m

The Behavioral

Persuasion in Politics
Eiilau
by HeinzUNIVERSITY
'STANFORD

mm^

^pl

if.

f^

w'^A^
'

C.3

iJtl^KBiM^r
ft^ifr

ii.^^Hwir

UNIVERSITY
OF FLORIDA

LIBRARIES

COLLEGE LIBRARY

THE
behavi oral

'persuasion
in politics
HEINZ EULAU
STANFORD UNIVERSITY

RANDOM HOUSE
New

York

TO

Harold D. Lasswell, Persuader

FIRST PRINTING

Copyright, 196^, by

Random House,

Inc.

rights reserved under International and Pan-American

Copyright Conventions. Published in New York by Random
House, Inc., and in Toronto, Canada, by Random House of
Canada, Limited.

All

Library of Congress Catalog Card

Manufactured

The

in the

Number:

62-21^^$

United States of America by

Colonial Press Inc., Clinton, Mass.

PREFACE

New ways of saying and doing things have

some and

terrified others.

to death a sage

named

The

citizens of

always tempted
Athens once put

Socrates, charging "corruption of

the young," and "neglect of the gods

when

the city

wor-

and the practice of religious novelties." Both charges

were, and remain, somewhat obscure. It seems that Socrates, among other crimes, had introduced a new, critical
ships,

method

into philosophical speculation.

among the

And

politics

was

topics he considered.

In politics, methods of study and objectives of inquiry

are particularly subject to controversy.

The

stakes

are

probably higher in politics than in any other field of human effort, and today that even includes religion. New
approaches and objectives are likely to arouse passions.
Indeed, so much so that some people believe disinterested
political inquiry to be impossible. Of course, the innovator is no longer forced to drink hemlock, at least not
in societies that take pride in free investigation. Instead

he becomes involved in interminable debate with

critics.

his

Preface

[vi]

Perhaps

by

outset

it is

best, therefore, to state

describing

what

this

book

is

my

case at the very

not.

not a tract in political polemics. There has been,

should be, a great deal of dispute about what I
shall call the behavioral persuasion in politics, but I mean

This

is

as there

it. More often than not, the debate has been

about the possibility of having a value-free science of
poHtics. The question is not unique to the study of politics, and though I think it is important, I have never found
the polemics in its wake very profitable.
Nor is this book a programmatic statement, an agenda
of things to do and how to do them. Although programming is an important element in the strategy of scientific
research, the purely programmatic phase of the behavioral

to ignore

persuasion in politics has passed. For the interested reader

I shall list

some

early programmatic statements in the bib-

liography to the Introduction.

Nor

book

this

is

a critical

inventory of political behavior

growing and changing area of ininventories are urgently needed. But since

research. In a rapidly
vestigation,

there

a great division of labor in research in political

is

behavior, the preparation of adequate inventories requires

the collaboration of several specialists. Again, the interested reader will find listings of useful inventories in the
bibliographies attached to the different chapters.
Finally, this

not a text about political behavior.

is

behavioral persuasion in the study of politics

is

The
quite

young, and though there has been a great deal of behavioral research on politics in recent years, there is not
yet enough. Various topics have been unevenly treated,
and the findings are very tentative. I must confess to
prejudice against texts in a field that is still so wide open.
Texts often give the impression of knowledge where, in
fact, it

is

lacking.

This book
tionary:

is,

hope, an essay, as defined

by

the dic-

"a literary composition, analytical or interpre-

tative, dealing

with

its

subject

or personal standpoint."

It

from

more or

less

limited

does not pretend to be wise

or erudite, nor to be systematic, comprehensive, or ex-

Preface

[vii]

have not reviewed the literature nor consulted

files, though I occasionally reread
on the subject. I wrote this book
a personal document, off the cuff, setting down what

haustive. I

my own voluminous
my earlier comments
as

came

to

mind

litical life, in

as it

derived from

my

experiences in po-

reading, in teaching, and in research.

Although there are no quotations or citations, I am

deeply indebted to the many scholars who make this intellectual venture possible. The behavioral study of politics

is

a truly

collaborative

effort.

The

bibliographies

which follow each chapter are meant to acknowledge this

debt. If I do not identify the particular sources in the
text, neither do I presume to originaHty. Much of what
I am merely trying to
and personal a way as I can.
However, the one source which I must acknowledge
is Harold Lasswell, to whom I have dedicated this essay.
His work has been a continuing source of stimulation and
suggestion. But I am not a disciple. Indeed, one of the
admirable things about Lasswell's teaching is that it makes
discipleship impossible. Perhaps the greatest compliment
I can pay him is to say that he has been, and still is, the
gadfly of American political science, the one Young Turk
who can be counted on to stay that way. His influence is
I

write here has been said before.

say

it

in as simple

pervasive.

This book

reflects

my

didactic approach to

political

about teaching, a good deal

about research. Yet the three middle chapters correspond
in title to the three undergraduate courses I offered at
Stanford University between 1958 and 1961. Of course,
the material covered in a college course over a period of
several months cannot be repeated in a chapter of only
a few thousand words. For both individual study and

behavior,

say very

little

class discussion these courses

of empirical research.

It

depended heavily on reports

has always seemed to

me

that

by which knowledge about politics, or anything else, is created is as important from the teaching
point of view as such knowledge itself. Students, by being
the process

allowed to read and study empirical research reports,

may

[viii]

Preface

how very little is as yet known about political behow difficult it is to come by reliable political
knowledge, and how tricky the problems and dilemmas of
learn

havior,

a science of politics really are.

This, in short, presents an orientation to the study of

politics that, in

some

respects, has

come

to be a

way

of

life.

Stanford, California

Heinz Eulau

CONTENTS

Preface
Introduction

THE ROOT

IS

MAN

BEHAVIORAL APPROACHES

Chapter One

Units of Analysis

Chapter

13

14

Levels of Analysis

19

Theory and Research

24

Behavioral Methods

31

THE SOCIAL MATRIX

Two

38

Role as a Basic Unit

39

The

46

Groups
The Horizontal Dimension
Vertical

Dimension

Strata

Chapter Three

THE CULTURAL CONTEXT

54
62

Cultural Patterns

63

Cultural Orientations

68

Contents

[x]

Cultural Changes

74

Political Culture

79

Chapter Four

THE PERSONAL

BASIS

85

Tersonality in Politics

86

Opinions and Attitudes

90

Values and Personality

95

Role and Personality

Chapter Five

100

BEHAVIORAL DILEMMAS

no

Observational Dilemmas

Macro-Micro Dilemmas

Dynamic Dilemmas
Epilogue

Index

THE GOAL

IS

MAN

112
.

123

127
133

139

THE BEHAVIORAL
PERSUASION IN POLITICS

INTRODUCTION

THE ROOT

The

root

is

man.

don't think

IS

it is

MAN

possible to say any-

man without
manhis acts, goals,

thing meaningful about the governance of

talking about the political behavior of

commitments, and values. Man has

and empires, created customs and institutions,
invented symbols and constitutions, made wars, revolutions, and peace. Politics is the study of why man finds
it necessary or desirable to build government, of how he
adapts government to his changing needs or demands, of
how and why he decides on public policies. Politics is
concerned with the conditions and consequences of human
drives, feelings, beliefs,

built nations

action.

A study of politics which leaves man out of

is

a rather barren politics.

that he can consider his

them by

Yet such

own

is

its

man

creations without measuring

himself. Political science has studied

ideas, values, customs,

equations

the propensity of

political

symbols, institutions, processes, and

policies without reference to their creators for a long time,

but the cost has been high. I do not want to belabor this
point. I mention it only because the simple question I
want to ask Why do people behave politically as they do?

Introduction

[4]

seems to

have explosive consequences for the study of

politics.

what

Just

question

is

many

political behavior?

times,

by

have been asked the

students as well as

by

colleagues,

and out of political science. Is it a field of study, a

method, or an approach? If it is a field, it must have content and boundaries. If it is a method, it must have rules.
If it is an approach, it must have direction. I cannot say
that it is one or the other. It is none of them alone, and it
is not all of them together. This leaves the questioner
confused, perhaps irritated, even hostile. It is not considin

ered a virtue to tolerate ambiguity. I would certainly not

argue that ambiguity is preferable to clarity, though it
may be preferable to false or easy answers.

The

difficulty begins

with

definitions

commit and

and reveal

his orientation.

definitions. If

constrain.

They

They

are

taken seriously,

orient their user

embedded

in his con-

cepts and his theorizing, are a source of sense, but also of

it is with "political behavior." For some
have asked my students to define politics.
Politics, they tell me, has something to do w^ith government, power, policy, influence, decision making, conflict,
or even "authoritative allocation of values." I cannot but
marvel at such ingenuity. But when I ask just what people
do when they act in ways to which these concepts presumably refer, there is a perplexing silence.

nonsense. So
years now,

wonder why

own

of our

this is so.

when I ask what

The students will

Evidently,

However,

sophistication.

we
this

are the victims

is

not the case

when they practice religion.

me that a man is religious when he

people do
tell

prays, attends mass, sings

hymns,

listens to

sermons, im-

merses himself in baptismal water, senses the presence of

divine guidance, abides

by

believes in immortality.

And

nomics:
ulates,

man

suggest that
rules

produces, buys,

consumes, and so on.

the

Ten Commandments,

there
sells,

Not

is

or

no trouble with eco-

exchanges, invests, specso with politics.

what makes man's behavior

political

When

is

that he

and obeys, persuades and compromises, promises

The Root

Is

Man

[5]

and bargains, coerces and represents,

and

fights

fears,

my

students are baffled.

Certainly these verbs do not define pohtics. But they

do refer to those of man's acts that are at the core of what

we

study when we talk about politics. And there are many

more. If human behavior is the root of politics, they are
more useful in studying political things than nouns like
authority, power, conflict, allocation, or government. It
seems to me that behavior comes first: ruling before gov-/
ernment, obeying before authority, voting before decision,

demanding before value, fearing before sanction, coercing

before power, persuading before influence, fighting before
conflict, believing

But such

is

political study,
first

and then

As we

before ideology.

we

the enterprise, whether

or political science, that

sense, rather

define politics, so

than sense

we behave

first

call it politics,

we must

define

and then

define.

our

politically, for

definitions of politics are themselves evidence of political

behavior.
serve and
that

man

They

determine, at least in part, what

how we

explain

it.

It

would be

we

silly to

ob-

deny

in politics, being a defining animal, has various

definitions of politics. I

am merely

pleading that in seek-

our definitions of politics, we

turn to what men do as they behave politically and why
they do it. Definitions unrelated to the behavior of man,*
in politics as in any other area of human activity, have nol
ing, clarifying, or refining

content.

The

is concerned with\
and the meanings he attaches to
his behavior. Politics asks about ancient traditions and
grandiose designs, about complex systems and intricate
processes, about fearful atrocities and superb achievements.
But as an eminent physicist once remarked, it is a subject
"more difficult than physics." The physical scientist seems
to have one great advantage over the political scientist:
whatever meanings he may give his objects of study, they
do not talk back to him. Atoms, neutrons, or electrons
do not care how they are defined; political actors do mind.

behavioral persuasion in politics

what man does

politically

Introduction

[6]

This
is

is

precisely

why

a political science that ignores

However, the

fact that

men

give meanings to their be-

havior need not be a handicap.

men

On

the contrary,

what

say about themselves and others represents an in-

source of information about behavior.

finitely rich

'

man

necessarily a very incomplete science of politics.

And

the meanings that people give to politics are appropriate

data for scientific analysis because people behave in terms

of these meanings. These meanings do not provide the

scientific

observer with the kind of definitions he needs

in order to

proceed with

his investigation.

He

must de-

velop his own. But, whatever definition of politics the

it cannot be altogether arbitrary.
be "meaningful" in terms of the meanings
that men give to their political behavior. In the language
of science, definitions must be operational. No matter how
concrete or abstract conceptually, they must be relevant
empirically. The meanings that political actors, consciously
or unconsciously, attribute to their own behavior are of
interest to the political scientist because they provide a
partial explanation of the motives for that behavior.
Defining political behavior is a delicate problem, partly
because people in politics define and interpret what they
do differently, and partly because political scientists are
by no means agreed on what they mean when they say
that they are studying political behavior. One way to avoid

political scientist adopts,

It

must

itself

the dilemma

is

to ignore

it,

offer a definition as succinctly

and go on from there. The researcher must do

this. He cannot get entangled in problems of definition if
he hopes to come up with a piece of research. The only
as possible,

criterion

is

that his definition suit his research objectives,

and
\

The

probably all that can be reasonably expected.

problem of definition must ultimately be solved by

this is

empirical research.

rigorous approach to definitions alone will not spur

progress in the study of politics. In fact

it

might well

stifle

new approach which requires not definitional

rigor so much as new categories and concepts with which
to explore new terrain. In returning to the behavior of
in infancy a

The Root

man

Is

Man

root of

as the

opened up new

[7]
politics,

the behavioral persuasion has

possibilities in the

study of

politics. If this

more problems than it has solved, including

those of definition, it is more of a challenge than a defeat.

has created

is

The return to the behavior of man as the root of politics

a new beginning. For in dealing with the conditions and

consequences of man's political conduct, the behavioral

by modern modes of

persuasion represents an attempt,

knowledge begun

analysis, to fulfill the quest for political

by

the classical political theorists.

suasion in politics, as

understand

bases of man's political experience in

of the past found nurture

rists

makes the

The

behavioral per-

is

a return to the

it,

which the great theo-

and sustenance.

so-called classic theories great are their

What
some-

times explicit, sometimes implicit assumptions about hu-

man

nature in politics.

The

theoretical constructions of

the polity found in the classics are "peopled systems,'^

model communities based on some notion of how men

behave politically as they do and why, in addition to fre-

how men ought to behave

and what the polity should look like. Of
course, the psychology, sociology, and even anthropology
involved in these images is, from the contemporary perspective, primitive, underdeveloped, and often mistaken,
r certainly do not advise a reading of the classics for the
purpose of learning about political behavior, not even of
the empirical Machiavelli. But this is not the point. The
quently being prescriptions for

in the polity

point

is

that classical political theory, as the

havioral persuasion in politics, has at

its

modern

of man, even if, counter to the behavioral persuasion,

predominantly prescriptive rather than descriptive.
It

may seem

be-

base the conduct

it is

startling that the behavioral persuasion is

a continuation of the classical tradition of political inquiry.

On

the face of things, the discontinuities between the an-

cient and the

modern approaches seem more significant

Modern modes of thought, criteria

than the continuities.

of validation, and methods of investigation are so radically

different that the link between classical political theory

and the behavioral persuasion would seem to be rather

Introduction

[8]

deal depends on what one means by

one means continued textual exegesis of the
classics as if they were sacred writings, the behavioral
persuasion does, in fact, make a radical break with political theory. But if by continuity one means, as I think
one should, the application of modes of thought and tech-

tenuous.

good

continuity. If

own

niques of inquiry appropriate to one's

time to the

problems of the time, then the behavioral persuasion is a direct and genuine descendant of the classical
tradition. The classical theorists, from Plato to Mill and
beyond, in building their models of the polity, sought to
bring to political inquiry the best conceptual and techpolitical

nical tools at their disposal.

who

adapt the

new

The modern

theories, methods,

political scientists

and techniques of

behavioral science to political analysis are in the tradition

of the classical political theorists.

Concern with the pohtical behavior of man has posed

the following questions: Does

it encourage dealing in trivia

while the "really important" problems of politics are neg-

Does

not lead to cultivating areas of research

is easy, regardless of the "significance"
of the political problem? What does political behavior re-

lected?

where

it

access to data

search contribute to the solution of the great issues of

politics?

These questions are not so much unwarranted

they are misdirected.

It is

much

perfectly true that

as

be-

on politics is concerned with simple

But a simple question is not necessarily a simple
matter. The line between asking simple but important
questions and asking trivial ones is often very narrow.
Moreover, there is much confusion as to just what "sig-

havioral research
questions.

nificance" means. If

it is

only defined in terms of the so-

called great issues, then politics

become

as

science

is

likely to

the handmaiden of policy, for better or for worse

(see the Epilogue).

issues criterion

But whatever other relevance the great-

may

define significance

have,

it

does not necessarily help to

from the point of view of

political

be-

havior research.

question

From

may

be simple, then, without being

the standpoint of empirical research,

it is

trivial.

trivial

only

The Root

Is

Man

[9]

does not yield answers that "significantly" add to

knowledge. In fact, many answers given to questions,
if

it

simple or complex, from the point of view of policy, do

not add to knowledge. By knowledge I mean, of course,
a set of verified statements about reality. If the statements
hang together and do not contradict each other, we have
knowledge. This is as true of politics as of any other area
of human endeavor.

Now,
is

that

the interesting thing about this chain of reasoning

it is

question

based on hindsight, although

trivial,

we

speak as

if

we had

when we

foresight.

call a

This

A trivial question

is

so

one
we assume has been answered already, or, if it has not been
trivial
answered, that it can be answered easily enough.
question, it is implied, is one that every fool can answer.
simply predict, from hindsight, that a trivial question
will not significantly contribute to new knowledge.
However this approach to knowledge does not lead anywhere. Knowledge is the process of knowing, always undergoing change. If this is so, we cannot call a question
trivial, for we cannot know whether it is trivial or not
until it has been asked and answered. Triviality is not a
matter of the kind of questions we ask, but of the answers
we get. Only after a question has been answered can we
say that it has been trivial. Without an answer, one can
call a question trivial only if one assumes that everything
worthy of being known is known already. It is sometimes
advantageous to assume that we don't know what we
think we know. In other words, it may sometimes be advisable to ask old questions as if they were fresh.
It is the function of science to understand and interpret
science of politics which
the world, not to change it.
deserves its name must build from the bottom up by
asking simple questions that can, in principle, be answered;
it cannot be built from the top down by asking questions
that, one has reason to suspect, cannot be answered at all,
at least not by the methods of science. An empirical discipline is built by the slow, modest, and piecemeal cumula- ^
tion of relevant theories and data. The great issues of

because knowledge

is

transmitted.

We

is

Introduction

lo]

as the conditions and consequences of freeor authority, are admittedly significant topics,
but they are topics compounded with a strong dose of
metaphysical discourse. I don't think that they are beyond
the reach of behavioral investigation, but before they can
be tackled, the groundwork must be laid.
There is little glory to be had in the patient analysis of
mass political behavior (and elite behavior is, indeed,
much more glamorous as a topic of inquiry). But the hundreds of studies of electoral behavior, some good, some
not, accumulated in the last sixty years, allow us to make
some statements about democracy that are true with a
reasonably high degree of probability, certainly higher
than if these studies had not been made. This is all one
can hope for in the present stage of political inquiry, an
early stage in spite of the great thinkers who have influenced our notions of significance. But we cannot decide
whether an Aristotle's concerns should be our concerns by
appealing to Aristotle.
can decide this only by questioning our own experiences in the world of politics in
which we live. If our experiences lead us back to the
great issues, all to the good; if they do not, little is lost.
In returning to man as the root of politics, the behav-

politics,

dom,

such

justice,

We

ioral persuasion reveals itself as a "radical" orientation in

the study of politics. But

eyed prophets nor blind

its

practitioners are neither wide-

apostles.

They

are self-consciously

sensitive to the difficulties involved in the behavioral study

The way ahead is never clear or straight. One

not always know the destination, and even if in sight,
one may never reach it. But it makes an exciting journey,
if not always a rewarding one.

of

politics.

may

What

the behavioral persuasion challenges in the tradi-

tional study of politics,

if

it

challenges anything,

is

the

comfortable assumption that theory is the same thing as

y knowledge. But theory is only a tool. If it is a tool, like
all tools it tends periodically to wear out and need replacement. The behavioral persuasion in politics is both theoi

retical

and empirical in direction.

Its

radicalism stems

from

Bibliography

[ii]

the conviction that a proposition

on being

tested,

it

may

be worn out when,

can be disproved.

BIBLIOGRAPHY
PROGRAMMATIC STATEMENTS OF THE
BEHAVIORAL PERSUASION
Catlin,

George E. G., Science

York: Knopf,

arid

Method

of Politics.

(New

1927.)

Harold D., The Analysis of Political Behavior: An

Empirical Approach. (New York: Oxford, 1948.)
Merriam, Charles E., New Aspects of Politics. (Chicago: University of Chicago Press, 1925.)

Lasswell,

Rice, Stuart A., Quantitative

Knopf,

Methods

in Politics.

(New

York:

1928.)

COLLECTIONS OF POLITICAL BEHAVIOR RESEARCH

Eulau, Heinz, Eldersveld, Samuel
(eds.), Political Behavior:

(New
Fiszman,

and Janowitz, Morris

J.,
Reader in Theory and Research.

York: Free Press, 1956.)

Joseph R. (ed.), The American

Political

Arena.

(Boston: Little, Brown, 1962.)

Uhner,

S.

Sidney

(ed.).

havior. (Chicago:

Introductory Readings in Political Be-

Rand McNally,

1961.)

Wahlke, John C, and Eulau, Heinz, (eds.). Legislative Behavior: A Reader in Theory and Research. (New York:
Free Press, 1959.)

INVENTORIES OF GENERAL POLITICAL BEHAVIOR

RESEARCH
David, The Study of Political Behavior. (London:
Hutchinson, 1958.)
Ranney, Austin (ed.), Essays on the Behavioral Study of
Politics. (Urbana: University of Illinois Press, 1962.)
Butler,

[i2]

Introduction

Research Frontiers in Politics and Government, Brookings

Lectures,

1955.

(Washington, D.C.:

Brookings Institution,

I955-)

Waldo, Dwight,
Trend Report.

Political
(Paris:

Science in the

UNESCO,

United States:

1956.)

CRITICAL COMMENTARIES ON THE BEHAVIORAL

PERSUASION
The American Science of Politics. (Berkeley:
University of California Press, 1959.)
Storing, Herbert J. (ed.), Essays on the Scientific Study of
Crick, Bernard,

Politics.

(New

York: Holt, Rinehart and Winston,

1962.)

CHAPTER ONE

BEHAVIORAL

APPROACHES

The

behavioral persuasion in politics has more than one

approach, and there are many voices that speak in its

name. But they

study of

man

all

have in

as the

common

commitment

to the

root of things political, or, to put

it

more

technically, to the individual person as the empirical^

unit

of analysis. Therefore I think

it

is

legitimate to

speak of "the behavioral persuasion in politics."

How-

ever these voices do not necessarily speak in harmony.

Behavioral researchers on politics

selves in

many

differ

among them-

respects: in their conception of the nature

of knowledge and

its

relation to reality; in their formula-

tion of the theoretical propositions guiding their investiga-

and tactics of research; in

problems and research sites and, finally,

tions; in the choice of strategies

their selection of

in their appraisal of their

own

of politics they are studying.

role in relation to the

It

world

seems preferable, there-

Behavioral Approaches

[14]

fore, to speak of behavioral approaches

when

comes to

it

technical matters.
I think that most behavioral researchers agree on at least
four major aspects of the behavioral enterprise in politics,

aspects that define the stance and style of those

scribe to the behavioral persuasion.

who

sub-

Each aspect involves

problems on which behavioral practitioners maydiffer, but which bind them precisely because they agree
that the problems require solutions. First there is the
''problem of the most adequate theoretical (as against empirical) units of analysis. The second problem is the level
of analysis on which political behavior research may be
most fruitfully conducted. Third, there is the question of
the proper relationship between theory and research.
Finally, there is the problem of what methodological requirements should be met.
set of

UNITS OF ANALYSIS
The

political

behavior of the individual person

is

the cen-

and crucial empirical datum of the behavioral approaches to politics. This does not mean that research is
tral

restricted to the individual person as the theoretical focus

of investigation. Indeed, most behavioral researchers are

not concerned with the individual political actor as such.

small group, an organization, a community, an elite, a
mass movement, or a nation may be the focus of behavioral inquiry and events, structures, functions, proc-

esses,

or relations

may

serve as categories of behavioral

analysis.

This does not imply that behavioral investigations dealing with units other than the individual actor "reduce"

them

to the level of the person.

when we
and

We mean

different things

say that an individual behaves in certain ways

that a nation behaves in certain ways.

nation's behavior in the

conduct of

its

To

speak of a

foreign relations

quite clearly, to engage in metaphorical license, just as

we speak of the behavior

The use of metaphors for

do when
stocks.

is,

we

of stars, storms, or

the purpose of an-

Units of Analysis

[15]

alogical thinking

is

feasible

and even necessary, but

it

can be dangerous because it may make for a good deal of

distortion. It is only too easy to speak of a group's or nation's behavior as if it were identical with that of the individual person. Consequently, inferences
such practice are likely to be false.

The

on the behavior
and transactions make
he is concerned with de-

political behaviorist concentrates

whose

of individuals

up

drawn from

interactions

collective behavior, even

if

scribing and explaining the actions of groups, organizations, or other large collectivities.

Groups, organizations,

or nations have no independent status apart from the con-

duct of the individuals who are related by behaving towards each other in certain ways. This does not mean that
groups, organizations, or other formations are not "real"

and meaningful units with structural properties and funcown. They certainly are. In fact, the great
bulk of problems interesting the political scientist contions of their

cerns the actions of such groups. But,

from the behavioral

perspective, these collectivities exist and behave the

way

they do only insofar as the people composing them act in

certain ways.
In choosing the individual as his empirical unit of
analysis, the political behaviorist does not deny the reality
of institutions. He merely asserts that institutions do not
and cannot exist physically apart from the persons who
inhabit them.
this

The

political behaviorist

is

likely to stress

point because institutional and behavioral analyses

have, at times, been treated as

each other.

more or

less

the people

if

they were opposed to

are not. Political institutions are never

from the patterns of behavior of

them or the regularities of their acinstitutions can and must be analyzed

different

who

tions. If this

in terms

They

be

create
so,

of the behavior of their molecular units, the

individuals whose relations to each other and behavior

towards each other are more or less rigidly structured.

We

speak of "the opinion of a court," "the decision

of a bureau," or "the action of a legislature" only be-

cause speaking in "institutional language" brings the great

[i6]

Behavioral Approaches

complexity of political life into a manageable perspective.

Behind the opinions, decisions, or actions of institutions
are human beings who decide, who have opinions, and

who

act.

When we
we

say that the legislature makes the law of the

mean

that most people affected by legislabe guided in their own conduct by what
a majority of the legislators have agreed they must do or
not do. The legislators themselves have acted as they did
because they entertained reasonably stable expectations
that most people whose behavior they are called on to
guide will accept their decisions and, through conforming to them, confirm what, in institutional language, we

land

really

tive action will

call

"the authority of the legislature,"

It is

precisely be-

cause the behavioral patterns involved in this set of interare relatively uniform, regular, and
can speak of the legislature as an institution.
Political institutions are behavior systems or systems of
action. Just as they cannot exist apart from the persons
whose behavior brings them into existence, so political behavior cannot exist apart from the network of interper-

personal

relations

stable that

sonal

we

that

relations

we

call

political

sumably, political behavior has

much

institutions.

Pre-

common

with

in

other types of social behavior, but the adjective "political"

warns us that the behavior

cial case.

The

difference

is

in question represents a spein the institutional setting of

the behavior.

An
be

attempt

political in

is

sometimes made to define what seems to

behavior by identifying some special char-

assumed to be inherent in behavior. For instance,

"power"
or "conflict" have been singled out as the characteristic
features of political behavior. But power or conflict relations, though admittedly frequent in politics, are not
acteristic

interpersonal behavior constituting relations of

exclusively political aspects of interpersonal actions.

We

can speak meaningfully of "economic power" or "racial

conflict" without attributing peculiarly political characteristics to them. Otherwise almost all behavior could
be called political and the term would lose whatever dis-

Units of Analysis

[17]

criminatory meaning

may

it

have. This does not

mean

that

power or conflict cannot serve as useful organizing concepts. As such they may enable us to examine institutional
settings for whatever generic features accompany the
patterns of interpersonal relations and behavior to which
they presumably

But

refer.

doubt that they are useful

in specifying the context of political behavior.

Political

seems to be a residual rather than a generic

term. This makes

it

futile to search for characteristic fea-

from an institutional or
environment that shapes and patterns certain

tures of political behavior apart

situational

types of interpersonal relations. Political behavior

is

in-

variably "electoral behavior," "administrative behavior,"

and so on. Or it is behavior relevant

making of public policies, training the young in the
norms of a group or society, recruiting people into political institutions, and so on. This implies that behavior in
one context or institutional setting differs from behavior
"judicial behavior,"

to the

in others,

and that the context or setting

is

the critical fac-

tqr in political behavior.

There may Be great

similarities in the

lators, bureaucrats, judges,

behavior of

legis-

party leaders, and statesmen,

or of agitators and revolutionaries, or of voters and nonvoters.

But our concern

havior that

we

is

with those characteristics of be-

expect to occur because the behavior

relevant to the institutional political order of

part.

For

It

it is

is

not a separate
would be meaningless to treat

this reason, political

field of investigation.

political

which

behavior

is

behavior apart from situational

conditions

in

the political order. Political scientists of the behavioral

persuasion, whatever their approach, are not doing this.

Rather, they select those behavioral patterns in the con-,

duct of individual actors that pertain to particular po-j

or situations. Institutions differ in the

litical institutions

structure of relations between individuals and the functions they

perform

in the political order. Inter-institu-

tional analysis of political behavior can reveal, therefore,

significant similarities or differences in political relations,

between groups

as

much

as

between

individuals.

This type

Behavioral Approaches

[i8]

On

the

analysis

of

of analysis does not take politics out of behavior.

contrary,

it

makes

possible

the

behavioral

institutions.

The

between institution and behavior is

complementary. Institutional arrangements,
norms, or functions express behavioral patterns that have
been stabilized through the passage of time. In turn, current behavior is necessarily circumscribed and directed by
relationship

necessarily

the past patterns

we

call institutions. Institutions differ a

which

great deal, of course, in the extent to

past be-

havior has been regularized and stabilized in terms of

structures, functions, norms, or goals.

As

a result, different

degrees of stability have a profound effect on what

is

pres-

ently mandatory, permissible, or prohibited behavior. In

other words, institutional parameters of behavior are

or

less rigid

or flexible.

more

great deal can be discovered

about these parameters by formal, institutional analysis.

Whether and how much actual behavior conforms to or
deviates from institutional prescriptions, regardless of

whether they refer to structure, functions, norms, or goals,

of immediate interest to the behavioral researcher. He
may discover politically significant discrepancies between
what is thought to be relevant behavior and what the beis

havior actually
If

is.

the student of political behavior ignores the institu-

tional sphere of politics, he imposes a severe handicap

on

He

sometimes does so because institutional

parameters are so taken for granted that their consequences for current political behavior seem minimal.
Whether they are or not is, of course, a matter of critical
inquiry. In his haste to account for political behavior in
his research.

social,

cultural,

searcher

may

or

personal

terms,

stitutional analysis.

behavioral re-

the

readily deceive himself

by

neglecting in-

great deal of political behavior can

be accounted for most economically by viewing it as a

result of institutional prescriptions. To take a very simple

and

obvious

fluenced

by

case,

the

electoral

behavior

institutionalized

is

certainly

in-

of

the

characteristics

party system or the representational system. Decision-

Levels of Analysis

[19]

making behavior

in a legislative body differs from decision-making behavior in a bureaucratic setting. Behavior

involving authority relationships is likely to be different
in differently structured formal institutions: let us say, in

tional

department or a social welfare agency. No politiwould be so foolish as to ignore the instituenvironment in which behavior occurs, even

though

his

a police

cal behaviorist

main focus of attention

is

the individual actor

rather than the institution.

LEVELS OF ANALYSIS
man is not a political animal by nature.
he is, he is only partly so. Of course, both the extent
and the intensity of man's political participation vary a

Despite the cliche,

If

good
These

deal

in

persons,

different

variations

may

groups,

and

cultures.

themselves be politically

condi-

tioned. Political institutions of different kinds set various

requirements

and

lirnits

on

political

behavior.

In

to-

man's perspective of himself and his

environment his identifications, demands, and expectations is more highly politicized than in more libertarian
systems. Most facets of life are determined by the polity.
But even in a totalitarian system there is likely to be some
talitarian systems,

variance. Political parameters of an institutional sort can-

not alone describe or explain variability in political behavior.

is

Because this is so, the behavioral persuasion in politics

not readily contained by the conventional, academic

subject-matter boundaries.
stems, at least initially,

interdisciplinary orientation

Its

from the very simple assumption

only one aspect of his total

very important aspect. If
this is so, differences in political behavior must be sought
in the whole human being. But no one discipline pretends
to know man as a whole. This does not mean that we can
know about political behavior only if we know about all
aspects of human behavior.
can reasonably assume that
that man's political behavior

behavior, and

by no means

is

We

political

behavior

is

more

closely related to

some

aspects

Behavioral Approaches

[20
of behavior than to others. Just

what

aspects of political

and other behavior are to be related in a particiilar investigation depend on the problem at hand and the theoretical scheme for solving it. Being interdisciplinary
means selecting from what is known about man those
of his behavioral characteristics investigated
'

by

different

behavioral sciences that seem to be especially relevant to

the solution of political problems.
If

one were to rank the reasons for the interdisciplinary

orientation of the behavioral persuasion in politics, the

V'

problems
of problems that the political behaviorist formulates cannot be
solved without recourse to the concepts and theories of
the several behavioral sciences. Problems have a way of
spilling over disciplinary boundaries. Attempts to solve
them in terms of a single discipline's concerns are likely
to be partial and unsatisfactory. For instance, elections
cannot be studied fruitfully on the assumption that the

borderline

character

would have

to be given high priority.

voter

is

of

political

behavior

The kind

rationally calculating, self-interested political

man. But one need not assume the opposite and deny the
rational interests of the voter as a family man, a
wage earner or a church member. The question of how
and why the voter decides as he does cannot be answered
by abstracting his behavior in the voting booth from his
total experience. His loyalties to family or class, his- identifications with group or party, his career demands and
expectations, his cultural milieu, and even his personality may have to be taken into account. The interdisciplinary orientation prevents neglect of the wider context
in which political behavior takes place. It calls attention
to the possible effects of social, cultural, and personal
y factors that, on the face of things, are not political as
such.

This formulation may give the impression that what is

behavior can somehow be separated out of
what is social, cultural, and personal. Actually, what we
call social, cultural, and personal pervades all behavior, including political behavior. Only for the purpose of inpolitical in

Levels of Analysis

we

quiry do

think of

[21]

what

social

is

tural but not political, or personal

analytically distinct. It

is

more

speak of levels of analysis

sonal levels^on

which

appropriate, therefore, to

^the

political

but not political, culbut not political as

social, cultural

and per-

may be

examined.

behavior

configurative analysis of political behavior cannot ig-

though a problem may be

on one level than on another (see
Chapters Two, Three, and Four).
nore any one of these

more

The study
acts,

is concerned with the

and expectations of man in
contexts. But there is little reason to assume that

of political behavior

attitudes, preferences,

political
His

levels,

effectively treated

behavior in politics

is

basically different

havior in other contexts. People

the social

life

of their

who

community or

from

his be-

do not participate

in

in voluntary associa-

be politically apathetic. The problem of the union leader seeking a maximum of concessions at the bargaining table does not seem to differ significantly from the problem of the legislator seeking a
maximum of advantages for his district. Both must satisfy
their respective voters to keep their jobs. Although participation in private groups and bargaining v/ith management are not political, there may be behavioral similarities
from one institutional context to another that are potions are also likely to

litically relevant.

Similarities in behavior sensitize the observer to critical

differences,

once

institutional contexts are varied.

For

in-

stance, superficially viewed, the behavior of a bureaucrat

in a public

agency appears to be very similar to the be-

havior of his opposite

/Much

generic
larities

number

in a private organization.

bureaucratic behavior can be explained in terms of

characteristics

that

stem from structural simi-

of different institutional settings. But there are dif-

ferences of a personal or social nature that cannot be ig-

nored.

bureaucrat working for a public agency

may

be

quite differently motivated and have different career ex-

pectations than the bureaucrat working for a private firm,

with determinable consequences for the organizaton of
which each is a member. Or the fact that the public

^'

Behavioral Approaches

[22,]

more likely to be scrutinized bythan the private bureaucrat's actions by stockholders may have behavioral consequences that only interdisciplinary inquiry, with a focus on different institubureaucrat's actions are
legislators

tions,

An
from

can harness.
interdisciplinary orientation has value quite apart

usefulness in a particular research situation. It

its

on v^hich research may be

from
formulating particularistic hypotheses where generic ones
may be more appropriate. It suggests to him the feasibility
points to the level of analysis

conducted most

fruitfully. It saves the researcher

of controlling different institutional contexts in order to

identify similar behavioral patterns, or of controlling be-

examine their

havioral patterns in order to

effects

in

different contexts.

Even

political

if

problems and patterns of political bedifferent from problems and be-

havior were totally

havioral patterns in other contexts, theoretical formiilations concerning

human behavior are unlikely to

Theory found useful in one

in splendid isolation.

coexist

context

will sooner or later fertilize investigations in other contexts.

theory's viability

is

best judged

interdisciplinary applicability. If

planation of behavior in
as

many

it

by

its

range of

contributes to the ex*

different settings, political

well as non-political, the knowledge gained will be

the

more

than

less

significant.

variance

is

theory that explains more rather

preferable.

havioral researcher in politics

is

For

this reason, the

be-

forever on the lookout

for theoretical models developed in the other social sci-

own

ences. If they apply to his

more
lem.

certain that his

And

just as

work

problems, he can be

deals

with

a significant

theory transcends conventional

all

the

prob-

discipli-

nary boundaries, so does method (see Chapter Five.)

How far one should go along interdisciplinary routes
depends largely on the particular research problem. Many
problems in politics can be dealt with without recourse to
theoretical

propositions

or

empirical findings

of

other

towards constituents
are likely to vary with the degree of competition in the
disciplines.

politician's

attitudes

Levels of Analysis
district

[23]

election. The more competthe more likely it is that he will be acces-

where he stands for

itive his district,

sible to conflicting interest groups, that he will seek to

mediate political conflicts in terms of his appraisal of the

relative strength of

competing groups, and that he will

try to explain his conduct in a

few groups

as possible. It

way

that will antagonize as

would be most uneconomical

and, in fact, expendable, to seek an explanation of such be-

on the level of the politician's personality. It

would contribute little, if anything, to the problem of
why politicians from competitive areas behave differently
from politicians in non-competitive areas.
havior

On

the other hand, one might pose this problem: are

men who

seek public office in a competitive political en-

vironment different

in their personalities

from those

who

seek office in a non-competitive arena? In other words,

does competition attract a different manner of man than

non-competition? Just what is it that makes some men
cherish and others avoid political battle? Answers to
these questions

would require the

investigation of political

attitudes in terms of personality.

The

behavioral persuasion in politics has been especially

and charged with "reduction-

criticized in this connection

ism":

that an interdisciplinary orientation inevitably^

i.e.,

reduces the political to the

that the jpolitical

not believe that

of

is

it

relevance

political

proaches. Just

termined by

what

is

cultural, or personal;

that

certainly an arbitrary

marks the behavioral

ap-

poHtical in behavior cannot be de-

criteria of

definition

sentialist

social,

taken out of political behavior. I do

is reduction, but rather an expansion

of

immanence.
politics

way

is

An
a

immanent or
convenient

es-

and

of limiting one's scope of in-

way of proceeding with an

But it is just because the traditional method
has been found wanting that it is no longer feasible to

quiry. This

is

the traditional

investigation.

draw the boundaries of a research project in politics too

it can be shown that explanation of things political is possible, if not necessary, on different levels of
rigidly. If

analysis,

including that of personality, the product of

Behavioral Approaches

[24]

inquiry

is

not the result of reduction but rather of expan-

sion of the pohtical arena.

The

proposition that politics

manent

but by

characteristics

channels

disciplinary

into

is

not defined by im-

problems

inevitably

interdisciplinary

discourse.

Once one approaches problems

its

analytically rather than

concretely, whether as a behaviorist or institutionalist, the

interdisciplinary orientation
special

explanations.

makes for generic rather than

is undoubt-

generic explanation

edly preferable to a special one. Indeed, the generic perspective permits us to locate any special perspective, to
see

term suggests,

as the

it,

as a special case.

For

instance,

kind of decision-making
behavior is probably more fruitful than politics thought
of as typically and uniquely power-centered behavior.
Decision-making takes place in many areas of social action, the purchase of goods, the choice of an occupapolitics considered as a special

tion, the
cases, as
is

a step

judging of a beauty contest. These are all special

political decision-making. The political scientist

is

ahead of himself

if,

in dealing

with

political de-

cision-making, he has a generic point of reference.

out

it,

he

is

likely to

be provincial and limited

Within per-

spective.

THEORY AND RESEARCH

No

piece of political behavior research

scribe the universe of politics,

reliable the

why

description.

The

is

content to de-

no matter how
goal

is

realistic

or

the explanation of

people behave politically as they do, and why,

as a

and systems function as they

do. There are many methods of explanation. Whatever

result,

they

political processes

are,

they require theorizing

activity.

(I

speak of

"theorizing activity" rather than "theory" because

it

frees

having to say what theory is. Any attempt to do

so in an essay of this sort would be foolish. My interest is
in what, generally, practitioners of the behavioral persuasion in politics do when they theorize, not in defining

me from

theory.)


Theory and Research

What
which

[25]

when they theorize by

when they seek explana-

behavioral researchers do

mean, very tentatively,

why people behave politically as they do
differs a good deal from one to the other. At one pole,
some would probably say that they are not theorizing at
all, but only describing what they see. They deceive
themselves, for what they see depends on how they see
it, and how they see it depends on images in their minds.
These images may be very diffuse and hardly deserve being called theories. But they orient the observer, innocent
as he may be of what he is doing and though he may be
protesting his theoretical innocence. His work will not get
him very far because, paradoxically, his very attempt at
only describing what he sees is suspect. Did he really see
all that could have been seen? What did he leave out?
How did he order what he saw? Did this really follow
that? On the whole, the overwhelming number of modtions

of

ern behavioral practitioners no longer plead theoretical

innocence.
plicit

More

make

often than not, they seek to

the assumptions and

ways of thought

their work. This is what I have called theorizing

though it does not necessarily entail theory.

At
ent,

ex-

that guide
activity,

the other pole are the builders of logically consist-

deductive

models

of

systems,

political

theories in the sense that "formal truth"

is

perhaps

the distinctive

content of the theorizing activity in which model builders are engaged.

There

are not

many
who

behavioral persuasion in politics

the

right

time

for

constructing

practitioners of the

believe that this

logically

ductive pictures of the political process.

deal of admiration for these efforts, but

closed,

is

de-

have a great

must confess to

some doubt, not because I question the practicality of

formal models or their suggestiveness in research, but because I suspect they are not as theoretically pure as their
creators insist. At least I cannot avoid the impression that
behind the most formal models there lurk quite explicit
images of empirical reality. In other words, just as puret
empiricism has theoretical components, so pure theory!
has empirical components.

Behavioral Approaches

[26]

my

In

opinion, this

not a drawback. Out of

is

this dual-

ity stems the conviction that, in the present stage of de-

velopment, theory and research are necessarily interdependent,

that

questions

theoretical

must be stated

in

operational terms for the purpose of fruitful empirical

research, and that, in turn, empirical findings should

be
brought to bear on the theoretical formulation of political problems. This does not deny the possibility of a
high road to theorizing about political behavior and a low
road. But I am not sure, if there are two roads, which is
high and which is low, and I have a hunch that there are
many roads in between that are more immediately viable. This is why it seems most feasible to attack the
problems of political behavior research on as broad a
theoretical front as possible. Whatever the weight given to
one or another, it seems quite clear that if the condition of
mutual interdependence between theory and research is
to be achieved, some theorizing activity must precede
empirical

work

if

the latter

is

to be theoretically relevant,

just as empirical considerations

must enter theoretical

ef-

hypotheses are to be tested by research.

Admittedly, theorizing must be sufficiently independ-

forts

if

ent of operations to give

But

it

room

in

which

to breathe.

cannot be altogether separate from empirical re-

it

search.

One might

of interdependence

some have, that the condition

met, and that a theory's operational

argue, as
is

can be tested in principle by

may be a necessary connot a sufficient one. I cannot see how one

can be appraised,

utility

if it

reference to empirical data. This

dition,

can

but

know

is

in principle

are testable.

edge.

For theory

Whatever

whether theoretical propositions

is not the same thing as knowl-

the plausibility or validity of theoretical

speculations, they are not truths, full, partial, or probable.

Theory

knowledge,

raw

is

not knowledge but a tool on the road to

not knowledge but only the

just as facts are

materials to be molded, through theorizing activity,

into statements acceptable as probably true, or at least

not

false,

because they have been tested in the process of

empirical research.

It

is

the theoretician's responsibility

Theory and Research

[27]

not only to assert that his propositions can be tested, but

to suggest how they can be tested. But not even this makes
his propositions empirical,

though

it

may make them em-

pirically relevant.
It is this

kind of thinking which produces the charac-

commitment

teristic

theorizing

the

of

and research

behavioral

persuasion

to

mutually interrelated activities. But if theorizing and research cannot do without

each other, they necessarily limit each other. As a result,
the behavioral persuasion is characterized by a healthy
as

respect for those problems that research

bound by

theory-

and theorizing bound by research entail. On the whole,

political behavior research has been limited to relatively
modest theoretical propositions. Theorizing activity has
been tempered by recognition of obstacles in the collection of data and technical limitiations in the treatment of
data. Above all, this theorizing of the "middle range," unlike theoretically

model

innocent empiricism or empirically blind

building, has been concerned with viable prob-

problems that are operationally manageable and

lems,

likely to yield returns of a cumulative sort.

The

theoretical quest of behavioral research

on

politics

complicated by having to satisfy two masters. On

the one hand, political behavior is only a special case of
is

human
litical

as

it

behavior. If this

so,

is

any theorizing about po-

behavior must take account of behavioral theory

develops in

all

behavior

political

the social sciences, and the findings of

cannot

research

about other aspects of

human

contradict

behavior.

From

findings
this per-

is really no place for an independent gentheory of politics. On the other hand, political behavior research is conducted within the large scale in-

spective, there
eral

stitutions

and processes of

special theory of politics

at least

be relevant,

individual.

politics. Insofar

on

as

there

this level of analysis,

it

is

must

not applicable, on the level of the

progress has been made as yet along

if

Not much

these lines.

Commitment

to the interdependence of theory and re-

search and to solving the problem of the relationship be-

Behavioral Approaches

[28]

tween
some

and general behavioral theory has

consequences for the development of

special political

interesting

political theory.

The

behavioral persuasion, in attending

to both theory and empirical research,

game

definitional

may

bring the

that has been played so long in the

However one

study of politics to some conclusion.

de-

fines politics, as the process of allocating values authoritatively, as a competitive struggle for

power,

as collective

making for the community, and so on, each

decision

definition includes more^-or less well articulated premises,

postulates, or assumptions about politics.

behavioral

theorizing,

Another

premises.

is

certainly,

to

is

One
lay

function of

bare

these

to clarify the empirical referents of

The notion that

an allocative -process, for instance, assumes that
resources are scarce and may be used in alternate ways. It
further assumes that the goals are multiple and that, thereconcepts, definitions, and propositions.

politics

is

fore, choices

must be made among them. This

also implies that political actors will disagree

ends are preferable,

be allocated
It is

as

well as over

how

in order to achieve agreed

evident that

we

on

resources are to
ends.

are not dealing here simply with a

primitive definition of politics, but with a

political

model of the

process borrowed from economics.

inspection,

it

definition

over what

also appears that the

On

closer

model makes assump-

human rationality and about the behavior of

human beings. And rational behavior is assumed

tions about

rational

to maximize preferred returns

sources. It

is

on the investment of

of politics as "authoritative allocation of values"

ily translated into
litical

re-

not accidental, therefore, that the definition

process as a

is

read-

an even more formal model of the pochain of inputs and outputs.

In subjecting the

model

to empirical testing, behavioral

research alone can give the theorist some feeling as to

what concept or definition is operationally useful and

what is not. The specification of the empirical data
needed to test a model may reveal which of rival definitions are serviceable and which are expendable. It may
show which definitions are empirically, and possibly

Theory and Research

[29]

necessary for each other. For instance, I

have found in some of my own work on legislative behavior and institutions that a definition of politics as allocation is insufficient unless it is implemented by some
theoretically,

definition of politics as conflictual behavior.

Moving simultaneously along both

theoretical

and em-

what

pirical paths quickly sensitizes the investigator to

even theories are expendable. For

instance, it seems that power, long accepted as the central
organizing concept of politics, is rapidly losing ground
from the point of view of its operational, if not analytical,
utility. Paradoxically, it is losing ground not because it is
abused, as it has been by some theoreticians in the past, but
definitions, concepts, or

because

it is

used.

For

proves increasingly

as it is

useless.

used in empirical research, it

In recent years, there has

been much research interest in community power strucnow

tures and in power relations in legislative bodies.
have a large body of research findings as to how,
presumably, decisions are made in villages, cities, and

We

metropolitan areas, as well as in legislatures.

research there

shows
I

itself

am

the

more

The more

elusive the concept of

power

to be.

many methodological
have been found to stand in

not thinking here of the

and technical
the

is,

way

difficulties that

of operationally defining, identifying, discover-

ing and measuring "power."

those engaged in

What

power research

find interesting

is

that

are increasingly forced

to rethink the concept as they face empirical situations

that defy
tions

its

traditional verbal uses.

made between

We

now

find distinc-

the weight, scope, and domain of

power, or between "monolithic" and "polylithic" power.

Whether these distinctions and elaborations will save the
concept as a theoretically useful one I cannot say. As used
to be the case with the concept of "sovereignty" (unlimited and limited, undivided and divided, shared and
what not), power is still a concept we cannot do much
with apparently, but which we do not dare to do without.

Theorizing depends on the problem to be solved.

Some

Behavioral Approaches

[30]

problems are more complex than others, and more may

be known about simpler ones. Political behavior involved
in the conduct of foreign policies in the international
arena is probably more complex than behavior in the domestic legislative process, and the latter is likely to be
more complex than a person's behavior in the voting
booth. How complexity is handled depends, in turn, on
the model used in the analysis of empirical data. Although
a simple model is preferable to a complicated one, it is
also likely to be empirically more exclusive. On the other
hand, an elaborate model or conceptual scheme may make
the problem technically unmanageable. Just as the analysis
of only two variables that are theoretically linked
in rather simple propositions may not explain very
much, a comprehensive scheme of a potentially all-inclusive range may defy the practicalities of research. The
most feasible alternative is to deal with modest propositions that require simultaneous manipulation of only a
few variables, but to do so in a larger conceptual system
that, though it cannot be tested directly, serves the very
useful purpose of guiding an investigation and giving it
theoretical significance.

Considerations of this kind have some further consequences for the development of behavioral theory in connection with problems of varying degrees of complexity.

The more complex

research

is

the empirical problem with which the

dealing, the

more

difficult access to relevant

and the more need there

on high levels of generality. On the other hand, the less complex the empirical
problem, the easier the collection of relevant data and the

behavioral data

is

likely to be,

will be for theoretical exploration

less

incentive to theorize. Therefore, theoretical formula-

be very specific and of relatively low generalThis makes plausible the theoretical unevenness of be-

tions will
ity.

havioral research in different, substantive areas of political

science. In recent years, the behavioral persuasion has gen-

erated a considerable

body

of theoretical

work

of high

generality in the study of international politics, but

so far

produced

little

hard empirical research.

it

On

has

the

Behavioral Methods

[31]

other hand, empirical studies of electoral behavior are

abundant, but this research has been limited to testing

very modest propositions of low theoretical generality,
and no comprehensive theory of the electoral process has
as yet been formulated.
This produces an interesting paradox. It v^ould seem
much more reasonable to apply and test theoretical models of high generality in relatively simple research areas
about which a good deal is already known and where access to data

is

relatively easy.

models of international

For

instance,

many

of the

could be tested with data

drawn from the political life of metropolitan areas. It
would also seem more reasonable to cope with behavior
in a

complex

setting

Some

propositions.

politics

by way of more modest theoretical

work now being done on the

of the

national level in underdeveloped areas seems to be of this

order.

complex and simple theobetween different empirical research

be the next major phase in political behavior

believe this exchange of

retical formulations
fields will

research,

if it

has not already begun. Students of interna-

tional political behavior

havior will find

it

and large-scale

institutional be-

increasingly profitable to deal with

propositions that have been found serviceable in studies of

voting or problem-solving behavior in small groups. Students of relatively small institutions (such as legislative

and local communities will draw

on the more comprehensive formulations of
communication theory, system theory, and other con-

bodies

or

courts)

increasingly

figurative models.

BEHAVIORAL METHODS
The

revolution in the behavioral sciences has been predominantly a technological revolution. Compared with
developments in the natural sciences, the gimmicks and
gadgets of behavioral science remain rather crude. But
compared with the tools available to the classical writers,
modern behavioral technology represents an enormous
advance. It would seem foolish not to apply this new

Behavioral Approaches

[32]

behavioral technology to the problems of politics. Yet,

for reasons difficult to pin down, the application of behavioral methods to politics has been halting and circum-

some

quarters, there has been fierce resistance to

our knowledge about politics through behavioral analysis. Why such resistance occurs is a matter
of interest to the sociologist of knowledge, and how it
might be overcome is a task for the psychoanalyst. The
spect. In

increasing

result

is

garded,
as

that the behavioral persuasion in politics

by

its

revolt

science.

havioral

As

opponents
the

against

classical

suggested earlier, this

persuasion

by

well as

as

revolted

its

tradition
is

is

re-

practitioners,
in

political

a mistake. If the be-

anything,

against

it

was

against the failure of academic political science to use the

modern technology

in the study of politics as, I believe,

the classical writers

would have used

it

had

it

been

avail-

able to them.

Resistance

to

the

application

of behavioral

must be distinguished from some very

methods

real difficulties of a

methodological sort that the behavioral persuasion faces

and with which I deal in the last chapter. I would now
like to ask if there are

not some "natural limits" to the

behavioral analysis of politics, limits that no technological

revolution can eradicate.

asked only

if

believe this question can be

one assumes, a

priori, that there

must be

immune

to scien-

aspects of political behavior intrinsically

But

one assumes the opposite, that politibehavior, can be observed by the

methods of behavioral science, the limits appear to be
technological ones. Scientific technology knows its present limits; it cannot predict its future limits. As technology advances, the range of phenomena amenable to
scientific analysis also expands. Therefore, it is really impossible to say that the data of politics are such that they
cannot be harnessed by any scientific methods and techniques. The presently available technology has made possible the production and processing of political data, or
tific analysis.

cal behavior, like

if

all

data relevant to political behavior, that was, until recently, unavailable to political science.

Behavioral Methods

This

[33]

not the place to review the areas of political behavior research investigated by the techniques of modern
behavioral science nor to present an inventory of research
methods. The interested reader may wish to consult the
is

I do want to emphasize the discovery

and amplification of knowledge about politics which the
invention of new methods and techniques makes possible.
The development of the probajbility sample survey as a
reliable instrument of data collection is perhaps most
noteworthy. It has made public opinion and electoral re-

bibliography. But

search a prolific source of political data. Indeed, for a

time, the study of political behavior was equated with
and restricted to the study of public opinion and voting.
But more recently, the extension of systematic surveys
to the study of specialized elites and institutionalized

groups

like legislative bodies or bureaucratic organizations

has greatly augmented the store of political

deepened

political

formants,

repetitively

analysis.

The

use

data

panels

of

of

and
in-

interviewed in successive waves,

has added a longitudinal dimension to behavioral inquiry

now

of politics. Political change can

be observed

at the

level of the individual actor.

Perhaps equally significant

the invention of metric

is

techniques such as scalogram and factor analysis, and their

application not only in survey materials, but also in the

treatment
cisions.

of

legislative

roll-call

Although the individual

in political

votes

or judicial

actor's behavior

action of this kind,

it

it

de-

known

cannot be analyzed

meaningfully from the perspective of politics

tive enterprise unless

is

as a collec-

can be ordered to reveal underly-

ing regularities and uniformities. Scale analysis, for instance,

makes

roll-calls

possible the ordering of discrete acts like

along a single dimension and the classification

of actors in terms of their scale positions.

provides for

It

such ordering and constitutes an

instrument of considerable analytic and predictive power.
criteria of reliability in

Similarly, factor analysis of judicial opinions, for instance,

makes

possible the discovery of the principal

of any set of interrelated decisions.

The

components

discovery of

Behavioral Approaches

[34I

and uniformities in

regularities

by

possible

judicial

made

behavior

factor analysis gives depth to the study of

courts as political institutions.

The

of

analysis

facilitated

by

political

observational

form and content of

as

has

been

recording both

action-in-interaction at the level of

Though few ongoing

the individual.

groups have

group behavior
techniques

("natural") political

yet been subjected to one or another ob-

servational technique, it promises rich returns once small

groups like legislative committees, administrative bodies,
or party councils allow themselves to be systematically
studied.

The

application of sociometric techniques that

choices among group members has

been more immediately useful. These techniques, feasible
not only in small but also in middle-sized groups, help
elicit inter-individual

in identifying the informal structure of leadership or fac-

They make

tions.

such

as

possible the study of

group properties

authority, communication, cohesion, morale, or

consensus at the level of the individual. These properties

are important not only in explaining the behavior of individuals in groups, but also the actions of a

group

as

whole.

Through techniques
broaden the range of

made

of this kind, then,

political

possible to

be
and to refine and

available to behavioral analysis

systematize political behavior research.

not meant to minimize the

jecting political behavior to

it is

phenomena

difficulties in

that can

However,
the

way

scientific inquiry. I

this is

of sub-

think that

the practitioners of the behavioral persuasion in politics

are

more

sensitive to the

problem of

a science of politics

than those who deny its possibility.

It does not seem very fruitful to specify what a
behavioral science of politics might look like "in the end'*
because science
In general,

the

is

an ongoing endeavor that has no end.

persuasion tries to develop

behavioral

rigorous research designs and to apply precise methods

of analysis to political behavior problems. In its methodological orientation

it

is

concerned with problems of

experimental or post facto design, reliability of instru-

Bibliography

[35]

ments, and criteria of validation, and other features of

scientific

procedure.

Its

propositions

reliable

function, as

see

it,

is

to produce

about politics by reducing error,

which

involves, the invention of appropriate tactics of re-

search,

and by measuring error that remains through the

application of relevant statistical techniques.

as

the present success

terprise in politics

may

of the behavioral

As

limited

scientific

en-

be, the alternatives are even less

satisfactory.

The

discriminating feature of the behavioral persua-

sion in politics

is,

above

sensitivity to error in its

its

all,

observations of politics and

its

mulated, universal "truths."

It

suspicion of a priori, for-

proceeds in terms of con-

tingencies and probabilities, rather than in terms of cer-

and

tainties

verities. It represents

persuasion as

have called

granted and accepts

when

it,

as valid

an attitude of mind, a
takes nothing for

that

only the results of

its

in-

would be unreasonable to assume that they

can be explained solely by the operation of chance. This
is a difficult standard to live by, perhaps more difficult in

quiries

it

politics than. in other fields of

itics

as

in

certainty, but

physics

must

and

human

action.

metaphysics,

man

For

in pol-

looks

for

settle for probability.

BIBLIOGRAPHY
SELECTED STUDIES OF POLITICAL BEHAVIOR
CONTEXTS

IN INSTITUTIONAL

Campbell, Angus, Converse, Philip

Stokes, Donald,

The American

E., Miller,

Voter.

Warren

E. and

(New York: Wiley,

i960.)

Who Governs? Democracy and Power in an

American City. (New Haven: Yale University Press, 1961.)
Kaufman, Herbert, The Forest Ranger: A Study in Administra-

Dahl, Robert A.,

Behavioral Approaches

[36]
tive Behavior.

(Baltimore: Johns

Hopkins University

Press,

i960.)

Key, V. O.,

Jr.,

American

State Politics.

(New York: Knopf,

1956.)

Seymour M., Trow, Martin, and Coleman, James, Union

Democracy. (New York: Free Press, 1956.)
Matthews, Donald R., United States Senators and Their World.
(Chapel Hill: University of North Carolina Press, i960.)
Schmidhauser, John R., The Supreme Court: Its Politics, Personalities and Procedures. (New York: Holt, Rinehart and
Winston, i960.)

Lipset,

INTERDISCIPLINARY FORMULATIONS
OF POLITICAL BEHAVIOR
Dahl, Robert A., and Lindblom, Charles E., Politics, Economics,

and Welfare. (New York: Harper, 1953.)

Harold D., Power and Personality.

Lasswell,

(New York: Nor-

ton, 1948.)

Alexander, The Governing of Men. (Princeton:

Princeton University Press, 1946.)
Snyder, Richard C, Bruck, Henry W., and Sapin, Burton,
Foreign Policy Decision Making. (New York: Free Press,

Leighton,

1962.)

THEOPvETICAL STATEMENTS OF THE

BEHAVIORAL PERSUASION
Dahl, Robert A.,

Preface to Democratic Theory. (Chicago:

University of Chicago Press, 1956.)

Easton, David, The Political System.

(New York: Knopf, 1953.)

Harold D., and Kaplan, Abraham, Power and Society.
(New Haven: Yale University Press, 1950.)
Mitchell, William, The American Polity. (New York: Free

Lasswell,

Press, 1962.)

APPLICATIONS OF BEHAVIORAL TECHNOLOGY

IN RESEARCH ON POLITICS
Berelson, Bernard R., Lazarsfeld, Paul F., and

McPhee, Wil-

Study of Opinion Formation in a Presidential Campaign. (Chicago: University of Chicago Press,

liam N., Voting:

I954-)

MacRae, Duncan,

Jr.,

Dimensions of Congressional Voting.

(Berkeley: University of California Press, 1958.)

Bibliography

[37]

Schubert, Glendon, Quantitative Analysis of Judicial Behavior.

(New York:

Free Press, 1959.)

B., The Congressional Party.

Truman, David
Wiley,

(New

York:

1959.)

Wahlke, John C, Eulau, Heinz, Buchanan, William, and Ferguson, Leroy C, The Legislative System: Explorations in Legislative Behavior.

(New York:

Wiley,

1962.)

CHAPTER TWO

THE SOCIAL
MATRIX

Man's

political

behavior and relations are only part of his

existence as a total

human

being, yet they are easily ab-

from the more comprehensive social environment. For some of the purposes of political analysis, this

stracted

is

both legitimate and necessary.

Not

all

of man's mani-

fold social relations are politically relevant, but enough

are to suggest that

we pay

attention to the social matrix of

of political behavior. Otherwise political description and

explanation

may be

readily distorted or seriously misin-

may be
with exclusively from the political point of view, or
a person who is not involved very much may be treated
as if he were totally alienated from the political arena.
The elite and mass models of contemporary political
society come close to making assumptions of this kind
about political behavior. Neither image is likely to do
terpreted.
dealt

person deeply involved in politics

Role as a Basic Unit

to

justice

political

[39]

reality.

other models of man,

men

related to other
a total

man, unlike some

Political

not an abstraction, but socially

is

ways

in a variety of

that

make him

human being.

The

of political behavior can best proceed,

analysis

therefore,

by

locating

man

as a political actor in the social

matrix of interpersonal relations.

human

implies that

The

notion of a matrix

behavior, including political action,

and transactions that orient people

towards each other, making them mutually responsive. I
find the concept of "role" most useful to capture these relations, but it is the web of all of man's interpersonal relations, the social matrix, that gives behavior, including poconsists of interactions

litical

behavior,

It is

characteristic structure.

its

convenient to think of this structure in terms of a

vertical

mension

and

horizontal

consists

dimension.

The

vertical

di-

of man's group relations and can be

analyzed in terms of the concept of "group" broadly interpreted to include

horizontal

strata that define the

terpersonal

all

dimension

interpersonal collectivities.

consists

those

upper and lower

intergroup

or

of

strata, then, will serve as

contacts.

The

differentiating

limits of either in-

Role,

group,

and

conceptual building blocks of

the social matrix.

ROLE AS A BASIC UNIT

In

its

simplest form, political behavior, like

behavior, involves a relationship between at least

man

beings. It

is

all

social

two hu-

impossible to conceive of political be-

havior on the part of a person that does not have direct,

indirect,

or symbolic consequences for another person.

The most

concept for analyzing a relationship

actors and for determining the political relevance of the behavior characteristic of the relationship is "role," for we are not interested in all of a
person's behavior but only in that aspect which is

between

suitable

at least

two

relevant to a political relationship.

The concept

of role

is

familiar to

most people.

We

The

[40]

Social Matrix

speak of the father's role, the teacher's role, the minister's

role, the judge's role,

and so on.

What we mean
by

in all of

and
he bebehave, or should behave in certain ways. In
man's social behavior or judging it, we do so

these instances

that a person

is

is

identified

his role

that, in interpersonal relations activating the role,

haves,

vi^ill

looking

at

frame of reference in which his role is critical. If we

do not know a person's role, his behavior appears to be
enigmatic. But a child ringing a doorbell is unlikely to be
in a

mistaken for a political "doorbell ringer." Political behavior, then, is always conduct in the performance of a
political role.

Out

of observations as simple as these, social scientists

have built a variety of theories about the origins, structure, functions, and meanings of social roles. Whatever its
uses in everyday language or scientific research, role
seems to commend itself as a basic unit of social and political analysis.

Role can be used

as a

of behavioral analysis:
personal. It

On

is

conceptual tool on

concept generic to

the social level,

it

On

three levels

all

the social sciences.

invites inquiry into the structure

of the interaction, connection, or

relationship.

all

the social, the cultural, and the

the cultural level,

bond

that constitutes a

it calls

attention to the

norms, expectations, rights, and duties that sanction the

maintenance of the relationship and attendant behavioral
patterns.

And on

the personal level,

it

alerts research to

the idiosyncratic definitions of the role held

actors in the relationship. Role

is

by

different

clearly a concept con-

with the analytic objective of the behavioral sciIt lays bare the m?^r-relatedness and inter-dependence of people.
On the social level, many of the most immediate interactions can be analyzed in terms of polar roles: husband
sistent

ences.

implies wife; student implies teacher; priest implies

com-

municant; leader implies follower; representative implies

constituent, and so on. The behavior of one actor in the
relationship is meaningful only insofar as it affects the
behavior of the other actor or

is

in response to the other's

Role as a Basic Unit

[41]

Whatever other

behavior.

acts a representative

may

per-

form, for instance, only those in the performance of his

constituent relationships are of immediate interest in political behavior analysis. I say of immediate interest be-

no

cause, in actuality,

single relationship

is

isolated

from

other social relationships in which the partners to the

focal relationship are likely to be involved.

Many

by unipolar

relationships are not structured

roles

roles,

most cases, a role is at the core of several otheri

making for a network of roles that can be very com-V

plex.

alone. In

legislator

"representative"

"enemy") to

"colleague" to his fellow legislators,

is

to

(or

"friend"

constituents,

his

lobbyists, "follower" to his party leaders,

"informant" to the press, and so on. Whatever role is

what emerges is a very
intricate structure of relations in which one role is im-

taken, simultaneously or seriatim,

plicated in several other roles.

may

role

example, the

be implicated in several networks. For

of a city is not only a chief executive,

mayor

a role that implicates

ships

in

related

many

to

other

his

him

in several other role relation-

position,

relationships

relevance to his political roles.

but he
of

is

also

more or

He may

involved

less

direct

be a husband and

alumnus of the local college, a member of the

Rotary Club, a lawyer, a churchgoer, an investor in a local business, and so on. Depending on circumstances, these
roles may complement each other, be mutually exclusive, or conflict. A network of roles reflects the complexity of social and political behavior patterns and warns
against treating any one role as if it were exclusive.
Analytically, each network of roles can be thought of
as a "role^ system." This has two corollaries. First, some
father, an

roles are

more

directly related to each other than are

other roles. For instance, the roles of husband and father

or legislator and representative are intimately connected.

Other
their

roles may be less so. The existence and degree of

mutual implication is always subject to empirical

determination.
likely to

The

legislator's role as a

be related to

his role as a

lawmaker

parent than

it is

is

less

to his

The

[42]

Social Matrix

occupational role as, say, an insurance agent. This does

not mean that the parent role is altogether irrelevant
in his legislative behavior.

tending public schools

is

legislator with children atprobably more interested in

school problems than a legislator

who

is

a bachelor.

The

notion of role system directs attention to the totality of

social behavior.

At

the same time,

it

points out the need

of specifying the boundaries of the particular system un-

der investigation.

The second corollary of role system implies that a

change in one role may have consequences for the actor's
other roles and, therefore, for the relationships in which

he is involved by virtue of his roles. (This must not be

confused with a change in position. Wheii a Senator be-

comes President,

his

new

make

Senatorial role

terminated. His

is

new roles that greatly affect

his other role relationships). As an example of role
change, take the representative who finds it impossible
position will

to accept instructions

on

for

from

his constituents

and increas-

own

judgment. In the technical

language of role analysis, this is a change from the "delegate" to the "trustee" role. It is likely that this change in
ingly

relies

his

the representational role will have consequences for the

legislator's

party-relevant roles.

a partisan follower into

The

structure

terned but also

He

might change from

an independent.

of role relationships
fluid.

One

is

not only pat-

source of change in role

is

change in the expectations of others in the role system.

Another source may be an actor's own redefinition of
his role. These possibilities suggest the importance of
treating role concepts from a cultural and personal standpoint as well as a social one.

On

the cultural level, role refers to those expectations

of a normative sort that actors in a relationship enter-

concerning each other's behavior. These are the

and duties that give both form and content to the
relationship can be maintained only as
relationship.
long as the participants are in agreement as to what each
actor must or must not do in the performance of his role.
tain

rights

Role as a Basic Unit

[43]

If there is disagreement over what kind of behavior

should be expected, the relationship is likely to disinte-

grate.

Expectations which define roles and give direction to

the behavior of actors in a role relationship are cultural
in

two ways. People do not continuously

define and re-

define their mutual relations and expectations. If a rela-

tionship had to be defined

if

anew with each

interaction, or

expectations had to be elaborated with every

counter, stable social

most of the

They

life

would be

new

en-

impossible. In fact,

crucial role relationships are well defined.

are well defined because expectations are widely

shared and transmitted through time. There

broad cultural consensus

as to

what the

is,

then, a

and duties
consensus on

rights

pertaining to social roles are, and there

is

the sanctions available to participants in a relationship

if

behavior should violate agreed-on norms.

There may be more or less agreement from one role to

from one culture to another. In Western culture,
there is a broad consensus as to what kind of behavior the
role of parent vis-a-vis the child calls for, though there
are differences in role conceptions from one subculture
another,

to another. But

work

if

a role

is

to identify.

located at the center of a net-

is more difficult
Only some minimum agreement might exist.

of roles, consensus

on expectations

For instance, it is difficult to say without inquiry just

what behavior is expected of the politician. For the politician

is

involved in a multitude of relationships, with

other politicians,

community opinion

leaders,

financial

patrons, spokesmen of interest groups, friends and neighbors,

government bureaucrats, and so on. Each

others,

have

its

set of these

themselves role-takers in the relationship,

own

particular expectations as to

how

may

the politi-

cian should conduct himself. Consensus cannot be taken

for granted.

Precisely

because

role

expectations

may

be widely

shared and relatively permanent, they give stability to the^

relationship.

Role relationships thus make for stable pat-

terns of behavior and minimize

what would otherwise

The

[44]

Social Matrix

have to be considered arbitrary behavior. Understanding

a role means that we know how a person should behave
and what he should do in the performance of a role. This
includes knowledge of probable sanctions and thus makes
accurate prediction in social relations possible. This ability

predict

to

within the limits

another's
set

by

behavior,

always,

of

course,

on the assump-

expectations and

tion that behavior will agree with the role, permits the

partners in a role relationship to shape their

in anticipation of the other's reactions. In
this

The

is

own conduct

some

respects,

kind of guessing game without beginning or end.

repetitiveness of the

game makes

for patterns of be-

havior that produce those uniformities of behavior whose

cultural-normative source

is

not especially

felt (for

more

on this subject, see Chapter Three).

Role analysis aids in discriminating between norms for
behavior and actual performance of a role. It may be argued that the best way to identify a man's role is to see
how he actually behaves. A role, it would seem, is best reconstructed from performance. But this procedure, apparently so objective, ignores an important aspect of be-

meaning. The same bit of action may have

meanings for different actors (and, of course,
different observers). Meanings are important in politics
because politics is eminently concerned with the consequences of behavior. These consequences require evaluation. Roles as normative expectations of an actor himself
concerning his conduct or of others provide meaningful
criteria of evaluation that would otherwise remain quite
arbitrary. For this reason, the distinction between the
normative and behavioral components of a role is analytically and empirically necessary.
Even if there is a wide consensus on roles, there is always a good deal of variation in their performance. This
may simply be due to the fact that a role is defined not
only by others' expectations but also by an actor's own
conception of his role. Admittedly no self-conception of
a role can be completely different from the conceptions
havior,

its

different

of others in the role relationship. In spite of differences in

Role as a Basic Unit

[45]

behavior, most conduct

is

recognized for what

cause roles can be identified. But though

is

be-

we may

see

it

ourselves as others see us and take appropriate roles, roles

are never taken in identically similar ways.
tion

may

The

explana-

be that two actors taking the same role

have somewhat different

self -conceptions of the role

may
be-

whom

they react are different actors

with different expectations. This interpretation remains
cause the others to

on the

level of social

viation
is

is

major,

and cultural

analysis.

minor, socio-cultural analysis

it is

necessary to find

is

And

if

the de-

sufficient. If it

more personal

clues.

Actors do bring idiosyncratic perceptions of the interpersonal situation, attitudes, and motivations to a role.
Role analysis does not preclude, but may require, investigation of role conceptions from the point of view of the
actor's personality.

roles

is

An

actor's

capacity to take certain

predicated on the possession of certain personal-

ity characteristics. Just

what

of theory and research.

these are

shall

is

a subtle

come back

to

problem
it

in the

fourth chapter. But whatever hypotheses are formulated

about personality in

politics,

they cannot ignore the wide

range and the great variety of possible political roles. It

would be quite erroneous to assume a one-to-one rela-

between

tionship

a political role

and a given personality

type, possibly treating personality as the independent and

role

as

the

autonomy of

dependent variable. This would deny the

on the social level.

Role conflict may stem from various conditions, but
two are noteworthy. These may actually be divergent
city councilman
expectations of a person's behavior.
may expect the city manager to guide and direct the
council's legislative business, while another councilman
analysis

may
tions.

expect him to abstain from policy recommenda-

Or

there

may be

disagreement between others' ex-

pectations and an actor's

own

conceptions of his role.

Moreover, the demands made from one role system to

another may be so intense that behavior in the performance of various roles cannot satisfy role requirements.
For example, involvement in the life of the Senate may

The

[46]

member's time that he cannot meet

so absorb a

tions as a representative of his state. In

role conflict

On

certain functionally necessary roles

may

democratic

On

so

may

role

politician's

strong leadership.

and even

Study of
gests a

some

how

number

roles are

deprive

group

the personal level, role conflict

disorganize behavior that

irregular,

the social level,

not be taken.
concerning the

expectations

conflicting

instance,

his obliga-

of these cases,

have dysfunctional consequences

likely to

is

all

of either a social or a personal sort.

For

Social Matrix

becomes highly

it

of

may

erratic,

irrational.

role conflict

is

of possibilities.

avoided or resolved sugonly list them. First,

I shall

more pervasive than

resolved in their favor.

others and conflict

is

Second, some roles are more

clearly defined than others,

which again aids the resolusome roles are more

tion of conflict in their favor. Third,

institutionalized than others, leaving the actor relatively

little

choice. Finally, roles are

so that, depending
flicting roles

more or

less

segmentalized

on circumstances, even potentially con-

can be taken.

THE VERTICAL DIMENSIONGROUPS

One might argue that "group" is an
The argument would be that group

expendable concept.
is

a loose, connota-

tive term, impossible or at least difficult to define operationally,

does not

and what cannot be defined for measurement

What exists in reality is not a group but a

exist.

pattern of interaction

cept of group

is

among

people. Therefore, the con-

useless because

it

cannot

tell

us anything

significant about reality.

The argument

is

fallacious.

We

cannot deny

that, for

certain purposes, individuals interacting with each other

cannot be treated

And

as

empirically joint units, that

is,

as

group, people can be treated analytically as interacting with other groups, even though, in
reality, individuals rather than groups are involved in
groups.

as a

concrete relationships.

Otherwise one could not speak


The

Vertical

Groups

Dimension

[47]

meaningfully of inter-group, inter-governmental or international relations.

This does not imply that in using group as an analytic

is, necessarily committed to one or another
of the so-called group approaches or to particular methods of observation and measurement. Moreover, because
we can think of groups as collective actors, it does not
follow that group is necessarily the most useful analytic
concept. But if the proof of the pudding is in the eating,
the usefulness of any particular formulation of political
behavior problems in group terms lies in the empirical
concept, one

results.

However, along with the vagueness of the concept,

is a problem in the enormous variety of groups that

there

may

be

politically relevant.

This does not represent an

insuperable obstacle to the use of the concept. For

it is

through conceptual articulation to reduce the

great number of concrete groups to manageable analytic
proportions. Once we distinguish between group as a
concrete structure and an analytic one, group, given its
proper label, may prove useful. If concrete groups
family, neighborhood, party, organization, or whatever
are thought of as analytic structures such as primary or
secondary groups, in-groups, or out-groups, formal or informal, reference or membership groups, two very practical results may be noted. First, retention of the concept
of group with analytic labels attached reduces empirical
complexity; and second, it permits generic considerations^'
about diverse, concrete groups that would otherwise be
possible

difficult to entertain.

Attaching analytic labels to a concept is not the same

defining it. It has an orienting function. It
directs the researcher's attention to what he wishes to
investigate. It points to certain patterns of behavior to be
described and explained, and it provides convenient, but
only convenient, handles for doing so. If on inquiry, the

thing as

label does not fit the facts, it can be abandoned. This

procedure saves us from claiming too much or too little

The

[48]

Social Matrix

for analysis of political behavior in group terms.

It does
not say that all political behavior is group behavior, as
some extreme group theorists would have it; nor does it

suggest the opposite, that group analysis of pohtical behavior is a dead end, as some anti-group theorists claim.
It

seems to

me

that a

model of

politics that ignores

group

altogether risks incompleteness.

But are role

exclusive?

analysis

The answer

is

and group

analysis not mutually

The

advantage of conceiv-

"no."

ing of political behavior as conduct in the performance

of political roles

is

that

it

concentrates not simply on the

on those othenvironment whose expectations contribute to his role-taking. These others may be
other individuals, but they may also be groups. Now the
role concept implies that if the others are a group, the
group is not something outside of the role-taking individual. Just as the role concept refers to the existence of
a relationship, so the concept of group refers to the particular character of the relationship. Though every group
can be analyzed in terms of the roles that members (a
role) take towards each other or towards non-members,
not every role relationship can be analyzed in terms of
individual person alone, but simultaneously
ers

the person's social

in

group.

Whether

the latter

is

possible

is

a function of the

properties of the relationship that must be discoverable

as

independent of the roles involved.

If

such properties

can be identified and differentiated, and if analytic labels

can be attached, group can be used as a meaningful unit
of analysis not only in treating groups as political actors,
but also in accounting for some of the roles taken by individual persons in the group and vis-a-vis other groups.
There is no intrinsic conflict between analysis of political behavior in terms of either role or group. Whether
one wishes to employ one or the other concept is largely
a question of the problem to be analyzed, a matter of theoretical preference or research strategy. For some purposes

it

is

certainly sufficient and economical to think of

behavior almost exclusively in group terms. For

instance, voting behavior can be and has been analyzed

political

The

Vertical

Groups

Dimension

[49]

way. Group analysis of electoral behavior does not

may want to know but, depending on
the problem one poses, it may be superior to either role
in this

exhaust what one

analysis or some other type of treatment. On the other

hand, the study of legislative behavior is not likely to get

very far without a more refined analysis of the roles that

each legislator can take as the member of a legislative
group. While his membership as such can tell us something about his legislative behavior in general, it is not
likely to tell us very much. Analysis of the legislator's
multiple roles may, or may not, lead us to the other
groups to which he belongs, with which he identifies,
or to which he refers in his legislative preformance.
The very vagueness of the concept of group forces the
investigator to discriminate initially

among

the types of

which he is interested. Clearly the relegroup in a person's political behavior is likely

relationship in

vance of a
to vary with the character of the interaction. Of critical
importance are the size of the group, its permanence, the
degree of intimacy or formality in intra-group relations,
the degree to which members identify with each other or
group symbols (solidarity), the extent to which attributes or attitudes are shared (homogeneity), the group's
tasks and the degree of specialization among the members, the formal system of coordination of individual
activities, and so on. Depending on specification of these
group properties, we can characterize groups as primary
or secondary groups, associations, organizations, communities, factions, cliques, parties, and so on. Whatever
the classification, political behavior is likely to vary with
the type of group in which the individual is involved. As
a result, a good deal can be said about his political behavior from knowledge of the type of groups that serves
concept of analysis.
properties probably account not only for stability in political behavior but also for change. That the
children from a Republican family may become Demoas a

Group

crats, that friends

social circle

may

who

are

no longer members of the same

part political ways, that people long

The

[50]

Social Matrix

apathetic may be activated by group influand so on, are empirical phenomena well enough
known. Changes in political behavior may be due to a
person's shifting from one group setting into another^
politically

ences,

new

since

group

patterns of political behavior often follow

affiliations

would be

new

or identifications.

on knowledge of
While
the group is a factor to be reckoned with, it is only one
factor. If the individual's behavior itself is viewed as a
component of the group, we can avoid the rather
mechanistic notion of group influence on behavior. In
It

a person's

folly to rely exclusively

group

affiliations to

predict his behavior.

the group's
we can view the re-

stead of thinking in one-directional terms

influence on or over the individual

between group and individual

lationship

transactional. Indeed people

may

as reciprocal

or

use political opinions or

new friends,
new community. This

behavior to gain entry into a group, to find

to

make themselves

may
may

at

home

involve conscious or unconscious behavior, and

or

may

not involve changes of

anticipatory socialization
ical

in a

is

may

whose views he

The group
form of

political

views, whatever

lead an individual to seek out groups

anticipates to

has at

positive

it

This kind of

particularly frequent in polit-

group membership. One's

their sources,

belief.

its

be congenial to his own.

powerful sanctions in the

disposal

commandments or

negative injunctions

to enforce conformity to group values, attitudes, and ex-

But we cannot assume that the individual is

simply the victim of forces he cannot resist and influence
in turn. Whatever patterns may be found in empirical reality, the group concept is neutral to the range of variations that may be characteristic of the relationship bepectations.

tween individual and group

in

different

group struc-

tures.

Groups range from the most casual and informal small

group to highly institutionalized and coercive organizations. If one adds the degree of political relevance that a
group may have for an individual's behavior, the task of
research appears formidable. Moreover, and paradoxi-

The

Vertical

Groups

Dimension

[51]

groups one might least expect to have political relevance are likely to be sources of lasting identifications,
demands, and expectations in the political order. The per-

cally,

vasiveness of group-anchored aspects of political behavior

by research on such varied suban agency of political socialization,

has been demonstrated

jects as the

family

as

on peer groups as reservoirs of political recruitment, on

neighborhoods as foci of political involvement, on communities as carriers of political values, on organizations
as

molders of

political opinion,

on nations

as recipients of

and so on.
important factor in determining the char-

political loyalties,

Not

the least

acter of the individual-group relationship

the simul-

is

taneity of groups in the individual's social environment.

Multiple group membership or overlapping group

mem-

bership facilitates or impedes social integration on levels above and beyond the group itself. Multiple group
membership interests us not only because of its conse-

quences for individual political behavior, but also for the

whole network of inter-group

relations that constitutes

An image of
atomized groups, like the image of atomized individuals,
is not a viable model of group analysis. The degree to
which groups are in harmony or conflict or to which they
co-exist without overt contact is largely a function of the
number of overlapping memberships and the intensity of
the vertical dimension of the social matrix.

a person's relationship to different groups.

Of more immediate
behavior

is

interest in the analysis of political

the possibility that an individual's relationship

to one

group somehow

group.

If

affects his relationship to another

the attitudes, values or goals of the groups are

harmonious, the individual's political behavior

affected in

ways

different

from

a situation in

is

surely

which they

is likely to be subject to
an unfortunate term because,
once more, it connotes a kind of individual helplessness
vis-a-vis the group. But whatever the expression, one individual-group relationship always seems to implicate
another. The notion of "cross-pressure" suggests the need

clash. In this case, the individual

"cross-pressures." This

is

The

[52]

Social Matrix

to disentangle often complex patterns of social and political relationships.

The

group interests and loyalties may

an important source of role conflict.
These conflicts engendered by cross-pressures become
be

at

possibility that

odds

also

is

most evident in critical situations where choices must

be made and where latent contradictions, arising out of
multiple group memberships not normally experienced,
make behavior erratic. For instance, on an issue like cen-

book or movie, a man's religious

and commitment might well come into conflict
with his membership in or identiflcation with an organization devoted to civil liberties. Attention to the consequences of multiple group relationships on the level of
the individual sensitizes the political analyst to a broad
range of behavioral problems that otherwise remain hidden. Of course, many roles stemming from relationships
with different groups need not conflict. And even if thy
do conflict, men may be only vaguely aware of the contradictions. It is not the fact of multiple group relationships, but its possible dysfunctional consequences for roletaking in situations where political decisions must be
made, that is of special interest from the behavioral point
sorship of a particular

affiliation

of view.

group's significance to a person

problem.

If

is a psychological
every person brings into the group relation-

ship certain predispositions and attitudes, group analysis

cannot be limited to group structure, processes, and goals,

but must accommodate these individual psychological variables.

Groups

are objects of

human

experience making

for different orientations. Perception of the group's rele-

vance to an individual's personal needs and goals or one

attractiveness as compared with other
groups, and so on, are factors that, in turn, affect the relations among groups as collective actors. Considerations
of this sort, stimulated by the concept of group on the individual level, may be a rich source of particular hypotheses about political behavior on the group level
group's greater

of analysis.

The

Vertical

Groups

Dimension

The group concept

[53]

psychologically treated

calls atten-

tion to certain symbolic associations that relate the individual to groups to which he does not belong. These
groups seem to be particularly important in politics,
where a great deal of behavior is only indirectly of an
interactional kind.
person's goals, values, attitudes, and

opinions

may be

derived from and shaped

by

identifica-

and comparisons which are not immediately traceable to his group memberships, but for which
the existence of groups constitutes an environment of
tions, perceptions,

The more generalized concepts of "reference

groups' and "reference group behavior" are potent exreference.

planatory tools of political analysis.

The
havior

group be-

orientations characteristic of reference

may

be

affective, as in identification; cognitive, as

in definitions of a situation; or evaluative, as in

compara-

judgment. Identification with a political party, a reform movement, or a persecuted minority may lead to a
wide range of political action not explicable in terms of
either a person's roles or group affiliations; yet the existence of groups in the individual's symbolic environtive

ment may be

critical.

Likewise an individual

political situations to himself

by

or opinions of groups he values.

comes

a source of perceptions

may

explain

referring to the attitudes

The

reference group be-

and standards,

definitional

or normative, that affect political behavior. Finally, ref-

erence groups can serve the individual as a frame of comparison for evaluating himself and others.

own
who

Comparing

his

circumstances with those of others, a city resident

pays high tax rates

may move

into an unincorpo-

low tax

rates. These
between affective, cognitive, or evaluative
reference group functions are, of course, only analytic.

rated suburban neighborhood with

distinctions

In reality, reference groups

may

serve

all

three functions

simultaneously.

In a political world removed from one's

own

direct ex-

what one perceives as real or significant is likely

vary from one group context to another. Groups pro-

perience,
to

vide the individual with those definitions of the situation

The

[54]

Social Matrix

without which behavior, in politics as elsewhere, would

be highly arbitrary. Seeing one's own political opinions
and values shared by others will validate political reality.
This is probably true of the person only occasionally
concerned with politics, and even more so of the person
continuously involved the politician, the party worker,

the legislator, the judge, the bureaucrat, and so on.

Group

"belongingness" or reference group orientations not only

reinforce a person's political perceptions and beliefs, but

they also tend to shield him from

which he might

we

find unpleasant. It

political

is

experiences

common

observa-

we

wish to read, or that we

avoid contact with people with whom we might disagree.
tion that

read only what

great deal of research has

shown

that, especially in

perceptually ambiguous situations, group-anchored

atti-

tudes are projected onto situations that, in reality, are

quite different

from what they

are perceived to be. Ref-

erence group analysis does not pretend to explain

some groups rather than others

why

are chosen as foci of ori-

entations. It seems likely that this choice is related to the

second dimension in terms of which political behavior in
the social matrix can be analyzed, the horizontal ordering

or stratification of

human relationships.

THE HORIZONTAL DIMENSIONSTRATA

In politics and out, people do not interact with each other

beyond an

Even within
some members rarely meet
contact with other members. This is

identifiable range of relationships.

a group,

if it is

or have

direct

large enough,

partly due to different persons performing functionally

differentiated roles that

do not require direct

relations.

In part, the roles are differently evaluated and interaction between the incumbents of such roles is not considif interaction does occur, what the incumbents of functionally differentiated and differently
evaluated roles should or should not do their rights and

ered desirable. Or,

obligations

is

more or

less

strictly

regulated,

for

ex-

ample, in a military organization or a feudalistic social

The

Horizontal Dimension

Strcda

[55]

formation. But even in as intimate and small a group

as the nuclear family, the parents (and sometimes even
the husband vis-a-vis his

from each other

v^^ife)

and children are

set off

in clearly discernible ways. Similarly in

the huge group formations

we

usually call societies, cer-

groups are differentiated from others,

thus creating a system of relationships that would in all
likelihood not exist if functional or valuational differentain individuals or

tiation

were

absent.

Just as the concept of group alerts one to the vertical

pattern of social relations, so the concept of stratum calls

attention to the horizontal levels of the social matrix
within which individual and group relations take place.
The notion that political behavior and social relations are
is, of course, one of the oldest, most
and persistent organizing ideas of political and
thought. Built around it one finds numerous and

generally stratified
central,
social

diverse concepts such as formulations of caste, class, or

status,

in short, formulations of social structure as the

The concept
and empirically more
valuationally neutral and does

horizontal dimension of the social matrix.

of stratum seems to

me

analytically

serviceable because it is
not commit the analyst of political behavior to a particular theory that might seek to account for the sources of
stratification, and its effects on political behavior and the
whole system of relations in which people are involved.
To think of social or political relations, whether interindividual or inter-group, as stratified means that behavior is bounded by upper and lower limits beyond which
it does not extend. In general, status refers to inter-individual and class to inter-group differentiation, though
these concepts have no stable and unambiguous empirical
referents on which observers might be agreed. The distinction between status and class as a distinction between
individual and group referents seems to be useful in looking at a functionally very complex matrix such as that of
the United States. Here incumbents of high status positions
in groups of relatively low class character may yet interact with their opposites in groups of high class character*

The

[56]

Social Matrix

For instance, labor union leaders interact with industrial

managers more often than with their own rank and file.
The example suggests the utility of the concept of

new

stratum. Evidently a

stratum, the "higher servants"

of labor, business, or government, that

group nor a
has emerged

class in

is

neither a status

the conventional uses of these terms,

as a result of

such differentiation and col-

laboration.

Whether one

thinks in terms of status or class, then,

on behavior, including

the limits imposed

by

havior,

political

be-

a person's or group's location in a stratum of

may be more or less rigid or flexible,

depending largely on the total number of horizontal
levels that can be identified. Within any horizontal order,
whether characterized in terms of class or status, individuals or groups in neighboring strata are more likely to
come into contact with each other than individuals or
groups in widely separated strata. In a highly stratified
system with many levels of status difiFerentiation, there
is little contact between incumbents of status positions
that are removed from each other by more than two or
the social matrix

three steps.

On

the other hand, in a small entrepreneur-

manager type of system with few intervening

boss

is

in continuous contact

with

all

levels,

the

employees, even

if

they occupy different status positions.

The same is likely to be the case in the larger social
order characterized by classes. In the United States as a
whole, classes are difficult to identify, though most observers speak of a middle class and a working class, or of
white and blue collar classes. Due to lack of sharp differentiation, there
class lines

and

is

a great deal of interaction across these

class

is

less likely to

be

a critical factor in

behavior than in societies with more highly stratsocial orders. Where class stratification produces

political
ified

many

levels,

there

tween

classes

and

is

more

political

interaction within than be-

behavior takes on a more per-

vasive class character.

The

determination of any system of stratification

empirical

question

guided

by

theoretical

is

an

speculations

The Horizontal Dimension

Strata

[57]

about relevant

criteria of differentiation.

stratification

grounded

is

The

notion of

in the assumption that patterns

of behavior in the performance of social roles, within or

without particular groups to which people belong, are

This assumption is shared by most theorists concerned with stratification, though they dissocially evaluated.

agree on the criteria of evaluation, the character of the

ensuing horizontal structure, the rigidity or permeability

of strata boundaries and, last but not

least,

the problem

of "whose evaluations should supply relevant criteria of

stratification. Who is to say whether and what kind of
system of stratification exists in the social matrix?
In general, two approaches to stratification pervade the
bulky body of writings on the topic. For better or worse,
they are called objective and subjective approaches. Ac-

cording to the objectivists, the existence of status or class

and the order of stratification what is high and
what is low are immanent in the social order and quite

strata

independent of the feelings, attitudes, or perceptions that

may have concerning them. It is the task of the
observer, using whatever objective indices his theoreti-

people

cal propositions suggest, to stratify a population.

do

so, as

Marx

did,

by

He may

locating a person in terms of his

position vis-a-vis the relations of production or, as

modern

social scientists do,

by applying such

some

criteria as

income, education, or occupation. Although people objectively located in a stratum are assumed to share a
"consciousness of kind," certain "interests," or other subaspects are considered dependent
be used in ascertaining the system
of stratification. Indeed they are to be explained by stratification independently and objectively arrived at.
The subjectivists, on the other hand, start with the assumption that precisely because social evaluation accomjective feelings, these

variables that cannot

panies every functionally differentiated relationship, those

in the horizontal order are themselves best qualified to

supply criteria of stratification. Again the criteria are

thought to be immanent. The observer is only supposed
to find out just how people appraise either their own or

The

[58]

Social Matrix

others' location in the social order. Various methods have

been used to discover the subjective meanings that people
give to their class or status, but they need not concern us
here. Controversy over the vaUdity and reliability of
stratum indicators, objective or subjective, has been acrimonious. Again this is not the place to go into the merits

Whatever method

controversy.

of this

is

chosen, the

critical issue is

the utility of the notion of stratification as

a handle for

analyzing political behavior in the social

matrix.

This

issue

can only be

search.

The

usefulness of

depends on

its

through empirical reany method of stratification

clarified

ability to predict

cal behavior that

may be

consequences for

politi-

traced to a person's location in

the horizontal dimension of the social matrix.

One

of

these predictions might relate to an individual's choice of

the groups with

whom

he

whom

identifies or to

he refers

in his political conduct.

must be remembered that

It

of the

many

political

stratification

variables in terms of

behavior

may

be explained.

which

Any

is

only one

differences in

hypothesis that

seeks to explain a particular pattern of political conduct

is

always circumscribed

stratification involves

ues

is

by

some

several

As

always necessarily relevant. On the societal level,

two nations may appear to be very similar

systems in

class

from

a structural perspective, yet the valuational

ings given to both class and politics

ent.

other variables.

valuation, a group's set of val-

may

be quite

meandiffer-

As a result, political behavior related to stratification

two nations may differ a great deal. For instance,

in the

recruitment practices into the poHtical

elite

may

vary.

might be largely recruited from the upper

social stratum in one system; in another, they might be
recruited from the lower strata. In the first system, public service might be looked on as an obligation; in the
second, as a sinecure to be exploited for personal advanCivil servants

tage.

For reasons such as these, hypotheses about the consequences of stratification for political behavior are un-

The Horizontal Dimension

Strata

[59]

be universally valid. The general proposition

that stratification and political behavior are significantly
related is probably viable only to the extent that analysis
likely to

is

what

contextual. Just

particular patterns of political

behavior will be found in any social matrix analyzed in

terms of stratification cannot be predicted from the structural characteristics of the matrix alone.

Insofar as politics
a

group

^whether

is

purposive activity through which

a national society at

one end of the

matrix or the nuclear family at the other end engages

in collective decision-making, it is generally considered
to have the attributes of self-consciousness as a group and

Both of these are very vague concepts. But we

whatever the particular system of stratification, the existence of strata makes for self-consciousness
and the specification of interests. The degree to which
interest.

assume

this

The

that,

so is always a matter of empirical determination.

notion that a stratum, whether a class or status

is

group, develops consciousness of kind stems from the observation that

it

is

a Jiecessary correlate of interaction.

People in constant contact with each other but cut off

from others are likely to see themselves as being the same
and to see the world around them in similar perspec-

They

view their interests in similar

behave in politically like ways to advance these interests. Out of speculations such as these
Marx built his theory of class consciousness, class interest,
and class conflict. But whether stratification makes, in
fact, for consciousness of kind and common interests and,
tive.

are also likely to

ways and they

will

therefore, for similar patterns of political behavior, can-

not be taken for granted.

It

must always be subjected to

empirical testing.

Much
political

the

in

social order
it

between

stratification

is

open or

closed, that

is,

and

the extent to which

permits or prohibits, facilitates or impedes social

bility.
is

relationship

behavior depends on the extent to which a given

The

many factors, ranging from material concustomary practices and ideological perspec-

a function of

ditions to

mo-

rigidity or flexibility of a stratification system

The

[6o]

The

tives.

Social Matrix

relatively high degree of social mobility in the

United States for both individuals and groups has impeded the development of class consciousness, at least in
the Marxian sense, and made politics relatively free of
class considerations, but only relatively so when compared with societies where stratification is more rigid
and pervasive. This is only another way of saying that,
compared with other systems, stratification is less relevant in the United States in explaining variance in political behavior.
It

does not

politics.

mean

that class

is

community

American
on voting behavior,

irrelevant in

great deal of research

and recruitment
have more to do
with politics than the American ideology of mobility

pressure

politics,

practices

shows

and equality led

politics,

may

that stratification

earlier observers to suspect.

hand, the same studies show that

it is

On

the other

relatively easy for

individuals or groups in American society to identify

with the values and symbols of the elite, even if they cannot actually change their place in the order of things.

Such

across

identification

strata

boundaries makes for

limited awareness of class interests in a collective sense,

with obvious consequences for

political behavior.

BIBLIOGRAPHY
ROLE ANALYSIS
Ward

and McEachem, Alexander W.,

(New York: Wiley, 1958.)
The Theory of Social Structure. (New York:

Gross, Neal, Mason,

S.,

Explorations in Role Analysis.

Nadel, S. F.,
Free Press, 1957.)

Wahlke, John C, Eulau, Heinz, Buchanan, William, and Ferguson,

LeRoy C, The

lative Behavior.

Legislative Syste?n: Explorations in Legis-

(New

York: Wiley,

1962.)

Bibliography

[61]

Labor Unions: Organizaon Professional Roles. (New York: Free

Wilensky, Harold

L., Intellectuals in

tional Pressures

Press, 1956.)

GROUP ANALYSIS
Bentley, Arthur F.,
Social

Pressures.

The

Study of
Process of Government,
University of Chicago Press,

(Chicago:

1908.)

Golembiewski, Robert T., The Small Group: An Analysis of

Research Concepts and Operations. (Chicago: University of
Chicago Press, 1962.)
Katz, Elihu, and Lazarsfeld, Paul F., Personal Influence. (New
York: Free Press, 1955.)
Olmstead, Michael S., The Small Group. (New York: Random
House, 1959.)

Truman, David B., The Governmental Process: Political Interests and Public Opinion. (New York: Knopf, 1951.)
Verba, Sidney, Small Groups and Political Behavior: A Study
of Leadership. (Princeton: Princeton University Press, 1961.)

STRATUM ANALYSIS
Barber, Bernard, Social Stratification.
Brace, 1957.)
Centers, Richard,

The Psychology

(New

York: Harcourt,

of Social Classes. (Princeton:

Princeton University Press, 1949.)

Eulau, Heinz, Class and Party in the Eisenhower Years.

York: Free

(New

Press, 1962.)

Janowitz, Morris,

The

Professional Soldier.

(New York:

Free

Press, i960.)

Marvick, Dwaine (ed.). Political Decision-makers: Recruitment

and Performance. (New York: Free Press, 1961.)
Namier, Sir Lewis, The Structure of Politics at the Accession
of George III. (New York: St. Martin's Press, 1957.)

CHAPTER THREE

THE CULTURAL

CONTEXT

Like the

air

we

breathe, culture,

close, so natural, so

much

We

are not aware of

it.

from one culture

into another.

pect of culture

that

is

notice

its

it only when we move

The most important as-

existence

existence of another culture.

tutes the

however defined, is so
what we are that we

a part of

One

is

predicated on the

culture always consti-

environment for another. Only awareness of

ferent cultures sensitizes us to the culture in

live,

ture

act, think, feel,

bound

Culture

and judge. Otherwise

dif-

which we

we

are cul-

we do not know who we are nor why.

is

all

pervasive.

It

permeates,

without our

awareness, every aspect of our daily living. Political be-

havior
unless

is

we

no exception.
treat

tural context.
cal behavior

it

We

culturally

cannot fully understand it

and locate it in the wider cul-

This enables us to see

is

how

similar to or different

havior in another culture.

The

our

from

own

politi-

political

be-

ideal of scientific objec-

Cultural Patterns

[63]

tivity requires cultural analysis of political behavior, if

for no other reason.

The
of

modes
in the

behavior in terms
and language-related

difficulty of analyzing political

culture

is

that

our

language

of thought, themselves products of culture, stand

way.

We

is this or that, or that it

not a thing nor can it act.
It does not exist outside or apart from behavior as something that can be identified independently of behavior or

say that culture

does this or that; but culture

is

the results thereof, regardless of whether these results are

symbolic or material. Culture is a mental construct abstracted from the behavior of people and their works.
Despite this, it is possible to speak of a culture because,
as I suggested, it makes sense to speak about culture only
if we can specify the existence of more than one culture.
At the risk of ambiguity, I shall continue to use the
term "culture" without defining it for the simple reason that, among the host of over 250 available definitions,,
I know of none that is neither too narrow nor too inclusive. Ambiguity is not necessarily a drawback here. As a
concept, culture is a synthetic rather than an analytic
construct, and I wish to use it to stimulate investigation,
not to conduct it. In an essay of this sort, conceptual refinement may be more of a handicap than an advantage.
More important when looking at political behavior in
cultural terms, we are still observing the same behavior
that we observe when we use social terms. It is not the
behavior involved that is different but our observational
standpoint.
are simply moving to another level of
analysis. Just as the notion of social matrix and related
concepts are useful in viewing politics in social terms, so
the notion of cultural context and related concepts
call our attention to the cultural environment of politics

We

and to

politics as a cultural

phenomenon

itself.

CULTURAL PATTERNS

We

speak of culture patterns when we observe similarithe behavior of people in the same culture that are

ties in

The Cultural Context

[64]

relevant to

Though

diverse,

functional aspects of social action.

economic, military, or

their religious,

political

practices serve quite different functions in a group's* existence, the patterns of

similar.

For

may be
group's

behavior involved

may

characteristic of different functional areas of a

child rearing, religious practices,

life:

organization, or political decision-making. If

patterns that

similar

time (or,

if

at least for

chance.
that

be very

instance, authoritarian patterns of behavior

are

and

pervasive

economic

we

stable

observe

through

they change, do so gradually), we can assume,

working purposes, that they are not due to

We attribute the similarities in behavioral patterns

functionally

are

specialized

to

the

working

of

culture.

we

Suppose

ask

why

so

many

people in the United

States are relatively uninvolved in politics?

compare

invites us to

with

political

can political

The

question

behavior in this country

we need not necescomparison in order to look at Ameribehavior in cultural terms. Even without it,
behavior elsewhere. But

make such

sarily

political

the concept of culture suggests that the patterns evident

in

mass

political

behavior

may

also

be present in other

functional areas of social behavior.

We

may

try initially to strengthen our observation.

It

impresses us that political interest and involvement rise

fall off even more
Moreover, the interest seems to be
more like a spectator's than an active participant's. The
idea of spectator interest makes us look for other functional areas where spectatorship seems to be the typical

rapidly as an election approaches, but

rapidly thereafter.

response pattern.
It

has often been suggested that American politics has

the characteristics of a game. Indeed, a perceptive journalist

once called

analogy
* I

am

itself is a

it

"the great American game."

cultural artifact,

The

and our language sup-

"group" in the most generic sense of the term as

a nation, a region, a community, a tribe, an
any collectivity whose members
organization, a social class, a committee
are in more or less direct and permanent contact with each other.
referring

using

to

society,

Cultural Patterns

[65]

Like baseball or any other game, politics is

It is played by "rules of the game" that must
be observed and professionals play the best game. Admirports

it.

"played."

ation for a job well done and emphasis

on winning are

We

part of the general behavioral pattern.

cheer the vic-

Those who also ran are the forgotten men of American public life. The game, in sports as
tor, forget the loser.

in politics,

is

The

partly for entertainment.

spectator pat-

tern of American mass behavior, mirrored in the lan-

guage of

sport,

tice, too, that

is

a pervasive feature of the culture.

we "watch"

(No-

the stock market and "play"

it.)

Admittedly, these are speculations based on the most

have introduced them
only to illustrate that behavior is likely to be patterned in
more or less consistent ways in different functional areas
of life within the same culture. This notion, in turn, leads
superficial kinds of observations. I

us to view behavior as an integrated whole rather than a

set of disparate items.

This does not mean that different

of behavior

functional aspects

are

necessarily

related;

may be quite disconnected. Politics and economics

may or may not interpenetrate. Politics and leisure may
or may not be related. Yet even if not connected structhey

turally, the

same or

similar cultural patterns will pervade

and analysis of these

different functional areas. Discovery

patterns

of course, only a

first

step in the cultural anal-

human behavior.

ysis of

To

is,

avoid any misunderstanding,

am

not saying that

behavior within a group that shares a culture

is

all

charac-

The concept

of cul-

ture merely suggests that viewing behavior as a

whole

terized

may

by

help

identical patterns. It

in

explaining

is

particular

Moreover, there are enormous

cultural analysis as

not.

items

difficulties

we move from

of
in

behavior.

employing

relatively small, func-

tionally undifferentiated, simple groups, such as preliterate societies or even an adolescent

plex,

functionally highly

urban gang, to com-

differentiated

modern

collec-

tivities.

Although some sweeping and ingenious characteriza-

The Cultural Context

[66]
tions have

been made of American culture in

ogeneity, they are

little

more than

all its

heter-

impressionistic.

The

complex cultures involves many theoretical

and methodological difficulties. But this is not an argu-

analysis of

ment

against the use of the concept of culture in

modern

settings. In the first place, if discontinuities in behavioral

patterns outweigh the continuities, this

datum

is itself

a cultural

to be taken into account in the analysis of political

behavior in a complex, cultural context. And second, the

very useful concept of subculture is predicated on the as-

sumption of the existence of a common culture. This

common culture may prove to be empirically elusive,
leading to the kind of impressionistic analysis that has

been made of American culture, but

assumed nevertheless. The technical

its

existence

must be

difficulty of discov-

ering cultural patterns in diverse areas of behavior must

not be confused with the absence of patterns.
The culture concept puts the accent on wholeness,
which guards us against taking behavior in a functional
arena be it sports, religion, the economy, or the polity
out of its cultural context. The failures of attempts to
transfer political institutions from one culture to another
are well known. In these cases, behavioral patterns could

not adjust or be adjusted to externally imposed arrangements. Similarly, within a group it is risky to view political institutions and processes outside of the cultural
context in which they are embededed. Hoiv individuals behave in functionally different areas of activity is likely to
be more continuous than what they do. For this reason,
the variance in the political behavior of different groups
may be illuminated by information about their behavioral
patterns in other functional activities, that
cultural
least

it

litical

comparison of

will help us avoid

similarities

and

is,

by

cross-

differences.

At

making generalizations about po-

behavior, whether parochial or universal in scope,

that are necessarily culture-bound.

The

from one funcfrom politics to economics,

crime, from crime to sports, and so on,

pervasiveness of cultural patterns

tional area of life to another,

from economics

to

Cultural Patterns

[67]

suggests that cultural analysis

may be

particularly use-

understanding public policies and the manipu-

in

ful

lative aspects of political behavior. All too often, social

problems in the real world, graft and corruption, apathy

and resistance to reform, social justice and peace, and so
on, are seen as if they were soluble only on the political
level of action. Cultural analysis suggests that these prob-

may

be deeply rooted in a group's total way of life,

more than treating them
by political means. Often these political remedies are
lems

and that

their solution requires

themselves so rigidly prescribed

by

cultural

do's

and

they must remain ineffective. Cultural analysis

can serve as a potent diagnostic tool of political behavior,
diagnosis being preliminary to prognosis and political
don't's that

therapy.

We

speak of cultural patterns of behavior

when they

are widely shared, rather than social or personal patterns.

The

we can
have never

degree to which they must be shared before

say that they are

is

an interesting question.

encountered a very satisfactory answer. Knowledge of

the extent to which cultural patterns pervade different
functional areas

of interest in defining the boundaries

is

of a culture. But

it

unlikely that in a world of close

is

intercultural contacts the boundaries can be clearly specified.

Patterns

weave

across cultural boundaries, causing

similarities in behavioral patterns.

tion of

how much

sharing

is

As

a result, the ques-

necessary for us to speak of

cultural patterns remains moot.

More

critical in appraising the cultural factor in politi-

cal behavior

is

the degree of stability in the patterns. This

stability of behavioral patterns

is

probably the most

tell-

ing evidence of the working of culture in politics. Moreover, the rate of change that can be observed in be-

havioral patterns
too,

we

change

is

itself

a cultural

may

reach a point where

to speak of patterns of behavior at

I

phenomenon. Here,

are facing subtle technical problems.

cannot say. In any

it is
all.

case, increasing

havior suggests cultural

crisis

The

rate of

hardly permissible

What

this

point

randomness

is,

in be-

and possibly cultural trans-

The Cultural Context

[68]

From

formation.

this

point of view, political institutions

For by definition, an
widely shared, regularized patterns
of behavior that are fairly stable through time. Institutions are, of course, only the most overt aspects of
are cultural products par excellence.

institution

cultural

a set of

is

patterns

in

we

In general, then,
political

behavior

political

behavior.

components of

are equally crucial

if

Covert patterns

a political culture.

can speak of cultural patterns in

the patterns are similar to other pat-

way of life, if they are widely shared

and change only slowly. How these patterns have come
about and how they are transmitted are questions of conterns in a group's

siderable interest to the cultural analyst of political be-

The most plausible hypothesis is that they are the

cumulative results of learning, and that they are transmitted from generation to generation in the process of
havior.

The

socialization.

patterns are a group's adaptations to

the environment in which

it lives,

and they change

as

new

forced to adapt to changing environmental conditions, including cultural changes in the engenerations

are

vironment of other
there

is,

the

more

cultures.

Political behavior research

the

processes

The more

cultural contact

cultural patterns are likely to change.

of political

is

increasingly concerned with

socialization

and the conse-

quences of socialization for the maintenance of political

institutions and the more informal patterns of political
conduct.

CULTURAL ORIENTATIONS
have suggested so far that we can look at patterns of poand other behavior as outside observers who infer
uniformities and regularities from what they see. If ive
name the patterns, as I have when speaking of an authoritarian pattern or a game pattern, the procedure is alI

litical

together arbitrary.

Our

observations as outsiders are, of

course, cultural products also

in the case of the

observer of the cultiu'e of science.

tic

of the scientific culture

is its

The

scientific

critical characteris-

cognitive intersubjectiv-

Cultural Orientations
ity:

that

[69]

among

the agreement

is,

scientists that the

ob-

phenomena are, in fact, what they are alleged to

To make this agreement possible, science has de-

served
be.

veloped

its

Presumably

own
if

patterns of behavior,

its

rules of

method.

these rules are adhered to, the same obser-

vations will be

ferences will be

made and, more important, the same indrawn from the observations. We say that

the methods of observation, the observations themselves,

and the inferences from the observations have meaning.

Likewise, the patterns of behavior in other functional

meanThese meanings enable people

behavior and coordinate their

areas of life at the core of cultural analysis have

ings to people themselves.

to understand each other's

conduct. Political behavior, as

has not only
actors

who

all

other

form and content but

play the

were observed only

game of

human

its

behavior,

meaning for the

politics. If political

in terms of

the observations so obtained

also

behavior

content and patterns,

would be very

partial

and

misleading. Political behavior, though similarly patterned,

may

have quite different meanings to the actors whose

behavior

is

observed.

Much

traditional political inquiry

it was limited to
and took the meaning content

has been purely formal in the sense that

the observation of patterns

of behavior for granted. Political institutions or constitutions

were described and

what

these patterns

the people involved was not investigated.

ings

were

and difmeant to
Rather mean-

their formal similarities

ferences were noted, but

ascribed, usually

on the

basis of the observer's

own cultural understandings.

Cultural analysis of political behavior cannot be limited
to objective observation of patterns. It

is

concerned with

the meanings that people give to these patterns, regardless

of whether they are formally institutionalized or not,

including the functions that they see as being served

politics.

Such

analysis

is

by

impossible unless the individual

person, as bearer of cultural meanings,

is

the unit of

highly literate and articulate cultures, the

meanings given to politics can be studied by analyzing
analysis.

In

The Cultural Context

[70]

documentary evidence. Content

analysis

now

serves as a

useful tool of inquiry into verbal symbols and their manifest as well as latent

Documentary

meanings. But there are

difficulties.

may

not be sufficiently explicit to

permit the discovery of meanings. And they may be
neither typical nor representative, but rather parataxic,

more

materials

indicative of the writer's idiosyncratic interpreta-

tions than of widely held cultural understanding.

The meanings
highly

conscious,

may be
They
they may

people give to their behavior

subconscious,

or

unconscious.

may be so sacred they forbid articulation, or

be highly secular and profane. Whether they are one or
the other or in between is likely to depend on a culture
group's exposure to other cultures.

The more

group, the more ingrained and traditional

its

isolated a

behavioral

patterns are likely to be and the more unconscious the

meanings given these patterns. At the other extreme,
groups in close cultural contact are likely to be highly
conscious of the meanings they attribute to their own behavioral patterns and those of their cultural neighbors.
Homogeneous groups take their patterns for granted;
heterogeneous groups are forever in search of meanings
that give their behavior symbolic significance.
Behavioral analysis of politics cannot afford to ignore
the meanings that people give their own behavior because
these meanings have consequences for their own political
actions and those of others. Meanings are the organizing
principles of behavior through which people make themselves at home in the world and orient themselves to action, including political action.

If,

like behavioral patterns,

meanings are widely shared and relatively

stable,

possible to construct a group's cultural self-image.

it

is

How

people orient themselves to the groups to which they

belong and those to which they do not belong gives meaning to their behavior. Indeed social behavior is impossible
without the existence of widely shared, stable, and agreed-

on meanings that serve as a frame of reference in conduct. Meaning, figuratively speaking, is the glue that

Cultural Orientations

[71]

binds people together in groups and divides them from

others.

In general, cultural analysis of political behavior can

between various types of meanings in

terms of a group's orientations to political action. Some
easily discriminate

of these orientations are cognitive. These are a group's

perceptions of reality in terms of M^hich

it

defines

its

en-

vironment. Such perceptions can range from highly scientific and sophisticated concepts at one extreme to the

most

from

tive beliefs.

How

make

and

superstitious

at the other,

irrational

scientific

notions

knowledge

people behave in

about reality

to the

most primi-

politics,

how

they

those collective decisions that guarantee their sur-

vival as a group,

is

powerfully influenced by their image

of the world they live

Diff^erences in

in.

image from one culture to another are

potent determinants of political conduct.

Whether

the

environment is seen as friendly or hostile, whether expectations about the future are optimistic or fatalistic,
whether group demands are pathologically exaggerated
or minimized, orientations of this sort are critical frames
of reference for political behavior. The reason for such
differences

is

of central interest to the cultural analyst.

However they may have come

them

research must treat

planatory variables.

The
own

as

about, political behavior

potentially

belief that

man

important excan shape his

world through his

eflrorts will lead to a different
kind of political action than the view of man at the mercy
of forces over which he has no control. A culture group
whose time perspective is in the past may find it difficult
to cope with the present, just as a group whose perspective is largely in the present may be unable to plan wisely
for the future.

Orientations to political action are also affective.

we

Here

deal with those widely shared emotional responses to

which we give names

These emoand they may be

directed in different degrees towards different objects.
tions

may be

like loyalty or morale.

positive

or negative,

The Cultural Context

[72]

Whether man
is

relates himself to others in terms of what

conventionally called love or hate, trust or fear, is

likely to

have significant consequences for

his

political

behavior. Insofar as one or another of these orientations

widely shared and more or less accentuated in a given

makes an enormous difference in politics. It will
produce notions about other political actors, whether inis

culture

dividuals or groups, as friends or enemies, as objects of

identification or withdrawal.

enced

Whether

politics

is

experi-

as a life-and-death struggle for existence, as a

com-

though friendly game, or as a cooperative enterprise for mutual aid, is likely to be a function of widespread affective components in the culture.
petitive

Finally orientations to political action are normative.

Included here are the evaluative standards by which
judgments are made about political behavior, the ethical
rules that guide behavior, as well as the goals considered

worthy of attainment. Some of these orientations may be

actually operative ideals whose observance is not in
doubt. Others may be ideal standards that differ from
what people actually do when they guide or evaluate

own and others' conduct. In politics, the degree of

agreement on ideal patterns of conduct is probably an important stabilizing and legitimizing factor in actual behavior. Conflict between ideal patterns and operative
their

ideals

becomes

a central source of political tensions that

call for solutions.

ideal

Behavioral norms and values, whether

or operative, are crucial items

on the behavioral

agenda of inquiry into politics.

Just as meanings are implicit in patterns of political behavior, and transmitted by the members of a group, so are
cognitive, affective and normative orientations. Insofar as
meanings and orientations constitute a more or less coherent and integrated set of symbolic expressions, they
come to be accepted as self-evident systems of belief.

By

belief I

mean

the non-logical, pre-rational

compo-

nents of the total cultural ideology. Belief in this sense

is

not rooted in reasoning and direct observation, and it is

highly resistant to disproof. Belief can only be given up

Cultural Orientations

[73]

exchange for another behef. Beliefs are the self-evident

knows to be true without
need for further proof. Putting these propositions in
doubt arouses hostility, causes pressure for conformity,
or leads to excommunication in rigid cultural contexts.
in

propositions that everybody

From

the cultural point of view, the analytical problem

not whether beliefs can, in fact, be tested by the

is

methods of

social science or at least dissected

by

the cri-

teria of logic. Cultural orientations to action differ in the

degree to which their components can or cannot meet

the requirements of proof or logic. Rather the point

is

that propositions of this order are believed because they

are

widely

shared,

relatively

stable,

and

successively

transmitted and, in turn, give direction to social and political action.

systems are characteristic

of even highly soand secularized cultures. That one's own form

of government is far superior to any other; that only a
particular economic system can guarantee a high standard of living; that war is rooted in human nature and cannot be prevented; that one group is out to destroy
another, these are some of the beliefs more or less widely
shared at one time or another, even in cultures that take
pride in rational conduct. Often these beliefs have only
symbolic functions with few consequences for actual political action. But under certain conditions, for instance in
Belief

phisticated

crisis situations
liefs

when

the group's survival

is

at stake, be-

have an important effect on the political order. In the

American South, the

belief that

childlike, overemotional, short

tive, docile, easily frightened,

on

Negroes are impulsive,

intelligence and initia-

or incredibly superstitious

has been the most pervasive single factor in political be-

That

this belief in "white supremacy" is neither

nor subject to disproof does not negate its effectiveness in giving meanings to political orientations as

havior.

rational

well as actions.
Analysis of political culture is not concerned, then,
with the truth of beliefs but with the functions they
perform. Precisely because beliefs are grounded in emo-

The Cultural Context

I74I

tion rather than logic and because they are experienced as

certainties rather than
litical
fies

probabiHties, orientations to po-

action grounded in a total system of belief that de-

disproof have a strong influence on political cohesion.

mutual idenand the dangers,

real or imaginary, of an environment that must be controlled. For this reason, threats to belief systems arouse
the most passionate and violent types of reaction. They
Belief systems serve to give people a sense of

tification, to protect

them

against doubts

are considered as equally necessary to the survival of the

individual as of the group. There is generally a discrepancy between the degree of sacredness with which belief
systems are invested and the realistic functions that the
systems and their component orientations may perform
as integrating or disorganizing elements in the total cul-

tural configuration of political behavior.

CULTURAL CHANGES
Apart from socialization, the sharing of political beliefs
and norms within a culture is due to those pressures towards conformity stemming from man's relatedness to
man. This interdependence gives rise to expectations that
become cultural norms, though at first they may be only
statistical regularities. The pressures to conform, then, are
not somehow external to behavior but characteristically
cultural

components of behavior. In

sures are greatly accentuated

by

politics, these pres-

coercive sanctions pur-

posefully designed to assure survival of the community,

institution, organization,

or small group. But the cultural

perspective also warns us against overestimating the coerciveness

of

political

analysis of politics as a

power. The more traditional

power phenomenon overlooked

the pressures towards conformity that are cultural and,

To

be a loyal
and
so on, are cultural expectations.
act as we do because
we have learned to do so without questioning it. If our
actions were widely questioned^ requiring enforcement
therefore, are not experienced as coercive.

citizen, a faithful partisan, a conscientious taxpayer,

We

Cultural Changes

by

[75]

loyalty oaths and other types of coercive assurance,

would likely be in crisis.

Emphasis on the pervasiveness of cultural factors in
political behavior must not be thought of as cultural dethe political culture

More

terminism.

often than not, a deterministic interpre-

human pursuits, is due to a

between culture as a kind of thing
and individual behavior. This fallacy leads to the notion
of culture as an independent force, governed by its own
inner laws in the face of which man is helpless.
On the contrary, since culture and all that the concept
tation of politics, as of other

fallacious distinction

implies
beliefs,

structural patterns, functional actions, meanings,

norms are abstracted from the behavior of man,

they are surely subject to change by man. If culture is

manmade rather than "natural," the individual is free to
change it. Whether he will probably depends on the degree to which cultural patterns, beliefs, and norms satisfy

man's needs and the adjustments he has to make visa-vis the natural environment. It is only when cultural
patterns and meanings become rigid that they turn out to
be obstacles to change needed in the face of new developments.
call this cultural lag. Political behavior pat-

We

terns

and

institutions are particularly subject to cultural

lag because they are sanctified

by customs and

traditions.

Patterns of political behavior are difficult to break. All

the same, the cultural perspective in the study of political

behavior

is

whom

of value not only to the political scientist

provides contextual relevance, but also to the

policy-maker who is constantly concerned with orienting behavior in the direction of preferred group goals.
Culture is not a conceptual strait jacket. Cultural
phenomena are only one set of factors affecting political
behavior. Choice, in politics as in economics or in private
decisions, though circumscribed by what is socially posfor

sible

it

and culturally permissible (and,

dividually acceptable),

is

ever present.

might add,

The

in-

reason for

enough: cultural patterns, meanings, or

is simple
norms, though widely shared, are rarely transmitted in
an absolutely uniform manner. This is, of course, truer
this

The Cultural Context

[76]

of complex groups than simple ones. In general, free-

dom

of choice in behavior

is

complex the group's web of

plex

many

groups with

likely to

be greater the more

But in com-

social relations.

where

subcultures

individuals

rarely share the same set of interpersonal relations, each

person tends to be exposed quite differently to a great

variety of cultural and subcultural cues that, in their particular combinations,

make

for a

more or

unique be-

less

havioral expression.

Moreover, the cultural stimuli impinging on

course of socialization into

new

inconsistent

complex,

particularly

in

Apart from contributing to the

experiences

man

in the

are

quite

modern

individual's

groups.

freedom of

choice, these inconsistencies are also sources of cultural

As
The

in a state of con-

transformation.

a result, culture

tinual change.

rate of change differs

culture,

The

in turn

affecting

changes themselves

sistencies,

social
arise

is

from culture

out

of

incon-

cultural

of norms between subcultures

conflicts

to

patterns and relations.

and

contacts with alien cultures. Political cultures are subject

to the same transformations.

The

transformations

may be

rapid or slow, depending on the cultural configuration as

and political interacand individual preferences. In turn, politics in modern secularized groups may itself become an element of
either integration or decomposition in the culture as a
a whole, the requirements of social

tions,

whole.
Culture, including political culture,
best observed

Whenever

when

group

is

is

dynamic. This

new group comes

is

into existence.

formed, whenever individuals come

into effective contact with each other, the

members

of the

problems of mutual adjustment to

each other and to non-members. It is difficult to imagine,

group face

similar

of course, that already available cultural solutions will

not be taken from the larger culture. But the group would
probably not have formed in the first place if all sought
after solutions

tural

were already embodied

in the group's cul-

model. In other words, the group's behavior

search of

itself is

in

not as yet consensually validated. Be-

Cultural Changes

[77]

havior in such a culturally emerging group

is

tentative

and exploratory. Mutual identifications and commitments to group goals are at first limited. Only gradually,
as

the

members come

to

know

each other, will rules

emerge for dealing with the problems that brought the

group into existence and for regulating the group's conduct. Analysis of the formation of groups can give us

much

dynamics of

insight into the

When

political culture.

conceived of as emergent, our attention is called to the time dimension of political behavior. Although most political behavior research is conculture

is

ducted in the present and often lacks historical depth,

and de-

cultural analysis cannot avoid noting the genesis

velopment of

patterns and beliefs in different

political

In tracing the emergence and evolution of traditional

eras.

the

patterns,

behaviorist

political

rather than historically oriented.

He

is

developmentally

is

not interested in

own sake. Rather he tries to

and possibly explain how present patterns of political behavior have come to be what they are. Developmental reconstruction is difficult. Dead men do not talk
and heavy reliance must be placed on documentary evidence that may not be representative of relevant past
patterns. If regularity in political behavior is to be more
past occurrences for their
trace

than an assumption, long-range

as

well as short-range

trends must be investigated, both to give time-dimen-

and to make comparison possible

sional depth to analysis

across

continuous

change

is

patterns

periods

time.

and their transformation

more rapid shifts in

result from purposive,

those
that

of

Because

cultural

relatively slow, the analysis of political culture

litical action,

in turn,

is

aids

in

pinpointing

individual or group behavior

and social action. Pohave an impact on the

political

likely to

more general culture. The active intervention of the

American Federal government in Southern affairs, for
and determinable consequences not
only for the politics of the South, but for Southern cul-

instance, has decisive

ture in general.

Cultural

analysis

is

necessarily

comparative,

even

if

TJie Cultural Context

[78]

not made intentionally. Deliberate coman important "only," of course

what otherwise remains latent and, as a result, may
cause misleading inferences. From the comparative stand-

comparison

is

parison only articulates

must

point, the discovery of similarities

the specification of

background of
served. This

abide

by

is

similarities

a point of

It

method

in the study of politics.

definition,

is

especially difficult to

For

politics,

commits the observer to the

clarification

logically precede

only against the

that differences can be ob-

differences.

almost

specification

of differences rather than similarities.

conflictive aspects of politics point

by
and

The

up how persons or

groups involved in a political relationship differ, not how

they are alike in characteristics or goals.
Looking at politics in cultural terms is a compensatory
remedy. If two things are to be compared, even though,
as in politics, their differences

seem

blatant,

sary to assume a tertium comparationis,

more or

cal notion,

less explicit,

of just

it

is

neces-

some hypothetiwhat it is that is

to be compared. This third element, a purely hypothetical

construct,

must contain aspects

being compared. Without

comparison

is

this

impossible.

If

common

to both things

assumption of commonness,
cultural analysis is under-

taken self-consciously with comparison in mind,

it

leads

to the discovery of similarities as a necessary condition

for seeing differences.

These considerations are

in line

with the conventional

assumption of behavioral science that uniformities can

be identified in the conduct of individuals and in the
functioning of institutions, and that these uniformities are

due to

common human and social factors which transcend

The fact that all groups have some form of de-

cultures.

cision-making structure through which collective problems can be solved is more significant than the fact that
the governing function is performed in one group by a
council of elders, in another group by an autocrat, or in a
third group by an assembly of all adults. But treatment of
a particular institution does not take us very far if analysis
is

limited to the discovery of similarities in functions per-

Political Culture

by

formed
analysis

the

all

[79]

different

more

structures.

makes

This

cultural

useful as a tool of political inquiry:

always assumes that between different groups states,

tribes, parties, households or other decision-making collectivities
there are also likely to be differences in political
behavior patterns, beliefs, or norms that are
culturally conditioned and, therefore, fully understandable only within the particular cultural context in which
they are embedded. In this regard, then, cultural analysis
is both a requisite of and corrective for comparative poit

inquiry.

litical

POLITICAL CULTURE
The concept
dox.

of a political culture is something of a parahave argued that cultural analysis cannot success-

on

fully focus
total

way

of

a particular functional area of a group's

life

without simultaneously focussing on

other areas. Yet the concept of political culture seems to

limit our attention to those patterns of behavior and those
orientations

that,

presumably,

characterize

one

only

aspect of a group's existence. Such a self-imposed setting

of

seem to make

boundaries would

it

impossible

to

specify and measure the salience of politics within the

larger culture and the consequences of different political

patterns and orientations

the task of analysis

culture

is

politicized.

would seem
untenable

as a

whole. But

As

which

the concept of political culture

to imply, a partial cultural analysis

if

from which

on the culture

to delineate the degree to

is

is

really

one accepts fully the premises of method

cultural inquiry proceeds.

But cultural

analysis, like all scientific inquiry,

always

involves a process of theoretical abstraction and empirical

isolation.

The concept

of political culture does not deny

that the political culture

In short,

it

is

embedded

in a larger culture.

must be interpreted to refer to

a subculture,

only for convenience sake that we do not speak

of political subculture (as we should) or political subsubculture, the notion sugsubculture (as we might).

and

it is

The Cultural Context

[8o]
gests,

shares certain patterns and orientations with the

from which it is analytically (but certainly

not empirically) set off by other patterns that are relatively unique to itself. There may even be a conflict of
larger culture

patterns between the political subculture and the

more

general culture. For instance, the hierarchical patterns of

military behavior

may

conflict with egalitarian patterns

in the general culture. Indeed, observable tensions in the

American military establishment may be traced to

between the military subculture and the general

conflict
culture.

Sacred orientations in the political culture, like its fetishes,

conflict with the generally secularized views of the
culture as a whole. Resulting tensions on the cultural

may

level of analysis

may

be mirrored on the social and per-

sonal levels, thus yielding further insights into individual

and inter-individual behavior.

Cultural

analysis

of

political

behavior

seems

rnore

group (using the term in its most generic sense

to refer to an institution, a local community, a social class,
viable in a

religious order, a geographical region, an age group,

and so on) than in the larger societal contexts where it

is usually employed. In fact, within the United States, for
instance, fruitful subcultural analyses of political behavior

have been made

groups as diverse as the Senate, a small

a slum area of a metropolitan city, the
military establishment, a democratic trade union, an oldage movement, a Japanese relocation center, and many
village

in

community,

other political groups or settings. Though

groups share the larger culture, each exhibits
terns of behavior, including political ones.

all
its

And

of these

own

pat-

there are

subcultures within subcultures as there are groups within

groups.

Even

the smallest social unit

the clique, the

culture. In

work

the nuclear family,

has own

team, the committee

most respects

its

a dual cultural function

is

per-

group transmits not only its

own peculiar ways of thinking and doing things, but also
serves as a kind of conveyor for transmitting the more

formed

in these groups: the

general cultural patterns.

Political Culture

The

[81]

existence of political subcultures in particular po-

whether formally institutionalized or not,

always an empirical question. It cannot be prejudged.
But the notion of subculture as the more general concept
litical situations,
is

of culture alerts the analyst of political behavior to the

postem from subcultural differences. In a

country like the United States with its high degree of social intercourse and mobility, both vertically in the group
order and horizontally in the stratification order, there is
possibility that otherwise non-explicable variances in

litical

may

action

a great deal of cultural

homogeneity among subcultures.

Differences in political behavior due to subcultural differ-

ences

may

be

subcultural

difficult to observe.

political

contrasts

But on close scrutiny,

will

not escape the be-

havioral microscope. Indeed, any empirical construct of

the

American

political culture will

have to be built on

minute, subcultural analyses of political behavior.

Political culture, then, refers to the patterns that

inferred

from the

political

can be

behavior of groups as well as

the beliefs, guiding principles, purposes, and values that

the individuals in a group, whatever

mon. Each

its size,

individual, of course, has his

hold in com-

own

conception

of the political culture that influences what he does or

does not do politically. For this reason, cultural analysis

alone, just as social analysis alone, does not exhaust

what

to be discovered about political behavior. Full analysis

is

must

also take

account of the personal basis of political

conduct. But to the extent that individual patterns and

meanings more or
all

it

implies

is

less coincide,

useful

the concept of culture and

handle of political behavior

research.

In this sense, political culture refers to the consensus of

numerous

need not be
any one person. On empirical research grounds, it is a concept probably more useful than
"constitution" in defining a group's political working
agreements. Political scientists have long sought to overovercome the formal-legal definitions of constitution
individuals, but a consensus that

precisely duplicated in

The Cultural Context

[82]

by speaking
metaphor

is

of the living constitution. This biological

more confusing than

constitution of a polity

havioral

patterns

is

clarifying.

The formal

record and a source of the be-

and norms that are components of

the political culture.

it

is

From

the behavioral point of view,

too restrictive to assume that a concept like living

constitution can approximate the richness of patterns, beliefs, and orientations that guide political behavior in government and out. Political culture may as yet be an am-

biguous concept, but it has the virtue of stimulating political behavior research along cross-cultural lines, within
a major culture as well as between subcultures,

which the

notion of living constitution does not.

Although culture may itself not be an operational tool
in a technical sense, it can, if properly constructed in a
specific empirical context, aid in predicting political be-

havior with a high degree of probability.

Of

course,

all

depends on the accuracy with which any concrete political culture has been formulated by the investigator. A
particular culture construct admittedly distorts reality in

order to capture reality.

The

uniformities and regularities

even though the procedure used in formulating the image of the culture is
anything but statistical. The construct of a particular political culture or subculture is always a summary expression of many individual behavioral patterns. Most people
within a group are likely to have some experience with
it

posits are necessarily statistical,

of the same patterns, though all may not be

exposed to all the prevailing patterns. In the study of political behavior within a single culture, widely shared patterns can, therefore, be considered as constants, providing
a base line for examining particular individual responses

many

to particular political situations.

On

the other hand, in

cross-cultural analysis of political behavior, differences in

observed patterns and orientations are

critical

variables

for studying responses in functionally similar political

situations. In either case, cultural analysis permits the investigator to transcend the boundaries of his

own

culture.

Bibliography

[83]

BIBLIOGRAPHY
CULTURAL ANALYSIS
The Cultural Background of Personality. (London: Routledge and Kegan Paul, 1947.)
Potter, David M., People of Plenty. (Chicago: University of

Linton, Ralph,

Chicago Press,

1954.)

The Little Community, and Peasant Society

and Culture. (Chicago: University of Chicago Press Phoenix

Redfield, Robert,

Books, i960.)

Thompson, Laura, Culture

in Crisis.

(New York:

Harper, 1950.)

VALUE ANALYSIS
Benedict, Ruth, Patterns of Culture.

(New

York: Houghton

Mifflin, 1934.)

de Grazia, Sebastian, The Political Cofmnunity: A Study of

Anomie. (Chicago: University of Chicago Press, 1948.)
Morris, Charles, Varieties of Human Value. (Chicago: University of Chicago Press, 1956.)
Stouffer, Samuel A., Cormrninism., Conformity , and Civil Liberties. (New York: Doubleday, 1955.)

CROSS-CULTURAL ANALYSIS
A., and Coleman, James D., The Politics of the
Developing Areas. (Princeton: Princeton University Press,

Almond, Gabriel
i960.)

Deutsch, Karl W., Nationalism and Social Communication.

(New York: Wiley,

Lipset,

Seymour M.,

1953.)

Political

Man. (Garden

City, N.Y.:

Double-

day, i960.)

Pye, Lucian W., Politics, Personality, and Nation Building.

(New Haven:

Yale University Press, 1962.)

SUBCULTURAL ANALYSIS
Banfield,

Edward C, The Moral

(New York:

Free Press, 1958.)

Basis of a

Backward Society.

The Cultural Context

[84]
Vidich, Arthur

J.,

and Bensman, Joseph, Small

Town

in

Mass

Society. (Princeton: Princeton University Press, 1958.)

W. Lloyd, and associates, Democracy in Jonesville.

York: Harper, 1949.)
Whyte, William Foote, Street Corner Society. (Chicago: University of Chicago Press, 1943.)
Wood, Robert C, Suburbia: Its People and Their Politics.
Warner,

(New

(Boston:

Houghton

Mifflin, 1959.)

CHAPTER FOUR

THE PERSONAL
BASIS

The

analysis of political behavior should be exhausted

once man's social relations and cultural milieu have been
accounted for. There would seem to be no room for looking at it on a personal basis. Indeed even if one knew nothing about a man's personality, a great deal of his political
behavior could be satisfactorily explained in social or
cultural terms alone. But throughout the history of political speculation from the Greeks to the present, there
has always been the realization, however vague or mistaken in detail, that a politics which does not consider
human nature can only be a very partial politics.
But to say that a man's personality has something to do
with his political behavior is not saying very much. It
only states an assumption: that what kind of personality

man

brings to politics makes a difference.

are these questions:

when

is

it

feasible

More

relevant

and worthwhile

The Personal

[86]

Basis

from the point of view of perwhat can personality study contribute to the
analysis of political behavior? These questions make the
assumption that differences in political behavior are due
to personality differences somewhat problematical. Perto study political behavior
sonality?

sonality

may

then,

to specify the boundaries of the functioning of

is

or

may

not make a difference.

personality in politics.

vary a good

deal,

It

The problem,

appears that the boundaries

from problem to problem and

may

situation

to situation.
If

we

speak, as

political behavior,

we
we

do, of the effect of personality

on

are presumably concerned with the

between two variables. This formulation is

not an altogether happy one. Personality, like culture, is
an abstraction from behavior, not something independent
of behavior. It is not directly accessible to empirical observation. Rather it is abstracted from and constructed out
of the very patterns of behavior and feeling, past and present, that it is intended to explain and interpret. In speaking of the effect of personality on political behavior or of
the relationship between them, we are once more using
relationship

conceptual shorthand.

prefer to speak of the personal

basis of political behavior because

it

suggests a distinction

between those aspects of political behavior that are personal and those that are social or cultural. It is the relationship between the personal aspects on the one hand and
the social or cultural aspects on the other that is the heart
of the matter.

PERSONALITY IN POLITICS
"Of what
chologist

use to a
is

man

is

his political

behavior?" the psy-

likely to ask. In other words, the psychologist

treats personality as the object of inquiry.

behavior

him
mill.

may

interest the psychologist,

Man's

when

it

political

interests

because it is empirical grist for his theoretical

Hopefully, something called personality comes out

at all,

of the mill.

behaves

as

Depending on the psychologist's theory, man

he does politically because

his

behavior

satisfies

Fersonality in Politics

[87]

personal needs, demands, and drives, or because

tensions and frustrations, or because

or because

deprivations,

and values.
This way of looking

reveals

it

at political

it

it

releases

compensates for

private

perceptions

behavior can

tell

us a

great deal about the quality of political participation, the

intensity of political preferences, or the individual's ori-

entation

to

political

action.

But that personality

presses itself in the political arena

ex-

of only incidental in-

is

he

terest to the psychologist. Insofar as

is

concerned

v^ith

the functioning of personality in social or institutional

contexts at

all, other arenas such as business, sports, or the

can serve his purposes equally well, perhaps even better because they are more readily accessible to psycholog-

arts

ical

investigation.

Whatever the

arena,

use to a

man

is

his

economic behavior,

the

therefore,

psychologist's basic question remains the same:

of

what

his athletic be-

havior, or his artistic behavior?

The

sumed

the needs, drives, predis-

to be shaped directly

positions, or fears

by

behavior

and hopes of the individual

itself is as-

actor.

The

arena has no existence independent of the psychological

processes of the participants. Political behavior, or any

other type of behavior,

rectly

from

For the

is

conceived of as springing di-

personality.

political scientist,

on the other hand, the func-

tioning of the political arena as a behavioral system

is

the critical problem to be explained. His approach

is

He does not ask about the consequences of political behavior for a man's personality.
Rather, he wants to know about the consequences of differences in personality for the performance of political
roles and the working of political institutions. The personal meanings and motivations, conscious or unconscious,
underlying a man's political participation or political preferences are of analytic interest because their discovery
may contribute to an explanation of the political as a
behavioral system. This system, it should be recalled, is
always a network of interpersonal behavior. Analysis of
the personal basis of political behavior can tell us, first of
necessarily different.

The Personal

[88]

how

Basis

why

and

a particular political actor relates himhe does.

Moreover, if it can be shown that political activity
serves a person to overcome a low estimate of himself,
all,

self to others as

that

him to remove feelings of isolation and

by making for personal involvement, or that

serves

it

loneliness

preferences

political

expressions

are

of

rebelliousness

against or submissiveness to authority, and so on, the in-

formation

may

be of inestimable value in appraising the

quality or durability of a political system. In the aggregate,

knowledge of the personal

havior

may

adaptability

basis

of political be-

illuminate the style of a political system,

to

environmental conditions,

its

ability

satisfy

human

But
ments

arising out of the interpersonal situation-

its

to

needs, and so on.

political

behavior

is

also

responsive to require-

and to
widely shared goals, values, and expectations. Conformity to these situational requirements and
cultural expectations will depend, in part, on an individual's capacity to respond, a capacity not unrelated to his
personality. It also hinges on social and cultural sanctions,
broadly interpreted as including both indulgences and
deprivations more or less consciously perceived and ex-

more or

less

perienced as guide lines in political behavior. Regardless

of their personal motivations and predispositions, people

tend to behave in more or

less regular and predictable

and out. And their expectations concerning their own conduct and that of others are relatively stable and uniform. Some may be more conscientious in what they do, more enthusiastic, more eager to

fashion, in politics

more ambitious than others. However, despite

such differences that undoubtedly stem from differences
in personality, political processes and institutions funcplease, or

tion in

For

ways

quite unrelated to variations in personality.

instance, the policies

and

fare department are affected

personality of

scribed

its

functions

activities of a public wel-

by more

chief or case workers.

of

the

department,

factors than the

The

legally pre-

the

professional

ethics of social workers, the realistic assessment of client

Personality in Politics

[89]

needs, the availability of financial resources, the traditions of the

likely to

community, and so on,

affect the

The problem

is

all

of these factors are

department's policies and

to decide just

when

to deal

activities.

v^^ith politi-

behavior from the standpoint of

its personal basis and

problem because we cannot
talk about everything at once. One solution would be to
argue, as has been argued, that in the end only personality matters, that only if man is unmasked, so to speak, can
his political behavior be significantly explained. This argument is patently untenable because, as I have suggested,
political processes and institutions function quite independently of variations in personality. Another solution
would be to treat the personal basis as a residue, some-

cal

how

to

do

it.

It is a subtle

thing to be investigated only after analysis

on the

social

and cultural levels has yielded whatever is sought.

The problem is more complicated, for a number of
reasons. But one reason stands out: unless we subscribe to
an exclusively individualistic psychology, we must assume that what is social, cultural, and personal in political behavior is so intertwined that either over- or underemphasis of any one distorts political reality. It does
not follow, of course, that each political act, sequence of
acts, or system of political behavior must always be
created on all three levels. These are not competing but
alternate ways of looking at the same behavior. They
may set limits to each other, but they need not conflict.
The personal basis of political behavior cannot be treated
in a priori fashion as either an exclusive or a residual level
of behavioral analysis.

One might

explore research strategies in two ways.

one might look at the conceptual linkage of the social, cultural, and personal dimensions of political behavior. To what extent and how do alternate types of analysis complement or limit each other?
Determination of
First,

may suggest what particular level of analysis

most appropriate for whatever theoretical problem is
to be solved. The second strategy is specifying the theoretical problem. One need not assume that all problems
the limits
is

The Personal

[go]

Basis

of political behavior require equally intense study of the

personal basis, but one

may

assume

stage of scientific development,

that, in the present

many problems may be

equally well solved in social or cultural terms.

OPINIONS AND ATTITUDES

can be conducted on a broad

Political behavior research

front.

At one

end, there

the biographical study of out-

is

standing individuals, statesmen, warlords, revolutionaries,

and so on, as well as the man on the street (though

one may question just what intensive analysis of the man
on the street as an individual would contribute to an un-

At

derstanding of the political system).

the emotional reactions of
crisis,

of mass

movements

are

of

intrinsic

extremes, the political behavior

(families,

the other end,

nations,

especially in

or the aspirations and behavior patterns

times of
these

whole

committees,

juries),

of

interest.

of

Between

small

politically

groups

attentive

publics (politicians, opinion leaders, journal readers), of

demographic aggregates (small businessmen, senior

zens,

farm

women),

of organized collectivities

organizations,

utilities),

or

of

citi-

(labor unions,
institutionalized

groups (legislatures, bureaucracies, military formations)

is a matter of inquiry. Because the individual is the basic
empirical unit of political behavior analysis, the very
practical question arises whether it is feasible, in an economic and effective sense, to study the political conduct
of nations, mass movements, large organizations, institutionalized groups, demographic aggregates and even relatively small groups in terms of their participants' personalities.
It

sounds banal to say that the greater the number of

individuals involved in a political system, the less sense

makes to seek

it

concerning the personal basis of

political behavior. It clearly is impossible to subject an adequate random sample of a national population or of a large
group like an army to the same kind of intensive personality study that is possible in the case of individual politifull data

Opinions and Attitudes

[91]

This need not mean that survey analysis

of the political behavior of large groups or aggregates,
even of national societies, cannot benefit from theoretical
formulations of personality. But in general, the kinds of
problems which interest the political scientist, involving
not only individuals as individuals but also more or less
complex collectivities, though possibly psychological in
formulation, do not readily lend themselves to analysis as
problems of personality.
So far I have intentionally shied away from defining
cal biography.

personality. It

is

difficult to

use concepts referring to the

personal basis without becoming entangled, wittingly or

some particular theory of personality. Almost every relevant psychological concept is invariably tied to

not, in

some

particular school of psychology or given different

meanings by different schools. There would be no point
in reviewing various definitions of personality here or
adopting one of them. This is not a plea for conceptual
fuzziness or eclecticism. It only means that since political
behavior analysis is concerned with the individual as the
empirical but not necessarily the theoretical unit of inquiry, it is advisable to avoid committing it to a particular theory of personality. For in the end such theory may
not be strategically appropriate for the kind of problem
to be solved.
The problem at hand is the best guide to the choice of

approach.

However

the personal basis

personality

we may

is

defined,

in studying

deal with unconscious motiva-

and thought processes, including fantasies; with

less visible reaction patterns (mechanisms of defense or cognitive dissonance); or with attitudes, perceptions, and preferences, rational or not, that can be more
readily verbalized and discovered. Just how these aspects
tions

more or

personality are related to each other, just what

emphases are to be given one or the other in constructing
personality, just what techniques are best suited to producing the necessary data, and so on, are matters of immediate concern to the psychologist. They need not occupy us here. Our objective is not to formulate homo

of

The Personal

[92]
poUticuSy

if

this

but to suggest strategies of

possible,

is

Basis

research on the personal basis of political behavior in

connection with quite diverse problems of politics.

In sum, analysis of political behavior in its personal

from

pects need not proceed

and integrated conception of personality.

It

as-

consistent

theoretically

may

select

those aspects that seem relevant to the problem and

fit

the population to be investigated to them. Research tools,

they are dictated by the

as

ceptual tools, as they are

size of the sample,

by

dictated

and con-

the problem at

hand, are likely to influence each other. Formulations of

behavior in a large sample survey, which involves

political

relatively short interviews with a population of consid-

erable heterogeneity,

must be necessarily limited to

quiry in terms of rather superficial perceptual or

tudinal

may

That

questions.

these

attitudes

or

inatti-

perceptions

be anchored in interpersonal relations rather than

the personality

the point.

beside

is

some personal meanings

They may

reveal

that the individual gives his po-

litical activity.

The

extent to

which the researcher may draw

infer-

ences from such data on the functioning of personality in

politics

is

a matter of scientific conscience.

One may wish

to interpret scaled responses to a handful of questionnaire

an authoritarian personality an
attitude, or simply a
feeling of efficacy. Indeed this has been done, but there
is the danger of over-interpreting this kind of data. The
cold fact is that a certain number of people have reitems

as

indicative

of

anomic character, a tough-minded

few

sponded in certain ways to

of an individual's responses

may be

may

The

suggestive,

pattern

though

it

may

not be evidence of personality in politics.

the other hand, if one's research deals with a single

or

On

questions.

few people, the personal basis of political

behavior can be fruitfully explored through a prolonged
interview of the Freudian tradition or, where interviewindividual or a

ing

is

impossible,

by

exhaustive analysis of documentary

data. In these cases, political behavior

ploy the intricate

set of

analysis will

em-

concepts from one or another

Opinions and Attitudes

[93]

theory of personality. This will undoubtedly yield a rich

harvest of knowledge about the relationship between personality and politics.

For

instance, the interpretation of

Hitler as a hysterical-paranoid personality

may

be

set

background of social and economic conditions in Germany and the emotional state of the German
masses, providing insight into the dynamics of large political systems under certain conditions of stress and
against the

strain.

Only very few

this

political scientists have ventured into

type of intensive personality research. This is due

partly to strategic difficulties inherent in the research

uation, partly to the fact that

few

possess the skill

Of

training necessary for deep analysis.

think the former

more

these

two

sit-

and

obsta-

Living persons of
political consequence are unlikely to submit to intensive
study. In the case of historical figures, the nature and relicles, I

ability of

is

critical.

documentary materials

is

always open to ques-

tion.

Somewhere between survey analysis and depth analyeiforts are made to combine intensive, though not

sis,

prolonged, interviews with projective

personality

inventories,

and

other

tests,

standardized

procedures.

These

techniques are well suited to inquiry into political be-

more widely applicable than

deep analysis, yet more thorough than survey research
can afford to be. They would seem to be especially useful in the study of small institutions, though so far they
have been employed only in the study of small aggregates
of isolated individuals. These techniques require a measure of diagnostic skills, too, but political scientists could
learn them.
havior because they are

Depending on the strategy employed then,

different

conceptual aspects of personality can be selected for investigation of the personal basis of political behavior
without commitment to some overarching conception of
personality in politics.
comprehensive model of political man may be more of an obstacle than an aid.
may

We

yet gain a good deal of insight into the personal basis of

The Personal

[94]

by examining

political behavior

and

affective orientations

Basis

those cognitive processes

through which the individual

seeks to cope with political

life,

including the situations

and expectations that define his political

The study of attitudes and perceptions
diately useful in discovering

how

roles.
is

political

more immebehavior af-

and systems than the

kind of intensive inquiry that seeks explanation in terms
of unconscious drives or other adaptive mechanisms of
fects large-scale political processes

the psyche.

Knowledge

of

how man

perceives himself as

how

he interprets the world of politics,

how he values what he sees, and how he acts politically
in pursuit of personal values can tell us a great deal about
a political actor,

his political behavior.

Overt behavior, including opinions, can be revealing

in

Most obviously, perhaps, it can tell us

something about government and politics as objects of
perceptions and attitudes. Does the world of politics
several respects.

enter a man's perceptual field at

all?

Is

the legislator

aware of the interest groups in his environment? Are political campaign issues important to the voter? Once we
establish perceptual relevance,

tive

components of

we may

political attitudes.

investigate affec-

Does the

legislator

groups are helpful or harmful? Does

the voter consider government as benign or evil, political
candidates as attractive or not, political issues as urgent
or not? Answers to these simple questions can tell us a
great deal about a person's ability to differentiate, the
values in terms of which he appraises reality, and his time
perspectives. If not all pressure groups are perceived as
powerful, which ones are and which ones not? What degree of influence is attributed to these groups? What crifeel that pressure

judgment are applied?

teria of

It is likely

that the values

involved in political judgments or the perceptions of ur-

gency
tute a

or

less

that reveal the political time perspective

more or

less

patterned hierarchy, as well

permanent syndrome of

While

consti-

as a

more

political predispositions.

different levels of the personal basis of political

behavior are probably interrelated, one need not assume a

Values and Personality

[95]

one-to-one correspondence.
great
behavior or opinion may be rooted in

many varieties of
common personal-

deep level of analysis, but the

behavior of diverse personalities may be identi-

ity characteristics at the

political
cal.

For

instance, distrust of authority

may

express

itself

in unconventional political opinions or radical behavior

or in compulsive demands for a just and powerful leader.

Deep-felt dependency needs may underlie an ideology
that glorifies obedience or one that insists

For

this reason,

we

on freedom.

should always take care in interpret-

ing opinions and attitudes, even if they seem to constitute

a consistent pattern that suggests a personality syndrome.

VALUES AND PERSONALITY

From

the research point of view, the conceptual plastic-

about the personal basis of political

behavior can be an operational blessing. It permits the
psychologist to construct his model of personality by
ity of formulations

selecting

from many

aspects those that

seem most

fruit-

ful in explaining a particular political behavior problem.

Most psychological
cal mental states,

nomenclature
havior.

They

as

theories of personality consider typi-

or mechanisms whatever the

most basic determinants of be-

drives,

the

consider perceptions, attitudes, or values as

more or less peripheral layers of personality. But

no intrinsic objection to choosing those elements

there

is

of per-

seem most useful for a special problem of

have already mentioned the advantage of paying special attention to perceptions and attitudes in the study of political behavior. It is also worthwhile to pay close attention to a person's values, whether
he holds them consciously or not.

sonality that

political behavior. I

Values can serve

as

central

organizing principles in

studying the personal basis of political behavior. Pre-

sumably

a person's values are sufficiently structured to

constitute a value system;

may

and

if

they do, the value system

well be the most stable component of political be-

havior. Therefore

it

is

a critical task of research to dis-

The Personal

[gS]

cover the content of a person's value system,

ment and degree of internal consistency, and

arrange-

its

its

to the culture that patterns the politics of the

Basis

relation

group to

which the individual belongs. More than any other aspect

of personality, personality-rooted values express the personal basis of political behavior.

To

conduct

ferentiate

as

this

type of analysis,

sharply

as

possible

and the cultural components in

of

The

values.

validated

discover

On

culturally

by being widely

it is

necessary to dif-

between the personal

an individual's scheme

transmitted

set

of

shared, though

we

values

is

can only

through studying the values of individuals.

it

the other hand, the value system of an individual,

though transmitted culturally, should be more or less

identifiable as unique in his political behavior. Unless the
distinction is made, political behavior analysis is likely to
be trapped in what
correlation: that

from

is,

would

call

a cultural value system. I believe that this error

made by

those

who

substitute

political character for political

in

the fallacy of cultural

the error of inferring a personal value

mind

litical

is

some pseudo-concept of
behavior analysis.

have

those broad characterizations of social and po-

behavior that are so popular because they are so

the organization man, the inside-dopester, or

flippant:

the status-seeker. Just because a culture like America's

can be characterized in terms of the high value given

personal

success,

achievement,

it

individual

mobility,

or

competitive

does not follow that these are the main

values or even the sole values that shape the political be-

havior of particular individuals. In the

first

place, there

and
widely shared. And second, without empirical investigation in each case, we cannot tell in what combinations
these values occur to constitute individual value systems.
are

other

In fact,

values

it is

that

are

culturally

transmitted

the special task of research to discover

how

widely shared and personally held values can shape a

system that gives the individual's political behavior its
relative stability and consistency.
Whatever the degree of political involvement, culture

iiiifu'w^a

Values and Personality

[97]

determines significant aspects of an individual's personality. It does this not only in the form of role expectations and other behavior norms, as well as perceptual cues
that orient the individual to his social

and material en-

vironment, but also in the form of values. The individual

internalizes these values in the course of socialization.
They thus become aspects of his personality. Personalities in

a given culture

group can probably be character-

ized in terms of values held in

cialization process

itself

is

common

because the so-

culturally patterned.

But

as

there are also discontinuities in a culture, and as social-

vary a great deal from subculture to suband internalized in varying combinations and degrees of intensity, depending
on an individual's peculiar circumstances. As a result, one
individual's value system is rarely identical to another's,
though enough similarities may exist to permit their
description in terms of types.
personal value system,
then, is not simply a replica of a cultural value system.
Otherwise it would be impossible to account, on the personal basis, for the great variety in political behavior
within the same culture.
Because values and value systems differ from one cultural context to another and from one person to another,
problems of political relevance on the personal basis of
behavior may be quite differently experienced and evaluization practices
culture,

values

are transmitted

ated in different milieus. In particular, the

personal

life

way

in

which

experiences affect political behavior in the

performance of roles is likely to depend on the culture's

value system and, accordingly, the values given to personal
experiences and correlate behavioral manifestations.
This may be illustrated by the possibly different consequences in different value contexts of primary identification as part of the socialization process. In most
cultures, children find their identity through identification with the parent of the same sex in a particular phase
of the growth process.

The

degree of success achieved in

must be appraised against the background

of culturally approved solutions. Now identification in
identification

The Personal

[q8]

Basis

the primary circle interests us from the standpoint of political

behavior because

responses

to

seems to entail an individual's

The kind of identification

it

authority.

achieved with the father as an authority figure seems to

be of consequence for the kind of identification with later
symbols of authority individuals like teachers, job
superiors, or political leaders, or groups such as community, institution, or nation.
Identification seems to have two main types of con-

sequences

for

rebelliousness

behavior:

political

against

authority.

submissiveness

However

or

to

culture's

value system will emphasize these pos3ible types of re-

sponse differently. In a culture that highly values authority as a set of expectations defining adult relationships,
is

it

likely that the process of identification in the nuclear

family will tend to polarize the alternatives of submission and rebellion toward

more extreme

solutions.

On

the other hand, in a culture that places relatively low

value on authority, the working out of child-parent identification

is less

likely to

have extreme solutions. In turn,

because the individual repeats the patterns experienced

growth, primary identification and

But
as normality in behavior is defined culturally rather than
personally, individuals do not necessarily experience a
system of authority relations, whatever its shape, as either
in different stages of

resultant solutions tend to reinforce cultural values.

especially indulgent or deprivational.

As

a result, authority relations in

some

political systems

engender a great deal of submissiveness or assertiveness

without negatively affecting the system's stability, while
in some other system, comparatively lax authority relations

also

fail

to

thority

are

adversely political stability.

affect

seems that in either

case,

sufficiently

differential evaluations

internalized

to

It

of au-

become strong

organizing principles at the personal basis of political behavior.

There

is

always an area of exploration for locating the

personal aspects of political behavior in the context of

cultural expectations. In other words,

it

is

important to

Values and Personality

determine just

more or
ues

less

how

[99]

a person's personal perspective

correspond

to

those

consensually

validated

pretations that are the "givens" of social conduct.

stance,

if

his

conscious identifications, demands and valinter-

For

in-

giving or withholding affection in childhood

has differential consequences for behavior in maturity,

empirical analysis has yet to determine just

how

affective

orientations experienced early in life are interpreted in

the cultural context. Discipline has different meanings in

and subcultures. In a culture where senot only widely practiced but also approved, its consequences are likely to be different from
those in a culture where stern discipline is restricted and
disapproved. But just what these consequences are can
never be left to deductive inference; they must be emdifferent cultures

vere disciplining

is

pirically verified.

For

this reason,

we

should distrust attempts at char-

whole nations in terms of

some pervasive set of personality attributes inferred
from the nature of socialization. Although national characterological interpretations of

acter

is

widely used term,

cal concept.

Many

it is

not a meaningful empiri-

of the constructions of national char-

acter that have been

made

are largely intuitive and often

based on casual observations, literary statements, or case

analyses of limited scope. If they are based on large
opinion or attitude surveys, the surveys

may

tell

us some-

thing about the distribution of attributes in a population,

but we should be even more suspicious of inferences

about a hypothetical, if not mythical, national character.
Anthropological researches in small, simple societies
have stimulated so-called national character studies. In
small societies, model traits can possibly be explained by
the exposure and limiting of children to a very narrow
range of experiences that are constantly reinforced by a
relatively homogeneous cultural value system. Early acquired characteristics at the personal basis of behavior
be important for the functioning of the social

may

system. But

we cannot assume that this happens in an

way in complex, modern systems. I doubt

equally simple

The Personal

[loo]

Basis

that a given political structure, let us say democracy,

is

predicated on the existence of a democratic character,

may mean. Of course, psychologically

healthy people are preferable to unhealthy people in any
whatever that

system. But definitions of mental health are difficult to

come

by. Equating mental health with democratic predis-

symptomizes a cultural myopia, pleasant

and self-congratulatory as it may be.
positions only

ROLE AND PERSONALITY

The

broad patmost tangible evidence of

investigation of specific roles rather than

terns of conduct provides the

the importance of the personal basis of political behavior.

This follows from our initial acceptance of role as the
basic theoretical unit of behavioral analysis.

We

can think of role

as that aspect of personality that

refers to an individual's social identity. If this

is

so, re-

search on political behavior must deal with the agreement

between
aspects.

political roles as aspects of personality

One may approach

of the relationship.

this

problem from

psychologist,

would

and other
end

either

think,

would

be interested in learning how harmony or conflict among

several concurrent role requirements affects the degree to
which the personality is capable of modifying itself in
response to particular role requirements.
to

know how

He would

want

the relationship between role and other as-

pects of personality varies with the special characteristics

of the individual.

which

a role

is

He would

needs and demands.

On the other hand,
sis,

somewhat

try to discover the degree to

central or peripheral to a person's private

if

role

is

the starting point of analy-

different concerns

come

into focus.

No

as-

sumption needs to be made about the direction of the

relationship whether other aspects of personality shape
role or are shaped by it. Nor is it necessary to assume ei-

ther a zero or a one-to-one relationship.

The

degree of

congruence between role and other aspects of personality

is likely to vary a good deal.

Role and Personality

[loi]

The complexity of a
The more complex

political

system

one cause of

is

a system, the greater the

this.

number

and variety of political roles. The sheer number of available roles and their heterogeneity are likely to reduce
congruence between role and requisite personality characteristics. In a democracy like the United States, there
are millions of citizens

making by voting

who

participate in public decision-

Hundreds of thousands
engage in other forms of political activity. Thousands of
others fill a great range of elective and appointive offices,
in government or political organizations. People with
quite different personality characteristics can function effectively in this sort of political system with its great variety of political roles. Only in the upper tiers of the
structure

political

in elections.

personality

are

characteristics

more

be relevant in recruiting personnel for particular roles and in shaping the performance of these roles.
But even in connection with these roles the democratic
quality of the political system is likely to reduce the degree of congruence between role and other facets of perlikely to

sonality.

For example, take an

is

elective official like the

The number

legislator.

large because the legislator

number

is

implicated in a great

of interpersonal relationships. All of those he

constituents,

with

deals

lobbyists, party leaders,

of multiple legislative roles

admin-

have different ex-

and so on
pectations of him. Therefore, conduct
istrators, colleagues, friends,

is

likely to

performance
be highly variable,

in the

despite similar personality characteristics

some

American

of simultaneously available roles

on the part of

legislators.

But

how

are personality characteristics in behavior re-

lated to recruitment into the legislative office?

think that legislators

who

One might

seek elective office and are will-

ing to expose themselves to the

common

experience of

repeatedly having to fight electoral battles might also

share certain personality characteristics. These characteristics

as an aggregate from those

not cherish the uncertainties of the electoral

might distinguish them

who do

The Personal Basis

[i02]
struggle

may

but seek access to important decision-making

by appointment. Even here

posts

personality

analysis

turn out to be negative for this reason: in political

recruitment processes where large numbers

participate

in

of

people

the choice of elective office-holders, the

norms and expectations of the community are probably

more relevant and decisive than the predispositions of

who

those

It may be less the candithan his image that provides a more

stand for election.

date's real personality

satisfactory explanation of electoral success.

Congruence is
which roles and

vary with the degree to

also likely to

role performance

standardized, and routinized.

are institutionalized,

The more

tionalized and role performance

is

a role

is

institu-

secured through

insti-

tutionalized sanctions, the less likely will behavior require

the mobilization of energies rooted in the personality.

no elbow room in which to funcThis does not mean that certain personality types
are not more suited than others for filling highly standardized roles. But it is probably the role that limits behavior
rather than the particular personality that is attracted
Personality variables have
tion.

to the role.

We
ence.

should not oversimplify the problem of congru-

For

instance, the definition of bureaucratic roles

is

highly standardized, and the performance of bureaucratic

roles depends on stable, disciplined behavior and adher-

ence to prescribed rules. The kind of conduct that is expected is insured by appropriate attitudes that are transmitted, learned, and reinforced through institutional
training.

Now

it is

quite likely that the bureaucratic or-

ganization attracts personalities with predispositions

or

less

and,

well suited to the kind of behavior that

in

fact,

rather

severely

enforced.

It

is

is

more

expected
probable,

though bureaucratic role performance

may not be dependent on personality, recruitment into
the bureaucratic role will be facilitated by appropriate
predispositions. This, in turn, can lead to rigidities in bu-

therefore,

that

reaucratic behavior that, paradoxically,

effective

may

performance of the bureaucratic

obstruct the

role.

In fact.

Role and Personality

[103]

congruence of role and other aspects of personality in the

may have dysfunctional consequences

recruitment phase

for the operation of a bureaucratic organization.

Congruence of role and personality

is

likely to vary,

too, with the range of available behavior alternatives. In

the American two-party system, for example, the citizen

can vote or abstain from voting.

If

he votes, he can vote a

straight ticket, split his ticket or perhaps vote for a third

party. In general, his behavior alternatives are quite restricted.

We

can explain performance of the citizen role

more economically in social or cultural than in personal

terms. That John Smith votes Republican because all of
his friends are
ests better

Republicans or because he sees his interby the Republican party is probably

served

sufficient explanation of his behavior. If personality analysis

should discover that he votes Republican because of

who was a Democrat,

Republfcan vote has this (conscious or unconscious) meaning for him, we have an interesting bit of
information that may be useful in understanding John
Smith as a personality. It does not add very much to our
knowledge of his political behavior. On the other hand,
if we should find that John Smith votes Republican
though all of his friends are Democrats, or that he votes
as he does though his interests would be better served,
even in his own opinion, by the Democrats, analysis of his
political behavior on the personality level may supply a
a deep-seated hatred of his father

and that

his

missing link in explanation. In short, analysis of political

behavior in terms of personality seems advantageous if
we are confronted with deviant conduct, deviant in a statistical,

not a pathological sense.

There are other situations in which the relationship

between role and personality may be accentuated and
where relatively high congruence has significant consequences for the functioning of a political system.
High congruence may occur in connection with
largely voluntaristic roles, that is, roles that are not solidly
institutionalized.

In Western democracies, for instance,

the roles of rebel, hermit, or prophet are

less likely to

be

The Tersonal

[104]

taken in response to others' expectations (though

ment

is

ever-present

by

Basis

this ele-

definition of role), but in re-

sponse to strong motivations at the level of personality.

These

roles are not only voluntaristic

tional point of

for

skills

view but

from an

also highly specialized.

institu-

They

call

that are rarely transmitted in a cultural sense

and they allow a considerable degree of freedom in performance. In an institutional setting like a legislature, this
category might include the role of maverick. Individuals
with appropriate personality predispositions will take
and shape these

roles. In their case, then, analysis of the

personal basis of political behavior in personality terms

may be very rewarding.

Personality may also be

the decisive factor in role taking

and performance where

person

expectations.
is

How

is

exposed to conflicting

role conflict of one sort or another

resolved can be analyzed in social or cultural terms

alone, of course.

But

it

may

be

also

a function of person-

For instance, withdrawal from political participation can be due to the neutralizing efl'ect
of cross-pressures, an explanation on the social level of
analysis. But it can also be due to anxieties occasioned by
ality predispositions.

the conflict that are associated with similar experiences in

the formative stages of personality development.

Role and personality may approximate a high degree of

congruence in perceptually ambiguous situations in which
the actions, expectations, and sanctions of others in a
role relationship are blurred, for whatever reason. If no
stable points of behavioral reference are available and no
directional cues are forthcoming, the actor, unable to
cope with ambiguity in any other way, will define his
role for himself by falling back on personal values and
experiences that may or may not be objectively relevant
in orienting himself to others.

While

tions to resolve personal strains

ambiguous

situation,

it

tends to

this

behavior func-

brought about by the

make

for a great deal of

projective thinking, stereotypic responses, and irrationality.

Finally, roles allowing for a high degree of discretion

Role and Personality

in

performance

[105]

may be

sonality characteristics.

highly congruent with other per-

The king who cannot do wrong,

the authoritarian or charismatic leader, and similar roles

are almost

by

stitutional

on the expectation

definition predicated

that there shall be

no

sanctions,

expectations. In the absence of in-

personality

has

much leeway

to

shape the role in socially and culturally unpredictable

ways. Of course, there are limits to what incumbents of

such roles can do. Autocrats have been toppled from
power and prophets have lost their glamour. But the limits
behavior in these roles are sufficiently vague to
allow the actor to shape his social environment in line
set to

with
is

needs and predispositions more than

his personality

usually the case.

These examples of relatively high congruence between

and personality might give the incorrect impression

role

that

analysis

them.

of political behavior

The examples

is

confined only to

are exceptions to support the gen-

eral proposition that in

complex

difficult to postulate theoretically

political systems

it

is

or discover empirically

personality attributes that are invariably and inevitably

linked to particular roles. On the other hand, it does not

follow that recruitment into or performance of certain
roles cannot at times entail the possession of appropriate

personality characteristics.

am merely

arguing that be-

havior in the performance of culturally validated roles

on the one hand, and behavior in response to personality

on the other, is sufficiently plastic to prohibit any kind of dogmatic assertion about the relationship between them. Some roles may attract certain personality types and limit access to other types. Other
roles may occasion behavior quite independent of any
personality characteristics. Congruence is a variable property of the relationship and always a matter of empirical
predispositions

determination.

The

relationship

between role and other aspects of

not easy to determine empirically because

the behavior involved is elusive. Take, for example, what
is sometimes called policy-planning. As conduct in the
personality

is

The Personal

[io6]

performance of
the appraisal of

Basis

a specific role, policy-planning requires

many

alternatives

and even greater num-

bers of consequences flowing from these alternatives un-

der conditions of more or

less

uncertainty. In terms of

personality, an actor in an uncertain situation

must be able

to tolerate ambiguity. Persons given to undue anxiety,

whatever

its

source in their

life

histories,

are probably

up to the expectations and demands made on

the policy planner as a role type. I would think it exunable to

tremely
of

live

difficult to separate

policy-planning

out those behavioral elements

stem

that

from the condition of

from subjective-

objective uncertainty and those that stem

personal qualities.
It is

equally difficult to

make

a distinction

between the

function of personality in role recruitment and role per-

formance. In some

cases,

as

that

of

the

organizer

as

enough to make

against administrator, the roles are clear

discovery of suitable personality characteristics relatively

easy.

Whatever they

are, the personality features neces-

sary for organizing are probably quite

different

from

those necessary for administering. Indeed, in these cases

high congruence between personality and role

is

deliber-

ately cultivated in the selection of personnel. It

may be

difficult to find a single

ality attributes

deemed

person

who

desirable to

possesses the personfill

both roles simul-

taneously.

example suggests how subtle the problem is.

sometimes equated with dominance as a
personality trait. It is fairly safe to say that a dependent
or withdrawing person is unsuited for the role. But
leadership does not necessarily require dominance at the
personal basis of behavior. Congruence between the role
and other aspects of personality is likely to be incomplete as well as variable, because performance of leadership takes place in response to social situations and cultural norms that may be quite independent of personality
needs. Indeed if complete congruence between the leader
role and dominance as a personality trait were to occur,
it would in all probability have disastrous consequences

One

final

Leadership

is

Bibliography

[107]

for political behavior and the political system, whatever

its

ideological

complexion.

The

resultant

tensions

be-

tween the demands of personality and social or cultural

requirements for behavior would make any sort of posystem unmanageable. Just as Hitler's hysterical
and paranoid personality came ultimately into conflict
with the realistic requirements of the system that he had
created in his own image and doomed it, so a totally
litical

involved democratic citizenry, involved in the sense of

being compulsively committed to participation in all decision-making,
polity. It

is,

would

probably

doom

democratic

therefore, neither necessary nor correct to

more or

power-oriented
depending on the
degree of power actually associated with a role, that only
persons with the appropriate increment of power-orientation in their personalities can fit into corresponding pointerpret

all political

roles as

less

in a psychological sense, or to expect,

litical roles.

BIBLIOGRAPHY
PERSONALITY ANALYSIS
Adorno, T. W., Frenkel-Brunswik,
Sanford, R. Nevitt,

York: Harper,

The

E.,

Levinson, Daniel

Authoritarian Personality.

J.,

and

(New

1950.)

Erikson, Erik H., Childhood and Society.

(New

York: Norton,

1950-)

Fluegel,

J.

C, Man, Morals and

Society. (London:

Duckworth,

I945-)

George, Alexander L., and George, Juliette L., Woodroiv

Wilson and Colonel House. (New York: John Day, 1956.)
Lasswell, Harold D., Psychopathology and Politics. (New York:
Viking Compass, i960.)
Wallas, Graham, Human Nature in Politics. (New York:
Knopf, 3rd. ed., 1921.)

The Personal

[io8]

POLITICAL OPINIONS AND ATTITUDES

J., The Psychology of Politics, (New

Eysenck, H.

Basis

York: Prae-

ger, I954-)

Herbert, Political Socialization: A Study in the Psychology of Political Behavior. (New York: Free Press, 1959.)
Lane, Robert E., Political Life: Why People Get Involved in
Politics. (New York: Free Press, 1959.)
Newcomb, Theodore M., Personality and Social Change. (New
York: Dryden, 1943.)
Saenger, Gerhart, The Social Psychology of Prejudice. (New
York: Harper, 1953.)
Smith, M. Brewster, Bruner, Jerome, and White, Robert W.,
Opinions and Personality. (New York: Wiley, 1956.)

Hyman,

POLITICAL VALUES IN PERSONALITY

W., Miller, Neal E., Mowrer,
O. H., and Sears, Robert R., Frustration ajid Aggression.
(New Haven: Yale University Press, 1939.)
Fromm, Erich, Escape from Freedom. (New York: Pvinehart,
Dollard, John, Doob, Leonard

1941-)

Lane, Robert E., Political Ideology:

mon Man

Believes

What He

Why

Does.

the Afnerican

(New

Com-

York: Free Press,

1962.)

Lewin, Kurt, Resolving Social Conflicts.

(New York:

Harper,

1948.)

Mead, Margaret, Soviet Attitudes Toward Authority. (New

York: McGraw-Hill Book Company, 1951.)
Money-Kyrle, R. E., Psychoanalysis and Politics: A Contribution to the Psychology of Politics and Morals. (New York:
Norton,

1951.)

Murphy, Gardner, In

the

Minds of Men. (New York:

Basic

Books, 1953.)

PERSONALITY AND POLITICAL ROLE

Cantril, Hadley, Human Nature and Political

Systems.

(New

Brunswick, N.J.: Rutgers University Press, 1961.)

Gerth, Hans, and Mills, C. Wright, Character and Social Structure. (New York: Harcourt, Brace, 1953.)
Gouldner, Alvin W. (ed.), Studies in Leadership. (New York:
Harper, 1950.)
Lasswell, Harold D., Power and Personality. (New York: Norton, 1948.)

Bibliography
Lipset,

Seymour M., and Lowenthal, Leo

[109]
(eds.),

Culture and

Social Character. (New York: Free Press, 1961.)

Riesman, David, Faces in the Crowd. (New Haven: Yale University Press, 1952.)

CHAPTER FIVE

BEHAVIORAL

DILEMMAS

The

behavioral persuasion in politics

by. Behavioral practitioners

make

is

difficult to

live

exacting scientific de-

They encounter great practical obdilemmas of behavioral inquiry are perplexing

and often frustrating and the increments to political
knowledge that any single piece of research may make are
very modest.
But certainly, as one reviews the developments of the
last fifteen years, roughly the period since the end of
mands on

themselves.

stacles; the

World War II, the contributions that behavioral research

has made to political knowledge are impressive. There
have been substantial advances in almost
tional

areas

tributions are

of

uneven

in

science.

are reasons for this unevenness.

volve

Of

all

the tradi-

the conboth quantity and quality. There

political

Many

course,

of these reasons in-

very technical, methodological matters that are

[m]

Behavioral Dilemmas
explain

to

difficult

in

simple language,

and trying to

them simply may be somewhat foolish. My purpose here is to highlight some of the difficulties and dilemmas that make behavioral research in politics a venturestate

some' enterprise, just

as the

prospects of solution represent

the continuing challenge of the behavioral persuasion.

The

behavioral persuasion

in

politics

status of science. Seeking this goal,

lems that characterize the

human

as

well

as

the

it

scientific

natural

aspires

shares

way

sciences.

all

of

to

the

the prob-

life,

in the

Scientists

and

philosophers of science have written about these problems

at length and I shall not treat them here. Nor shall I deal
with those perennial questions such as what we mean
by science when we speak of a science of politics, whether
such a science is possible and, if it is, just how it is similar

to or different

from the other

sciences, behavioral or nat-

is both
and desirable, and I shall start from there.
Another disclaimer is in order. I shall not deal with the
ever-present mistakes and errors that inevitably occur in
any research effort: errors of omission at the inception
of a project, errors of commission in the process of re-

ural. I take

it

for granted that a science of politics

possible

search, errors of exaggeration in the use of over-refined

research tools (where validity

is

sacrificed to precision),

or errors of underestimation in the treatment of data

is handcuffed in the name of
not deal with the endless frustrations experienced in chasing after respondents in a sometimes hopeless attempt to maintain the probabilistic purity

(where the imagination

validity).

And

I shall

of a sample design; in spending strenuous and often

seemingly wasted hours to increase the reliability of a
content analysis; in trying to develop alternate codes for
the classification of

raw

data that will not violate either

one's theory or the data; or in simply

programming a

meaningful analysis. The possibility of error exists all

along the way.
Perhaps the first lesson which the behavioral researcher
in politics should learn is to be candid about the limitations of his work. This does not mean he should permit

Behavioral Dilemmas

[ii2]

sloppy work, but the researcher himself should point out

flaws in his work that may be due to mishaps in the field
discrepancies in the data, mistakes in analysis, and so on.
For nobody knows as well as he where and what the flaws
are. There are usually legitimate reasons for difficulties
and mistakes that are of scholarly interest.
I
shall not attempt, then, either a systematic or
complete discussion of all the dilemmas that I sense in

behavioral research on politics. In the spirit of the essay

form

have chosen,

the things that

own

come

shall

mind

to

as

only mention a few of

they are grounded in my

experience in research.

OBSERVATIONAL DILEMMAS
Behavior

is

through which man moves in

datum of observation but a datum

a series of acts

time and space.

It is a

that does not constitute a self-evident fact. Instead

fact that

process

it is

given meaning by the observer in the very

of observation. Otherwise observation itself
is

would be meaningless. And if observation had no meaning, it would not occur. Observation is itself a form of
behavior that involves giving certain types of meaning to
the object of observation, depending on who the observer
is. This is true of the mother who observes her baby, of
the tourist

who

observes a

omer who observes

political scientist

the

who

new

landscape, of the astron-

movement

of the stars, and of the

observes the action of a legislature

or the behavior of voters.

Observation is as much giving meaning to the behavior

is observed as being engaged in meaningful behavior.
The observer necessarily observes himself in the act of
observing others. The extent of his success determines, at
least in part, his success in observing others. This double
that

aspect of observation

makes

it

a scientific challenge.

Error in observation means that the observed act or besay

havioral pattern has been given false meaning.
that the observed behavior has been misinterpreted.

We

Whether

the misinterpretation

is

willful or not, the false

Observational Dilemmas

[113]

due to the behavior of the observer. The

sources of error in observation giving false meaning to
the object of observation are manifold. Their discovery has been a matter of interest to the philosopher, the
theologian, the psychoanalyst and the scientist. Whatever
the approach to the discovery of error, there is agreement
that the reduction of error involves changes in the behavior of the observer, not of the observed. If the history
of humanity is a history of errors, as some believe, the history of science is a continuing effort to reduce errors in

meaning

is

observation.

The problem
ences

is

havior.
is

of reducing error in the behavioral sci-

compounded by

The

the complexity of

inordinately complex, even

neurological

plex

is

human

be-

smallest observable unit of behavior, the act,

phenomenon

if

alone.

treated as a biological or

How much

more com-

the series of acts that constitutes the behavior of

human organism as a whole! And this complexity is

compounded when we come to deal with man's behavior in his social relations. The observer's task would
the

almost seem hopeless

aspect of

human

if

it

were not for one

behavior: the fact that

man

significant

himself gives

meaning to his multiple actions. Of course, a man may err

in the meanings he gives to his own behavior. There may
be quite a discrepancy between what he is doing and what
he thinks he is doing. But it seems quite clear that the
human dilemma need not be the observer's dilemma. The
observer has one advantage over the observed: he can
check and, in fact, must check his own meanings against
those that the observed gives his
spect,

then,

the behavioral

own

behavior. In this re-

approach in

politics

has

decided advantage over those that describe and analyze

actions without considering

what

these actions

mean

to

the actors.

But

this creates

another problem.

If

our observations

meet the test of inter-subjective agreement between observer and observed, the meanings given to behavior by the observer and those given by
the observed must be captured in a single structure of
of political behavior are to

Behavioral Dilemmas

114]
leanings that

is

internally consistent. This does not

On

that their language must be alike.

scientific observer

language that

is

sufficiently

contain the language of the observed. But

abstract to

the

clearly,

must use

mean

the contrary, the

meanings,

though

expressed

different

in

languages, must be consistent.

The meanings

people give to their political behavior

are critical data for scientific observation precisely be-

from the standpoint of the behavioral persuasion,

no "behavior as such" in a purely physical or

cause,

there

is

mechanistic sense. Observation

communicative act in
are mutually imWhether these communicative acts are expressed
is

which both observer and observed

plicated.

visually, sensorially, or verbally, observation

is

impossible

without communication, and communication is impossible

without attention to meanings. This creates many problems of method in observation, for the existence of the
communicative network not only binds but also affects
both observer and observed. The necessity of attention
to meanings is quite evident in the uses that the interview,
which is the main tool of behavioral analysis, can be put
to. The observer's questions and the observed's responses
must be mutually meaningful. The communicative nature
of observation is most obvious here. But the ability to
hide an observer behind a one-way mirror or a hidden
microphone does not negate the communicative aspect
of observation. Without unity of meanings between observer and observed, there would be no communication
and hence no observation.
Neglect of this two-way flow of meanings can seriously

damage the
is

now

reliability of behavioral research.

Though

it

generally recognized that research must attend

to latent as well as manifest meanings of political behavior,

progress has been

little

made

as

yet in publicly defining

the observational standpoint and the meanings of the investigator.

plicit

and

am

thinking of something

more than an

ex-

statement of the observer's theoretical framework

its

scientific

ment of

meaning-content, or even

a clear state-

his goal values or policy preferences.

What

Observational Dilemmas

[115]

have in mind is those often preconscious or unconscious

meanings that may affect his observational stance. For
even though two observers may vs^ork with the same formal-theoretical frame of reference or with the same goal
values, they may still make different observations. I am
not demanding that every behavioral researcher should
expose himself to intensive psychoanalysis. I am only suggesting that sustained self-observation

on the observer's

part as he relates himself to the observed will consider-

ably reduce w^hat

feel to

be

very genuine dilemma of

behavior research.
Training in self-observation

political

is

only the beginning of

the behavioral quest for reliable observation. Part of the

observational dilemma in politics, at least at present,

backward

training in research

generally.

Compared with the

is

the

methods and techniques

training

in

other

dis-

not utilize the technological

opportunities now available. Admittedly, the behavioral
approaches in the study of politics are not exclusive alternatives to the more traditional historical, legal, or institutional approaches in political science. In the last
ciplines, this trainings does

fifteen years,

havioral

who

most of those

have come to the be-

persuasion in politics have been largely self-

taught. Trained in the traditional techniques of political

science, they

had to develop new

skills.

Though much

may have
them, their background in classical political theory, historical knowledge, and institutional concerns sensitized the new group of behaviorists
to the observational dilemma. I sometimes feel that some
of the traditional curriculum in political science

been

a waste of time for

of the younger practitioners of the behavioral persuasion

in politics, impatient
full

mastery of the

with

new

this

background, yet not in

techniques, lack the sophistication

As a result, techknow-how is confused with

that behavioral observation requires.

nological and methodological

knowledge

that, in the case of political behavior,

must

in-

clude sensitivity to the dilemmas of observation.

The roots of the behavioral persuasion in the traditional
foundations of political science are as important in the re-

Behavioral Dilemmas

[ii6]

searcher's self -orientation as the novel developments in-

theory and method taken over from the other behavioral

sciences.
cannot vi^rite off traditional approaches without understanding them. This means that the political behaviorist's training will be prolonged. Not only is he expected to master what is viable in the tradition of political
science (over which there will be disagreement), but he
is also expected to be familiar with the advances made
in the neighboring behavioral sciences, with statistics and

We

other empirical research techniques. These are demanding

requirements, and relatively

few

are willing or able to

survive such strenuous training. For this reason alone, the

practice of behavioral research in political science will re-

main

a very distinct and separate enterprise, at least in the

immediate future.
The observation of political behavior is expensive as
well as elusive. Even a very small study is extraordinarily
time-consuming. The more spectacular analyses of individual political behavior reported in the last few years,

analyses of electoral decision-making, legislative conduct,

community

involved teams of researchers,

and analytical work, and considerable expenditures. At the same time, the financial resources
or

politics,

several years of field

presently

available

The

limited.

result

for
is

political

that

many

behavior research

are

behavioral studies appear

unimportant. But a study's size does not determine

its

sig-

nificance.

Just the opposite

built

by

may

be

theory, methods and data.

matter

true.

An

empirical science

is

the slow, modest, and piecemeal cumulation of

how

The importance

of a study, no

big or small, must be judged in the total con-

text of relevant research.

The dilemma

of behavioral re-

not scarcity of studies but absence of a

cumulative body of theory within which new studies can
be accommodated and digested so that as research proceeds, one can speak of an expansion of knowledge.
An abundance of studies does not make for cumulative
knowledge. Studies may be inventoried and codified, but
search in politics

even

if

is

they are reinterpreted in terms of some compre-

Observational Dilemmas

[117]

hensive, theoretical schema, codification

for cumulation. Reinterpretation

original intent of a study,

knowledge

is

cription that

confounded.

and the dilemma of cumulative

On

the other hand, the pres-

vi^ithin

the same theoretical frame-

a counsel of perfection that

is

substitute

violence to the

research on a given substantive problem

all

should be undertaken

work

no

is

may do

is

quite unrealistic in

the present stage of development in political behavior re-

There are too many competing theoretical approaches that are plausible and possible, and it will take a
good deal more empirical work before any one theoretical
search.

orientation can be expected to seize the field.

success

is

not

legislated. It is

theory's

proved or disproved by re-

search.
difiicult to

It is

disprove a hypothesis in political re-

To

be acceptable, it would seem to require much

more replication the independent empirical testing of
theoretically derived propositions than is practiced at
present. Without a great number of retests, it is impossible to assess the reliability of proof or disproof; there is
search.

much

professional prejudice against repeat performances.

So-called originality rather than replication receives the

kudos of the

men

of learning.

It is difficult

to find financial

support for replicative studies. Graduate students doing

doctoral research are admonished to come up with fresh
contributions to knowledge. This is only part of the
dilemma. The difficulty of maintaining sufficient controls
over external conditions is more important in" the long
run. If conditions vary, changes in hypothesized relationships cannot be observed and measured. Replication seems
worthwhile only if the analysis proceeds from hypotheses
about the constancy of relationships among variables

under

like conditions.

Again,

am

not proposing perfection.

am merely

hinting at perfection to pinpoint a dilemma. Conditions

in political life are rarely alike, but this
despair. It

similar

to

is

probably enough

make

replication

if

is

no reason for

conditions are sufficiently

worthwhile.

properly, comparative analysis serves in

If

undertaken

many

respects

Behavioral Dilemmas

[ii8]
as a substitute for
sible in

more rigorous

replication than

is

pos-

laboratory experiments.

do not believe experimental replication is impossible

and I disagree with those
who say that a rigorous science of politics is beyond our
reach because human beings cannot be manipulated to suit
the experimental needs of the researcher. However, I do
not want to minimize the difficulties of experimenting
with political behavior or institutions. Some very useful
experiments in the stimulation of voting or with small
groups, for instance, have been made in both laboratory
and field settings. The difficulties of experimenting with
people in laboratory or real-life situations are not due
to obstacles inherent in the human condition. Medical research would not have made the advances it has made
if people had not agreed to submit to experimentation.
I

in political behavior research

believe the time will

advantages

of

come when people

experimentation

will realize the

the

for

prediction

of

policy consequences and will cooperate with behavioral

researchers in the study of politics, as they already do

cooperate in what

is

"human

called

relations research"

in industry.

Experimentation in political behavior research will do

more than facilitate replication and cumulation. It will

also make available new data that can only be guessed at
now. It is fashionable, especially among traditionally oriented

political

to

scientists,

belittle

the

psychologists

whose favorite research site seems to be the college

campus where students are available in abundance for
testing or experiment.

students have

Yet out of the studies with college

come new methods

of inquiry

now

used

by

and
analyzing data that, until recently, seemed forever unobtainable. The application of scale and factor analyses to

behavioral practitioners

legislative

roll-calls

or

in

politics

judicial

in

discovering

decisions

has

yielded

data that were previously inaccessible, and these techniques

have revealed important behavioral patterns that are related to institutional outputs. Sociometric techniques, first

invented in the study of small,

artificial

groups,

now make

Observational Dilemmas
it

[119]

possible to identify in great detail informal influence

and processes
whole communities.
structures

in large

institutions

and even

In other words, the availability of behavioral data

itself facilitated

vances.

The

by new methodological and

is

technical ad-

study of elections and electoral behavior was

limited for a long time to the use of aggregate voting rec-

ords alone.

It

was the systematic interview survey and

probability sampling that brought the study of elections,

which

are surely critical democratic institutions, to the

point where genuine explanation of variances in behavior

became possible. We now take the probability sample survey so for granted that we easily forget the lack of behavioral knowledge about elections only a few years ago.
At the same time, behavioral research must be modest
in appraising its data. Much unnecessary argument follows the lack of candor about limitations in the new kinds
of data that are being made available. In view of the difficulties with direct observational techniques when it
comes to the study of large groups, behavioral research
must rely, of necessity, on opinions, perceptions, and
verbal reports of behavior.

The

behavioral persuasion in

cannot claim more for these data than what they

are. For instance, sometimes research seeks to discover the
influence attributed to certain individuals in a group, or-

politics

ganization, or

community on people or

decisions.

These

data cannot be taken as evidence that these individuals

are actually influential in general or in particular situations.

Statements about influence or power based on meas-

ures of attribution are always inferential and this should

be admitted frankly. On the other hand, I cannot accept

argument that influence attribution is altogether
meaningless, and that it should be tested independently
in some objective, factual sense. It seems to me that if
such independent testing of influence were possible, the
discovery of influence attribution would be unnecessary

the

in the first place.

The

case studies

very

it is

point

is,

of course, that apart

difficult, if

cover influence or power

from

not impossible, to

as real, generic

dis-

phenomena.

Behavioral Dilemmas

[i2o]

Moreover,

perceptions

observable behavior

or not.

If I

all

may

their

have

consequences

perceive another person as influential,

well behave towards him as

for

own, whether factually true

I

may

he were in fact influential.

If I find out that he is not, or less so than I expected, I
will have "misbehaved," again with identifiable consequences for the political relationship involved. I will have
to change my perception. Undoubtedly, attributions of influence are based on just such experiences and, for that
if

reason too, constitute "definitions of the situation" that

have a very real existence of their own and, therefore, are

proper objects of behavioral investigation. Even a situation that is misperceived may be real in its consequences.
Perceptions may or may not correspond to reality, but
as the determination of reality is elusive, they may have
to serve as substitutes. They simply should not be majde
as statements about reality. The methodological problem
is not the degree of correspondence between perception
and reality. The problem is rather to decide just what degree of inter-subjective agreement among informants or
observers should be accepted as constituting sufficient ev-

idence for making inferential statements about reality.

This

is

very

much

like agreeing

on what

constitutes an ac-

ceptable level of confidence in probability

statistics.

No

needed to cope with the problem of the relationship between perceptions and reality.
There are many problems which interest behavioral researchers in politics for which data continue to be scarce
injection of metaphysics

is

or inaccessible. Tool research, that

is

research not

imme-

concerned with new knowledge but with fashioning new techniques for gaining knowledge, is one way to
come to grips with this. Tool research has not been very
high on the agenda of political science and, so far, political scientists have mainly borrowed techniques from the
other behavioral sciences. But there are problems unique
diately

to political science that require technological ingenuity.

Because political behavior invariably occurs in

institu-

tional settings, the size of the institution necessarily deter-

Observational Dilemmas

[121]

mines the number of subjects for behavioral inquiry. The

behavior researcher cannot ignore a small institution just because it is peopled by few individuals, for it
may be a very important sub-system of the political
system. But because the subjects are few, as in a legislative
body, the number of analytical categories is likely to exceed available data. Multivariate analysis necessary to do
justice to theoretical propositions cannot be pursued very
far. The researcher has the choice of either reducing the
number of his analytical categories or inventing an alternative way to handle his data. For instance, in a study of
political

the legislative as a behavioral system,

found

it

neces-

sary to develop a two-stage type of analysis that linked to

each other the most frequently observed pairs of related

variables, in this case role orientations, in a configuration

of pairs until
hausted.

make

The

all

most frequent

pair relationships

were

ex-

researcher faces the dilemma of having to

and more unrefined analysis than his themight call for. But even such relatively

a simpler

oretical ambitions

gross analysis

is

preferable to purely impressionistic link-

ing of variables and treatment of complex phenomena.

As

a final

example of

this

type of dilemmas, there

is

the gap between case analysis and systematic analysis. Al-

though

a great

many

cases have

ferent areas of politics,

been accumulated

in dif-

from the administrative process

and legislative decision-making to party and pressuregroup politics, the grounds for selecting the cases are

somewhat obscure. In

general, the substantive interests of

the researcher and his convenience, that

is,

accessibility of

documents and persons, seem to be the guiding criteria.

These are hardly scientific criteria.
Most of the cases now available deal with exciting,
spectacular, and perhaps critical situations rather than
with more modal situations. As a result, the degree to
which the cases are typical or not and of what universe of
cases they might be representative is not clear at all. Then
there

is

Would

the question about the reliability of case studies.

a second researcher,

working with the same mate-

Behavioral Dilemmas

[i22]

and dealing with the same informants, make factual

rials

discoveries reasonably similar to those

of the

first

re-

searcher, so that the findings cannot be considered ran-

dom?

studies.

am speaking of findings, not interpretation in

Few cases are cast in a theoretical framework

controls

makes

it

what goes

case
that

what is left out. This

draw inferences from a

into a case and

extremely dangerous to

variety of even presumably similar cases. Significantly,

although cases are said to be rich sources of hypotheses

about politics that are to be subjected to more rigorous

and systematic research in the future, few follow-up studof a replicative character are ever made.
In principle, there is no generic difference between the
case method, the comparative method, and the statistical
method. The difference between these methods does not
lie in the data or in how the data have been collected. The
difference Hes in the nature of the inferences that can be
ies

made from

we

the evidence in the use of these methods. If

grant that the inferences

are

made

probability statements rather

the difference

lies in

in aU three

methods

than eternal

verities,

the level of confidence for the truth

value of the statements that can be legitimately accepted

as appropriate.

methods are

way

would argue,

therefore, that

quantitative. If so, the inferences

all

three

made by

of statistical techniques can be accepted with greater

confidence than inferences

made by comparative modes

more rehable than inferen-

of analysis, and the latter are

statements based on consideration of a single case.

This does not mean, of course, that one method

tial

is

some absolute sense.

The method one chooses depends on a great many considerations, each of which influences the strategy of research designs. There may be disagreement with regard
to a particular piece of research as to which method is
most appropriate. All things considered, the researcher
will choose the method that will yield the most satisfactory results. How to make the right decision is on another
necessarily better than the other in

agenda of methodological discourse.

Macro-micro Dilemmas

[123]

MACRO-MICRO DILEMMAS
The

great issues of politics,

justice,

war and

peace, freedom and

order and revolution, and so on, require as

minute investigation

as

more modest problems.

much

Political

science is necessarily concerned with these issues. It has

been in the tradition of the discipline to cope with them
in molar rather than molecular fashion. As a result, there
has been a misunderstanding between those interested in
the great issues and those who settle for solving simpler
problems. I do not believe that the broad perspective
characteristic of large-scale policy studies is at odds with
the microscopic inquiries of the behavioral persuasion.

Indeed, linking different levels of analysis, the levels of

individual or small-group behavior

and the global

levels

of institution, community, or nation, constitutes a major

unsolved item on the methodological agenda of the behavioral persuasion, and a challenging one.

The
clear.

solution of the macro-micro

The problem

is, first

of

all,

problem

is

far

a theoretical one.

from
Con-

cepts of sufficiently high and abstract generality are

needed to accommodate the multitude of levels on which
political discourse can be conducted within a single theoretical schema. The theory must also be satisfactorily
operational to permit empirical research on the level of
the individual political actor. Recently schemata centering
in concepts like decision-making and conflict have made
progress along these lines. But not surprisingly, empirical
research has been conducted in connection with issues
where access to behavioral data is relatively easy. For instance, propositions derived from conflict theory can be

more conveniently

tested in research on metropolitan

than international problems. Perhaps the issues of metropolitan conflict are pettier than those of international
conflict, but because a great deal of behavioral research
has been concerned with small rather than large issues, it
does not follow that the latter are ipso facto beyond the
range of behavioral inquiry.

Behavioral Dilemmas

[124]

Concentration of empirical research attention on areas

where behavioral data

on macro-analytical

analysis

makes sense

are accessible

methodological

the

respects

where data

are easy to

come

if

one

behavioral

has given the

levels. If this

impression that behavioral investigation

in areas

of

difficulties

is

possible only

by, the impression

Because the problem of linkis not simply theoretical but methodological, it would be sheer folly, at
this stage of development, to seek solutions in areas where
is

understandable but

false.

ing macro and micro levels of analysis

data are difficult to gather. Therefore, the behavioral re-

searcher

is

well advised on strategic grounds to seek solu-

tions of the

macro-micro problem in research sites where

is viable. This self-imposed limitation

behavioral research

of research attention

is

proof of strength, not weakness,

in the behavioral persuasion in politics.

The macro-micro problem can be visualized as a

dilemma of diverse, though continuous, observational
standpoints.

From

tween macro

units

this

perspective,

the

and micro units of

distinction

analysis

be-

becomes

Small units like the individual or the small group,

relative.

and large units like the organization or the nation-state,

can be treated not as polar but as continuous variables.
If this is so, all units, small or large, should be subject to
ordering on a single continuum. Research can cut into
this continuum anywhere along the line. In fact, this is
what traditional analysis has largely done in concentrating on institutions as units of analysis. From the observational

standpoint of a political system conceived as a

macro

unit, institutions

On

the other hand,

appear as relatively small

from the standpoint of

units.

individual be-

havior, an institution, even one as small as a committee,

looms as a fairly large unit. A nation-state is a macro unit
if looked at from below, and a micro unit if looked on
from above (the standpoint of a regional system or the
global community of states). The task of theory and re-

search

mental

is

to link these diverse units in terms of the seg-

relations

discoverable

between

them,

whether

Macro-micro Dilemmas

[125]

vertical or horizontal, as parts of continuous sets of inter-

locking behavioral chains.

As

have already suggested, such research

is

predi-

cated on the availability of sufficiently general and abstract

concepts that permit theoretical and operational

between and among different levels of analysis.

Role analysis or system analysis, alone or in combination,
is well suited to this task. For example, if position is taken
as the basic micro unit of any institutional system, and
act as the basic micro unit of any behavioral system, act
and position can be linked, and with them the macro unit
called institution and the micro unit called individual, by
virtue of the role that an actor (the micro unit) takes in
the performance of acts that are relevant to the position
he occupies in an institution (the macro unit). Put
linkages

somewhat

differently, the

macro unit

institution

pears as a role system that can be analyzed

now

ap-

on the micro

level of the individual actor.

Theoretically speaking, any unit of political analysis,

whether large or small, depending on one's observational
standpoint, can be thought of as a behavioral system of
action related to other behavioral systems of action.
if

theoretical discourse

straction

can

we

is

Only

raised to this high level of ab-

speak meaningfully of a relationship

between an individual and an institution, or between a

group and a government, or between a nation-state and
the world community. Hopefully, such reconceptualization of concretely diverse units makes behavioral analysis
possible on all levels and permits valid inferences from
one level to another, without danger of traditional fallacies. To mention just two of these: there is the fallacy
of extrapolation, i.e., treating micro systems as analogues
of macro systems and extending findings from the microto the macro-level of analysis; and there is the fallacy of
personification, reducing macro phenomena to the micro
level of analysis, as in some formulations of national character.

The problem

is

not one of immanent conflict between

Behavioral Dilemmas

[i26]
individual and group, or

between group and state, as

sometimes made out to be in normative political doctrine. Rather, from whatever point on the macro-micro
continuum one proceeds, the task of research is to build,
by patiently linking one unit with another, the total chain
it is

of interrelations that link individual to individual, individual to group, group to group, group to organization, or-

ganization to organization, and so on, until one gives the

network continuous order. The use of behavioral

on some
theoretical construct of the total system in which the inentire

data in political analysis depends quite clearly

dividual actor

ing

of

linkage.
analysis

is

the basic empirical unit.

conceptual

The

alone,

but

am
of

not speakoperational

availability of the great issues to behavioral

clearly

is

linkage

contingent

macro-micro dilemma.
The problem of using both
in behavioral analysis

is

on the solution of the

discrete

closely related.

and aggregate data

What may be true

who

of aggregates need not be true of the individuals

compose them. There

make an

to

is

a simple reason for this. It

incorrect inference

is

when moving from

easy

state-

ments about the behavior of aggregates, such as a national

electorate, to the behavior of any one person within the
aggregate. In other words,

any one person

possible to
if

it is

impossible to predict

how

though

it is

in an aggregate will behave,

make

probability statements about his behavior

we know enough

about the behavioral parameters of

the aggregate to which he belongs.

The use of aggregated data conceals much of the variance in the behavior of individual political actors that
the use of discrete data reveals. This .does not mean that
statements based on aggregated data cannot be trusted
and accepted. Aggregate data are often the only kind of
behavioral information available for the purpose of

ing statements about groups or large

we

should not

want

we

to

stiU

make

make

mak-

collectivities.

this necessit}^ into

a virtue. If

But

we

behavioral statements about large systems,

have the problem that aggregate data are eviOn the other hand, even if individual

dently not sufficient.

Dynamic Dilemmas

[127]

data are available and are combined to permit statements

about units larger than the individual, such aggregation
may still do violence to findings about individual behavior. It has the advantage of showing how great the variance is that aggregate or broad institutional language conceals. But just what we mean empirically when we speak
of a group's loyalty, a party's cohesion, or an organization's morale remains unanswered. Do we speak about a
group property that is independent of the behavior of
any one individual belonging to a group, party, or organization? Or are we really only speaking about the aggregated characteristics of individuals? The extent to
which certain problems of politics can be subjected to
behavioral analysis depends on answers to these questions.

DYNA:MIC DILEMMAS
By

all odds, the most troublesome of methodological

dilemmas in the study of political behavior is the analysis
of change through time. Most behavioral research in politics is cross-sectional, dealing with individuals, samples,
or institutionalized groups at one point in time. Treatment of political change has been left largely to the historians of institutions, with the result that historical and

institutional study in political science are often considered

identical.

This

is

unfortunate because, from the perspec-

tive of a science of politics rather than a history of poli-

the analysis of change through time has as

its goal not

but the discovery of causal relationships. Emphasis on change as a series of conspicuous,

tics,

historical reconstruction

successive events often leads to spurious causal interpretation.

Behavioral study relies on talk, and dead

talk.

Of

course,

it is

men do

not

possible to analyze changes in past

behavioral patterns, or at least their symbolic expressions,

analysis where documentary evidence is available. But this type of work is enormously
time-consuming, and unless one is interested in historical

by applying content

Behavioral Dilemmas

[i28]

patterns for purposes other than their

own

sake, the sci-

not likely to be rewarding. As a

result, be-

havioral analysis of politics has been largely

a-historic

entific yield

is

(though not anti-historic). The defense, or better, the rationalization, has been that, in any case, the task of behavioral research is to establish functional rather than
causal relations between variables. This is a rather disingenuous avoidance of the causal challenge.
There is nothing intrinsically a-historic in the behavioral persuasion in politics. In fact, as cross-sectional stud-

accumulate, for instance, in the voting

ies

field,

it

be-

comes increasingly possible to study change and infer

causation from the presence or absence of correlations.
Comparison and correlation of cross-sectional studies

made

at different points in

time are subject, of course, to

the same limitations that characterize trend analyses of

They permit identification of net changes

behavior of the cross-sections, but they do not al-

aggregate data.
in the

low

of

analysis

internal

individual persons.

The

changes

may be compensatory

of individuals

the

in

behavior

of

behavioral changes characteristic

in the cross-section

changes in one direction may

be offset by changes in the other. But the marginal re-

as in the aggregate; that

is,

may indicate no, or little, change. Moreover, if there

too brief a period for which cross-sectional data is

sults
is

available, the

tions

is

time series constructed from the cross-secunique events and

likely to reflect a sequence of

again makes for spurious inferences about causation.

However, in the absence of genuinely longitudinal infor-

comparison

mation,
data are
tional

still

more

analysis

and
reliable

based

correlation

of

cross-sectional

than global historical-institu-

on spectacular occurrences. The

through time and through com-

analysis of cross-sections

parison

is,

of course, possible only

if

the data collected

in different periods are genuinely comparable.

Longitudinal treatment of behavioral data

reliable in causal analysis than the use

is

much more

of cross-section

through time. This type of analysis may range from a

minute-to-minute account of the behavior of a single in-

Dynamic Dilemmas

[129]

dividual to the use of aggregate data

analysis

is

At

tremes.

a well delimited territory.

where the unit of

These are the ex-

the present time, however, longitudinal analy-

of political behavior at the individual level has been

sis

applied most successfully through the panel method. In

method, a panel of the same respondents is interviewed at different points in time. This makes it possible to observe changes in behavior that might otherwise
this

be ascribed to intervening events like an election or a

crisis.
can then ask respondents about these events

We

directly

and identify the sequence of cause and

peated interviewing of the same respondents

effect.
is

Re-

expen-

and for this reason alone the panel method has not
been used as widely as it might. It has been used
most often in before- and after- election studies, that is,
over a relatively short period of time. But there is no
reason to suppose that, given sufficient resources, it could
not be used in the study of political behavior over a numsive,

ber of years.

Apart from cost, the main difficulty with the panel

method of longitudinal analysis involves sampling. In
order to investigate a sample over an extended period of
time, the characteristics of the population-to-be

that

is,

the population concerning which statements are to be

made

must be specified. Loss of subjects, or drop-out,

and changes in the attributes of the respondents in the
course of the study, such as age or economic status, make
this specification of the
cult. It is

probably for

population-to-be extremely

this

diffi-

reason that panel studies have

been short-range, involving from two tp four interview

waves. This avoids technical difficulties stemming from
changes in both interviewers and respondents. Other difficulties

that will adversely influence spontaneity in

an-

swers to the same questions are "sample bias," arising out

of differences in the characteristics of those willing to be

interviewed several times and those unwilling; and "reinterview bias," that

is,

undue

self -consciousness

on

re-

spondents' part about being interviewed repeatedly.

Even

in the face of these problems, the panel

method

is

Behavioral Dilemmas

[130]

the most promising technique for studying political behavior through time. Its application belies the contention
that the behavioral persuasion in politics
anti-historical. It certainly

is

is

necessarily

the most dependable tech-

nique for studying the process of cause and effect, for it

permits description of the direction, degree, and character of change.

This suggests a subtle interplay between techniques of

developmental analysis, logistic requirements of research,

and

theoretical

formulations.

If

cross-sectional

analyses

are framed in terms of models that are a-historical, such

as structural-functional
or input- output formulations,
adoption of these theoretical models has, in turn, tended

to limit behavioral studies to cross-sectional techniques.

On the other hand, theoretical models of change through

time are rarely satisfied by cross-sectional surveys taken
at different points in time. But the expensiveness of longitudinal research and other tactical difficulties have militated against much theoretical concern with political
change at the level of the individual person.
should not confuse longitudinal analysis or comparison of cross-sections through time (trend analysis)
with those historical reconstructions of past events which
characterize the chronological treatment of large-scale
political phenomena, or with the equally historical re-

We

as practiced

by

historians or psychoanalysts

do

constructions of individual
psychoanalysis.

Whether

life histories

the reconstructing, they provide important clues in un-

derstanding of change, but as statements about causes and

consequences they are necessarily inferential.
Developmental analysis is perhaps the most exciting
approach to the study of political change through time.
Using this approach, the researcher deals with political
behavior in the present by studying cross-sections in
terms of whatever theoretical behavior model he may
choose, but he does so against the background of both
historical reconstructions and trend analyses, as well as
against the foreground of whatever images he may build

Bibliography

[131]

for the future. These are developmental constructs that

emphasize the "from where, to where" sequence of

and hypothetical events. Moreover, developmental
sis seems well suited to bridge the empirical world
political scientist and the normative world of the

actual

analyof the

policy

planner.

BIBLIOGRAPHY
OBSERVATION, MEANING, COMMUNICATION
Boulding, Kenneth,

Michigan

The Image. (Ann Arbor:

University of

Press, 1956.)

Cherry, Colin,

On Human

Communication.

(New

York: Wiley,

I957-)

Kecskemeti, Paul, Meaning, Communication and Value. (Chicago: University of Chicago Press, 1952.)

Mead, George H., Mind,

Self

and Society. (Chicago: University

of Chicago Press, 1934.)

Osgood, Charles E., Suci,

George J., and Tannenbaum,

Percy H., The Measurem,ent of Meaning. (Urbana: University of Illinois Press, 1961.)

METHOD, EVIDENCE, INFERENCE

Buchler, Justus,

The Concept

of Method.

(New

York: Colum-

bia University Press, 1961.)

Chapin, F. Stuart, Experimental Designs in Sociological Research. (New York: Harper, 1955.)
Durkheim, Emile, The Rules of Sociological Method.
York: Free Press, 1958.)

Francis,

Roy

G.,

The Rhetoric

(New

of Science. (Minneapolis: Uni-

versity of Minnesota Press, 1961.)

Greenwood, Ernest, Experimental Sociology: A Study in

Method. (New York: King's Crown, 1945.)
Lemer, Daniel (ed.). Evidence and Inference. (New York:
Free Press, 1958.)

Beliavioral

[132]

Dilemmas

DESIGN, TECHNIQUE, STATISTICS

Chernoff, Herman, and Moses, Lincoln E., Elementary Decision

Theory. (New York: Wiley, 1959.)

Goldberg, Samuel, Probability: An Introduction. (Englewood
Cliffs,

NJ.: Prentice-Hall, i960.)

Herbert, Survey Design and Analysis.

(New York:
Free Press, 1955.)
Lazarsfeld, Paul F., and Rosenberg, Morris, The Language of
Social Research. (New York: Free Press, 1955.)
de Sola Pool, Ithiel (ed.). Trends in Content Analysis. (Urbana:
University of Illinois Press, 1959.)
Hyman,

Selltiz,

Stuart

Claire,

Jahoda,

Marie, Deutsch,

W., Research Methods

Morton, and Cook,

(New York:

in Social Relations.

Holt, 1959.)
Wallis,

W.

Approach.

Allen, and Roberts,

(New

HISTORY, CHANGE,
Allport,

Harry

York: Free Press,

Gordon W.,

V., Statistics:

A New

1956.)

DEVELOPMENT

BecoTning.

(New Haven:

Yale University

Press, 1955.)

Benson, Lee, Turner and Beard. (New York: Free Press, i960.)
Goldfarb, Nathan, Longitudinal Statistical Analysis. (New

York: Free Press, i960.)

Harold D., The World Revolution of Our Time.

Lasswell,

(Stanford: Stanford University Press, 1951.)

Maclver, Robert M., Social Causation. (Boston: Ginn, 1942.)
Teggart, Frederick J., Theory and Processes of History. (Berkeley: University of California Press, i960.)

EPILOGUE

THE GOAL

The

goal

is

man. This

is

MAN

IS

the ultimate justification for pur-

suing the behavioral persuasion in politics, as

other

human
is

of any

made

sense to me. This, of

and as such it cannot be
demonstrated to be either true or false. In

the sake of science has never

course,

it is

enterprise. Like art for art's sake, science for

a statement of value,

scientifically

this regard, then,

my

understanding of the behavioral peron a premise that may be accepted

suasion in politics rests

or rejected but that cannot be proved or disproved.

scientific study of man in politics
not saying very much, for men disagree on the nature of man in politics. Which is the man
in whose service the behavioral persuasion finds its reason
just man?
for existence? Is he a democratic man?
power seeking man? Is he a man who must be controlled
because he is brutish and nasty? Or is he a man who must

But to say that the

has

man

as its goal

is

be liberated from the shackles of oppression to

live a dig-

These are philosophical questions better left

to the philosophers. Whatever answer one chooses, there
is likely to be a corresponding predicament. The most we
nified life?

Epilogue

[134]

can say
of

man

is

that different

men have

different conceptions

and, as a result, give different meanings to

what

they do and why they do it.

I do not believe that the philosophical predicament
need be a scientific predicament. Whatever philosophical
views different scientists may hold about man and the
reality of

man, they need not interfere with their work in

field. For there the validity of

the laboratory or in the

theoretical propositions about human behavior, from

whatever philosophical position derived, is a matter of inter-subjective agreement, not absolutist assertion, and the
reliability of observations is a function of measures that
are inter-subjectively agreed on as well. The very existence of any scientific enterprise is predicated on inter-subjectively consensual rather than subjectively philosophical

notions about man, reality, or universe.

Is the man in whose service scientists do their work a
mere phantom then? Clearly not. For if "man" does not
exist, there are many men, acting men, valuing men, goalseeking men, and among them the scientist himself. If
this is so, and if science is justified by its services to men,
is

it

possible for science to be value-free,

with the values

men

unconcerned

cherish and the goals they seek?

Can

maximize whatever particular

values particular men prefer be neutral in a world of men
who disagree? The issue has been debated for decades,
with little agreement.
This should not be a source of despair. It does not
a

science

mean

seeks

that

to

that a science of

man

is

havioral scientists cannot agree

tween empirical methods and

tions.

human

Indeed,
goals

believe

that

impossible because be-

on the

relationship be-

their normative

implica-

the very multiplicity of

and values makes the

scientific

enterprise,

including the behavioral persuasion in politics, the chal-

lenge that
of values

is is. For it is because of the seeming multiplicity

which men hold that behavioral science cannot

and must not avoid dealing with the preferences, values,

and goals of men. It does not follow that science cannot
investigate preferences, values, or policy objectives be-

The Goal

Is

Man

[135]

cannot tell us what goal is best or what action is

just. I do not believe that these things are any more inscrutable
than more mundane matters of political
cause

it

behavior.

While

ethics

is

a legitimate pursuit in

its

own

right,

it

does not have a copyright on propositions of value. Logicians have long occupied themselves with the problem of
consistency in the order of values. Behavioral scientists,

might be able to contribute something

This something, I daresay, is nothing less than
inquiry into the problem of the universality of values, a
universality that any ethical philosopher must assume if
he wishes his propositions of how men should conduct
themselves to be accepted. I do not believe for a moment
that the ethical philosopher engages in his search for what
is wise, right, or just for its own sake any more than the
in politics and out,
as well.

work for the sake

scientist does his

of science.

In seeking to discover empirically whether and what

universal

by men anywhere, the befrom the same assumption as

philosopher. Whatever success may accompany

values

are

held

havioral scientist proceeds

the ethical

it seems to me that he is in a better position

than the ethical philosopher to pass judgment, not, of

his search,

course, in his role as scientist, but in his role as moralist.

common humanity,
inhumanity to man, the
behavioral scientist cannot escape the task of determining
what is human and what is not. But unlike the ethical philosopher he can say, "I have been there."
The question of whether a value-free science of politics
is possible must not be confused with the question of
whether a value-free science is desirable. The former is a
problem of fact that, in the end, can be answered only
through empirical research into the nature of science as a
form of human activity. The latter is a problem of value
that will remain open as long as scientists themselves give
different answers. I think it will remain an open question
for generations to come.
If a value-free science of politics is possible, it can be
As long

as

there

just as there

is

is

talk about man's

talk about man's

Epilogue

[136]

put to the service of good

well as slavery, of

well as

as

evil,

of freedom as

well as death. In this respect a

life as

science of politics only shares the supreme dilemma of

the sciences, natural and behavioral.

all

would be most

It

presumptuous to assume that political science has at its

disposal a knowledge of good and evil, of justice and injustice, of right and wrong. But I believe that the position
that a value-free science of politics is undesirable makes
just that
It

presumption.

does not follow that a value-free science of pol-

itics is

undesirable because

Values

may

it

may

be

difficult to achieve.

creep into investigations of politics at almost

any stage of the research process, from the selection of

the problem to the interpretation of the findings.
believe that
biases will

any

make

clarification

and elsewhere,

study any more scientific. I

some behavioral scientists, in

his

rather strange that

feel that

if

don't

of the researcher's value

find

it

politics

they only lay bare possible value

biases, their research will gain in scientific stature. In that

case, so the

argument goes, any

scientific appraisal of the

research can discount the value bias and determine the

degree of objectivity that has been reached. This argu-

ment

strikes

me

as rather curious,

because

if

the scientist

is

be up to him
rather than to his critics to control them. If, on the other
hand, he persists in proceeding with his scientific work
though he knows it is biased, I can only conclude that he
must find his practice desirable. He says, in fact, that he
cannot eliminate his biases and that, therefore, he might as
well live with them.
This stance should not be mistaken for the approach to
the problem of values and scientific research on politics
called policy science. The policy science approach does
not assume that a value-free scientific study of politics
is impossible because men pursue values through politics.
Indeed, it sharply distinguishes between propositions of
fact that are believed to be subject to scientific-empirical
inquiry, and propositions of value for which empirical
science has as yet no answer. But it does not deny that
cognizant of his biases,

it

would seem

to

The Goal

Is

Man

scientific research

[137]

on propositions of

policy objectives; indeed,

as

all

science,

ever goals

it

fact cannot serve

asserts that political science,

should be put in the service of what-

men pursue

in politics.

In seeking to place the scientific study of politics at the

disposal of the search for policy objectives, policy science

need not violate the canons of

it is

scientific

method. In

fact,

genuinely sensitive to the tensions and the subtle bal-

ance that exist between fact and value. Rather than contwo realms, it keeps them apart. If it did not

fuse the

the sharp differentiation, we would soon have a

"democratic political science," a "communist political science," an "anarchist political science," a "Catholic political
science," and so on. Curiously enough, this absurdity is
not altogether a phantasy. It is the inevitable outcome of
the position that if only values are clarified, scientific research can move on its merry way.
It is more than likely that as long as men seek to impose
their own values on other men, scientists will have to face

make

the question of whether to place their services at the disposal of the combatants. Their answer will undoubtedly

depend on the values they cherish as individuals with a

moral conscience, but it will also depend on the institutional structure in which they are involved. The decision
to serve group, class, nation, church, or world community,
as one sees these collectivities, is always a matter of personal ethics. But whatever the choice, the scientific way
of life is always dangerous. Even if the scientist sees his
work as being in the service of goals that he himself cherishes, there is nothing in his science that prevents its being
used for ends of which he disapproves. In this sense, at
least, science is value-free. I don't think the scientist can
escape this dilemma of having his work misused without
giving up his calling. Only if he places himself at the service of those whose values he disagrees with does he commit intellectual treason.

Epilogue

[138]

BIBLIOGRAPHY
Ayer, A. J., Logical Positivism. (New York: Free Press, 1959.)
Bronowski, J., Science and Human Values. (New York: Harper,
1956.)

The Case For Modem Man. (New York:

Harper, 1955.)
Leighton, Alexander H., Human Relations in a Changing World.
(New York: Dutton, 1949.)
Lerner, Daniel (ed.), The Human Meaning of the Social Sciences. (New York: Meridian, 1959.)
Frankel, Charles,

Lerner, Daniel, and Lasswell, Harold D.,

The

Policy Sciences.

(Stanford: Stanford University Press, 1951.)

Lynd, Robert

S.,

Knowledge for What?

(Princeton: Princeton

University Press, 1939.)

Myrdal, Gunnar, Value in Social Theory. (London: Routledge

and Kegan Paul, 1958.)
Northrop, F. S. C, The Logic of the Sciences and the Humanities. (New York: Macmillan, 1947.)

Weber, Max, The Methodology of the

York: Free Press,
Zetterberg,

Hans

Social Sciences.

(New

1949.)

L., Social

York: Bedminster,

1962.)

Theory and Social

Practice.

(New

INDEX
Act, 113, 125

Aggregate analysis, 126-7

Analogy and metaphor, 14-15
Anticipatory socialization, 50
Authoritarian patterns, 64
Authority, 34

Behavior and meaning, 6

Behavioral analysis of institution, 15-18

Behavioral methods, 31-5

Behavioral research, i lo-i i
Behavioral techniques, 11 8- 19
Beliefs, 72-4

Belief systems, 74
Biographical study, 90

Bureaucratic behavior, 21-2

Case analysis, 12 1-2

Causal challenge, 128

Change, 127-8

and

Choice, 75-6

political action, 77

as observational standpoint,

Class, 55-6

Class consciousness, 59-60

Comparative analysis, 77-8

Complexity, 113
and roles, loi
as research problem, 30
Conformity, 74

Congruence
1

oif

role

63

Culture conflict, 80

Cumulation in science,

Data

11 6- 17

accessibility, 123-4

Decision-making, 24, 123

Definition of situation, 53-4,

Conflict, 16-17, 123

sonality,

Cross-cultural comparison, 66
Cross-pressures, 51-2
and role conflict, 52
Cross-sections, 128
Cultural analysis and public
policy, 67
Cultural change, 74-9
Cultural consensus, 81
Cultural context, 79
Cultural crisis, 67
Cultural determinism, 75
Cultural dynamics, 76-7
Cultural lag, 75
Cultural meanings, 69-70
Cultural orientations, 68-74
and political behavior, 71-2
Cultural patterns, 63-8
Cultural stability, 67
Cultural transmissions, 80
Culture and personality, 97

and per-

01 -3

Constitution and culture, 81-2

Content analysis, 70
Contextual relevance, 75
Cross-cultural analysis, 82

120
Definitions, 4-6, 23

Democracy

and

personality,

lOI

Description
24

and explanation,

Index

[i4o]

Developmental

analysis,

77,

130-1

Deviant case

100

analysis, 103

Discretionary

role

and per-

Models, 30
Multiple group

Electoral behavior, 31, 49

Electoral research, 119
Empirical science, 116

consequences

membership,

52
,

Multivariate analysis,

1 1

Fallacy of cultural correlation,

96

Great issues of
Group, 46-54

Misperception,
of, 120

sonality, 104-5

Experimentation,

Macro-micro linkage, 125-6

Mental health and democracy,

National character, 99
Natural state groups, 34
Observation, 68, 112
and error in, 113

politics, 123

as

communication, 114

analysis of, 34
and individual, 51

Observational dilemmas,

and role, 48
approaches to, 47

Opinions and

properties of, 49, 127

Panel method,

112-

22
attitudes, 90-5

33, 129-30
Perceptions, 94-5
Personality in politics, 23, 86,

Hypotheses, 22
Identification, 97-8

91

and authority, 98
Inference, 122
Input and output, 28
Institutional research, 120-1
Interdisciplinary
orientation,
19-22
International politics, research

and
and

culture, 99
political
.

study,

85-6,

89,

90-1

and

120, 134
Interview, 114

87-8,

and problem, 91-2

and psychology, 86-7, 100
and recruitment, 101-2, 105
Personality

in, 30-1

Inter-subjectivity, 68-9, 11 3-14,

science,

91. 93,
and political system, 88

questionnaire

method,

92

Policy planning and personalJudicial behavior analysis, 33-4

and personality, 106

Legislative

behavior,

ity, 105-6

Policy science, 136-7

Political behavior, 3-5,

Leadership, 34
16,

42,

10, 16-

17, 20-1, 23, 27

Longitudinal analysis, 128-9

and classical theory, 7-8

and context, 17
and democracy, 10
and great issues, 9-10
and research significance,

Macro-micro continuum, 124

Macro-micro dilemmas, 123-7

and
and

49, lOI

Levels of analysis, 19-24

and configurative analysis,
21

8-9
role, 39-40
stratification, 58-9

Index

[141]

and voting, 33
as technological revolution,

as basic unit, 39-46, 100

as micro-unit, 125

Political change, 33
Political character, 96

Role change, 42
Role conflict, 45-6
and personality, 104
Role expectations, 42-3
Role meanings, 44
Role network, 41
Role sanctions, 44
Role system, 41-2

Political competition, 22-3

Roll-call analysis, 33

Political culture, 79-82

Rules of the game, 6$

31-2

concerns

of, 21

cultural analysis of, 64

group analysis of, 46-54

and change

stability

in,

49-

Political institutions, 15
as cultural products, 68
as

systems of action, 16

Political

man, 39

conceptions

of, 133-4

Political science,

3, $-6, 8-10

and

behavioral
115-16

Political

theory

persuasion,

as decision-making, 24

as

power behavior, 24

Position, 125

Power,

16,

Socialization, 68, 97

Social matrix, 38-60

Social mobility, 59-60

Sovereignty, 29

and t)ehav-

ioral science, 7-8, 27-8

Pohtics, 3-5, 10, 13, 16
as allocation, 28

'

Science, 9
Self-observation, 115
Small groups, 3

Status, ss
Strata, 54-60
Stratification, 57-8
Sub-culture, 66, 79-80, 81

Theory and problem formulation, 22-3

and research,

17,29,74

as attribution, 119-20

Probability sample survey, 33

Problem, 24, 30
Prolonged interview, 92
Rationality, 20, 28
Reality, 120

24-3

as theorizing activity, 25
as tool, lo-ii, 26

Time, 77
Tolerance of ambiguity, 106
Tool research, 120
Units of analysis, 14-19

Recruitment and personality,

collectivities as, 15

persons

as,

106

13-15

Reductionism, 23
Reference group, 53-4

Values and personality, 95-100

Values in science, 134-6

Replication,
Role, 39-46

Voluntaristic

11 7-1

Role analysis, 45
and group, 48
and personality,

roles

and per-

sonality, 103-4

Voting and personality,

45, 100-7

Wholeness, 66

103

no.4^

The

behavioral persuasion

in

p main

320.8D727no.42C,3

lata D331D

smi

l|