar of justice.

By far the most expensive aspects of law enforcement are po-
lice and punishment. An extra police officer assigned to a partic-
ular task will cost anywhere from ten to twenty-five thou-
sand dollars per year. And giving the policeman this job
means that one less policeman will be available for other duty.
The nonmonetary costs of extra policing include effects of a “po-
lice environment” on community atmosphere and on attitudes of
individuals subjected to increased police contacts. The monetary
costs of imprisonment to the prison administration are substan-
tial: from $620 to $2600 per prisoner per year in California, de-
pending on conditions and methods of accounting.?® Nonmone-
tary costs could include extra crime caused by the aggressions
or inability to pursue gainful employment that might be side
effects of protracted imprisonment. And, as discussed earlier,
when extra punishment is administered for exclusively deterrent
purposes, the harm suffered by the offender must also be consid-
ered as a cost by the official.

A rational crime-control policy takes account of all program
costs in determining whether programs are worth the resources
necessary to administer them and in choosing between alterna-
tive methods of achieving crime prevention. But since all costs
are not of the same type, difficult problems of comparison are
inevitable. How much money is it worth to prevent ten rapes?
How many fewer housebreakings justify keeping ten men in jail
for two extra years? How many robberies equal one burglary?
It is far easier to outline the types of cost that should be consid-
ered in evaluating alternative methods of crime control than to
proclaim with any confidence the priorities to be accorded to
different types of cost.

38 Lamson and Crowther, Memorandum, Assembly Office of Research, California Assembly,
1969 (unpublished). Supplementary data on costs were obtained from Mr. Lamson.

26
But if the successful establishment of priorities in choosing
among different types of costs proves elusive, a number of more
modest points can be made about the role of cost considerations
in deterrence. Perhaps the most important point is the most
basic: the study of costs in crime prevention is necessary for policy
research purposes as well as accounting purposes. In fact, fo-
cusing on issues of cost might provide a useful starting point
for building a simple model of the questions that evaluative re-
search should answer if it is to provide policy guides in crime
control.

An example of recent policy change might illustrate this last
point. In California, between 1963 and 1968, the average time
served by the prisoner released from the California Adult Au-
thority increased from 30 to 36 months.” Whether this change was
the result of a change in conscious policy on the part of the Adult
Authority or sentencing judges, or an unplanned occurrence,
it represents an objective shift in punishment policy that should
be considered on its merits. Are the longer prison sentences worth-
while? In order to provide a satisfactory answer to that ques-
tion, research must seek to establish: (1) the types and amount
of program costs; (2) the nature of the program’s effect on
crime; (3) the extent of program effects; and (4) the cost and
relative effectiveness of alternative methods of achieving the
same objectives.

Program Costs

As noted above, imprisonment generates a number of different
types of cost: monetary and nonmonetary, direct (such as dol-
lars expended by the State) and indirect (diminished ability of
longer-term ex-prisoners to adjust when released), costs ab-
sorbed by the State and those absorbed by the inmates. Cost
study begins with the enumeration of those costs easiest to quan-
tify and determine—the direct monetary costs absorbed by the
State, In California, the cost of maintaining one extra prisoner
for one year is $620 when there is surplus space available. The
average Adult Authority prisoner population is 30,000. The mon-
etary cost of the extra time served amounts to about $9,000,000
over the period of 36 months (one-half man-year x 30,000 pris-
oners X $620).'' These figures already exist or would be extremely
easy to compile in most jurisdictions in the United States, but
they usually command the attention of budgetary authorities for
housekeeping purposes while failing to play an important role
in basic decisions about crime-control policy.

39 Personal communication from Robin Lamson, November 6, 1969. See also, Lamson and
Crowther, supra note 388.

4 See Lamson and Crowther, supra note 88.

41 Ibid.

27
The statistics presented have already quantified a second type
of cost produced by the extra punishment: the additional con-
finement absorbed by the present prison population, an average
of six months for 30,000 men. Other types of possible cost, such
as the effect of overcrowding and decreased ability on the part
of ex-prisoners to adjust to outside life, cannot be estimated. And
there is no way of adding up the costs to the State and costs
absorbed by the prisoners into a single index number.

But even this incomplete picture of program costs provides
a useful point of departure. The utility of committing resources
to erime-control research should be judged against the total cost
of crime-control programs, and it is clear that the stakes are high.
Because the stakes are so high, the need for evaluation extends
not only to decisions about instituting new crime-control pro-
grams or committing more resources to police or corrections, but
also to existing patterns of resource allocation and existing strat-
egies of control. So little is known about present programs that
it would not be inaccurate to call most present erime-control pol-
icies expensive experiments that all too often do not permit ade-
quate evaluation. A necessary first step in evaluation is to deter-
mine, as fully as we can, exactly how much our present policies
cost.

The Nature of the Program’s Effect on Crime—
the “Whether or Not” Question

Thus far the analysis has only considered the liabilities in-
curred as a result of the change in sentence length. The motiva-
tion for longer sentences is the hope that they will reduce crime,
and a necessary step in evaluation is determining whether the
strategy is achieving its goal. The extra prison sentences could
be reducing crime in at least three ways: (a) incapacitation of
offenders for the extra six months could reduce the crime rate be-
cause the offenders would have committed crimes during those
months if they had been released, (b) the longer prison sentence
could reduce crime by persuading those prisoners serving the
longer sentence that future efforts at crime are not worth the
risk of penalties (marginal special deterrence), and (c) the long-
er prison sentences could deter from crime some potential crim-
inals on the outside who would not have been deterred by the
shorter sentences being administered in 1963 (marginal general
deterrence). The longer sentences may have been motivated sole-
ly by a desire for a greater degree of marginal general deter-
rence; like most crime control strategies, however, they may af-
fect the crime rate in a number of different ways. And any
rational analysis of the results of the program must seek to

28
account for all of the different mechanisms that may have an
effect on crime.

The different types of effect must be studied in different ways.
In order to test hypotheses about the reduction of crime among
the convicts serving extra sentences, the records of these con-
victs after release should be compared to those of other similar
convicts who served shorter sentences. The test for marginal
general deterrence would focus on the crime rate of the rest of
the population. If an investigation revealed that no reduction in
crime was being achieved by the shift in policy, the evaluation
would be completed; if investigation revealed that the extra im-
prisonment was responsible for some decrease in criminal activ-
ity, evaluation of the program would just be starting.

The Extent of Program Effects—the “How Much” Question

The emphasis in Question No. 1 on the cost of a particular
crime-control measure makes evident the point that, to properly
evaluate increased sentences, it is not sufficient to determine
whether the program decreases crime. It is also necessary to ask
how much of an effect is attributable to the new policy before
judgments can be made about the desirability of its continuation.
At this point the official has traveled a considerable distance
from the question of whether punishment deters crime, for in the
case of California prison sentences he is asking about marginal
changes in punishment rather than the difference between the
threat of punishment and none, and he is interested in the quan-
tity of an effect rather than its quality. It is at this point that
the official and the theoretician may often part company, because
the same set of findings may have strikingly different implica-
tions for researchers and policy planners. For example, in the
Chicago False Fire Alarm Project, it appeared that eight-hour-
a-day surveillance by teams of two firemen could reduce an in-
dividual school’s false alarm rate if the surveillance resulted in
apprehending a child pulling an alarm. However, after each of
sixteen schools had been exposed to two man-weeks of surveillance
it turned out that apprehensions occurred at only four schools,
and the effect of the apprehension on the false alarm rate was
slight and transient, amounting to an expenditure of several hun-
dred dollars for each false alarm prevented. Thus, while the test
may have provided valuable information, and while surveillance
makes sense in a limited number of circumstances, the fire depart-
ment would not be seriously tempted to continue the program
of regular school surveillance past the experimental period.

Another example to illustrate the same principle comes from
Professor Beutel’s investigation of the administration of crimi-

29
nal laws dealing with the writing of bad checks in Nebraska.
His basic findings were that, in spite of an unusually severe pen-
alty structure, Nebraska experienced about the same rate of bad-
check writing as a similar State with less stringent policies,*?
and those areas of the State that punished particular types of
bad checks more stringently than others did not have a lower rate
of bad checks.** Since the study involved comparing areas that
might have experienced different rates of bad-check writing in
the absence of the difference in punishment policy, it would be
difficult to prove that the stringent Nebraska policy had no de-
terrent effect at all. However, since the cost of the Nebraska bad-
check policy was so great," the small measure of deterrence that
might have been missed was of no practical importance.

The Availability of Alternative Means

Even programs with crime-prevention results worth their cost
should be subordinated to alternative programs that can achieve
better results at the same cost, or similar results at lesser cost.
The dollar spent in the prison system in pursuit of crime reduc-
tion is more usefully thought of as part of a budget for crime
reduction rather than simply or exclusively as part of a prison-
system budget. Money spent on prison sentences to reduce crime
could also be spent on other changes within the prison system
(such as better job training) aimed at reducing crime. The same
funds could be spent through other agencies to achieve crime
reduction—for example, through more intensive policing—and
perhaps with greater success. Rational resource allocation in
crime control, as the President’s Commission on Law Enforce-
ment and Administration of Justice pointed out,*¢ cuts across de-
partmental lines to seek the least cost means to achieve the
goals sought in common by the different levels of the criminal
justice system. We hear this chestnut of wisdom often, yet rep-
etition will be necessary until we act as if the principle is
understood.

Cost Models and Research Strategy

The evaluation model presented above is a shopping list that
includes questions that present research techniques often are
not equipped to answer. But the perspective gained by structur-
ing a model of crime-control research procedure can prove help-

12 Beutel, Experimental Jurisprudence.

43 Ibid., at 355.

4 See Ibid., at 351.

4 According to Professor Beutel, the administration of bad-check laws in Nebraska costs more
than making all such checks good. See Ibid., at 406-07.

46 President’s Commission on Crime and the Administration of Justice, The Challenge of
Crime in a Free Society (1967), at 7.

30
ful even if the information necessary to answer significant ques-
tions is often unavailable. It is difficult, even in abstracto, to
design research that could determine with any confidence the
extent to which the increase in California sentences has influenced
the crime rate through marginal general deterrence. But even
when the extent of crime prevention cannot be precisely esti-
mated, cost studies can start official discussions about how much of
an impact on crime would be sufficient to justify particular pro-
grams, and this can simplify and help shape the design of re-
search to suit official needs. As in Nebraska, once the cost of
a program is recognized, the researcher can stop looking for
needles in the haystack and confine his search to the existence
of possible skyscrapers.

Finally, a checklist of the questions that comprehensive policy
research must answer can be valuable for what it tells us we do
not know. Once missing links between particular research find-
ings and our rather simple evaluative model are established,
those areas where findings cannot be translated into policy man-
dates can be more clearly isolated.

31
Part 111. The Deterrent Effect—A Survey of Issues

When discussing deterrence as a motive of crime-control pol-
icies, the proper focus of inquiry is on the attitudes and behavior
of officials; when discussing the likelihood and magnitude of a
deterrent effect, emphasis shifts to characteristics and reactions
of potential criminals. Part II dealt with the view of threats
subscribed to by officials. This section deals with the responses
of threatened audiences, the major portion of the analysis
being concerned with the general effects of threats: behavior
that can be expected from people even if they are not
caught and punished. Much of this discussion of general effects
will be relevant as well to special deterrence, but the combina-
tion of actual punishment and further threat raises some special
issues that are separately discussed in a brief concluding section.

Section 1. General Deterrence

The legal threat is both a distinctive class of phenomena,
within which all members share certain similar characteristics,
and a composite grouping which includes communications that
differ greatly in content and effectiveness. Exploring the concept
of general deterrence thus becomes a two-pronged search: 1. for
generalizations that apply to all threat situations, and 2. for
explanations of the differences observable in the effectiveness of
various legal threats.

Since threat responses are distributed over a spectrum that
ranges from total lack of effect to almost full compliance, it
appears at the outset that the differences between legal threats
are more important than their similarities. But if present knowl-
edge about threats in general is not sufficient to shed much light
on the utility of particular legal threats, the initial consideration
of a few generalizations about threats can still provide a frame-
work for considering more important patterns of difference.

One common element of all threats (whose defining character-
istic is the announcement that unpleasant consequences will be
attached to particular behavior) is that they give their audience
a new reason for avoiding the threatened behavior. The power
of unpleasant consequences as a reason for avoiding behavior will
vary, of course, but the existence of an additional barrier to com-
mission of the threatened behavior is constant. Whether there
are any other effects produced by all threats is debatable, but

32
investigators have reported that with the creation of the new
barrier, two further changes in outlook on the part of a threat-
ened audience may be expected.

Constraint

According to Kurt Lewin, “Threat of punishment always and
necessarily gives rise to the structure of a constraint situation.” 1
Even though the threat of consequences does not physically im-
pair the ability of audience members to engage in threatened
behavior, the prospect of punishment creates an emotional bar-
rier that makes the threatened individual feel less free, with
respect to the threatened behavior, than he would feel with no
threat outstanding.

“Forbidden Fruit Is Always Sweeter”

According to some observers, existence of the threat as a
barrier to committing a particular act causes memers of
a threatened audience to revise attitudes toward the desirability
of the threatened behavior. People who felt that the threatened
behavior was desirable in the absence of the threat will consider
the behavior more desirable, even if they are persuaded by the
prospect of consequences to heed the threat. To those who
originally were neutral about the desirability of a threatened
behavior, it will acquire some positive value, apart from the
unpleasantness of threatened consequences.?

It is probably unsafe to assume that an atmosphere of con-
straint and the “forbidden fruit” effect are inevitable products
of threats. How a threat can produce feelings of constraint
among those who were never tempted to engage in the threatened
behavior is difficult to understand, unless the fact of the threat
causes a significant reevaluation of the behavior. And where
members of an audience have considered behavior to be highly
undesirable in the absence of threat, it is not clear that the
addition of a threat will make much difference in the way they
view the behavior. Then, too, intervening variables would over-
whelm any attempt to validate this prediction in many
circumstances.

Even if constraint and reevaluation of the threatened behavior
could be added to the creation of a new reason for avoiding it,
the general model of the psychological dimensions of threats
could not lead to predictions about behavior because the feel-

1 Lewin, Reward and punishment, in A Dynamic Theory of Personality—Selected Papers of
Kurt Lewin (1951), at 129.

2 See Brehm, A Theory of Psychological Recactance. “[Wlhen a person believes himself
free to engage in a given behavior, he will experience psychological reactance if that freedom
is eliminated or threatened with elimination.” There is some basis for thinking that while
threatening behavior may init'ally lead to a hivher valuation of that behavior, persons who
refrain from the threatened behavior because of the threat will afterward think of the behav-
ior as less desirable. See Brehm and Cohen Euaploration in Cognitive Dissonance, at 40-41.

33
ings toward the threatened behavior produced by threats lead in
opposite directions, leaving the behavior more dangerous at the
same time that it is more attractive, and no general statement
can be made about which pull will be stronger.

At the same time, the “forbidden fruit” effect may exist in a
wide variety of situations where members of an audience were
not previously disinclined toward the behavior threatened, and
the above psychological model of threat-effects can provide in-
formation about the emotional climate in which threats operate,
even though it cannot predict the outcome of threats. For those
who are tempted to commit threatened acts, threats produce
reasons for avoiding them—reasons that compete with the at-
tractiveness of the behavior, and with any new attraction that is
added by threat. Where a legal threat is relevant to predicting
behavior, it will produce conflicts. How such conflicts are re-
solved is a function of the differences between threats, but the
existence of conflict is a common theme.

Still, the critical work in any discussion of the effectiveness
of threats is the search for those variations in circumstance
that account for the great differences noted in the results of
threats. Toward this end, the following sections discuss (a)
differences among men, (b) types of threatened behavior,
(c) threat as communication, (d) applicability and credibility,
(e) variations in threatened consequences, (f) variations in severi-
ty of consequences, (g) the moral quality of threatened behavior,
and (h) the effects of group pressures.

A. Differences Among Men

One natural place to search for the explanation of different
patterns of response to threats is within the great range of
differences in psychology, sentiment, and status that set men
apart from each other in a large and complex social order. Ex-
planations of differential deterability that are based on dif-
ferences among men—unlike hypotheses about differences in de-
terrence caused by variation in the nature of the threatened
behavior—seek to establish general patterns of threat sensitivity
that follow a particular person from one situation to another.
Where it is known that Mr. Smith’s relative immunity to the
threat of sanctions while filling out his income tax is attributable
to a personality characteristic, it is likely that the same charac-
teristic will operate when he is exposed to the temptation to
steal, or to speed. If differences among men were the only ex-
planation of differential deterability, society could be neatly
divided into deterable and nondeterable segments. But since
differences among men can only be a part of the complex of

34
factors conditioning the effectiveness of threats in different sit-
uations, general patterns of threat responsiveness based on per-
sonal differences will predict responses with only partial effec-
tiveness. Knowing that Smith is relatively unconcerned about
the threat of punishment for tax cheating, and that this lack of
concern is related to a general tendency, will not tell us that he
is more likely than not to be undeterred by the threat of punish-
ment if confronted with the opportunity to commit armed rob-
bery. Such information will, however, mean that Smith is more
likely than others without this characteristic to be unimpressed
with the threat of punishment for robbery. But the chances of
deterrence in the second situation may still be overwhelmingly
high.
Differences in Personality Type

A great number of different personality types co-exist in the
same social order. Various schools of thought in the psychic
sciences use different schemes of classification to describe varia-
tions in personality type. Most of the personality differences that
have been mentioned in connection with the effect of sanctions
are based on simple distinctions not related to systematic ex-
planations of human personality.
The Future vs. the Present

One such relatively simple distinction is between those in-
dividuals for whom the future is an important part of present
thinking and those who are less “future-oriented.” The future-
oriented individual will be happy to forgo a large gain today
in order to enjoy a larger one next week. The person without a
significant sense of the importance of next week will accept a
lesser but immediate reward.’ In the case of threat response, it
is the converse of this effect that is significant: since the un-
pleasantness threatened will always come after the commission
of the threatened act, and is often quite distant in both
time and probability from the temptation to act,? the future-
oriented person will pay greater heed to the possibility of un-
pleasantness communicated by a threat, while the less future-
oriented person will be more willing to place a heavy mortgage
on events removed in time and space. Observation has established
that deviant subcultures place less emphasis on future events
than representative members of the middle class, a finding that
lends credence to the view that these groups are generally less

1 See Mischel, Preference for Delayed Reinforcement: An Experimental Study of a Cultural
Observation, J. Abn. Soc. Psych., 56 (1958).

2 “Punishment is psychologically farther off than satisfaction of . . . desire (so that) a cer-
tain intellectual maturity is a necessary condition of the effectiveness of threatened punish-

ment.” Lewin, Reward and Punishment, in A Dynamic Theory of Pcrsonality—Selected Papers
of Kurt Lewin (1951), at 168.

35
deterable, and could support the assumption that this feature
of personality accounts, in part, for the higher rates of criminal
behavior among some subcultures.?

The problem with attributing higher crime rates to lesser de-
terability and lesser deterability to the lack of emphasis on the
future among some subcultures is that of separating cause from
consequence. The relative lack of emphasis on the future may
simply be a reflection of the fact that people in some positions
have less to look forward to than people in others. If a tendency
to discount the future is the result of environmental conditions,
then these environmental conditions and not any separate trait
of personality are responsible for a different view of threats
among members of the culture of poverty. Moreover, even if it
is plausible to assume that a short-term world view contributes
to lessened degrees of threat responsiveness, it is difficult to jump
from that finding to the further conclusion that such a difference
is responsible for much of the increased rate of crime observed
among such groups. Too many variables—each a plausible ex-
planation for higher rates of crime in its own right—intervene
to spoil the respectability of such an assumption. And we are
left with the unavoidable possibility that a discounting of the
future is the result of a number of other psychic and physical
conditions that account for the diminished effectiveness of
threats.

“Optimists” vs. “Pessimists”’

Unless members of a threatened audience choose to ignore
threats “with a policeman at their elbow,” reacting to a threat
may involve weighing the chances of being apprehended and
punished. To a large extent, people weighing the chances of
being apprehended and punished will be influenced by the ob-
jective probability of being discovered. But most potential law-
breakers will not have complete information about the risk of
apprehension, so that subjective judgments will play an impor-
tant role in the individual's estimate of his chances of escaping
without apprehension and punishment. Moreover, psychological
research has established that even when it is possible for a person
to know the objective probabilities of achieving a particular
goal, a person’s estimate of his own chances of success will
often differ from objective probabilities.* It is thus plausible
to suppose that people who more often underestimate the
chances of being apprehended if they engage in threatened
behavior will be less responsive to threats than persons
who are more likely to overestimate the probability of the

3 Mischel, supra note 1, and sce Lewis, La Vida, xlii-lii.
4 See Cohen, Chance, Skill and Luck, Pelican Books (1960).

36
threatened consequences occurring. The worrier will tend to over-
estimate the objective probability of being caught, if he is not
completely informed about the probability of apprehension, while
the optimist will tend to underestimate apprehension chances.
Even if both the worrier and the optimist know what the chances
are, in general, of being caught, the worrier will tend to feel
that his personal chances of successful evasion are no better than
anybody else’s, and may be worse, while the optimist will feel
that his personal chances of successfully escaping threatened
consequences are much better than the general average. As a re-
sult, the optimist will prove to be a far more difficult person to
deter.

In one recent study addressed to the optimist vs. pessimist
theory, Daniel Claster asked a sample of delinquent boys and a
sample of nondelinquent boys to estimate (a) the general chances
of being apprehended and punished, and (b) their personal
chances of being apprehended and punished, in relation to a varie-
ty of offenses.” The two samples gave similar answers about the
general chances of being apprehended and convicted, and about
their personal chances of being convicted of offenses if apprehend-
ed, but the delinquent boys perceived their personal chances of
arrest if they committed crimes to be significantly lower than the
personal chances estimated by nondelinquents. These results might
be evidence of the delinquents possessing, as Dr. Claster describes
it, a “magical immunity mechanism” ¢ that serves to neutralize
fear of punishment, and makes this group less susceptible to de-
terrence. It is curious, however, that the same mechanism does
not appear in answers about chances of conviction after arrest.’
Moreover, it is possible that the delinquents’ estimates of personal
apprehension chances were accurate, and based on personal ex-
perience not possessed by the nondelinquent boys. So, available
data fall short of establishing that personality traits associated
with differential estimates of risk play a significant role in se-
lecting those members of society who fail to be deterred by the
threat of punishment.

Even if studies showed that criminals were more optimistic
than the facts warranted and less worrisome than law-abiding
citizens, there still remains the problem, alluded to earlier, of
separating cause from consequence. Dr. Claster spoke of the op-
timism expressed by delinquents as being an “immunity mecha-
nism” that “protected” delinquents from fear. Does this mecha-
nism exist before the boy embarks on a delinquent career, or is

5 Claster, Comparison of Risk Perception between Delinquents and Non-Delinquents, 58 J.
Crim. L. C. & P. S. 80 (1967).

 Ibid., at 86.
7 Ibid.

37
it a result of ignoring threats, i.e.,, a point of view that develops
after deterrence has failed, and can therefore hardly explain the
initial failure?

Finally, it should be noted that the “optimist” who acquires
that title because he tends to underestimate risks of apprehension
is not necessarilv a person with a generally rosy view of the
world around him. If the hypothesis holds true that persons
without a strong sense of a personal future doubt that the future
is promising, then the most threat-resistant personality of all
would be a man who tends to underestimate risks, yet holds a
generally pessimistic view of his chances for future success.

Planning vs. Impulse

Threats seek to create a strong association between threatened
behavior and threatened consequences in the mind of the po-
tential deviant, but such an association may require that members
of a threatened audience devote time and thought to the meaning
of their acts. In this sense, the process of deterrence by threat
may be one of accretion.® The person who acts on impulse is less
likely to spend time thinking of consequences, and if the associa-
tion between act and consequence cannot be fully realized in
the span of time consumed by an impulse decision, such actions
will be insulated from the accretive impact of punishment
threats. If decisions to act are the product of a longer period
of gestation, thoughts of unpleasant consequences to follow will
have more time to chip away at temptations to deviate. Where
impulsive actions are characteristic of an individual’s personal-
ity, he will, according to this theory, be generally less susceptible
to threats than the person whose actions are usually preceded
by planning.

The distinction between impulsive and premeditative personal-
ity styles differs in theory from that between future-oriented
and present-oriented personalities because the emphasis of the
former is on the amount of time available for reflection, rather
than on the relative value of present and future gratification.
In practice, however, impulsive conduct and a lack of future
orientation may most often occur together. Here the problem of
Cause vs. Consequence recurs again, for an impulsive life-style
may be the result, as well as part of the cause, of paying little
heed to the threat of punishment.

8 The possibility that threat communication and influence may result from cumulative impact
over time was noted in another connection, by Barmack and Payne, when observing that the
delay between an experimental countermeasure and significant results may have been due to
the fact ‘that the spread of information about the countermeasure is an accretive process
which necessitates repeated announcements . . . (my italics). Barmack and Payne, The Lack-
land Accident Countermeasure Experiment, 40 Highway Research Board Proceedings, 65138
(1961).

38
Conflict vs. Harmony

Some speculations about personality differences that may af-
fect threat-responsiveness grow out of more ambitious theories
of human behavior. Forty years ago Alexander and Staub theo-
rized that deterrence functions differently in the different per-
sonality types posited by the Freudian system.? According to these
writers, anyone suffering from organic brain damage is totally
beyond the reach of sanction threats, for he is often unable to
understand and act in relation to threats.® Three other offender
types (the neurotic criminal, the normal criminal, and the acute
offender) are discussed in greater detail. The neurotic criminal, to
Alexander and Staub, is a persistent offender with psychological
conflicts that are the basis of this criminal acts. With this
type of offender, crime is a symptom of personality disorder,
consisting of impulsive acts far removed from ego controls,
and presumably far rmoved from susceptibility to punishment
threats." The normal criminal they regard as a psychologi-
cally normal individual with a persistent history of criminal
behavior that probably results from early identification with
criminal prototypes. Criminal behavior is quite consistent with
the normal criminal’s ego structure, and thus does not lead to
inner conflict. The acute offender is described as a psychologically
normal individual without any long history of crime who is
driven to an isolated criminal act by a crisis situation.

Of these classes, Alexander and Staub consider the normal
criminal most susceptible to deterrence: “We admit that . . .
fear of painful consequences . . . may prevent [normal crimi-
nals from] committing anti-social acts, or may at least reduce
the number of the latter.” 2

The authors make no specific reference to deterrence and the
acute offender, but presumably the relative effectiveness of
threats would be hampered by the fact that so many other bar-
riers to criminal activity that normally operate will have been
overpowered by the force of circumstances before the acute of-
fender is confronted by the possibility of crime. Alexander and
Staub appear to doubt the effectiveness of penalty threats on
neurotic criminals because inner conflict, which militates against
rationality of responses, is the cause of this type of criminality,
and the degree of ego involvement in crime is low.

These observations, it should be recalled, are based on different
personality types among criminals, rather than among the gener-

9 Alexander and Staub, The Criminal, the Judge, and the Public, rev. ed., Free Press of
Glencoe (1956).

1 Ibid., at 82.

1 Ibid., at 42-46.

2 Ibid., at 210.

39
al population.’® Thus, this theory does not hold that neurotics are
less deterable than other personality types—indeed, they could
be more deterable as a class—but only that those individuals
who commit crime as the result of neurotic conflicts are seldom
deterred by threat of punishment.

Of further interest would be estimates of how many criminals
fall into each of the personality types outlined by Alexander and
Staub, and what types of crime different personality types com-
mit. Such estimates would be helpful, both as a necessary step
toward testing the hypothesis that inner conflict precludes de-
terrence and as a method of gauging the potential importance of
this type of insight on penal policy.

Crime, Deterrence, and Personality Characteristics

There are a number of other personality dimensions that ob-
servers have suggested may be related to the propensity of in-
dividuals to commit crime: it seems reasonable to suppose that
aggressive personalities will be more likely than passive person-
alities to act out impulses of all kinds, including criminal im-
pulses; individualistic personalities may be more likely than
group-centered personalities to deviate from group norms; and
the suggestion has been made that certain musculature-related
psychological types are more prone to crime than others.

All of these predictions, if correct, will mean that a greater
proportion of the crime-prone personality types will commit
threatened acts more often than those individuals displaying dif-
ferent personality traits, although the rate of deviance among
the crime-prone may still be quite small. But the higher rate of
noncompliance noted in such groups is not necessarily tied to dif-
ferent patterns of threat response. It may simply be that pres-
sures toward deviancy, or barriers (other than deterrence)
to deviancy are distributed unevenly among personality types.
Isolating facets of personality that are associated with high
involvement in criminal activity will tell us much about the type
of people on whom the operation of threat and punishment is
of the most critical importance to a legal system. But if
crime-related personality traits are not accompanied by psychic
mechanisms that affect threat responses, the crime-prone individ-
ual may still be just as susceptible to threats as anybody else. For
example, a disproportionate number of aggressive personality
types will commit aggravated assault each year. But the threat of
punishment may be equally effective on aggressive and passive
personality types, as illustrated by table 1.

"1 For an elaboration of the difficulties encountered in generalizing from a sample of impris-

oned criminals to other groups, see the discussion of The Warden's Survey, supra Section
II A.

40
Table 1
Hypothetical Percentage Distribution of Aggressive and Passive
Personality Types by History of Criminal Conduct and Effect
of Threats.
Passive Aggressive
Would not have committed as-
sault in any event 97 70

Would have committed assault, but
were deterred by threat of conse-

 

 

quences i 2 20
Committed assault in spite of
threat of consequences _ 1 10
Total 100% 100%

Knowing only the proportion of known offenders that are ag-
gressive-personality types, it would be tempting to conclude that
threats have less of an effect on the behavior of aggressive per-
sonalities than on passive personalities. When the percentage of
persons of each type that would have committed the crime but
were deterred is also known, we find that threats were equally
effective among both groups, deterring two out of three potential
assaulters. And the deterrent effect among aggressives resulted
in ten times as much crime prevention as among passives.

If a particular personality trait were associated with deter-
ability, it would give us something like table 2.

Table 2
Hypothetical Percentage Distribution of Future-Oriented Per-
sonality Types by History of Criminal Conduct and Effect of

Threats.
Future- Non future-

Oriented Oriented

Would not have committed assault

in any event = __ 80 80
Would have committed assault, but

were deterred by threat of conse-

 

 

quences ____________ 17 14
Committed assault in spite of
threat of consequences oo 3 6
Total 1009, 1009,

In this case, a difference in threat responsiveness appears to
explain the difference in crime rate. But the hypothetical tables
raise rather clearly some of the problems encountered in trying

41
to relate personality traits to deterability, and deterability to
differences in crime rates. In the first place, personality traits do
not come one to a customer—a particular individual will have
several different traits that should be accounted for in trying to
relate his behavior to elements of his personality. And people
who differ in one personality trait may differ in others as well.
Thus, if a greater proportion of aggressive persons than passive
persons were lacking a significant sense of the future, table 1
would have appeared to show that aggressive persons were less
deterable than passive types, even though aggressiveness per se
did not affect deterability.'*

A second problem is pointed up by the fact that the tables
presented above are hypothetical. There is, unfortunately, no
direct method of determining what proportion of a population,
let alone what proportion of a particular group within a popula-
tion with a particular personality type, “would have committed
assault, but were deterred by threat of consequences.” For this
reason, the tables serve as an illustration of the difference between
crime-related and more specifically deterrence-related personality
traits, rather than a method of researching the influence of char-
acter traits.

But a more basic objection may exist to the entire discussion
of the relationship between personality traits and deterrence. It
may be argued that even if personality differences do influence
deterability, such an influence is not relevant to legal policy be-
cause a threat is usually issued generally, and its audience will
be composed of a mix of personality types, so that “tailoring”
threats to particular personality types would be impossible.

In considering this objection, two points can be made in favor
of research into the relationship between personality and deter-
rence. First, there are reasons for supposing that any knowledge
that can be gained about the relationship between threat-response
and personality-type can be useful, even if all categories of crime
are committed by a mixture of different personality types. If the
deterrent threat issued to the community-at-large can be altered
to exert greater influence over a particular personality type with-
out impairing its effectiveness among others, this will strengthen
the effectiveness of the threat. If threats to the community-at-
large are not altered, the way in which threats are communicated
to particular groups and individuals can be changed. Moreover,
even if the general deterrent threat of the law cannot be special-
ly designed for particular types of offenders, the form of special-
deterrent threats probably can be tailored to individual person-
ality differences.

M See Zeisel, Say It With Figures, 5th ed., at 132-46.

42
A second answer to the objection that personality differences
are not relevant to penal policy is that, while crime as a general
category may reflect participation by the whole spectrum of per-
sonality types, some specific crimes may be committed much more
often by particular personality types, and with respect to those
crimes, the relationship between personality type and threat-
responsiveness may be valuable. While larceny may be said to be
the province of no single personality type, except perhaps the “lar-
cenous personality,” can the same be said for public exhibitionism
or even violent assault? Even supposing that no remedies exist
for the immunity to threats associated with particular personal-
ity types, if a certain personality type dominates the crime
statistics and population of potential offenders in one or two
specific offenses, it would appear useful to know that the de-
terrent threat is not achieving its objective.

Yet it cannot be denied that our present knowledge about
differences in human personality has less to offer in terms of
penal policy than we would hope. Data about the relationship
between deterrence and personality differences is tentative, and
even if more becomes known, such further knowledge will prob-
ably prove far more useful in the design of correctional pro-
cedures than in the realm of general deterrence.

Differences in Attitude

A second area where individual differences may lead to dif-
ferent patterns of threat-response is found in those patterns of
differences in attitude that are prominent features of plural
societies. This analysis uses the term “attitude” to describe nor-
mative feelings about different features of a person’s environ-
ment; therefore, attitudes and personality characteristics are not
mutually exclusive classifications. In some cases it might be proper
to view an attitude as a personality trait in its own right; in
others, an attitude will be the product of a particular personality
trait; in many cases, differences in attitude between men are
simply the product of differences in experience.

This section discusses differences in socialization and in specific
attitudes toward authority that may help to explain why some
men are deterred by threats and others are not.

Socialization

Within the same society, the life experience and therefore the
learning processes of individuals may vary greatly. Some men will
grow to adulthood with strong loyalties to the social system, as
well as to the smaller family, clan, institutional and community
groups in which they live most of their lives. Others develop

43
strong loyalties to the smaller groups, but are less committed to
the larger social entity. Some may grow to reject both the ideol-
ogy of their particular living groups and that of the larger
social system. The intensity of personal commitment to the goals
of a society is a significant predictor of law-abiding behavior for
one reason that has nothing to do with simple deterrence: the
strongly socialized individual will obey commands out of a desire
to do right, quite independent of the specific consequences of
wrongdoing. Of less importance, but still significant, a strongly
socialized individual may be more likely to comply with threats
because he is more sensitive to the negative aspects of threatened
consequences than his less patriotic neighbor. Social disapproval,
which is an important part of most threatened consequences,
will seem to the strongly socialized individual an experience well
worth avoiding, whereas less socialized men might be less im-
pressed. Because the desire to obey commands in order to do the
right thing and a greater sensitivity to consequences involving
community disapproval will occur together, it is extremely dif-
ficult to estimate how much of an observed difference between
strongly and weakly socialized individuals is attributable to the
greater sensitivity of the former to the negative aspects of
threatened consequences. Perhaps the crucial test of such an
effect would come in an area covered by a legal threat where the
chances of apprehension were known to be nil. Under such cir-
cumstances, the desire to be law-abiding for its own sake would
presumably still be operative, while a greater sensitivity to the
negative aspects of threatened consequences would not.

While no criminal behaviors exist in our society for which the
chances of apprehension are absolutely nil, it might be suggested
that areas of behavior where the chances of apprehension are
extremely low would serve as adequate substitutes. It could also
be argued, though, that since it is part of human nature to avoid
low risks of very high costs, those who fear the law’s sanctions
most will be influenced by that fear even where the probability
of risk is statistically minimal. Thus, sensitivity to sanctions may
influence rates of compliance even in such areas as the require-
ment (once difficult to enforce) that interest and dividends be
reported on individual income tax forms.*®

In the absence of conclusive data, a plausible thesis is that
where a threatened behavior is considered to be a serious breach
of society’s moral code, the major explanation for the higher
rate of compliant behavior is the strongly socialized citizen’s
sense of right and wrong, rather than his special sensitivity to
the negative aspects of threatened consequences. Where a threat-

18 See Schwartz and Orleans, on legal sanctions, 34 Chi. L. Rev. 274 (1967).

4
ened behavior is considered a less drastic breach of the moral
code, a special sensitivity to the negative aspects of threatened
consequences may play a noteworthy part in explaining the dif-
ference in behavior between these two groups. How much of a
role such considerations play, either in absolute terms or relative
to the socialized individual’s stronger drives toward doing right,
cannot be presently estimated.

Attitudes toward Authority: The “Authoritarian Personality”

Some people identify with authority more intensely than
others, and thus find it easier to obey orders just for the sake
of obeying orders. It is plausible to suppose that this type of
person will be more likely to refrain from committing a partic-
ular behavior because it is threatened than a person whose
specific attitudes toward authority are either more ambiva-
lent or more flexible. The person who obeys orders for their own
sake will do so whether or not he considers the threatened be-
havior as seriously wrong (in the absence of the threat) and
largely independent of his chances of being apprehended. While
people with less respect for authority may see an element of
challenge in a threat and be tempted to try to “get away with
something,” the authoritarian personality will comply with the
threat, and will be more likely to change his opinion of the
rightness or wrongness of the behavior threatened solely because
it is threatened.

These predictions are based on studies of the attitude char-
acteristics of the authoritarian personality begun during the mid-
1940s.’* While some overlap exists between the theory that strong
socialization produces greater compliance with commands and
these predictions about the behavior of the authoritarian per-
sonality, important differences can also be noted. The author-
itarian personality is not necessarily more strongly socialized
than the less authoritarian personality. He is more likely to ex-
press racial and religious prejudice than the nonauthoritarian,
even though these prejudices are not officially condoned.” The
authoritarian is no more likely to hold strong moral views about
the rightness or wrongness of conduct than the nonauthoritarian,
and his opinion, if not firmly held, is easier to change.’® Finally,
the authoritarian might respond with obedience, not only to legal
commands, but to orders from those in most high-status positions.
Such a tendency could lead to cooperation, for example, with oc-
cupying powers or other high-status individuals in positions of

1% Adorno, Frankel-Brunswick, Levinson, and Sanford, The Authoritarian Personality (1950).
17 See Ibid.
18 Crutchfield, Conformity and Character, 10 Amer. Psychologist 191 (1955).

45
power whose authority does not stem from the social value system
native to the authoritarian.

The experimental evidence linking the “authoritarian person-
ality” to possible differences in response to threats is circum-
stantial because the study of authoritarian attitude patterns has
been limited to finding differences of attitude among different
groups, and relating one type of attitude response with other
attitudes or with attitude change. Because attitude surveys elicit
only verbal responses, it is important to distinguish between such
expressions of sentiment and tendencies to act. Available evidence
on the authoritarian personality relates to sentiment, while any
conclusion about differential deterability must describe a tendency
to act.

Attitude toward Authority: The “Anti-Authoritarian Personality”

Of special importance to the student of criminal sanctions is
the possible existence of a small segment of the population that
rebels against the whole concept of obeying orders. Rather than
being characterized as authoritarian or nonauthoritarian, this
class is properly thought of as anti-authoritarian. A person with
this sort of attitude toward authority views each new threat as
an invitation to defiance. The psychological prediction that
threats create an atmosphere of restraint and make the threaten-
ed behavior initially more attractive than it would be in the ab-
sence of the threat applies to this group with special force. Unless
this class is substantially affected by a wish to avoid unpleasant
consequences, the net effect of a threat would be to make the
threatened behavior more attractive to the anti-authoritarian. For
this group, then, the chances of apprehension and the extent of
unpleasantness threatened are of particular importance in deter-
mining threat response, because there are very few other bar-
riers to committing the threatened acts. To some extent, the
attractiveness of a behavior to an anti-authoritarian may increase
as the consequences of apprehension are escalated, because the
greater penalty is evidence of a greater degree of dare attached
to the legal threat. But, with all forms of consequence except
some expressions of social disapproval, the anti-authoritarian per-
sonality probably will show no special appetite for unpleasant-
ness. For this reason, it is likely that the probability of apprehen-
sion and the magnitude of unpleasantness threatened will serve
as behavioral controls, either by keeping anti-authoritarian sen-
timents from expressing themselves in conduct or by governing
what types of conduct (presumably those threatened less severely
or with less credibility) will be chosen for expressing antisocial
drives.

The “anti-authoritarian personality” serves as an example of

46
the special role that relatively small groups may have in criminal
studies. In normal times, the anti-authoritarian personality
would not be of great interest to the psychologist trying to ex-
plain the outcome of presidential elections because this group is
so small. But recurrent and serious criminality is the product of
a relatively small segment of the community. And if small sub-
groups are disproportionately represented in the ranks of crim-
inals and potential criminals, they may account for a major share
of crime. If the socialization process keeps most people law abid-
ing, the effectiveness of simple deterrent threats must be judged
in terms of that part of the population which remains potentially
criminal after socialization has done its work. In many cases,
this residue may consist of small groups who differ substantially
from the rest of the population in many respects.

Status

Perhaps the most dramatic relationship one finds in studying
the characteristics of people arrested and imprisoned is that
which exists between social status and crime. Known criminals
in our society have less education, lower social standing, less
opportunity to become socially mobile, and less money than non-
offenders. As a corollary, all that is known about crime in our
society suggests that the well placed and well off commit fewer
and less serious offenses. While various forms of bias built into
the reporting of crime and the enforcement of criminal laws may
account for some of the difference noted between classes, it would
be fanciful to argue that all of this difference can be attributed to
bias. A more plausible explanation of most of the difference in
criminality noted between economic and social classes is that
crime is relatively more attractive to persons who have little
to gain through legitimate patterns of conduct, and relatively less
attractive to the man who has legitimate means available to satis-
fy most of his desires. The wealthy feel the temptation to commit
crime less often because they have more of what they want, and
because experience has taught them that legitimate means are
open to them for obtaining their objectives. The poor have
achieved fewer of their wants, and the prospects for future
achievement through legitimate channels seem remote. Under such
circumstances, criminal means to achieve some of the things all
men desire will seem rational to the poor far more often than to
the wealthy, and the temptation to rebel will be felt more in-
tensely by those with less to show for adherence to the predomi-
nant system of values.

1 See Zimring and Hawkins, Deterrence and Marginal Groups, 5 J. Research in Crime and
Deling. 100 (1968).

47
If differences in perceived opportunity through legitimate
means were the complete explanation for the relationship between
crime and social and economic status, there would be no basis
for concluding that low-status groups are any less amenable to
the influence of threat and punishment than those of higher
standing. As noted above in the discussion of crime and personal-
ity traits, finding that one group is tempted by criminal alter-
natives more often than another does not show that such a group
is less sensitive to threats. There are, however, reasons for be-
lieving that some relationship does exist between social standing
and the effectiveness of threats. First, it seems likely that those
who attain high status will possess many of the personality char-
acteristics that may be associated with maximum threat influence,
such as a sense of the significance of the future and strong loyal-
ties to a social system that has been responsible for much of their
success. By the same token, lower-status groups will contain a
disproportionate number of less consistently socialized and more
present-oriented people. If this theory holds true, an association
between deterability and social status will exist, but can be at-
tributed to factors other than the difference in status.

A second possibility is that personal success makes an individ-
ual more susceptible to the influence of threats because success
determines the amount of investment in society an individual puts
at risk when engaging in a threatened behavior. The greater the
investment at risk, the more likely it is that he will obey com-
mands. In this sense it could be said that the man who has every-
thing also has everything to lose, as well as little to gain, from
the commission of forbidden acts. The other side of the coin is
equally important:

Deterrence does not threaten those whose lot in life is
already miserable beyond the point of hope.?°

It follows that one of the most important ideas about those
factors that condition differences among men in relation to threat
and punishment is that those with a greater stake in life con-
tinuing as it has been will have more to fear from the legal threat
of unpleasantness and social disapproval. But this observation
cannot be made into a complete repudiation of the possibility of
deterrence among low-status groups, because bad conditions can
always get worse. Even the poorest and least free of men in our
society have life itself and some measure of personal freedom at
stake when considering the prospect of criminal acts. Very few
men have ever indicated a preference for prison over the worst
of our slums, and the prospect of social stigma would not be wel-

20 Packer, The Limits of the Criminal Sanction, at 45.

48
comed by most of those with the lowest stake in maintaining the
status quo.

Further, it cannot be said that threat responsiveness and in-
vestment in the status quo rise and fall together in any sort of
calibrated, arithmetic formula, particularly since a person's
opinion of the worth of his present life opportunities, rather than
status as an objective fact, is the significant factor in predicting
response to threats. Those with extremely high positions in life
may have motives for obeying the law stronger than any other
group. But there is no reason to suppose that a lower middle-class
person would consider his investment in the social system as any
less worth protecting than that of a wealthier neighbor. The
“investment” theory of deterrence is stated most powerfully
when it is put forward as a contrast between the great mass of
people with much to lose if apprehended for law violation and
those who feel they have so little that fear of loss is of less
importance.

B. Types of Threatened Behavior

It has often been observed that the threat of consequences
appears to operate more effectively in controlling some types of
criminal conduct than others. While evidence on the number of
potential criminals of any kind who are deterred by threats is
sparse, there is reason to feel that the jealous husband bent on
homicide is less likely to be deterred by legal threat than the
housewife considering pilferage as a method of reducing her
grocery bill, or the bank clerk who views embezzlement as a
method of securing funds to invest in a beefsteak mine.! Varia-
tions in the content of legal threats may also have greater or
lesser impact, depending on the crime: the rate of illegal parking
is believed to be more sensitive to changes in the credibility and
severity of threats than the rate of incest.

That deterrent threats function differently with respect to var-
ious types of threatened behaviors should come as no surprise,
since crime is a separate and unitary class of behavior only in
the legal sense. In all other respects, classes of criminal activity
differ from each other as importantly as any kind of criminal
behavior can be said to differ from lawful conduct. Of particular
significance in seeking to explain why legal threats function with
varying degrees of effectiveness are differences in:

1. the kinds of people drawn to different types of criminal
activity;

! See, e.g., Chambliss, Types of Deviance and the Effectiveness of Legal Sanctions, 1967 Wisc.
L. Rev. 703, at 704-12.

49
2. the pre-barriers, other than legal threats, that may con-
dition the rate of a particular crime;

3. the motivational situation typically associated with a par-
ticular crime;

4. the environment usually surrounding decisions to commit
a particular crime.

Differences among Potential Criminals

The type of person most likely to be a bank robber will differ
significantly from the potential bank embezzler if only because
the opportunities and skills necessary for the commission of these
crimes call for different sorts of people.? The potential embezzler
is apt to be of higher status, more socialized, and more respectful
of authority than the potential robber, and the class of potential
embezzlers may have different characteristic personality traits.’
If social and personality differences influence susceptibility to
the influence of threats, differences in the type of potential crim-
inal are one possible source of the variance in threat effective-
ness noted by type of crime.

Differences in Pre-Barriers, Other Than Threats

There are a number of reasons, other than the threat of legal
consequences, to refrain from homicide, including personal ab-
horrence of killing and powerful social values that consider such
behavior reprehensible. Apart from the legal consequences, there
are very few reasons to refrain from parking in a prohibited zone
when there is something to be gained from this course of action.
This point may be illustrated by a hypothetical table comparing
the behavior of persons tempted to commit homicide with those
tempted to park illegally.

Table 1

Hypothetical Distribution of 1000 Persons Tempted To Commit
Homicide and Illegal Parking.

Homicide Parking
Restrained from offense as result
of other barriers __________________ 990 50
Would have committed offense, but
were deterred by legal threat ~______ B 800
Committed offense _________________ 5 150
Total CL ~_ 1000 1000

2 Personality differences that might affect threat responses are discussed in subsection A,
supra.

3 See, e.g., Zimring and Hawkins, Deterrence and Marginal Groups, 5 J. of Research in
Crime and Deling. 100 (1968) for a discussion of possible differences in the composition of
criminals and potential criminals among various crimes.

50
In table 1, the complex of legal and social controls was more
successful in restraining potential killers than in preventing il-
legal parking: a total of five persons committed homicide, com-
pared with 150 parking violators. But the legal threat deterred
only half of the potential homicides that were not prevented by
other considerations, while deterring more than 80 percent of
potential parking violators who would have committed that of-
fense in the absence of threat. In the case of illegal parking, it
is important to note that there is nothing unusual about the 950
people who were not restrained from considering this course of
conduct as the result of factors other than deterrence—there are
very few social forces operating to make people fear or abhor
such minor deviations from social norms. On the other hand, the
small minority of people who continue to consider homicide in
spite of the powerful social and personal feelings about killing
are a special group, either powerfully motivated toward homicide
or insensitive to social values, or both. And the characteristics
that set this group apart from most others will ordinarily make
them less responsive to the threat of consequences. Thus, while it
may be true that “the effectiveness of deterrence varies in inverse
proportion to the moral seriousness of the crime,” * the increased
effectiveness of deterrence noted in relation to minor offenses may
be due to the impact of threats on persons for whom a deterrent
threat is not necessary to prevent more serious criminality.

Differences in Motivation Associated with Particular Crimes

For the individual who does not need a deterrent, the threat
of consequences can be received and digested without tension or
conflict, but processes other than threat will be responsible for
conformity to legal commands. The notion of deterrence suggests
that threats play a dynamic role in the behavior of some people,
establishing barriers to conduct that do not exist without the
threat and changing tendencies to violate the law into patterns
of conformity. The motivational conflict produced by threats is
the competition between the reasons for avoiding threatened be-
havior provided by the threat and the motives for engaging in that
behavior which lead to the need for external controls. Discussing
the credibility and severity of a threat helps to predict threat
response by providing information about only one side of such
a conflict: the drives to avoid criminal alternatives that emanate
from a particular threat. It is equally important to know what
kind of drives, and of what intensity, are responsible for pres-
sures toward deviation, for the balance between what is lost if
crimes are committed and what will be lost if they are not com-

4 Morris, The Habitual Criminal (1951) at 13.

51
mitted will vary with the importance of the drives toward par-
ticular criminal acts. Thus, the same type of threat operating on
the same person may provide sufficient influence to deter behavior
that is weakly motivated but will fail to deter behavior that is
strongly motivated.

The strength of a person’s motivation toward a particular
crime would appear to be a function of the importance of the
drive he seeks to satisfy by criminal means and the availability
of alternative means for satisfying the same drive. The strength
of the drive that motivates criminal behavior has an obvious effect
on amenability to deterrence: the man who steals because he is
hungry or addicted to drugs is less likely to be deterred than the
woman who shoplifts as a result of whim, or the parking violator
who is trying to save money or time. For certain people in certain
situations, the drive toward committing a crime may be so strong
that no threat will stop them.

When other means exist to satisfy the drives that lead to
crime, the prospect for deterrence is brighter than when ac-
ceptable substitutes do not exist: the man who can either steal
or work to eat is surely more responsive to legal threats than
the man for whom stealing is the only way to eat.

Different types of crime may be associated with different
levels of motivation, and this could account for some of the
variation in the success of legal threats. Possessing and taking
illegal drugs is, for the addicted person, the only available means
of fulfilling strong drives. To the extent that drug-taking is
preponderantly the behavior of addicts, it would seem likely
that the threat of consequences will have less effect on the rate
of this crime than on others.” For the drug addict, shoplifting
may be one of many means of obtaining money to support his
habit; for other shoplifters the crime may be the result of a
weakly motivated desire for more of the better things of life.
The legal threat can operate with greater force on the weakly
motivated shoplifter and result in keeping many members of this
group law-abiding. If the drug addict has unthreatened alterna-
tive means to obtain money, the threat aimed at shoplifting may
deter him from shoplifting even though he continues to violate
the law by taking drugs. If his only alternatives for obtaining
drug money are all illegal, the credibility and severity of a threat
directed at shoplifting may motivate the addict to use other
illegal means of obtaining money. To the extent that a large
number of potential shoplifters are either weakly motivated to-
ward the drives for which they might steal or have alternative
means of drive-satisfaction available, the threat of consequences

5 Chambliss, supra note 1, at 707-08.

52
will be more likely to have a significant influence on the rate
of this crime.

The Emotional Climate of Crime

It seems reasonable to suppose that the context in which de-
cisions to commit or refrain from crime are made will have
some bearing on the relative importance of threats in the de-
cisional process. Decisions that are made very quickly, as a re-
action to a sudden impulse, may be less susceptible to the influence
of threats than decisions that are made over longer periods of
time. Decisions about criminal conduct that are made when a
person is in a condition of great emotional arousal may be less
amenable to threats than decisions that occur when the potential
criminal is less aroused, because very high degrees of emotional
arousal may eclipse thoughts of future consequences by riveting
all of the potential criminal’s attention on his present predica-
ment. The threat of consequences would appear to be of lessened
effectiveness when decisions about criminal conduct are made
under the influence of drugs or alcohol.

If particular types of crime are associated with different types
of decisional environments, differences in the emotional context
of crime could help to explain why some threats are more effective
than others. The presumed effect of alcohol becomes important
if alcohol is associated with a particular type of crime, such as
homicide, more often than with others. To the same effect, finding
that the degree of motivation influences the effectiveness of
threats becomes important if it is determined that some crimes
(such as homicide and certain deviate sex offenses) are usually
associated with high degrees of motivation, while others (such
as illegal parking or shoplifting) are weakly motivated in most
cases. The relationship between personal motives or the context
of crime and a specific type of crime will never be a complete
one: there are, to be sure, sober potential killers and strongly
motivated shoplifters. Often the association of particular motiva-
tions or emotional states with crime will prove more helpful in
distinguishing between the different types of offender represented
in the same offense group than in distinguishing between types
of offenses. For example, the distinction between the amateur
and professional shoplifter may be as important in predicting
threat response as any distinction between shoplifters as a group
and pickpockets.® But if a particular level of motivation is more
often associated with one crime than another, the relationship
between motivation and crime need not be complete to explain
differences in the impact of legal threats.

6 See Cameron, The Booster and the Snitch (1966).

53
The above typology of factors that might be associated with
difference in the efficacy of threats in different areas of crime
control appears to be consistent with predictions made by others
about deterrent effectiveness in different types of crime. Profes-
sor Morris’ prediction that deterrence is inversely related to the
moral seriousness of the threatened behavior is explicable on
the basis that minor crimes are more weakly motivated, are
associated with lessened degrees of emotional arousal, are com-
mitted by broader cross sections of the general population, and
are characterized by fewer prebarriers other than legal threats.
The observation that crimes committed for material gain are
more susceptible to deterrence than crimes of passion, such as
homicide and assault, can be viewed as consistent with the
analysis above (though, of course, not validated) because crimes
of passion are associated with higher levels of emotional arousal,
and it is more likely that there are substitute methods of ob-
taining money (either lawfully or unlawfully) than substitute
means for achieving the ends sought in aggressive crimes. It
may also be the case that many potential property criminals are
not powerfully motivated.

The above analysis is largely consistent with the theory
that instrumental crime is more susceptible to deterrence than
expressive crime (where the crime itself is what the criminal
seeks).” An example of an instrumental crime would be stealing
from a department store in order to obtain money to buy drugs;
an expressive act would be the consuming of drugs. As suggested
above, this prediction would hold true in many cases because it ~
is more likely that substitutes exist for instrumental behavior
(where other means may exist for the attainment of an end
which is external to the criminal act) than for expressive be-
havior (where the act itself is what the potential criminal wants).
However, the distinction between instrumental and expressive
crime would not seem to be as clear where no substitute exists
for a specific criminal means of obtaining an end. A pregnant
woman seeking an abortion is trying to commit an instrumental
act, but this is her only means of avoiding childbirth. The em-
phasis on substitution in the present analysis would suggest that
this behavior would be difficult to deter, while the distinction
between instrumental and expressive conduct would tend in the
opposite direction. With respect to a doctor who is tempted to
perform abortions, both the emphasis on available substitutes
and the distinction between instrumental and expressive behav-
ior would suggest that deterrence is more likely.

Suggesting that deterrence is more difficult in some situations

7 Chambliss, supra note 1, at 712-17.

54
than in others is not the equivalent of asserting that deterrence
does not function at all for particular offenses. Even under un-
favorable conditions, the threat of consequences may influence
many people most of the time. For the same reason, saying that
deterrence is more difficult with respect to some offenses does
not mean that changes in the credibility or severity of threats
will not result in marginal deterrence in these cases. But it does
appear that marginal deterrent effects of changes in the condi-
tions of a legal threat will be greatest when the conditions for
deterrence are most favorable. Increasing the perceived credi-
bility of the threat regarding homicide and illegal parking may
thus reduce the proportion of potential offenders of both kinds
who become actual offenders. But similar changes in the content
of the two threats will achieve a more substantial increase in
deterrence among weakly rather than strongly motivated per-
sons, in situations where the degree of emotional arousal as-
sociated with decisions about crime is low rather than high and
where the threat is the only major barrier to crime, rather than
one of many social and moral reasons to avoid a particular type
of activity.
Because the classification of offense types in accordance with
the likelihood of deterrence cannot produce positive statements
about the presence or absence of a deterrent effect, this type of
prediction cannot be directly translated into policy. Even if
marginal increments in the credibility of threats aimed at illegal
parking will reduce that rate more substantially than a similar
investment in enforcement of the law against homicide, the
social importance of preventing homicide may dictate that re-
sources would be more wisely invested in deterring homicide. And
simply because legal threats operate with less effect on some
types of potential criminals does not mean that other forms of
social control will achieve better results. For this reason, Profes-
sor Chambliss appears to be swimming in rather deep water
when he suggests,
A truly rational system of justice is one that will maxi-
mize its effectiveness by imposing criminal sanctions
where these act as an effective deterrent, and at the
same time develop alternatives to punishment where it
is found to be ineffective. The implication of this fore-
going analysis would be that the legal system will have
little effect in reducing the frequency of such things
as excessive drinking, drug use, most murders, most sex
offenses and aggravated assault. For these behaviors,
alternative mechanisms of social control must be insti-
tuted.®

“81bid., at 714.

55
Unfortunately, even if predictions about the conditions that
militate for or against deterrence in particular types of offenses
are completely accurate (and there is little present basis for this
conclusion), such predictions can only provide statements about
the relative likelihood of deterrence, and these will be insufficient
foundations for priorities in penal policy. Much more specific
empirical research is necessary to determine whether particular
countermeasures achieve a degree of deterrence sufficient to jus-
tify their costs.

C. Threat as Communication

If a threat is to achieve a generally deterrent effect, it must
alter conduct as the result of communication to members of a
threatened audience. In this sense, the deterrent threat is best
viewed as a form of advertising, and such a perspective high-
lights the significance of differences in the way that threats are
communicated. In simple terms, this significance in the relation-
ship between threat-communication and threat-response has the
following basis:

1. If threats are to have any effect on members of an
audience, information about the threat must be communicated
to that audience.

2. The way in which threats are communicated may also
affect the willingness of people to comply with their terms.

Public Knowledge as a Threshold Requirement

Different types of information are necessary conditions to
different categories of crime prevention.

1. Unless members of an audience know that a behavior
is prohibited, they will not refrain from the behavior because
it is forbidden.

2. Unless it is known that the legal system will punish those
apprehended for committing a threatened act, the existence
of threatened consequences cannot have a generally deterrent
effect on the rate of that behavior.

3. Unless differences in the level of consequences threatened
are known by members of a threatened audience, a marginal
deterrent effect cannot be attributed to raising penalties.

Research is just beginning into the extent of public knowl-
edge of the existence and terms of legal threats. Accordingly,
very little is known about the extent of such knowledge or the
sources through which people obtain what information they have.
It is reasonable to suppose that most people understand the gen-
eral boundary in society between criminal and noncriminal
behavior; however, many people do not know that some forms

56
of behavior which are not viewed as major infractions of the
community’s moral code are unlawful, and even greater numbers
of people appear to be unaware of certain changes in the scope
of the criminal law.

An example of widespread ignorance of a criminal law that
has been on the books for many years was found in a survey of
public knowledge of crime and penalties in Nebraska, where
writing checks with insufficient funds on deposit to cover them
constitutes a misdemeanor.! In a public opinion poll, 41 percent of
adult males in Nebraska declared that this behavior would not
be criminal if arrangements to pay the obligation were eventually
made.2 Another public opinion poll, taken after England had
abolished prohibitions against attempted suicide, showed that
76 percent of those polled were unaware that a change had taken
place.* Both of these cases may be considered exceptional—in
Nebraska because the law was rarely enforced, and in England
because of the exotic and morally charged nature of the subject.

It is highly unlikely that public ignorance of the existence of
legal prohibitions is widespread with respect to most areas of
serious criminality. What is much more likely is that people are
not informed about the specific penalties provided. Surveys in
California * and Nebraska ® have established the fact that the
public knows little about the legislatively proscribed minimum
and maximum penalties for a variety of crimes. In California,
only inmates of adult correctional facilities were likely to give
the correct answer more than a quarter of the time to a num-
ber of multiple-choice possibilities. Public knowledge of changes
in the minimum and maximum punishments provided by law
was also found to be quite low.®

Findings of this type do not indicate that deterrence is not
possible among those who do not know the specific penalty
provided for violations of criminal law; as long as it is widely
known that behavior is punished, this information alone can
lead to significant amounts of deterrence. In fact, it has been
suggested that a threat which produces uncertainty about what
penalties are imposed for violation may be a more effective de-
terrent than complete knowledge about sanctions.” However, if

1 See Nebraska Revised Statutes § 28-1213, 1943.

2 Poll conducted in 1968-69 by the Center for Studies in Criminal Justice at the University
of Chicago, in collaboration with Professors Harvey Perlman and Allan Booth of the Univer-
sity of Nebraska.

3 See Walker and Argyle, Does the Law Affect Moral Judgments? 4 Brit. J. Crim. 570 (1964);
see also Walker, Morality and the Criminal Law, 11 Howard Journal 214 (1964).

4 Psychiatry Research Associates, Public Knowledge of Criminal Penalties: A Research Re-
port (1968).

5 Poll conducted in 1968-69 by the Center for Studies in Criminal Justice, see supra note 2.

6 Public Knowledge of Criminal Penalties, supra note 4.

7 Morris, The Habitual Criminal, at 12.

57
people are insensitive to shifts in the level of punishment
threatened for particular acts, increasing punishment levels will
not increase the general deterrent effectiveness of threats.

But finding that most people do not know specific provisions
in the penal code is not conclusive evidence that higher penalties
do not produce marginal general deterrence, for two reasons:

1. higher penalties might affect perceptions without pro-
ducing correct answers;

2. even if only 10 percent of a population knows the specific
penalty for a crime, that 10 percent of the general public
could be the overwhelming majority of those for whom the
threat of a particular level of punishment is necessary.

In order to determine whether higher penalties might lead those
who cannot correctly estimate the penalty for an offense to think
it is higher than they would if the penalty were reduced, com-
parison must be made between areas where the penalties differ
widely, to see if public estimates of the severity of punishment
also differ. And in order to determine whether potential crim-
inals are more likely to know specific penalty levels than the
average citizen, high-risk groups must be asked about penalty
levels, and the results of this type of poll should be compared
with answers given by the general public.

To date, no comparison of public penalty estimates in high
and low penalty areas has been made, but preliminary studies
have investigated whether high-risk groups are better informed
than the general public. In California it was found that Adult
Authority prison inmates, the highest-risk group of all, showed
a high level of knowledge about maximum and minimum penal-
ties for most types of serious crime.® In addition, college students
were able to reply correctly more often than any other group
that penalties for possession of marihuana had not been changed
recently.? On the other hand, inmates of Youth Authority penal
facilities were not better informed than the general public, even
though this group was much more likely to commit crime in the
future. Equally revealing was the fact that students from high
schools in high-delinquency areas knew just as little about mini-

8 The California data show the following contrast in percentage of sample with correct
answers:

Public Prisoners
First-Degree Robbery 8% 85%
Assault 36% 59%
Rape with Injury 16% 43%
Forgery 17% 50%

(Unpublished data from the California poll conducted by Social Psychiatry Research Associates.)

9 65 percent of the college students correctly answered that no change in penalty had oc-
curred, compared to 47 percent of adult prisoners and 38 percent of high school students in a
low-delinquency area, the second and third highest correct-answer groups on this question. Un-
published data taken from the California poll conducted by Social Psychiatry Research Asso-
ciates (1968), supra note 4.

58
mum and maximum adult penalties as students from high schools
in low-delinquency areas.’ With the exception of the college-
student answers on marihuana, it would thus appear that many
potential offenders have levels of information as low as those
demonstrated by the general public, but that firsthand experience
with punishment in adult prison facilities is a great teacher.

These tentative findings cast some doubt on the breadth of
any marginal general deterrent effect attributable to higher
minimum and maximum sentencing provisions. More important
perhaps for present purposes, such findings show that communi-
cation of threats is neither automatic nor complete in the criminal
justice system. For this reason, variations in the way threats
are communicated may lead to different levels of public aware-
ness, and this in turn may affect the likelihood and degree of
deterrence.

If information is to play a role in deterring individuals from
criminal conduct, they must have access to it and must remem-
ber it at the time that decisions about criminal conduct are
made. With this task as a goal, effective communication will
require delivering the message to a threatened audience, pre-
senting the message in ways that will make members of the
threatened audience pay attention and remember the informa-
tion being conveyed, and, if possible, associating the information
contained in the threat with the threatened behavior in the
minds of audience members so that their recollection of the
terms of the threat will be greatest when it is most needed—at
the time when criminal alternatives are considered.

Passing a new law will, by itself, have no effect on the per-
ception of those whose conduct the law seeks to alter. Formal
and informal channels of communication—including newspaper
and television accounts of the new threat and stories of those
punished for law breaking—will bring the message to members
of a threatened audience, but the California prison survey sug-
gests that word of mouth, in many circumstances, may be the most
successful means of communication. Some research findings
indicate that, in order to reach and influence most members of
the threatened audience, the threat will have to be repeated a
number of times, with each repetition not only increasing the
number of people exposed to the threat but heightening the
comprehension of those who have been previously exposed.?

10 See Public Knowledge of Cridimal Penalties, supra note 4.

1 Ibid.

12 See Barmack and Payne, The Lackland Accident Countermeasure Experiment, 40 Highway
Research Board Proceedings, 513 (1961). The Lackland experiment is one of the few investiga-
tions of propaganda techniques used to alter conduct in a field setting. The authors report a
three-month lag between the introduction of the countermeasure (an information and per-
suasion campaign with some changes in the threatened consequences for accident records) and

59
In many situations, authorities seek to capture the attention
of a threatened audience by using examples of those punished
for law violations. Judges sentence offenders with the objective
in mind that potential criminals will hear of the sentence and
that this message will result in general deterrence.’ In recent
years, the administration of some punishments, notably the death
penalty, has galvanized public attention. In other countries, the
names and punishments meted out to tax offenders are published,
both as a punishment for violation and as a mechanism for
drawing the attention of the community to the existence of
punishment for the violation. While the aim of such practices is
to enhance deterrence by drawing the attention of members of
the threatened audience to the punishment precept, there is little
evidence to indicate whether they have succeeded. Two studies
of the homicide rate just before and just after death sentences
in Philadelphia and executions in California do not provide evi-
dence of any net reduction in homicide rate that could be attrib-
uted to publicity about the executions.’* Other programs for
drawing attention to punishment have not been evaluated. The
capital punishment case may be exceptional, because with this
type of crime the emotional arousal and power of suggestion
associated with the death penalty and with the offense of homi-
cide may actually create homicidal tensions at the same time
as it may deter some potential offenders.’ In other cases, drawing
attention to episodes of punishment may have counter-deterrent
effects on those individuals who imagined the penalty to be
worse than it was. More often, however, it would seem likely
that publicity of this kind, if it has any effect, will tend in the
direction of deterring crime.

A special problem in the communication of information about
threat and punishment is that of bringing these considerations
to the attention of potential law violators at the time that deci-
sions about law violation are to be made. The moment of truth
in the life of a potential deviant seldom takes place in the local

signs of a significant effect. One explanation of this lag is that ‘‘the spread of information
about the countermeasure is an accretive process which necessitates repeated announcements

. and word-of-mouth communication until a substantial majority of the population be-
comes aware of the countermeasure.”

13 A Federal judge in Chicago was quoted by the Chicago Sun Times in 1967 as stating, at
the time of sentence, that the penalty he regularly fixes for draft-law violators had been re-
cently increased as the result of increased violation. An even more explicit case of courts sen-
tencing offenders on the assumption that potential offenders were listening closely was one
chapter of the story of the Great Train Robbery in Great Britain. See 80 L. Q. Rev. 473 (1964).

1 See the studies of Savitz and Graves, reprinted in Bedau, The Death Penalty in America,
at 315-32.

15 In Dr. Graves’ study, the homicide rate before executions was reported to be above aver-
age, while the rate just after executions was reported to be below average, consistent with the
theory that increased arousal attributable to the atmosphere leading up to the execution could
have affected the rate of homicide. See Ibid. at 329. Given the great number of factors that
could influence the apparent findings, the study must be viewed as merely suggestive.

60
law library. A threatening agency must either try to get its
message through at the precise time that such decisions are
made, or seek to plant the association between crime and punish-
ment so firmly in the minds of potential offenders that the
thought of penalties will naturally accompany the considerations
of the threatened behavior. On occasion, it is possible to place
reminders of prohibition and penalty at the scene: speed limit
signs and other posted notices in public places serve as notice
that certain behavior is forbidden, but also act as reminders
that unpleasant consequences will accompany apprehension. The
campaign in Michigan to post signs that read

DRUNK DRIVERS GO TO JAIL

made explicit this second function of on-site legal commands.
These examples show rather clearly, however, that most legal
threats cannot be directly inserted by will of the threatening
agency at the site of possible law violation. Where this type of
reminder is impractical, the only adequate substitute will be an
association on the part of members of the threatened audience.
And such an attempt is likely to produce more consumer resist-
ance than even the most unpalatable of commercial products—
it is perhaps best compared to the task of getting people to
associate chocolate candy with dentists.

Threat as a Persuasive Appeal

If the first task of a threatening agency is communication of
information, its second task is persuasion. The threatening agency
is not a dispassionate participant in the communication process,
with no stake in how the information affects behavior. Each
message about the content of a threat is an agency appeal de-
signed to reduce the quantity of threatened behavior. And since
“information alone seldom provides sufficient impetus to change
attitudes or actions toward a given object,” *¢ it is important to
consider those aspects of communication other than information
that might influence the ways in which people respond to threats.

The presentation of legal threats has involved a number
of persuasion techniques, perhaps the most common being the
heavy emphasis on a. the unpleasantness associated with being
caught and b. the high probability that violation will lead to
apprehension. Emphasis is confined in most cases to the maxi-
mum penalty possible for an offense, and appeals read like the
aforementioned one about drunk drivers, or

LITTERING IS PUNISHABLE BY $500 FINE

16 Leventhal, Singer, and Jones, Effects of Fear and Specificity of Recommendation upon At-
titudes and Behavior, 1 J. Personality and Psych. 20 (1965).

61
or perhaps like this

FALSE STATEMENTS ON THIS DOCUMENT
CONSTITUTE PERJURY
AND ARE PUNISHABLE BY IMPRISONMENT

Often an agency will seek to highlight the probability of ap-
prehension by telling members of the threatened audience about
the means available for detecting violations; two frequent in-
stances of this appeal are

SPEED TIMED BY RADAR
and

OUR CAMERAS ARE MAKING
A CONFIDENTIAL RECORD
OF THIS TRANSACTION

Private detective agencies, in apparent pursuit of general deter-
rence, post notices in conspicuous places to the effect that they
service the particular shop or building. Less commonly, agency
appeals try to downgrade the attractiveness of threatened be-
haviors or seek to secure compliance by means of appeals to the
honesty or patriotism of the threatened audience.”

One issue relating to the presentation of threat-appeals stems
from whether a high degree of emotional arousal on the part of
a threatened audience will create a greater deterrent effect.
In the belief that “shaking up” an audience will focus attention
on the foolishness of bad driving, many local traffic courts and
traffic-safety programs show films that depict violence on the
highway in the most graphic of terms.’* No legal-threat cam-
paigns have ever been subjected to evaluative research in order to
determine whether such emotion-laden approaches enhance de-
terrence, but a number of psychological experiments have been
conducted in closely analogous circumstances, with mixed re-
sults. The pioneer test of the effect of “fear-arousing” communi-
cations was published by Janis and Feshbach in 1953.1 In that
study, randomly selected groups of high school students heard
three different types of lectures on dental hygiene: one group
heard a lecture designed to have a strong fear appeal, with fre-
quent references to pain and harm to teeth; the second group
heard a presentation with a moderate amount of threat appeal;
"The most famous recent campaigns of this type are “Every Litter Bit Hurts” and “Only
You Can Prevent Forest Fires.”

18 The advertising description of one such film, “Signal 30,” stresses the blood-and-gore as-
pects (in color) as educational. The picture has been used in a number of driver-improvement
courses administered by traffic courts.

1 Janis and Feshbach, Effects of Fear-Arousing Communications, 48 J. Abnormal and Social
Psych. 78 (1953).

62
and the third group heard a lecture that focused on other rea-
sons why dental hygiene was a good idea. A higher percentage
of those students confronted with the strong threat-appeal re-
ported newly awakened worries about dental hygiene, but a lower
percentage of them reported conforming to the recommendations
of the lecture than in either of the other two groups. When the
experiment was repeated, no significant difference in effect was
found when the three different threat intensities were
compared.?°

Subsequent research involving variations in the “fear-arous-
ing” content of appeals to use automobile seat belts or take
inoculations tended to show that results vary. Some experiments
produce significant differences; others do not. A number of
explanations have been suggested for the variety of research
results. Janis and Feshbach suggest that one reason why strong
fear-arousing messages might reduce conformity is that such
tactics arouse aggression as well as anxiety.’ Another theory is
that scare techniques might work better than others on groups
that are not afraid to begin with, but that such appeals may
produce a counter-reaction in those who already are anxious
about the subject of the communication.

Experiments to date on the “fear-arousing” content of ap-
peals are not of central importance to the student of deterrent
threats because:

1. the findings are far from conclusive;

2. the appeals and audiences used for experimental pur-
poses differ from the setting of most criminal threats. In
particular, the criminal threat is made by an agency with the
capability of creating unpleasantness, unlike the dental hygiene
lecture, where the persuader has no power to make his proph-
ecies come true.*?

However, these tentative explorations do call into question the
efficacy of fear-arousal as a consistently successful method of
securing compliance. And, collectively, such research makes the
case for testing the effectiveness of fear-arousing appeals in
criminal law and related areas.?

20 Moltz and Thistlethwaite, Attitude Modification and Anxiety Reduction, 50 J. Abnormal
and Social Psych. 231 (1955).

21 See Hovland, Janis, and Kelly, Communication and Persuasion, at 83 (1953).

22 Ibid., at 82: “[Wlhere the communicator has the power to administer severe punishments
and the audience has already learned not to ignore his threats, strong fear appeals are likely
to induce a high degree of conformity.”

23 Perhaps the best place to start such an investigation would be in the related areas of
traffic safety and traffic law. The National Safety Council has conducted holiday safety cam-
paigns for some time that stress the number of deaths expected in a given holiday period,
and these campaigns have been criticized in recent years. Traffic courts use fear-arousing ma-
terials as a part of driver-education programs they compel some offenders to attend. Large
numbers of drivers are exposed to such appeals, and the stakes in the traffic-safety area are
higher than anywhere else in the criminal justice system in terms of lives lost and property
damage.

63
Personalizing Threats

In the model of simple deterrence put forward by early thinkers,
there were two assumptions:

1. that the threat of the criminal law was always issued

by the threatening agency to society in general by passing a

law;

2. that the information and persuasive appeals that created

a deterrent effect were communicated to all men in similar

ways.

Yet appeals to refrain from criminal conduct can be issued in
many different ways: to society in general; to particular groups;
to an audience that has been physically assembled to hear about
a threat; or even to individuals, by means of letters or personal
interviews. Personal interviews are, of course, a much more ex-
pensive method of communicating information and appeals than
issuing threats to a general audience. For this reason, such
intensive appeals will typically occur only among special high-
risk groups (usually prior offenders) or in experiments designed
to establish whether different types of communication affect
behavior.

There is some basis for believing that the more personalized
a particular communication may be, the greater the chances that
it will affect behavior. Since the advent of the computer, some
commercial firms go to great lengths to personalize advertising
appeals—printing the name and address of the recipient on the
outside of the envelope in a mailing campaign and addressing
the recipient by name a number of times in the course of pre-
senting their appeal.** The theory behind this practice may be
that the public is less likely to dismiss personalized communica-
tions and may react more favorably to appeals when there is
evidence that the firm is thinking of them personally in relation
to the appeal. With respect to threats, this type of practice
might influence behavior by producing the impression among
members of an audience that somebody is paying special attention
to their conduct.

Field experiments designed to test the effects of particular
types of appeals have employed personalized letters and inter-
views with mixed success. In one test of whether normative
appeals (PAYING TAXES HELPS YOUR COUNTRY) and
threatening appeals (COMPUTERS CAN CATCH YOU) would
affect tax-paying behavior, both appeals being issued by a non-
governmental agency, there was a significant increase in the
"For example, Publishers Clearing House, a magazine sales firm, uses name and local resi-
dence references twice in the same page of one 1969 contest promotion letter. Reader's Digest

used six different references to the recipient’s name and address, besides referring to two other
local residents by name, in a 1969 contest promotion.

64
amount of income reported and the amount of taxes actually
paid by subjects of both appeals.?® This tax experiment can hardly
be viewed as conclusive evidence that personalized appeals have
more influence than general appeals, inasmuch as no controlled
comparisons were made with similar groups that had been ex-
posed to general appeals with the same content. But the possi-
bility that personalization helps to explain the results obtained
in this experiment suggests that the effects of personalization
should be tested. And the possibility that personalization con-
tributed to the effect of the experiment is an obstacle to any
conclusion that less personalized appeals of the same type as
those used in the tax study would have had the same effects.

Personal interviews are extremely expensive, but most other
forms of communication are far less costly. In cases where de-
ficiencies of information or attention may impair the efficacy of
threats, correcting these deficiencies will usually prove less ex-
pensive than other changes designed to increase threat effective-
ness. To the extent that changes in communication can improve
the effectiveness of threats, and that extent is presently un-
known, a failure to take communication into account will result
in more costly attempts at reducing crime.

In sum, there is no reason to believe that a complete understand-
ing of the effects of threats on behavior is attainable by consider-
ing the legal system as a branch of the advertising profession;
neither is there much danger that considering threat as communi-
cation will lead to overemphasis by officials on this aspect of
deterrence. Variation in the substance of threats is surely more
important than variation in form. But analysis that focuses on
threat as communication makes explicit the fact that the percep-
tions of an audience, rather than the threat as intended by the
agency, will determine the degree to which legal threats will
achieve desired goals. Similarly, a communication perspective
clarifies the role that assumptions about public knowledge play
in theories which stress the importance of variations in prob-
ability of apprehension and severity of threatened consequences
in predicting the outcome of threats.

D. Applicability and Credibility

If a legal threat is to prove effective as a simple deterrent,
members of the threatened audience must believe that the threat-
ening agency is both willing and able to impose consequences on
them for committing the threatened act. Such a belief depends on
audience awareness of a threat, as discussed above, and on two
further conditions:

28 Schwartz and Orleans, On Legal Sanctions, 34 U. Chi. L. Rev. 274 (1967).

65
1. those threatened must believe that the threatening agen-
cy wishes to apprehend and punish them personally if they
commit the threatened act;

2. those threatened must believe that the agency is capable
of discovering and punishing violations of the threat.

These two conditions refer, respectively, to the applicability and
credibility of threats.
Applicability

Even if law-enforcement agencies are capable of enforcing
threats, members of an audience will not fear the imposition of
consequences unless they are persuaded that the threat is meant
to apply to them. If a threat has been generally unenforced for
a period of time, and this is known by a threatened audience,
it may result in the belief that the threat is insincere. When a
threat is issued to society in general, but only certain classes are
caught and punished, members of a class that go unpunished may
develop an elitist theory of the legal threat; they recognize the
threat as real, but feel it is directed at others.

In one study of penalties and enforcement practices at a large
midwestern university, it was found that campus police for some
time had ticketed all cars found in “NO PARKING” zones, but
nothing had been done to collect fines from faculty members
who had been ticketed but did not voluntarily pay. When the
parking problem grew worse, the university publicized a new
policy, increasing the fine for multiple offenders and announcing
that parking fines would henceforth be deducted from faculty
paychecks. After the new policy was announced, the rate of pro-
hibited parking by faculty members decreased, but this decrease
was largely confined to those who had been the most persistent
offenders before the change—those who had not paid the fines
under the old system. Thus, it appears that announcing that
faculty members would be fined, rather than the increase in fine,
was responsible for most of the decrease in prohibited parking.
Before the change in policy, nonpaying faculty members could
reason that their cars received tickets because the police could
not distinguish them from student cars, and that the nonenforce-
ment of the fines showed that the university did not mean the
parking restrictions to apply to them. For such a group, the new
policy might have been the first notice that the university
intended to restrict faculty parking in the prohibited zones.

In a similar vein, when laws against gambling or holding un-
licensed parades are not enforced against church bingo or parading
Elks, those groups may continue to violate the terms of such

! Chambliss. The Deterrent Influence of Punishment, 12 Crime and Deling. 70 (1966).
2 Ibid., at 72-176.

66
laws without fear of apprehension and without notice that their
behavior is considered wrong by the threatening agency. Under
these circumstances, the threat of consequences will not produce
even the moral impetus to abide by the law that might accompany
knowledge that the conduct was considered wrongful.

Credibility

Even if a threatening agency is willing to punish wrongdoers,
it may not be able to do so in all cases. Yet if audience members
are to be deterred from committing a threatened behavior be-
cause they fear the imposition of consequences, they must believe
that the threatening agency is capable of catching and punishing
some offenders. Moreover, if some people have strong drives to-
ward a threatened behavior or little fear of the consequences
threatened, their estimate of the specific probability of getting
caught may have a significant effect on whether they decide to
risk the consequences. If this is the case, the greater the perceived
likelihood of being caught, the more often processes of simple
deterrence will persuade a potential offender to refrain from
crime.

There is almost universal agreement among commentators that
the opinion of a potential criminal about his chances of appre-
hension is important in predicting the effectiveness of threats.’
Most often, this consensus is expressed in the theory that the ob-
jective probability of apprehension will influence the efficacy of
a legal threat. On occasion, writers express a faith in the mar-
ginal deterrent effect of comprehensive enforcement, while doubt-
ing that increased penalties pay similar dividends, as when Dr.
Temple predicts:

The effectiveness of a deterrent is derived less from its
severity than from its certainty.*

Though the objective probability that commission of a crime
will lead to imposition of consequences is of unquestioned impor-
tance to deterrence and will help explain why some threats oper-
ate more effectively than others, a complete emphasis on objec-
tive probabilities may be misleading for two reasons:

1. if audience members perceive that a threat applies to
them, the moral message that the behavior is forbidden can
influence law-abiding people to refrain from crime even if
the probability of apprehension is quite low;

2. predictions based on objective probabilities of apprehen-
sion may be inaccurate because subjective judgments about
3 See, e.g., Andenaes, Punishment and the Problem of General Prevention, 114 U. Pa. L.

Rev. 149, at 160-164; Temple, The Ethics of Penal Action, First Clark Hall Lecture, 1934, at

33; Shaw, The Crime of Imprisonment (1946 ed.).
4 Temple, supra note 3, at 33.

67
personal chances of being caught, rather than the cold mathe-
matics of crime, will determine how an individual responds to
the threat of consequences—and the relationship between such
judgments and objective probabilities is imperfect.

People may underestimate or overestimate the probability that
they will be caught because of incomplete information about
impunity rates for a particular crime. In addition, factors other
than objective probability, such as personal feelings of optimism
or pessimism, or propaganda about the chances of apprehension,
may play an important role in personal judgments about the ele-
ment of risk in a specific crime.” Emphasis on the objective prob-
abilities of apprehension may also conceal important insights
about how and when changes in enforcement affect the percep-
tions of a threatened audience, for such changes may have a
gradual rather than an immediate effect on subjective judgments
about risk, and change in the objective chances of apprehension
may affect some people differently than others.

Three Introductory Studies

Research to date supports the theory that the perceived credi-
bility of a threat serves to condition its deterrent effectiveness,
but there are indications that the process is more complex than
many early writers believed. In Denmark, an unusual opportu-
nity to study the effects of decreasing the apparent probability
of apprehension occurred when the German Occupation Govern-
ment arrested the Copenhagen Police Force en masse during the
Second World War. Reacting to this emergency, citizen groups
in Copenhagen established vigilante groups and increased the
penalties threatened for major crimes. During this policeless pe-
riod, the rates of many crimes where the offender could not be
identified by the victim (such as robbery, burglary, and larceny)
increased approximately tenfold. Other crimes, where the of-
fender could be identified without police work, did not increase in
frequency to any spectacular degree.’ It is unknown to what
extent the large increases in crime were due to the same number
of criminals committing many more crimes rather than to a great-
er number of people deciding to commit crimes.

In Detroit, police dissatisfaction with wages and working con-
ditions led to a more restricted form of “police holiday.” During
a period lasting a minimum of seven months in 1967, the same
number of police continued on traffic patrol, but issued about
half as many tickets to motorists for moving violations. If De-

5 See the discussion and references cited infra subsection A, notes 4-7, and accompanying text.

® Trolle, Seven Months Without Police, cited from Andenaes, The General Preventive Effects
of Punishment, 114 U. Pa. L. Rev. 949, at 962 (1951). Additional information about Profes-
sor Trolle’s volume (not available in English) was provided by Professor Andenaes.

68
troit’s motor-vehicle accident rate is an accurate index of the
number of traffic offenses committed, the number of actual vio-
lations did not increase, because the accident rate actually de-
clined.” Although violations of traffic laws are found to have oc-
curred in many accidents, it is far from clear that accident rates
are an acceptable index of rates of traffic-law violation.® But the
fact that accident rates did not rise at all is some indication
that Detroit’s police holiday did not have a spectacular effect on
the rate of traffic violations.

One reason why Detroit’s experience may have been different
from that of Copenhagen is that the same number of police re-
mained on the street in Detroit. Thus, the cues (mainly police
presence) that many citizens had of the extent of traffic enforce-
ment went unchanged, even though the objective probabilities
had changed markedly. In Copenhagen the rise in the crime rate
might even have reflected some overestimate of the actual decrease
in impunity rates because the traditional police presence had been
completely eliminated.

The British experience with the Road Safety Act of 1967 may
provide another example of changes in the perceived credibility
of a new legal policy differing somewhat from the objective
change in risks of apprehension. Immediately after the much pub-
licized Act went into effect, providing for a new blood-alcohol
percentage definition of drunk-driving and equipping the police
with portable “breath test” kits with which to screen sus-
pected drunk drivers, the rate of nighttime traffic fatalities
dropped sharply, indicating a substantial decrease in the rate of
drunk-driving. As time passed, the rate of traffic fatalities in-
creased somewhat from the low levels experienced right after the

7 Detroit police reports for 1966-67 show the following comparisons in moving violations:

1966 1967 % Change
Jan.—April 163,873 152,423 -1
May-Dec. 265,234 128,617 —52

Statistics on fatal accidents, obtained from the Detroit Police Traffic Division, show no trend
upward during late 1967:

Motor Vehicle Accident Fatalities

1966 1967 % Change
Jan.—April 79 62 —22
May-Dec. 160 156 —- 3
Total accident figures decreased dramatically after April of 1967:
1966 1967 % Change
Jan.—April 26,121 23,973 —- 8
May-Dec. 63,119 45,101 —29

However, this decrease might be another result of a slowdown in police reporting.

8 For a description of efforts to investigate the relationship between accidents and viola-
tions in California, see Coppin and Peck, The 1964 California Driver Record Study (1967).

9 See Great Britain Ministry of Transport, Press Notice No. 892, December 19, 1967; No. 7,
February 8, 1968.

69
law went into effect,’® which may indicate that drivers had reeval-
uated the impact of the change in enforcement on their personal
chances of being apprehended for drunk-driving. Whether the
burst of publicity and the element of unknown risk had caused
drivers to overreact to the change in the law at the outset, or
whether the passage of time and personal experience produced an
underreaction to the no-longer-new change in enforcement pol-
icy, is impossible to tell, but the role of subjective factors in the
deterrent effect of the Road Safety Act is established if either
of these interpretations explains the pattern of accident fatali-
ties after the law’s introduction.

Certainty and Credibility

The different types of stimuli that may influence the perceived
credibility of a legal threat include:

1. actual changes in the probability of apprehension that
become known through the publication of general statistics on
crime clearances;

2. actual or merely apparent changes in probability of ap-
prehension that are communicated to an audience through pub-
licity about inspection or apprehension of particular persons
(BANK CASHIER JAILED FOR EMBEZZLEMENT) or
groups (ROUNDUP OF DRUG SELLERS UNDERWAY);

3. publicity about new methods of enforcement (COMPUT-
ERS ENLISTED IN WAR AGAINST FALSE ALARMS);

4. perceptions of changes in the level of enforcement that
come from the personal or word-of-mouth experience of audi-
ence members with the symbols of enforcement, such as more
police on the street;

1 The Ministry of Transport reports that October-May serious and fatal nighttime acci-
dents decreased 37 percent in the months following the effective date of the law but increased
20 percent from a much lower base in the period October 1968 through May 1969. Personal
communication from J. Betts to the author, October 1969.

11 A somewhat more ambiguous indication that increases in certainty of apprehension deter
crime is the finding that States with a higher number of prison admissions per 100 reported
offenses tend to have a lower crime rate. Whether this means that crime is deterred by more
efficient police work, or simply suggests that it is easier to apprehend a greater percentage
of criminals when the crime rate is low, is unclear. Only the former interpretation would be
evidence of the marginal deterrent effect of increases in threat credibility. Professor Tittle,
who reports the finding that higher proportionate enforcement is correlated with lower crime
rates, proposes one test of whether low crime rates are a cause or effect of relatively high
probabilities of apprehension: he obtained the correlation between probability of apprehension
and the ratio of police to population, and found that as probability of apprehension tends to
increase, the ratio of police employees to population tends to decrease (the correlation obtained
was —.53). The data do not, however, negate the possibility that greater probability of ap-
prehension is the result of lower crime rates making police work easier, because the police
burden may best be seen as the ratio of police to crimes rather than population, and a high-
crime-rate area may have a high ratio of police to population and a low ratio of police to crimes,
while a less crime-prone area may have fewer police per thousand citizens but more police
per thousand crimes. There is also the disquieting possibility that apparent “certainty” is
achieved by differences in reporting practices. See Tittle, Crime Rates and Legal Sanctions,
16 Social Problems 409 (1969).

70
5. personal or word-of-mouth experience of audience mem-
bers with apprehension.

Of all these influences, it seems probable that personal experi-
ence (whenever present) exerts the most powerful effect on
perceptions of credibility. If a man drives under the influence of
alcohol once and is apprehended by police, certainly this experi-
ence is likely to increase his estimate of the risk of apprehension
he runs by repetition of such conduct in the future. The same
kind of change in perception can be expected if a person is
stopped by police in a check for drunk-driving but gets off be-
cause he was not under the influence of alcohol at the time. News
that friends or relatives have been apprehended or inspected will
have a similar effect, though perhaps will exert less influence
than a first-person experience. By the same token, committing
a crime without being apprehended, or having personal knowl-
edge of people who commit offenses without detection, will de-
crease personal estimates of the risks involved in a particular
crime.

A Swedish survey of drivers provides evidence that personal
experience with apprehension machinery not only increases per-
ceived risk of apprehension but decreases the rate of drunk-
driving among those who have been exposed to such experience. In
that survey, drivers who reported having been stopped and screen-
ed by Swedish traffic control police had less subsequent drunk-
driving than unchecked drivers.'? There is less reason to assume
that personal experience with increases in the number of police,
and nothing more, can reduce crime. In New York City the rate
of street crime decreased after the saturation of a small high-
crime-rate area with extra police.’* But because the New York
experiment succeeded in increasing the objective likelihood of
detection while increasing the presence of enforcement person-
nel, this study does not necessarily show that apparent enforce-
ment has an independent influence on the credibility of threats.*

In the aggregate, there is a direct relationship between in-
creased objective probability of apprehension and increases in
the perceived credibility of threats that result from personal
experience with apprehension: the total number of persons ex-
posed to apprehension will increase and decrease in relation to
the total amount of enforcement. For the whole of a particular
audience, then, changes in perceived credibility based on this
type of experience will occur only with changes in actual prob-
ability of apprehension. With respect to police presence, the

12 Klette, On the Functioning of the Swedish Legislation Concerning Drunken Driving (un-
published, 1966).

13 New York City Police Department, Operation 25 (1955).
14 See Ibid., at 9-15.

71
relationship is somewhat less direct. In rare cases such as
occurred in Detroit, the probability of apprehension may change
without any corresponding change in the degree of police pres-
ence or in methods of detection. More commonly, increased po-
lice presence will not increase the probability of apprehending
offenders on a one-to-one basis: doubling the police presence in
a community (and thereby doubling exposure to police) will not
necessarily double the number of arrests or the clearance rate
for street crimes. But the objective probability of apprehension
will tend to increase or decrease along with variations in the
presence of enforcement machinery.

There is no necessary relationship between publicity about
new methods of enforcement or prosecution of violators and
increases in the actual probability of apprehension. If these an-
nouncements affect the subjective judgments about risk of ap-
prehension made by potentional criminals, they represent an op-
portunity to increase threat credibility without increasing en-
forcement. But such an opportunity is limited by the powerful
effects of personal experience. Where potential criminals are like-
ly to gain personal experience of apprehension risks, either
through committing crimes and escaping without detection or
through personal knowledge of the successful criminality of oth-
ers, personal knowledge will chip away at the facade of credibil-
ity over a period of time. The publicity campaign might increase
the perceived credibility of the legal threat in the short run
under such circumstances, but perceptions of risk would decrease
as personal experience accumulated. One side effect of this type
of process might be that audience members exposed to publicity
campaigns and subsequently persuaded by other information that
the risk of detection was not as represented might discount any
future announcements of breakthroughs in crime prevention. If
so, publicity without substance may decrease the general credi-
bility of legal threats among some potential criminals.

In situations where many potential criminals are not exposed
to personal experience with enforcement machinery, the role of
general publicity campaigns in influencing public perceptions
of the risk of detection is more significant; such campaigns may
persuade large segments of the population that the risks of crime
are greater than is actually the case. Very few citizens have any
personal knowledge contradicting the maxim, “You can’t get away
with murder.” Given an area where the crime rate is low and
examples of undetected crime are not widely known, the facade
of enforcement can establish permanent impressions of threat
credibility in excess of actual clearance rates. And law-abiding
members of society will spend most of their lives dependent on

72
secondary sources for their impressions of the efficiency of most
categories of criminal-law enforcement.

The fact that law-abiding members of society depend heavily
on secondary sources for perceptions of threat credibility does
not, by itself, mean that publicity campaigns that affect such
perceptions will decrease the crime rate, for widely represented
among such groups are people who do not need a deterrent. It
is only when potiential criminals are not heavily exposed to per-
sonal experience that second- and third-hand impressions of the
credibility of threats will have a permanent influence on the per-
ceptions of those for whom the credible legal threat is most nec-
essary.

If it is true that there is some danger of undermining the
credibility of legal threats by using extensive publicity even
though there has been no change in the substance of law enforce-
ment, it is also true that dramatic increases in the efficiency of
law enforcement will fail to achieve maximum impact without
extensive publicity. For some potentional offenders, personal ex-
perience with the enforcement of a law will not occur until months
or years after a change in enforcement takes place, and such an
experience may be the result of law violation that could have been
prevented. The credibility of legal threats is enhanced more quick-
ly, and among a greater number of potential offenders, if pub-
licity accompanies increases in the ability of a legal system to ap-
prehend offenders.

E. Variations in Threatened Consequences

Since the prospect of unpleasant consequences is the fuel on
which the machine of deterrence is supposed to run, differences
in the character and extent of threatened consequences may ex-
plain why some threats influence behavior more than others. Con-
sidering the role of consequences in the deterrent equation raises
questions about:

1. the range of occurrences that should be defined as the
consequence threatened by a particular communication;

2. the different kinds of unpleasantness that threats may
communicate;

3. the effects of changes in the severity of threatened con-
sequences on crime.

The Scope of Threatened Consequences

If one were to ask a judge to describe the penalty for driving
while intoxicated, he would probably confine his answer to the
fine or sentence of imprisonment provided in the penal code of
his jurisdiction. If the question were rephrased as “Why might

73
people wish to avoid being convicted of this offense?’ the same
official might respond more volubly. He might mention the un-
pleasantness associated with arrest and trial for driving while
intoxicated, a process that many people would fear more than the
sentence of the court. He might speculate on why jail is an ex-
perience that people wish to avoid, describing the unpleasantness
of the surroundings and restrictions imposed by jails, but also
mentioning the effect of conviction and jailing, not only on a
person’s standing in his community, but in terms of people who
are of particular importance to the offender.

This expanded description relates back to the definition of
threatened consequences in two ways.

First, it demonstrates that some of the incidents of the ad-
ministration of criminal justice, other than formally prescribed
punishment, should be considered threatened consequences. In
the case of driving while intoxicated, the legal system intends
that the process of apprehension should be unpleasant, and mem-
bers of the threatened audience will share this view. There would
appear to be no good reason for restricting the definition of
threatened consequences so as to exclude official actions that may
be as significant as formally prescribed punishments, simply be-
cause they are incidents of processes leading to conviction.

Second, the importance of the effects of jail on personal stand-
ing in the community highlights the fact that official actions can
set off societal reactions that may provide potential offenders with
more reason to avoid conviction than the officially imposed un-
pleasantness of punishment. Because such reactions emanate
from sources other than the threatening agency, and may
not be an intended result of punishment, they cannot be con-
sidered to be part of the consequences threatened by an agency
for specific offenses. However, these secondary reactions must be
studied if the effectiveness of particular threatened consequences
is to be understood. For example, the social disapproval that jail
might precipitate will surely have different degrees of influence
on skid row drunks and potential anti-trust violators,> just as
they will operate with more force on certain types of shoplifters
than on others.? For this reason, the effectiveness of the threat
of jail as a means of deterring criminal behavior will vary with
the attitudes of different groups toward a result of jail that is

! With respect to the power of apprehension and interrogation, Cameron notes: “Among
pilferers who are apprehended and interrogated by . . . store police but set free without for-
mal charge, there is very little or no recidivism.” Cameron, The Booster and the Snitch
(1966) at 151.

% See, e.g., Lovald and Stub, The Revolving Door: Reactions of Chronic Drunkenness Offend-
ers to Court Sanctions, 59 J. Crim. L., C. and P.S. 525 (1968).

3 Cameron, supra note 1, at 164. “The contrast in behavior between the pilferer and the
recognized and self-admitted thief is striking.”

74
outside the boundary of formal legal punishment. And while the
definition of legal consequences should not include unofficial com-
munity responses, any study of the effects of legal consequences
must encompass all of the meanings that the threat of conse-
quences conveys to members of a threatened audience.

Types of Unpleasantness

A great variety of threatened deprivations may function as
reasons why members of an audience might wish to refrain from
committing threatened offenses. The different types of un-
pleasantness conveyed by threats may condition different patterns
of threat response, and thus be related to some of the variations in
threat effectiveness. The major forms of unpleasantness conveyed
by threats include economic deprivation, loss of privileges, con-
finement, physical punishment, and social stigma.

Economic Deprivation

The threat of economic disadvantage through such mechanisms
as fines or lost deposits is often used in the regulation of minor
offenses, and large fines are provided as optional forms of punish-
ment for more serious offenses. In addition, the award of money
damages as punishment or compensation in civil cases is often
seen as an attempt by the legal system to deter noncriminal be-
havior through the threat of economic deprivation. When the
main force of a deterrent threat is the unpleasantness of a fine or
cost, the operation of the legal threat appears to be close to that
of a price system. The measure of threatened punishment can be
rather precisely calibrated by dollar amounts, and both the of-
fender and the legal system may tend toward viewing the fine as
the “price” of the threatened behavior, a point of view that might
cause those who have paid the price to feel no moral compunction
about their offense, and cause those who have collected the fine
to attach less moral significance to the offense being punished in
this manner than if other forms of punishment had been used.

In one sense, the saliency of economic threats increases with
wealth : the threat of economic deprivation only operates on those
who have economic resources to lose, and elevation of the level of
fine threatened is of significance only to those who have sufficient
resources to pay the lower fine. In a second sense, the significance
of a fine threat is inversely proportional to financial means, for
the threat of a $10 fine would seem to be more severe to a man
with only $100 than to a man with $1,000. The threat of small
fines would thus seem more important to less wealthy members
of an audience, while the threat of very large fines should have
more influence on those with the means to pay them than on
persons without significant accumulations of capital.

75
Economic deprivations can be used to threaten activities that
are motivated by the desire for economic gain (in which case
the threat of punishment is peculiarly “in kind”) or to threaten
activities that are motivated by other desires. While studies have
shown the threat of small fines to be effective in requesting the
frequency of certain types of behavior, the studies to date have
dealt with minor offenses that are not strongly motivated, and
there is no reason to suppose that economic threats are of any
unique efficacy. However, economic deprivation is one form of
punishment that promises a direct gain to the punishing agency
in the form of the money taken from offenders by the act of
punishment. By contrast, most other forms of punishment result
in significant costs to the punishing agency.’ In this respect the
existence of a fine serves as an incentive to criminal-law enforce-
ment, where the costs of other forms of punishment might serve
as a brake on widespread enforcement.¢

Loss of Privileges

In many situations, conviction for an offense will result in the
offender losing privileges normally available to most members of
society. For some offenses of minor and intermediate severity,
the loss of privilege may be the most unpleasant of consequences
threatened, and will be specifically related to the offense (as
when serious or persistent traffic violations lead to the revocation
of the offender’s license to drive). With quite serious offenses, the
loss of privileges will be a relatively minor part of the total
punishment, and will not necessarily be related to the offense
leading to the loss (as when convicted felons are deprived of
voting rights, eligibility for public office, or the opportunity to
lawfully possess a firearm). When the lost privilege relates specifi-
cally to the offense leading to punishment, as in the driving
example, the threat of consequences may be of special signifi-
cance to potential offenders, because this group is more likely than
others to place importance on the privilege. Thus, the propensity
of persistent traffic offenders to commit such offenses is some
evidence either of frequent driving or of the special emotional

4 Studies to date of fine threats have related to illegal parking, traffic offenses, and non-
criminal conduct, while the effect of fine threats on major crimes has not been studied. By
contrast, most of the studies of imprisonment as a deterrent have concerned crimes such as
homicide, robbery, and assault. In one study, however, fines were found to operate as a greater
deterrent to recidivism of chronic drunks than jail sentences. See Lovald and Stub, supra
note 2.

® Thus, Professor Shelling defines a threat to be the announcement of “an act that one
party would have no incentive to perform, but that is designed to deter through its promise
of mutual harm.” Shelling, An Essay on Bargaining, 41 Am. Economic Rev. 281 (1956), at
282.

® The incentive to enforcement presented by fine systems is most evident when it produces
minimum ticket-quotas for parking and traffic police. While other incentives to enforcement
exist in other areas of crime control, it is only when the threatening agency stands to profit
in individual cases of conviction that increases in the level of enforcement will result in any
direct financial gain.

76
significance of driving to this particular group. The loss of priv-
ileges that are not specifically related to the offense, particularly
when attached to conviction for major crimes, may have less
significance to potential criminals than to most other citizens. It
is, for example, highly probable that the average citizen places
more value on his opportunity to vote or run for office than does
the potential armed robber.

The loss of privileges unrelated to the offense is not generally
discussed or evaluated as a deterrent mechanism when attached
to other major penalties for serious crimes. However, the ban on
the lawful possession of firearms by felons,” while not designed
as a deterrent to potential felons becoming actual felons, is viewed
as a means of preventing convicted felons from obtaining guns
that might be used in future crimes. Since firearms remain read-
ily available in most areas of the United States to convicted
felons through unlawful channels, the principal aim of this
type of regulation would appear to be that of deterring felons
from possessing and thus being tempted to use firearms. This
approach has not as yet been the subject of statistical evaluation.
The loss of driving privileges for persistent or serious traffic
violations is intended both as a general deterrent addressed at
large to all potential traffic violators and as a mechanism to re-
duce driving (and therefore the risk of violations) among per-
sistent offenders who have such privileges revoked. The effect of
such threats on the general rate of traffic violations has not been
studied, but there has been a considerable discussion and some
evaluation of license revocation as a means of reducing driving
violations among persons whose licenses are revoked. In this re-
gard, it has been argued that the essentially deterrent thrust of
revocation (since revocation generally does not make it physically
more difficult for an individual to gain access to a car) is more
effective than fines on persistent traffic offenders.’ Professor Pack-
er has, however, expressed some doubts about the efficacy of rev-
ocation as a means of treating traffic offenders, because the
practice employs a weak threat of detection where the incentive
to drive without a license is strong.*®

718 U.S.C. (1968) 924 (c).

8 See Newton and Zimring, Firearms and Violence in American Life (1969) at Chapters
1, 12-14.

9 Middendorf, The Effectiveness of Punishment and Other Measures of Treatment (1967).

10 “Unfortunately, we have no assurance that license suspension or revocation would be an
effective sanction. Failing some dramatic technological advance that would make it possible to
identify unlicensed drivers without stopping and questioning them, there is every reason to
believe that large numbers of people would simply accept the risk of being caught and would
continue to drive. The use of criminal punishment would simply be postponed one step rather
than avoided. There would be, if anything, a decrease in the deterrent efficacy of the law, be-
cause an unlicensed driver is less likely to be apprehended than one who is drunk.” Packer,
The Limits of the Criminal Sanction (1968) at 255.

The main thrust of Packer’s point deals with revocation as a deterrent to future violations

77
The criticism of driver’s license revocation as a special deter-
rent points up a general problem with the use of loss of priv-
ileges as a means of deterring further offenses by those subject
to revocation. The real meaning of loss of privileges (unless high-
ly reliable means for monitoring the use of the privileges exist)
is the addition of a legal threat to an existing one, where the
punished group has already ignored the existing threat. The same
type of criticism could be leveled at the threat of lost privileges
as a marginal general deterrent, because in order to outperform
the threat of fines or jail sentences, the loss of privileges would
have to influence the behavior of some people who are not im-
pressed by the threat of other sanctions for traffic offenses. In
the case of license revocation, however, there is no reason to re-
ject out of hand the notion that some traffic offenders will take
the threat of revocation more seriously than the threat of other
punishment for traffic offenses. Thus, if driving without a license
is either more severely threatened than other traffic crimes or
easier to detect, it may prove more effective than alternative
sanctions as a general deterrent to traffic violation, and even as a
special threat conveyed by way of revocation to convicted traffic
offenders. As is the case in so many controversies, a feasible pro-
gram of further research is the surest method of assessing the
pitfalls and potential of loss of privileges as a countermeasure
in a limited but important sector of the criminal justice system.

Confinement

Terms of penal confinement, ranging from a few days to the
remainder of an offender’s life, are one of many alternative sanc-
tions for minor crimes and, in this country, the usual punishment
threatened for major infractions. Apart from the fear of im-
prisonment because it is an incident for stigmatization, which

 

by those who have had their licenses revoked, and not with the general deterrent efficacy of
the threat of revocation, for he adds, “The apparatus of the criminal law, for all its terror,
usually produces a less onerous punishment than license revocation would be perceived as
being.” And although the difficulty with apprehension might dilute the public perception of the
disadvantages of revocation, it would probably not do so completely.

With respect to the credibility of license revocation, the quoted statement produces
a rare opportunity to disagree with Professor Packer’s analysis. A driver who is both drunk
and driving without a license is no less likely to be apprehended than one who is merely
drunk, and far easier to convict of a traffic violation. The penalties for driving after license
revocation appear to be the highest meted out to traffic offenders of any kind, according to
a study of sentencing in the Detroit traffic court. (Zimring, Punishment Theory and Traffic
Offenders, unpublished, 1966). And the driver without a license can be apprehended for this
offense whenever he is stopped by a police officer, not merely when he is drunk. General traf-
fic violations of the kind that betray the unlicensed driver occur far more often than drunk
arrests.

Thus, while many people, perhaps even a majority of the class, may continue to drive
after suspension or revocation, the revocation procedure and enforcement are not so tooth-
less as to preclude the possibility that many others will not.

Professor Crampton doubts the effectiveness of license revocation in the more limited area
of drunk-driving control, because many drunk-driving offenders are addicted to alcohol, al-
though this does not necessarily mean they are also addicted to driving. See Crampton, Driver
Behavior and Legal Sanctions: A Study of Deterrence, 67 U. Mich. L. Rev. 421 (1969).

78
will be discussed more generally under a separate heading, there
is much in the experience of penal confinement worthy of avoid-
ing, including loss of freedom to act and make choices, invasion
of privacy, rigid discipline, and living conditions far below or-
dinary standards of society. To paraphrase Lincoln on the expe-
rience of being tarred and feathered, if it weren’t for the honor
of imprisonment, there would still be powerful reasons for poten-
tial offenders to do without it.

Prison life also has a few characteristics that might offset its
manifest disadvantages for some people—the prisoner is fed and
sheltered, albeit substandardly, and is relieved of the often ago-
nizing responsibility to make any decisions about the course of his
daily life, and of the burden of fending for himself.’* But the dis-
advantages of penal confinement outweigh the advantages by an
overwhelming margin, and it is only among those who have been
rendered dependent by previous exposure to imprisonment that
any inclination toward a prison environment has been noted.

The quality of a person’s life-style in society may have some
bearing on the extent to which imprisonment is considered un-
desirable, because the relative deprivation of confinement should
be greatest in cases where the contrast between life on the out-
side and the prospect of prison is most extreme. To this effect,
mass media in the United States have found the contrast in
living conditions between prior life and prison to be noteworthy
in the case of James R. Hoffa, former president of the Inter-
national Brotherhood of Teamsters; the same prison conditions
provoke little comment when notorious criminals of lesser wealth
serve their sentences. But the relationship between life opportuni-
ty in society and the relative unattractiveness of the prospect of
penal confinement can be dangerously overstated: the most dis-
advantaged of our citizens stands to lose a precious measure of
personal freedom through penal confinement, and his ability to
distinguish between life in prison and life in society may be con-
siderably greater than that of dispassionate observers. With total
lack of personal liberty as the common denominator of penal con-
finement, the prospect of prison may well function with equal
effectiveness as a disincentive among rich and poor.

By the same token, what is essentially prison-like about pris-
ons will probably mean that the unattractiveness of prison as a
general deterrent is relatively insensitive to many changes in the
conditions of penal confinement. If this is the case, conjugal visits,
"1See Morris, The Habitual Criminal (1951) at Chapter 6 for a description of some per-
sonality types apparently suited to the discipline and dependency of prison life more ade-
quately than to the rigors of life in society.

12 See Clemmer, Observations on Imprisonment as a Source of Criminality, 41 J. Crim. L.,

C. and P.S. 311 (1950), for a discussion of the role which experience with prison might
play in developing patterns of dependency.

79
protein diets, and television privileges can be added or subtract-
ed from the basic environment of penal confinement without
affecting the fundamental aversion to loss of freedom experienced
by the vast majority of our citizens.

The perceptions of prison life held by those who have experi-
enced periods of confinement, and members of groups who have
not, are worthy of serious study by psychologists, but have not
as yet received a great deal of attention. It would be of interest,
and some importance, to determine whether potential criminals
view penal confinement in the same way as legislators and penal
authorities—as period of days or years during which experi-
ences will occur—or whether some potential criminals see confine-
ment most prominently as a series of deprivations, with the period
of time of secondary importance. It would be natural for a legis-
lator to assume, when making decisions about penalties for de-
terrent purposes, that 20 years in prison is twice as much as ten
years. But there are reasons to suppose that many potential crim-
inals (particularly those without any experience with confine-
ment) view the prospect of prison in a different light. Apart from
normal tendencies to discount future pain and pleasure, the in-
tensity of confinement as an experience rather than the length
of the experience might dominate the perception of most
potential offenders, with only those who are case-hardened to the
deprivations of prison life measuring the unpleasantness of pos-
sible future confinement in a time-frame. To the extent that time
assumes secondary importance in the perception of confinement
as a prospect worth avoiding, a 20-year term is not in any mean-

ingful sense twice the deprivation that a potential offender sees
in ten.

Physical Punishment

Every healthy organism seeks to protect itself from unpleasant
physical sensations, and to stay alive. Legal systems have, as long
as deterrence has been a goal of criminal sanctions, sought to
mobilize strong drives to avoid pain and death as a means to
reduce crime. The imposition of physical punishment is a basic
or primitive criminal sanction in at least three senses:

1. The drive to avoid pain and preserve life is a common
denominator for all men—if a man has nothing else to lose,
he still has an interest in preserving his physical security.

2. The imposition of physical punishment requires less ad-
vanced technology and fewer resources than the imposition of
penal confinement over a period of time.

3. The drive to avoid physical punishment need not be based
on any feelings of loyalty toward the agency threatening this
type of consequence, whereas the drive to avoid being branded

80
deviant is only frightening to the extent that a threatened
audience has some respect for the social institution making
that judgment.

It has also been suggested that the imposition of physical pun-
ishment is primitive in a pejorative sense, because such punish-
ments show little regard for the interests of the offender, and
brutalize society at large. In this regard, some have maintained
that the threat and imposition of brutal punishments impede the
moral development of societies that resort to them, and may un-
determine the moralizing effects of punishment. Thus, while it
can be argued that the willful taking of life merits capital pun-
ishment on other grounds, there are those who see a peculiar
inconsistency in showing society’s abhorrence of killing by taking
the life of the offender.

Agencies will resort to physical punishments in certain situa-
tions where alternative foundations for deterrence may not be
readily available. The threat of physical punishment is used ex-
tensively, for example, when the threatening agency does not
command the loyalty of its audience, as in the case of occupying
powers trying to secure the obedience of an enemy population, or
governments attempting to subdue internal rebellion. In most
Western nations the use of extreme physical punishment is com-
monly restricted to offenses that are considered particularly rep-
rehensible, or to offenders for whom other mechanisms appear
to be ineffective, as, for instance, where corporal and capital
threats are reserved for persons who are already confined in
penal institutions.

Because the drive to avoid pain is basic to all men, and the
drive to preserve life is not only basic but of ultimate impor-
tance, there is some reason to view such consequences as likely to
produce the broadest and most extensive of deterrent effects, but
there is little in present research to indicate that either corporal
or capital punishment has any unique deterrent efficacy when
other harsh measures, such as long-term imprisonment, are also
available. Modern discussions of corporal punishment have dis-
carded the notion that such punishments have salutary effects on
those actually punished. Little is available in the way of research
on the relative effectiveness of corporal punishment as a general
deterrent; if the threat of such punishment outperforms the
threat of other penal measures, the evidence of this phenomenon
is so far undiscovered.

Research into the effectiveness of capital punishment versus
protracted imprisonment as deterrents to criminal homicide is
abundant. Studies have compared homicide rates of neighboring
States with similar social environments but contrasting penalties

81
for homicide, and concluded that executions have no discernible
effect on homicide rates.’® Research efforts have focused on the
same jurisdiction before and after the death penalty was either
introduced or abolished without finding any change in the rate
of homicide that appeared to be attributable to change in penal
policy.’ Other strategies of research have been used, and produce
no finding of the special effectiveness of the threat of death.!s
If the threat of death has no marginal deterrent value when
compared with protracted imprisonment, it still may be true that
capital punishment would serve as an effective marginal deterrent
in other types of crime more susceptible to the influence of
variations in threatened punishment, and that the death penalty,
even in homicide, would function as an effective marginal deter-
rent when compared with much lower terms of imprisonment or
lesser penal measures. What these studies do establish, however,
is that the special effectiveness of even the most extreme of phys-

ical punishment as a means of securing obedience to law cannot
be assumed.

Stigmatization

Most men place a high value on the regard of their friends and
neighbors, and on their standing in society. The discovery that
a person has committed a crime can change the community stand-
ing of an individual because social feelings about the crime will
attach to social judgments about the criminal. In addition, the
consequences of apprehension for criminal behavior may carry
with them important measures of social reprobation. Imprison-
ment results in the socially unpleasant status of prisoner, and the
label is likely to last longer than the sentence of imprisonment.
Since one of the most frightening prospects of social life is what
Lewin called “the danger of exclusion from the group,” the pos-
sibility that criminal activity might lead to stigmatization, a
process that involves elements of symbolic and actual exclusion,
may serve as a significant deterrent to crime.

The relationship between legal punishment and social stig-
matization varies from case to case. Some types of trial-and-
punishment proceedings are specifically designed to bring down
the wrath of society on the offender, the institutions of public
display in the pillory and the publication of names of guilty of-
fenders being not the least examples of such intentions. Many of
the conditions of prison life in modern times have been described
as ceremonies of degradation meant to convey social disapproval.
In other cases, stigmatization attaches to knowledge that an of-

13 See the papers collected in Bedau, The Death Penalty in America (1964), at 258-332.
14 See Ibid., at 274-84.
1 See Ibid., at 284-382.

82
fender has been apprehended for a particular crime, or punish-
ed in a particular way, without any extra efforts on the part of
the punishing agency. Indeed, stigmatization may cling to punish-
ment and conviction even when the legal system tries to dis-
courage it. But if social stigmatization can exist independent of
legal efforts to achieve or accentuate it, the prospect of stigmati-
zation must still be counted as a formidable element of the mes-
sage communicated by many legal threats.

The prospect of stigmatization functions differently than the
threat of other types of consequences (such as pain or confine-
ment) because it is of special importance to those who attach
great significance to the judgments of others, and because the
sting of community attitudes is difficult to administer to offenders
in calibrated doses. Among members of a threatened audience,
the fear of stigmatization will thus be greatest among those who
are most socialized, those presumably least in need of a deterrent
in the majority of cases. When stigma attaches to an individual
as the result of punishment for a serious crime, it is more likely
to be a total destruction of reputation than a small step down
the ladder. And when an offender has been stigmatized once, the
effects of this process will tend to stay with him for many years.
Having earned the status of criminal, the offender might feel he
has little to lose from further stigmatization and, correspondingly,
he will have less incentive to avoid consequences that bring with
them the threat of stigma.

If the threat of particular consequences functions as a deter-
rent because it carries the prospect of stigma, as well as other
deprivations, it is essential to recognize that variations in the
severity of the nonstigma penalty are not necessarily accompa-
nied by upward or downward shifts in the stigma value of pun-
ishment of similar magnitude. Doubling a prison sentence from
five to ten years is not likely to double the social consequences of
conviction. To the extent that fear of stigmatization, rather than
fear of time in prison, influences the behavior of a threatened
audience, this change cannot be expected to exert a great mar-
ginal deterrent effect. And since the threat of further social stig-
ma may mean little to the man who has already been branded a
criminal, the same threat of consequences may have less meaning
to those who have already been stigmatized than to other classes
of potential criminals.

F. Variations in Severity of Consequences

Since the power of a legal threat to function as a simple de-
terrent comes from the unpleasantness of the consequences threat-
ened, one natural strategy for increasing the deterrent efficacy

83
of threats is to increase the severity of threatened consequences.
The theory of increased penalties as a marginal deterrent is sim-
ple and straightforward: all other things being equal, an increase
in the severity of consequences threatened should reduce the num-
ber of people willing to run the risk of committing a particular
criminal act, in much the same way that increases in the price
of a product will decrease the public demand for it.

If variations in the severity of threatened consequences were
indistinguishable in effect from the influence of price increases,
there would be reason to assume that all increases in severity
would tend to reduce the number of threatened offenses com-
mitted,’ but there is no basis for concluding that the extent of
the marginal deterrent effect produced would be substantial in
all cases. Let us suppose that apples are selling for 20 cents a
pound and close substitutes such as oranges or pears are 25 cents
a pound; doubling the price of apples would have a much more
significant effect on the quantity demanded than if apples were
20 dollars a pound, with close substitutes selling for 25 cents,
and the price of apples were subsequently doubled. And doubling
the price of apples would tend to have less of an effect on the
quantity of apples demanded if there are no close substitute
commodities available, than if a number of other types of food
are available to consumers that have the same attributes that
make apples desirable. For similar reasons, when penalties for
criminal activity that many people find attractive are quite low,
thereby making crime a reasonable alternative to legitimate means
of obtaining gratification for many persons, even a high probabil-
ity of apprehension may leave a high rate of the threatened be-
havior, and increases in the severity of threatened consequences
can be expected to have a more substantial marginal deterrent
effect than if the level of consequences threatened is already
quite high in relation to the benefits obtainable through crim-
inal means. Increases in severity would also be expected to pro-
duce more dramatic marginal deterrent effects when noncrimi-
nal or less severely threatened alternatives exist for getting that
which makes the threatened behavior attractive than when no
such alternatives exist.

Elements of the comparison between effects of price changes
and severity of threatened consequences take on an air of unreal-
ity because some of the examples of commodity pricing used
above are rather ludicrous: it is probable that if apples were sell-
ing for 80 times as much as other forms of fruit, doubling their

16 See Stigler, The Optimum Enforcement of Laws (unpublished, 1969), at 3: “The offender
is deterred by the expected punishment, which is (as a first approximation) the probability
of punishment times the punishment—$100 if the probability of conviction is 1/10 and the
fine $1000. Hence, increasing the punishment would seem always to increase the deterrence.”

84
price would have no effect on sales because that quantity would
be zero at either price. But the very whimsicality of that example
may point up differences between the function of price changes
and changes in the level of threatened consequences that force
some modification of the theory of increase in severity as a mar-
ginal deterrent. One important difference between penalty and
price is that it can be assumed that a potential buyer will know
the price of a product before making a purchase decision; no
such assumption can be made about the potential criminal’s knowl-
edge of the consequences threatened for a particular behavior.
Ignorance about the severity of threatened consequences, which
is widespread among potential criminals,” may lead to irration-
al conduct, on the order of buying $40 apples, and will negate
the possibility that change in consequences will operate as a mar-
ginal deterrent for the ignorant.

A second difference between penalty and price is that the price
of a product is a disadvantage to purchase that will occur with
certainty, while the consequences threatened for an offense will
occur only some of the time. This factor, by itself, would not nec-
essarily impinge on the marginal deterrent effect of increases
in the severity of threatened consequences in cases where the
risk of apprehension remains constant, for if potential offenders
believe that their chances of apprehension cannot be dismissed,
the risk of a high penalty provides more incentive to avoid crime
than the risk of a low penalty. However, if potential offenders
use the uncertainty of apprehension as an excuse for believing
that they will not be caught, they will be immune to the influence
of increases in the severity of threatened consequences. And this
perception of personal immunity may be widespread among those
whose behavior must be affected if increases in threatened con-
sequences are to operate as a marginal deterrent, both when the
risk of apprehension is generally quite low, and when the threat-
ened consequences are so high at the outset that only the fool-
hardy would find it worthwhile to run the risk of apprehension
and punishment.

A number of studies on the impact of penalty increases under-
score the point that wholesale importation of economic models
and unitary assumptions about the effects of penalty shifts are
not consistent with existing data. Efforts to discern whether the
death penalty reduces homicide more than protracted imprison-
ment, mentioned earlier, support the conclusion that the addition
or elimination of the death penalty has no effect on homicide rates.
Studies in single jurisdictions before and after upward shifts in

17 See subsection C, notes 4-6 and accompanying text, supra.
penalties provided for rape,’ marihuana possession,'® and assaults
on police officers 2° have found no change in the trend of crime
rates that could be attributed to the change in penalty, although
studies of crime-trends in a single jurisdiction are less reliable
than the series of studies on capital punishment and homicide.
Two efforts to study a rarely enforced Nebraska law provid-
ing much stiffer maximum sentences for bad checks over $35
than for smaller checks failed to find any difference in the fre-
quency of the larger checks that could be attributed to the dif-
ference in threatened consequences.’ Chambliss found that park-
ing offenses decreased after increases in the fines threatened for
persistent offenders, although other changes in policy toward
parking offenders probably accounted for the major share of the
decrease.??

Rusche and Kirchheimer confronted the issue of marginal de-
terrence in a broader framework by analyzing rates of a number
of different crimes in England, France, Italy, and Germany from
1910 to 1928.2 In England, the authors note a small decrease in
crime, together with a trend toward greater leniency. In France,
rates of most crimes remained stable, while punishment levels
decreased.” In Italy and Germany, punishment for major crimes
increased, but the crime rate fluctuated without any apparent re-
lationship to punishment levels.?” As the authors of this study rec-
ognized, the countries being compared differed substantially in
attributes other than punishment policy, making it impossible
to draw positive conclusions about the effect of higher penalties
on crime rates. But the figures from this study “provide no basis
for assuming that the policy of punishment affects criminal-
yn =

Gibbs 2" and Tittle 2* analyzed crime statistics and punishment
data from the various States in the United States in an effort
to determine whether variations in the probability of apprehen-

18 Schwartz, The Effect in Philadelphia of Pennsylvania's Increased Penalties for Rape and
Attempted Rape, 59 J. Crim. L., C. and P.S. 509 (1968).

1 California Assembly Office of Research, Crime and Penalties in California (1968), at 12.
The California statistics on marihuana deal only with the number of reported arrests, be-
cause no reliable estimates of the gross number of crimes are available for this crime with-
out a victim willing to report the offense to the police.

20 Ibid., at 10-12.

21 See Beutel, Experimental Jurisprudence (1957). In 1967, a study supported by the Center
for Studies in Criminal Justice at the University of Chicago used different methods to come
to the same conclusion. See Zimring, Bad Checks in Nebraska, A Study in Complex Threats
(unpublished, 1968).

22 Chambliss, The Deterrent Influence of Punishment, 12 Crime and Deling. 70 (1966).

2 Rusche and Kirchheimer, Punishment and Social Structure (1939).

24 Ibid., at 193-200.

8 Ibal., at 200-205.

28 [bal., at 200.

Gibbs, Crime, Punishment, and Deterrence, 48 Southwest Social Science Quarterly 515
(1968).

28 Tittle, Crime Rates and Legal Sanctions, 16 Social Problems 409 (1969).

86
sion or severity of punishment were related to variations in
the rate of particular crimes. The Gibbs study related only to
homicide, while the Tittle study involved homicide, assault, lar-
ceny, burglary, robbery, sex offenses, and auto theft. For one
crime, homicide, higher than average severity of sentences
was significantly correlated with lower than average crime rates
when figures for all States were analyzed without distinctions
based on region of the country.” For the other offenses, no sig-
nificant correlation between severity of sentences and the crime
rate was found, apparently leading to the remarkable conclu-
sion that severity of sentence influenced the rates of the crime
thought to be among the least deterable of offenses, but did not
similarly affect such traditional property crimes as larceny and
burglary.’ However, when severity of punishment and crime
rates were compared by grouping States into regions, States with
penalties higher than the average for their regions did no better
than those with lower penalties in homicide, or in the other cate-
gories of crime where nationwide totals had also shown no appar-
ent relationship.” It thus appears that regional differences in
crime that are unrelated to severity of punishment were responsi-
ble for the nationwide correlation between punishment levels and
crime.??

The studies to date illustrate some of the difficulties involved
in evaluating the impact of severity of consequences on crime

2 Ibid., at 415.

 Ibid., at 415-417.

3 The procedure followed to control for regional effects in Professor Tittle’s data was (1)
the division of the country into five regions—East, South, Border, Midwest, and West; (2)
the calculation for each region of the mean rate of the particular crime and the mean length
of sentence for that crime; and (3) the rating of each State with respect to whether it was
above or below the average for its region in rate of particular crime and in severity of sanc-
tion. It was found that while, for instance, the correlation between average prison sentence
and homicide rate for the Nation as a whole was —.46, in only 21 of the 46 States for which
data were available did a State with higher than regional average penalties show a lower
than regional average rate or, to the same effect, did a State with lower than regional
average penalties experience a higher than regional average homicide rate. At the same time,
in 25 of the 46 States, mean sentences and mean crime rates were both higher or lower than
the regional average. I conducted this analysis of the data, with Professor Tittle’s assistance.

32 The principal problem with using an undifferentiated national sample of States is that
areas, such as the Deep South, that have very high rates of homicide appear to regard intra-
racial killing as less serious than other areas of the country. Thus, the South, with a rate
of homicide of .09 per thousand during the period studied, compared with .03 for the Mid-
west and .02 for the East, had an average prison sentence of 89.6 months for homicide,
compared with 199.4 for the Midwest and 214.5 for the East, appearing to suggest that
higher penalties create less homicide. But when we ask whether those States in the South
with higher penalties do any better than their neighbors, the data suggest they do not. The
five Southern States with lower than mean sentences for homicide have slightly lower homi-
cide rates than the four higher than mean punishment States. The relationship between
severity and rate, undoubtedly present in the nation as a whole, is apparently not a casual
one. See note 31 supra for data on all 46 States controlled for region, and supra note 11,
subsection D, for a discussion of the relationship between probability of punishment and
crime rates discussed by Tittle. A further problem arises because homicide is also threatened
with the death penalty in a majority of states, and is universally threatened in the South,
which otherwise has the lowest average penalty for homicide. The Gibbs and Tittle data exclude
death sentences.

87
rates. Comparative study is dangerous, because factors other than
the crime rate may cause crime rates to differ, and areas that
differ in punishment policy may also differ in these other fac-
tors. Studying the crime rate before and after a change in pen-
alty is helpful, but it is difficult to determine on the basis of
such studies what the crime rate would have been in the pe-
riod following the change if the change had not occurred—trends
before the change are of some assistance, but trends change.
In spite of these methodological problems, existing studies do
support a number of conclusions.

First, the cross-cultural and interstate comparisons of sentence
severity and the crime rate show that moderate variations in
the severity of punishment for serious crimes is not a major fac-
tor in explaining the differences in crime rate that are observed
in different jurisdictions. While such studies are far from per-
fect, they do tend to set an upper limit on the amount of varia-
tion in the crime rate that might be accountable to punishment
policy.

Second, the specific investigations of capital punishment for
homicide, and those of upward shifts in the minimum and maxi-
mum terms of imprisonment for other major crimes, have tend-
ed to produce negative conclusions. It would be tempting to gen-
eralize from these specific case studies to the conclusion that in-
creases in severity hold little promise as marginal deterrents,
but such a generalization would be unwarranted. Many of the
studies of upward shifts in penalty, such as those of the change
in punishment for possession of marihuana and assault of police
officers in California, base their conclusions on data that cannot
completely contradict the possibility that increased penalties had
a marginal deterrent effect.?® Moreover, many of these studies—
including the California studies, the Nebraska bad-check sound-
ing, and probably the Philadelphia rape analysis—focused on
changes of penalty where public knowledge of penalties was low.
There is no reason to believe, out of hand, that similar results
would be obtained under more favorable conditions. Also impor-
tant, all of these studies involved moderate shifts in legislatively

3 The California marihuana study presents the weakest evidence, because it rests on arrest
trends rather than changes in the crime rate. And all of the single jurisdiction trend studies
leave open the possibility that trends can change, and that shift in penalty may be evidence
of a change in trend. However, the cumulative significance of a number of single jurisdic-
tion ‘‘before and after” studies, if all fail to reveal changes that could be attributed to the
change in punishment policy, is less vulnerable to attack than such studies when considered
one at a time. And when comparative and retrospective techniques are combined and fre-
quently repeated, as was the case in the capital punishment studies, the use of imperfect
research techniques can lead to the construction of an impressive case for or against marginal
deterrence in a particular comparison of penalty threats.

3 See Social Psychology Research Associates, Public Knowledge of Criminal Penalties: A
Research Report (1968), for data on California. The Nebraska data are discussed in subsec-
tion C, supra.

88
prescribed punishments for offenses that were already severely
punished. It cannot be concluded that larger jumps in penalty
will have no effect on crime rates, particularly when the base pen-
alty for an offense is quite low.

But available data do suggest that increases in legislatively
provided penalties for major crimes have little impact as a mar-
ginal general deterrent in many situations where officials place
great faith in such increases, and this trend could serve as the
basis for experimental downward shifts in prison sentences for
major crimes where present penalty levels reflect deterrent in-
tentions that may not be producing results.

The failure of many increases in severity of consequences to
function as a significant marginal deterrent to major crimes is
far from inexplicable. Earlier discussion has alluded to the wide-
spread lack of information about penalties among many classes
of potential criminals, to the persistent optimism of many po-
tential offenders, and to the fact that increases in punishment
that convey stigma do not necessarily result in increased stig-
matization. In addition to these factors, it is likely that increases
in the severity of threatened consequences are more or less sig-
nificant, depending on the relationship between size of penalty
increase and size of base penalty. If this is the case, raising a
threatened prison sentence from five to seven years will have less
impact than raising a threatened penalty from one month to one
year, even though the former shift represents more than twice
as much additional punishment. In each case, if the higher pen-
alty threat is to have a marginal deterrent effect, it will do so
by persuading those who would commit crimes if confronted
with the old threat to refrain from criminal conduct because of
the increase in threatened penalty. In the case of the five-
year base penalty, any marginal general deterrence will have
to occur among people who were willing to risk five years
in prison, and this group is confronted with only a 40 percent
increase in threatened penalty. When a penalty is raised from
one month to one year, the relative increase is 1100 percent, and
the population among whom marginal general deterrence must
operate is not necessarily as intransigent as in the case of the
five-year base penalty. There may be people who are just as im-
mune to the threat of consequences in the group of potential crim-
inals threatened by the one-month penalty, and for them the
increase in penalty from one to 12 months will have no effect.
But since the base penalty is lower, there may be a large num-
ber of people willing to risk the month in jail who are not as
strongly motivated or as difficult to influence as the group of peo-
ple willing to risk five years. When the large relative increase

89
in punishment is put into effect, it may function as a marginal
deterrent by substantially reducing the number of crimes com-
mitted by these less impervious potential offenders.

If the relative magnitude of increases in marginal deterrence
is important, upward shifts in the level of threatened conse-
quences will result in diminishing marginal returns as the base
penalty increases. This does not mean, of course, that increases
in threatened penalties will fail to produce any marginal deter-
rence when base penalties are large, for it is possible that such
increases will produce relatively small deterrent returns that are
of great social importance. Thus, it cannot be said that the dimin-
ishing effects of increases in severity in weeding out potential
criminals is a reason for rejecting increases in relatively high
base penalties as proper policy in all cases.

The Uses of Variations in Severity:
The Fortress and the Stepladder

If potential criminals are responsive to variations in the se-
verity of threatened consequences, how can the legal system make
the best use of variations in severity to achieve social defense?

One answer is that, since the goal of all legal threats is to keep
the population law abiding, the potential effectiveness of varia-
tions in severity of threatening consequences should be used to
create the widest possible distinction between criminal and non-
criminal behavior by threatening all types of serious crime with
penalties which are as severe as possible. The aim of this strat-
egy is to create a walled fortress around criminal activity by
using the full power of threatened consequences to keep potential
criminals from becoming actual criminals.

Another possible strategy would be to threaten all serious
crimes with major penalties, but to save a considerable amount
of variation in threatened penalties to underscore distinctions
between types of crime, as well as between serious crime and law-
abiding behavior. This theory sees the existence of threats of
conviction and imprisonment as the mechanism for trying to
keep all of the population law-abiding, but seeks to use variations
in the severity of major penalties to create a stepladder effect,
threatening those crimes which are considered most serious with
substantially greater consequences than less serious but still ma-
jor infractions of the criminal code. Thus, while simple larceny,
burglary, robbery, and armed robbery are all serious threats to
the security of property in society, it is nontheless true that they
are not equally serious crimes. An act of burglary is more se-
rious than simple larceny because it involves the invasion of a
person’s home or business. Robbery is more serious than burg-

90
lary because the robber uses force against his victim, which in-
creases personal fear and may lead to injury or death. Armed
robbery is more serious than strong-arm robbery because the
presence of a weapon increases the risk that a victim will be
killed. One method of using variations in severity of threatened
consequences is to make the distinction between penalties for
different types of crime as great as possible, in order to pressure
offenders to commit less serious offenses when total conformity to
the law is not obtainable.

Because the range of penalty options in society is not infinite
and will be limited by considerations other than deterrence, there
is tension between an emphasis on “fortress” and “stepladder”
strategies in crime-control policy. To the extent that a fortress
approach utilizes severity of consequences as a means of securing
greater conformity to law, it requires high penalties for serious
crimes of all kinds. Since the range of punishment options is
limited, and since increases in threatened consequences are more
significant when base penalties are low, the greatest contrast
between types of crime is achieved when base penalties are low-
est. Table 1 compares hypothetical sentence structures for a va-
riety of property crimes.

Table 1
Two Hypothetical Sentence Structures
Strategy “A” Strategy “B”
Years % increase Years % increase
in from penal- in from penal-
prison ty for next prison ty for next
most serious most serious
crime crime
Larceny 5 oo 1 oo
Burglary 10 100 3 200
Robbery 20 100(?) 9 200
Armed robbery 40 100(?) 27 200(?)
Robbery murder LIFE ? LIFE ?

Both schedules are mixes of fortress and stepladder strategies,
as are all penalty structures.

The table assumes that the legal system has an equal stake
in differentiating between property offenses at each point on the
ascending scale, which is probably not true. But it illustrates
that Schedule B can maintain 200 percent increases in penalty
between classes of property offense without running out of punish-
ment options, while Schedule A can only maintain 100 percent
increases. Moreover, while Schedule B confronts the practical

91
problem of whether life imprisonment really exceeds 27 years
by any significant degree when differentiating armed robbery
from robbery murder, Schedule A encounters this problem ear-
lier, because assumptions about longevity are involved in the dis-
tinction between robbery and armed robbery. It will probably
be wise in both sentence structures to increase the contrast be-
tween penalties for armed robbery and robbery murder, but this
can only be done by decreasing the gap between penalties for
other crimes of graduated seriousness, or by lowering the base
penalty for larceny.

The extent to which social policy is better served by sentence
structures that emphasize fortress or stepladder effects depends
on data about the degree to which each strategy is effective in
particular situations, and also on questions of relative value that
empirical research cannot fully answer. On the degree of rela-
tive effectiveness of fortress and stepladder strategies it is im-
portant to note that because an increase in the penalty for burg-
lary has reduced the rate of that crime does not mean that all
who have been deterred from burglary now obey the law—many
may have taken up other forms of crime, where the rewards are
higher or the risks and penalties are lower. It may be easier to
scale down the dangerousness of crime than to completely elimi-
nate criminal tendencies among an audience of people who would
violate laws at a given base penalty.

Beyond the empirical questions of the effects of increases in
severity of punishment on potential criminals, a question of rela-
tive values may have to be resolved before judgments can be
made about the efficacy of fortress and stepladder approaches. As-
sume that Schedule A in Table 1 produces 20 percent less larceny
and 5 percent more burglary, robbery, and armed robbery than
Schedule B. Any judgment about the relative effectiveness of the
two approaches will necessarily involve determining how many
extra larcenies it is worth to society to prevent the added danger
of one robbery." Empirical research can help define the particular
types of harm generated by each type of offense, but the ultimate
issue is one of values.

G. The Moral Quality of Threatened Behavior:
Conscientious Objection and Deterrence

In most cases when society as a whole views a particular
criminal act as morally wrong, those who commit that act will
consider it either morally wrong, morally neutral, or permissible
because it is the means to an end that is judged to be morally ac-
ceptable. Only in the last of these conditions will the threat of

35 See the discussion of this point in Part II, Section 3.

92
consequences go against the moral sensibilities of the threatened
audience, and a sense of moral objection to the threat might not
occur even then unless alternative means to the same end are
not available. There are, however, a few cases where the commu-
nity threatens activity that members of the threatened audience
regard as morally imperative. In the United States the laws
against polygamy as applied to the Mormons,' laws prohibiting
civil rights demonstrations,> and various provisions of the Selec-
tive Service Act as applied against pacifists serve as outstanding
examples of this type of occurrence. In such circumstances, if the
members of a threatened audience are otherwise in accord with
the legal system, the criminalization of morally valued activity
will set the stage for conflict between general feelings of loyalty
toward society and allegiance to the particular behavior. Often
this conflict will be between the individual’s loyalty to the all-
inclusive society and his loyalty to the particular subculture that
subscribes to criminal conduct. In some cases, as with the Selec-
tive Service Act since 1965, the moral value placed on the crim-
inal conduct may be largely the product of individual feelings,
and a group will have grown around the belief, rather than vice
versa.

In cases of conflict where loyalty to the parent society emerges
a clear victor, the moral value of the criminalized act will usual-
ly fade and will not permanently complicate the task of predict-
ing how the threat’s audience will respond to its terms. But if
members of a threatened audience retain the strong opinion that
the threatened behavior is morally right, predicting their
response to the terms of the threat is complicated, and the
existence of that particular threat may lead to more basic attitude
and behavior changes on the part of some conscientious objec-
tors. Those who retain their former attitudes about the threaten-
ed behavior will continue to have strong drives toward engaging
in it but will be confronted with the prospect of social stigma
and other types of punishment if apprehended.

While a normal response to unpleasant consequences is to avoid
them, the moral value placed on the act may cause the moral
quality of punishment threats to become inverted. Under such
circumstances the social stigma attached to persons apprehended
for committing the threatened offense may be viewed by
those who value the behavior as evidence of righteousness.® Pain
and deprivation suffered on account of having committed the

1 See Reynolds v. United States, 98 U.S. 145 (1878) ; Mormon Church v. United States, 136
U.S. 1 (1890).

2 See Cox v. Louisiana, 379 U.S. 536 (1965).

3 An allusion to the inverted moral quality of imprisonment can be found in the title of
Martin Luther King’s “Letter from a Birmingham Jail.”

93
morally valued act may be translated into symbols of the sin-
cerity of personal convictions and evidence of strength of char-
acter. And though powerful psychological forces will still be at
work influencing individuals to avoid painful consequences, the
process of inversion will limit their effectiveness.

If a group of those who place moral value on threatened be-
havior exists to give support to the individual confronted with
the choice of compliance or punishment, it is far more likely that
people will invert the moral quality of punishment and actively
rebel against the criminalization of the morally valued act. But
people operating without the support of such groups have, un-
der some circumstances, defied such threats, and history pays
special recognition to some members of this class. Moreover, it
is possible that people placing moral value on a particular form
of threatened behavior will coalesce to provide support for in-
version of consequences and defiance of the law. The formation
and activities of such a group would appear to be the basis of the
prosecution case in the recent Boston trial for conspiracy of
nine members of an ad hoc anti-draft movement.*

The threat and punishment of behavior judged to have moral
value may also result in more generalized opposition to the legal
system and a greater variety of criminal activities on the part
of individuals who rebel at the enforcement of what they con-
sider an unjust law. Inverting the moralizing quality of punish-
ment for a particular act may result in a loss of respect for law
in more general terms. This loss of respect would appear to low-
er barriers to the commission of other types of criminal behavior
and could result in higher rates of many types of crime, par-
ticularly if other crimes can be related to the ideological goals
that underlie the initial rebellion. Some traces of this tendency
can be seen in the recent behavior of some “revolutionary” mili-
tants in the anti-war and civil rights movements, where feel-
ings of frustration and anger over a series of laws thought to be
unjust were generalized over a period of time into a loss of respect
for the legal system, leading to the advocacy of a greater variety
of illegal tactics.

Yet the behavior of a large number of groups morally opposed
to particular laws is surprisingly law-abiding. The Jehovah's
Witnesses, the Quakers and, more recently, the vast bulk of those
in the pacifist and Civil Rights movements of the 1960s do not
generalize defiance of one law into other types of criminal con-
duct, and members of such groups are very rarely involved in
criminal conduct for which no personal moral justification exists.
The limits to lawbreaking observed among these groups may

4 See United States v. Spock, 416 Fed. 2d 165 (1st Cir. 1969).

94
result from the fact that opposition to particular laws on moral
grounds appears most often in groups that have strongly de-
veloped moral codes. For these groups the process of inversion
is self-limiting because it will take place only when the group’s
moral judgment is contrary to that of the legal system. For
all other types of crime, the moral quality of the threat of
punishment will not be inverted, and both the group sense of
morality and a legal threat will stand as barriers to the com-
mission of crime. It is only when such a group becomes totally
revolutionary that opposition to particular laws becomes a gen-
eral moral license to crime, and even then, group behavior may be
restricted by the constraints of “revolutionary morality.” 5

When authorities are frustrated by conscientious objection to
a particular legal prohibition, one method of dealing with such
opposition is to increase the penalties threatened for violation.
As noted before, some of the sting of threatened consequences
will be removed by the process of inversion, but such increases
in penalty may still provide members of the threatened audience
with reasons for refraining from the threatened behavior.
Whether penalty increases will reduce the rate of the particular
threatened behavior through increased fear of punishment more
than it will increase defiance by solidifying opposition is an
empirical question; results vary from case to case. One pos-
sibility is that such penalty shifts will reduce the rate of de-
viancy per 100 objectors, but that harsh penalties will increase
the resistance and hostility of those who remain defiant and
may also recruit new sympathizers to the cause of the conscien-
tious objectors in all but the most highly polarized of social
climates. This pattern of response appears to have been one after-
math of the harsh official and unofficial punishments meted out
to civil rights demonstrators in the South in the early 1960s.

Less powerful side effects can occur when legal threats are
attached to behavior that has been customary but not supported
by strong moral feelings. Resistance to the prohibition of alco-
holic beverages, for example, may be widespread when the use
of alcohol has been an accepted custom, but such resistance will
seldom initially assume the character of a crusade. However,
when harsh penalties are used to suppress customary behaviors,
or when the prohibition of a behavior is perceived by a group
as an attempt to persecute the group rather than deal with the
behavior, moral indignation and many of its consequences may

5 Common criminals have seldom become heroic figures in ideological movements, and rather
stern moral codes usually have been incorporated in revolutionary ideologies. To this effect,
Lenin is reputed to have complained that if the German socialists were to board buses in order
to nationalize them, they would first pay the fare.

95
appear, complicating the task of predicting the effect of threat
and punishment.

H. Moderating Individual Propensities through Group Pressure

An understanding of group processes is an essential precondi-
tion to understanding patterns of individual differences in threat
response, because the small family and peer groups that constitute
the most important part of our lives shape and modify the at-
titudes and values that are of importance in predicting how an
individual will respond to a threat. Beyond this basic degree of
group influence, discussed earlier, there are two situations in
which group pressures may determine threat responses:

1. Often a person will be forced to make a decision about his
response to a threat that might become visible to other mem-
bers of his group, in which case strong pressures to conform to
group expectations will be generated by the prospect that his
decision will become known.

2. In a smaller number of cases, groups rather than individ-
uals are the relevant unit of criminal participation; in such
cases, the reaction to legal threats will be collective rather than
individual.

When an individual must make a decision about responding to
a threat that may become visible to other group members, the
individual’s fear of disapproval by group members will supply
him with an additional reason to heed a legal threat if group
values include respect for the law and disapproval of criminal
conduct. In some cases, the deviant act itself might be visible to
the group if committed, as when obedience to the law demands
certain public gestures or disobedience requires a conspicuous
change in the individual's life style. In most other cases, the
danger of deviant conduct becoming visible to group members
will come from the risk of detection or formal punishment. In
such situations, the risk of detection includes the risk of visibility,
and the prospect of visibility is one of the significant forms of
unpleasantness an individual may seek to avoid. But when group
values are in conflict with those that underlie the legal threat,
the visibility of an individual’s response to a threat results in
pressure to defy the threat in order to preserve or secure standing
with the group. Against the prospect that deviant behavior will
result in punishment, the threatened individual must balance the
likelihood that failure to defy a threat will lower his standing
among the group and result in his being branded disloyal or cow-
ardly. It would thus seem much easier, and much less humiliating,
to secure the compliance of a misbehaving child when this process

* Andenaes, The General Preventive Effects of Punishment, 114 U. of Pa. L. Rev. 949, at 961.

96
is not visible to his deviant peers rather than in the presence of a
group with negative opinions of the value of conformity, and it is
likely that in this respect the child is the father of the man.

Group processes are of crucial importance even more directly
in the large number of cases where decisions to obey or disobey
the legal threat are made by small groups rather than individ-
uals. Studies of juvenile offenders can be taken as evidence that
much, perhaps even most, of juvenile crime is group activity.
Offenses such as vandalism, gang-fighting,* false fire alarm pull-
ing, pack rape, and many of the property crimes committed by
juveniles are typically committed by groups of youths. While group
crime is less common among older offenders, it is far from un-
heard of, and mob violence is by definition a group phenomenon.
When threat response is collective, individual members of a group
may find themselves willing to risk unpleasantness that would be
sufficient to deter them if each alone were responsible for the
decision, because the significance and immediacy of group pres-
sures toward participating in criminal activity overwhelm
ordinary tendencies toward caution. Because the pressures gener-
ated by the group are powerful, it may be that legal threat must
reach a particular proportion of the group, analogous to a critical
mass, before any of the group members will be restrained by
threat from engaging in criminal conduct. And deterring a large
proportion of a group contemplating criminal activity may be
difficult because fear of the law may not be considered a re-
spectable emotion, and those who feel it may be unwilling to make
their fears known. If these suppositions are correct, effective
communication among group members will often be necessary if
threats are to influence collective decisions.

Section 2. Special Deterrence

Most adults have been exposed to minor punishments for park-
ing and traffic offenses; very few people have been subjected to
major penalties. Those who have been subjected to legal conse-
quences, large or small, may react differently to legal threats of
all kinds, and particularly to legal threats concerning the be-
havior that led to previous punishment. For this reason, a sep-
arate discussion of some issues relating to special deterrence
is appropriate.

Any difference in threat response between previously punished
offenders and the rest of the members of a threatened audience
could result from:

a. the fact that apprehended and punished offenders, partic-

2 See, e.g., Carney, Mattick, and Calloway, Action on the Streets, at 60-66.

97
ularly serious offenders, are a special high-risk group, inde-
pendent of their experience with the legal system;

b. changes in attitudes, expectations, or status that the ex-
perience of punishment brings about;

c. a mixture of both.

The fact that apprehended offenders are a special group in
the population, apart from whatever effects apprehension and
punishment might have, is a necessary beginning to any discus-
sion of special deterrence. When half of all those who have been
convicted of larceny are arrested for larceny a second time, while
less than one percent of all other members of a threatened
audience are ever apprehended, some might be tempted to con-
clude that the apprehension and punishment of thieves does not
deter them from further offenses. To the extent that this con-
clusion rests on a comparison with the criminality of the general
population, it is, of course, faulty, because those punished for
larceny are a specially selected group of property criminals, while
the vast majority of the rest of the population never has and
never will commit a serious property offense. If such a com-
parison is invalid, the only data available on the effect of con-
viction and punishment on larceny offenders is that half are
subsequently reconvicted. It cannot be said that this percentage
is just as high as it would have been if these violators had not
been previously convicted, because they are part of a group that
might otherwise have experienced 80 percent recidivism. It also
cannot be said that this percentage is lower than it would other-
wise be, since we cannot assume that all of those who commit
larceny once will do so again, or that all of those who have
been apprehended for larceny once will be caught in the future
if they do repeat their crime.

The fact that those subjected to punishment are a special
high-risk group of potential future offenders establishes the im-
portance of studying the reactions of this group to punishment
and the threat of punishment, because previously punished per-
sons will be responsible for a far larger per capita amount of
crime than the rest of the population. The share of crime that can
be attributed to previously punished individuals will vary with the
type of crime: previously convicted killers form a very small
percentage of homicide arrests," but previously convicted robbers,
narcotics sellers, and forgers are believed to be responsible for a

1 An FBI study confined to persons with histories of arrest for serious crimes found that
922 out of 194,550 were arrested for homicide in a two-year period. Of those repeat criminals
arrested for homicide, 7 percent, or about 65, had been arrested for homicide once before
in five years, and none more than once. FBI, Uniform Crime Reports—1967, at 35.

98
large share of the total amount of these crimes.? And those who
have been convicted and punished for some major crime are
much more likely than the rest of the population to become
involved in other forms of serious crime later in their careers.®

Understanding the responses of previously punished offenders
to legal threats involves consideration of all of the variables
alluded to in the previous discussion of general deterrence, as
well as recognition of the fact that apprehension and punish-
ment will probably produce important changes in the attitudes
and circumstances of offenders—changes that affect the pro-
pensity of such persons to commit crimes and their suscepti-
bility to the influence of legal threats. Apprehension for com-
mitting a crime may cause offenders to revise upward their esti-
mate of the probability of being apprehended again; it may also
produce a public exposure of the individual's crime that can
either sensitize him to the moral gravity of his act or harden
his resistance to community judgments.

Punishment of offenders may cause a number of changes in
attitude that affect their propensity to commit crimes in the
future. Some of these changes are closely related to the effect
of threats and others are not, but all are important to com-
prehending the role of special deterrence in crime prevention.
Punishment effects that might condition future criminality in-
clude changes in:

a. the offender’s attitude toward threatened consequences;
b. the offender’s attitude toward the threatened behavior;
c. the offender’s attitude toward society ;

d. the offender’s ability to live within the law.

Each of these facets can be affected by the nature, extent, timing,
and social context of a particular punishment.

Changes in Offender’s Attitude Toward Punishment

Commentators have suggested that experience with punish-
ment may produce in its subjects changes in attitudes toward
punishment that both increase and decrease the degree to which
punishment is considered worth avoiding. Lewin, whose research

2 Of the repeat criminals arrested for robbery, 30 percent had been arrested for robbery at
least once before in five years; of the repeat criminals arrested for narcotics crimes, 56 per-
cent had been arrested for narcotics offenses at least one other time previously in five years,
and 36 percent had been arrested two or more times previously in five years; of the repeat
criminals arrested for bogus checks, 48 percent had at least one previous charge, with 27 per-
cent experiencing two or more. See Ibid.

3 Thus, while previously convicted killers do not account for much homicide, those with
serious arrest records of any type are responsible for quite a bit. Using the FBI figures on
homicide, among 194,550 persons with arrest records, 922, or slightly less than 5 percent of
all homicide arrests nationally, could be attributed to this sample during 1966-67. See also,
Willit, Criminal on the Road, at 214-20, wherein serious traffic offenders are found to have
much higher than average involvement with nontraffic offenses.

99
involved the effects of punishment on children, found important
changes in each direction. He concluded that,

When punishment has been previously administered,
the negative valence of the punishment (to its subject)
is very strong.*

One reason for this effect is that punishment leads to an “increase
in the degree of reality of the punishment,” * by acquainting the
subject with the sting of unpleasantness at first hand. This per-
sonal experience brings the unpleasantness of punishment much
closer to the potential offender than threats which communicate
a threat of unpleasantness that is “farther off” psychologically.
But Lewin also states that the experience of punishment makes
subjects “become callous to the punishment and thus less sensi-
tive to threats.” ©

In part, such an effect could be the result of the prospect of a
particular punishment losing its uniqueness to the already pun-
ished subject. This would be particularly true when punishment
assumes most of its negative value from censure, because once
a person has lost a considerable amount of standing in the com-
munity, the threat of future censure conveys little by way of
further deprivation. In some cases, exposure to censure will
cause an individual to react defensively by rejecting the values
of the group that rebukes him, as when Lewin predicts that
“severe punishment is apt to lead to a revolution in the child’s
ideology.” * Thus, while punishment may operate both to increase
and decrease the amount of anxiety its subjects experience about
future punishment, it seems more likely that increased anxiety
about future punishment would be dominant when highly social-
ized persons are subjected to minor punishments, and more
likely that callousness to punishment would assume a major role
in the punishment of less socialized persons and in the adminis-
tration of penal measures that depend on feelings of strong
social loyalties or fear of the unknown.

Research into the effects of punishment on attitudes toward
the unpleasantness of threatened consequences is not plentiful.
Cameron found that often-punished “professional” shoplifters
appear to take the stigma of arrest and the prospect of conviction
in stride, while amateurs apprehended for the first time appear
to display new awareness of the seriousness of being caught.?
In some part, this greater sensitivity on the part of less experi-

* Lewin, Reward and Punishment, in A Dynamic Theory of Personality—Selected Papers of
Kurt Lewin, 1951, at 135.

5 Ibid., at 162.

° Ibid., at 185.

7 Ibid., at 138.

® Cameron, The Booster and the Snitch (1964), at 163-65.

100
enced “snitches” may be a function of higher degrees of sociali-
zation, and to some extent it may be the result of having more
to lose in the way of community standing. Once apprehended, the
snitch who is set free without formal charges almost invariably
refrains from future shoplifting.?

One opinion study sought to determine whether prison inmates
were more or less sensitive to variations in the threat of censure
than college students, when making predictions about whether
a hypothetical bank teller would embezzle funds.'® Sensitivity to
the risk of censure and other factors were measured by the
variations in prediction that occurred when the risk of appre-
hension and other conditions were altered. Strangely, the study
appears to give some support both to the theory that prisoners
are more sensitive to the risk of censure-bearing punishment
than to other factors and to the theory that the prisoners have
less “ethical risk sensitivity” than college students; prisoners
showed more sensitivity to variations in what was termed the
risk of censure than to factors such as variations in the amount
that could be gained, but showed less sensitivity to censure in
general than the college students.’ But this cannot be taken as
evidence that experience with prison has an effect one way or
the other on sensitivity to censure, because the prisoners may
have had different value-orientations than the college students
long before they ever went to prison. Moreover, the prisoners
may have been more interested in projecting a socially accept-
able image when filling out the questionnaire than in divulging
the factors which influence their personal conduct when con-
fronted with the opportunity to commit erimes.*?

Changes in Attitude Toward the Threatened Behavior

When punishment is considered as a possibility by members
of a threatened audience who have not experienced it, it is
psychologically “farther off than [the] desired goal” ** attainable
by committing the threatened offense. Once punishment is ad-
ministered, the increased reality of the punishment experience
may lead to changes in the subject’s evaluation of the behavior
that led to his discomfort. The experience of punishment may
act, in Professor Andenaes’ words, “as a moral eye-opener,”
bringing home to the offender the fact that his behavior is con-
sidered seriously wrong and thereby reducing the probability

9? Ibid., at 151.

10 Rettig, Ethical Risk Sensitivity in Male Prisoners, 4 Brit. J. Crim. 582 (1964).

1 Ibid., at 587-88.

12 This possibility is discussed in detail in Hawkins and Zimring, Deterrence and Survey
Research (unpublished 1968), at 43-44.

13 Lewin, supra note 4, at 163.

101
that the offense will be repeated. Failing moral revelation, the
experience of punishment may make clearer to the offender the
disadvantages that accompany the offense and by association
give him more unpleasant feelings about the behavior, particu-
larly if the punishment occurs soon after commission of the
threatened behavior. On the other hand, the experience of pun-
ishment may cause offenders to experience conflict about the wis-
dom of their behavior and may lead the individual to rationalize
that, after all, the opportunity to commit the threatened act was
worth the punishment. In the process of such rationalization, the
threatened behavior may acquire a value to subjects higher than
it had prior to punishment.!*

This tendency to revalue behavior that has led to discomfort,
in order to rationalize decisions, has been noted in laboratory
experiments where the measure of sacrifice or discomfort was
small: in one case, girls who were forced to undergo a protracted
initiation in order to hear a dull sex lecture rated the lecture
more interesting than girls who were subjected only to a mild
“initiation.” * It is not clear whether more severe punishment
would lead to a greater degree of the “it was worth it” effect
or whether the whole process of rationalization would break
down where the unpleasantness suffered is too great to permit a
realistic judgment that the act could be worth the punishment.
Moreover, because those caught and punished for committing
offenses know that apprehension is uncertain, punished offend-
ers can always find ways to rationalize without changing their
opinion of the value of the threatened behavior: the offender can
say to himself, “I shouldn’t have been caught”’—or, more ominous-
ly, “Next time I won’t be caught”’—and continue to see his initial
offense as justified because the unpleasantness suffered was sim-
ply the result of bad luck.

In the criminal process, present penalty structures are high
enough so that it would be difficult for most punished offenders
to conclude that criminal conduct was worth the price they paid,
and it seems reasonable to suppose that punishment will create
negative associations with particular forms of threatened be-
havior, and with lawbreaking as a general concept, that out-
weigh any tendency to view the pleasures of criminal conduct
through rose-colored glasses. However, it is not clear that punish-
ment will lead to significant changes in attitude in all cases. The
prospects for deflating the attractiveness of criminal alterna-
tives are probably best when the consequences of apprehension

14 See generally Festinger and Aronson, The Arousal and Reduction of Dissonance in Social
Contexts, in Cartwright and Zander, ed., Group Dynamics, Research and Theory (1960);
Cohen and Brehm, Explorations in Cognitive Dissonance (1964).

15 Ibid.

102
activate latent moral judgments in the offender (as in the case
of amateur shoplifters) and where drives toward the threatened
behavior are initially weak. When the drive to commit an offense
is strong and the reality of punishment does not bring the of-
fender’s moral sensibilities to bear on his criminal propensities,
the likelihood of attitude-change is lessened.

Changes in Socialization

When the imposition of consequences for violation of a legal
threat is an important event in a man’s life, it may lead to
changes in his attitudes toward society that influence his future
behavior. This possibility is recognized by those who administer
penal facilities and has led to the advocacy of rehabilitation as
one goal of correctional administration. The theory of rehabili-
tation suggests that some forms of threatened consequences af-
ford the opportunity to reorient the values of their subjects
by discouraging commitment to antisocial values and encourag-
ing loyalty to prevailing social norms. The notion of rehabilita-
tion as a result of punishment differs from narrower theories
of punishment as a means of discouraging criminal conduct by
associating unpleasantness with criminality in the mind of the
offender, because rehabilitation also involves the transmission
of positive social norms, which would tend to reduce criminality
independent of the negative value of punishment.

It is also possible that the experience of punishment will lead
to less favorable attitudes toward social norms on the part of
punished offenders. The punishment process often generates hos-
tility on the part of its subjects and creates some pressure to
reject prevailing norms in order to protect self-esteem when
punishment has conveyed a rejection of the offender by the
social order. And since prison subjects the offender to constant
association with others who have antisocial values and criminal
skills, the net result of the punishment experience on an of-
fender’s attitudes may be an increased identification with de-
viant values. Isolating the factors that determine whether ex-
perience with punishment will produce greater or lesser loyalty
toward prevailing social norms is a task beyond the scope of
this discussion. It is sufficient for present purposes to note that
either the rehabilitation or further alienation of punished offend-
ers will have a significant impact on their future conduct.

Changes in Ability To Function

Some types of threatened consequences will affect the ability
of offenders to live comfortably within the law. Social stigma
and punishment-induced feelings of dependency can limit the

103
opportunities available to punished offenders and impair the
capacity of offenders to take advantage of such opportunities as
remain. In other cases, job training and the treatment of physcial
and mental conditions in a prison setting can make the punished
offender more capable of functioning effectively in society than
he had been prior to serving his sentence. If propensity to commit
crimes is related to opportunities to obtain gratification through
legitimate means, the changes in status and ability that different
punishment regimes may produce will affect recidivism rates
among punished offenders.

Many of the changes that punishment can produce are not
centrally related to the effect of threats on future conduct. If
rehabilitation reduces recividism because many former offenders
wish to conform to prevailing social norms, it would be unwise to
attribute this improvement wholly to greater sensitivity to threats
or to expect that punishment per se is responsible for the decrease
in subsequent criminality. Other possible punishment effects on
future criminality, such as changes in the offender’s attitude to-
ward threatened consequences, will be the result of changes in
patterns of threat response.

The only method of testing punishment as a special deterrent
is the study of future criminality among punished offenders, but
any such test will show the results not only of special deterrence
but of the many other positive and negative influences that
punishment might have on future criminality, because a partic-
ular form of punishment will produce a mix of effects. It will
be possible to test higher versus lower prison sentences or prison
sentences versus probation by randomly varying the penalties
given to convicted offenders, or simulating this experiment by
using base expectancy tables. But if prison sentences reduce sub-
sequent criminality more than probation, does this mean that
prison is a more effective special deterrent, or that it creates
more opportunity for rehabilitation? If prison and probation
lead to the same results, does this mean that imprisonment does
not perform as a marginal special deterrent, or that the negative
aspects of imprisonment balance off its greater special deterrent
force? How much of the difference in recidivism, if there is any,
is attributable to changes in attitude toward the threatened
behavior and how much to changes in attitude toward punish-
ment? At times it will be possible to experimentally vary one or
two separate factors of possible importance in predicting future
criminality, but when large variations in punishment policy are
tested (as when sentences involving penal confinement are com-
pared with those that do not), it will be difficult indeed to un-

104
ravel the different types of variables that might explain dif-
ferences in recidivism.

Yet, in policy terms, the problem of unraveling the various
effects of different types of punishment is more apparent than
real, for the crucial question is always whether a program has a
desirable net effect on future criminality, not whether some
aspects of punishment that are inseparably bound up with forces
pulling in the opposite direction would tend to reduce crime if
they could be isolated. In cases where punishment effects can be
isolated by varying penal sentences, they can be tested individu-
ally to determine whether their presence has any effect on
recidivism. If a set of punishment components cannot be sep-
arated, the only significant question is the effect of the set as
a whole, rather than the effect of certain parts, and variations
in punishment will provide the opportunity for such a test.

In one respect, the study of special effects is easier than that
of general deterrence because different groups can be exposed to
different treatments, using methods that control for variations
other than punishment that might undermine the significance of
results. Once punishment-groups have been selected and treated,
keeping track of their criminal records will provide much more
reliable guidance to the student of special effects than can be
obtained by studying general deterrence through analysis of
movements in the total crime rate.

And while it is difficult to institute more than one legal threat
for a particular behavior in a jurisdiction, a number of different
types of threatened consequences can be imposed on convicted
offenders in ways that generate reliable knowledge about the
relative effects of different types of punishment. While such
studies cannot discern the total amount of special deterrence
attributable to a particular punishment, they can tell us about
the much more important marginal effects that particular poli-
cies might have.

The major barrier to controlled study of different punishment
regimes, other than inertia, is the objection that such methods
require treating groups of similar offenders in different ways,
a practice that many see as ethically obnoxious.’ Yet detailed
study of present patterns of sentencing in our criminal courts
supports the conclusion that the variation in punishments meted
out for the same offenses and same types of offenders by different
judges is already extreme '” and produces many of the costs asso-
ciated with unequal treatment, without any of the benefits of
controlled study.

18 For a full discussion of this issue, see Morris, Impediments to Penal Reform, 33 U. Chi.

L. Rev. 627 (1966), at 645-55.
17 See Green, Judicial Attitudes in Sentencing, at 67-71.

105
A substantial number of studies dealing with the subsequent
criminal records of released prisoners bear on the marginal
special effects of increased severity of sentences. Simple recidi-
vism comparisons have tended to show that among persons
convicted of the same crimes, those treated more leniently have
lower rates of subsequent criminality than those punished more
severely,’ but when such comparisons are controlled for differ-
ences in the offender groups other than type of punishment,

the dominant feature of the results is that overall dif-
ferences between various methods of treatment are
small or nonexistent.??

If the overall rates of recidivism do not vary with severity of
punishment, some prediction studies provide evidence that this
apparent lack of difference could be the result of more severe
punishment producing significant positive effects in some types of
offenders and significant negative effects on others that tend to
balance out.?°

But it is one thing to say that severe punishment is ineffective
as a marginal special deterrent and quite another to suggest
that punishment does not produce special deterrent effects. The
studies of increases or decreases in length of penal confinement
have no impact on the argument that the experience of appre-
hension, conviction, and punishment has a powerful restraining
influence on those who encounter them. An FBI study of more
than 17,000 persons released from custody in 1963 found that
while more lenient penalties were associated with lower rates
of rearrest, those who had been acquitted or had charges against
them dismissed experienced a 91 percent rearrest rate within
four years, compared to a rearrest rate of less than 60 percent
for persons convicted and punished.?* This finding is more sug-
gestive than conclusive, because the FBI study did not relate
release and rearrest data to specific types of crime or age and
because of possible bias in the sample. But this type of finding,
and the shoplifting study alluded to earlier,?> suggest that ap-
prehension and conviction may substantially reduce future
criminality.

Taken as a whole, studies of recidivism establish that those
subjected to punishment for major crimes commit many more
crimes after their release than other groups in the population,
but fewer perhaps than they would if they had not been caught.

18 See, e.g., California Department of Corrections, Parole Outcome and Time Served for
First Releases Committed for Robbery and Burglary—1965 Releases (unpublished, 1968).

1 Andenaes, Does Punishment Deter Crime?, 11 Crim. L. Quarterly 76 (1968), at 93.

20 See Warren, The Case for Differential Treatment of Delinquents, Annals of the American
Academy of Political and Social Science 47 (January 1969), at 48.

21 FBI Uniform Crime Reports—1967, at 817.

22 Cameron, supra note 8.

106
Those studies that suggest the lack of a marginal special deter-
rent effect associated with longer prison sentences provide a
basis for official experiments with reducing punishment for
major crimes. When findings on the lack of marginal special
effects are integrated with parallel studies on the marginal
general effects of increases in threatened punishment for major
crime, they make an impressive case for the reduction of the
present scale of major criminal sanctions in the United States.

107
Concluding Reflections

One of many dangers in a survey treatment of a topic such
as deterrence is the lack of focus that is likely to accompany the
attempt to provide broad coverage. After touching on some of
the many factors that condition the effectiveness of legal threats,
it would perhaps be appropriate to search out basic themes which
underlie the preceding analysis.

The first such theme is an overwhelming sense of the speci-
ficity of findings in research in deterrence. A particular legal
threat involves a mix of factors—communication, enforcement,
type and extent of threatened consequences, type of behavior,
social attitudes—and the mix will be different for different threats.
Since it is difficult to know precisely why a policy does or does
not achieve results, it is dangerous to generalize from a partic-
ular finding to propositions about the marginal deterrent effects
of other types of threat, or of the same threat with different
levels of enforcement or communication. It will thus require a
large number of studies of different types of threat before plausi-
ble generalizations about marginal deterrence emerge from re-
search.

A second recurrent theme is the difference in texture be-
tween issues of absolute and marginal deterrence. The larger
concept of deterrence encountered when discussing whether and
how threatening behavior will reduce its rate is more complex and
more difficult to study than the narrow issue of whether partic-
ular changes in the conditions of a legal threat produce results
worth their cost. In the foreseeable future, research opportunities
in deterrence will be spread unevenly over the range of issues—
we will have ample opportunity to study the short-range mar-
ginal effects of modest changes in threat composition that nor-
mally take place in the process of administering the criminal law,
but little opportunity to study large jumps, up and down the scale
of punishment, or the results of decriminalizing behavior that
society has regarded as dangerous. Since generalization is difficult,
uneven research opportunities make for uneven progress in our
understanding of deterrence. Future research will reveal more
about marginal deterrent effects than about deterrence, more
about small changes than large changes. We will be putting this
knowledge to use long before basic issues about the effects of
threat and punishment have approached a resolution.

108
A program of research in the coming years should include
experiments in different methods of threat communication; the
evaluation of procedures such as intensive police patrol and com-
puter monitoring that increase probability of apprehension; the
experimental reduction of penalties for major crime; and the
experimental introduction of new ways of treating offenders in
areas such as traffic crime, where the political climate permits
changes in the kind as well as extent of threatened consequences.
As more becomes known, experimental variations in the condi-
tions of threat and punishment can be increased if advances
in knowledge make public and official attitudes about crime con-
trol more flexible. But the relationship between increased knowl-
edge and changes in official attitudes is by no means automatic.
A communication gap already exists between the level of in-
formation available about the effects of legal threats and of-
ficial beliefs. And any real progress in penal policy will depend
as much on closing this gap as on our willingness and ability
to increase knowledge.

vr U.S. GOVERNMENT PRINTING OFFICE: 1971 0—388-686

109
 
C0293690k8
 

DEPARTMENT OF
HEALTH, EDUCATION, AND WELFARE

PUBLIC HEALTH SERVICE
HEALTH SERVICES AND MENTAL HEALTH ADMINISTRATION
NATIONAL INSTITUTE OF MENTAL HEALTH
5454 WISCONSIN AVENUE

CHEVY CHASE. MARYLAND 20015 POSTAGE AND FEES PAI
U.S. DEPARTMENT OF H.E

OFFICIAL BUSINESS
Penalty for private use $300.

 

Public Health Service Publication No. 2056

U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE
PUBLIC HEALTH SERVICE-HEALTH SERVICES AND MENTAL HEALTH ADMINISTRATION

NATIONAL INSTITUTE OF MENTAL HEALTH