COMMUNICATION THEORY; THREE APPLICATIONS IN INTERNATIONAL RELATIONS...
Transcript of COMMUNICATION THEORY; THREE APPLICATIONS IN INTERNATIONAL RELATIONS...
Communication theory: threeapplications in international relations
Item Type text; Thesis-Reproduction (electronic)
Authors Svob, Robert Stanley, 1943-
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this materialis made possible by the University Libraries, University of Arizona.Further transmission, reproduction or presentation (such aspublic display or performance) of protected items is prohibitedexcept with permission of the author.
Download date 01/06/2018 06:41:57
Link to Item http://hdl.handle.net/10150/555120
COMMUNICATION THEORY; THREE APPLICATIONS
IN INTERNATIONAL RELATIONS
by
Robert Stanley Svob, J r0
A Thesis Submitted to the Faculty o f the
DEPARTMENT OF GOVERNMENT
In P a rtia l F u lfillm en t of the Requirements For the Degree of
MASTER OF ARTS
In the Graduate College
THE UNIVERSITY OF ARIZONA
1 9 7 2
STATEMENT BY AUTHOR
This thesis has been submitted in p a rtia l fu lf i l lm e n t of requirements fo r an advanced degree at The U niversity of Arizona and is deposited in the U niversity Library to be made availab le to borrowers under rules of the Library.
B rie f quotations from this thesis are allowable without special permission, provided that accurate acknowledgment of source is made. Requests fo r permission fo r extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his judgment the proposed use of the m aterial is in the in terests of scholarship. In a l l other instances, however, permission must be obtained from the author.
SIGNED:
APPROVAL BY THESIS DIRECTOR
This thesis has been approved on the date shown below:
CLIFTON E. WILSON Professor of Government
Date
DEDICATION
This thesis is dedicated to the l i f e and memory of the la te
Dr0 Martin Luther King* Dr. King knew and lived the tru th : That an
investment of love is the only way to reduce the entropy of hate.
"Free a t la s t , free a t la s t: Great God Almighty, I"m freeat la s t ."
Martin Luther King, J r .
TABLE OF CONTENTS
Page
LIST OF ILLUSTRATIONS . . . . . . . . . . . . . . . . . . . . . vl
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . vi i
1. INTRODUCTION . . . . . . . . . . . . . . . . . . . 1
Purpose ......................... 1Methods . . . . . . . . . . . . . . . . . . . . . . . . . 3Sources . . . . . . . . . . . . . . . . . . . . . . . . . 5
2. BASIC CONCEPTS OF COMMUNICATION THEORY . . . . . . . . . . . 7
A Visual Aid to O rientation . . . . . . . . . . . . . . . 8A Technical D efin itio n of Information . . . . . . . . . . 10
Choice and Information . . . . . . . . . . . . . . . 14Information, Order, and Organization . . . . . . . . 15
The Source . . . . . . . . . . . . . . . . . . . . . . . 17Measuring Information . . . . . . . . . . . . . . . . 18
The Message . . . . . . . . . . . . . . . . . . . . . . . 19The Transm itter . . . . . . . . . . . . . . . . . . . . . 21
Modulation . . . . . . . . . . . . . . . . . . . . . 21Encoding . . . . . . . . . . . . . . . ..................... . 23
The Receiver . . . . . . . . . . . . . . . . . . . . . . 25Prediction and F ilte r in g . . .......................... . . . . . 26
The Channel . . . . . . . . . . . . . . . . . . . . . . . 27The Destination . . . . . . . . . . . . . ..................... . . 30Noise . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Man-Made Noise and Inherent Noise . . . . . . . . . . 31Combatting Noise . . . . . . . . . . . . ..................... 32
Feedback . . . . . . . . . . . . . . . . . . 32Dec i s ion-Feedback . . . . . . . . . . . . . . . . . . 35Info rma t i on-Feedba c k . . . . . . . . . . . . . . . . 35
Goal-Oriented Feedback . . . . . . . . . . . . . . . . . 36
3o THE COMMUNICATION MODEL: THREE APPLICATIONSIN INTERNATIONAL RELATIONS . . . . . . . . . . . . . . . 41
Deutsch: In tegration . . . . . . . . . . . . . . . . . . 45Deutsch: Analysis . . . . . . . . . . . . . . . . . . . 52Russett: Responsiveness . . . . . . . . . . . . . . . . 59
iv
V
TABLE OF CONTENTS— Continued
Page
Russett; Analysis . . . . . . . . . . . . . . . . . . . 67McClelland: Event In teraction . . . . . . . . . . . . . 73McClelland: Analysis . . . . . . . . . . . . . . . . . . 78
h o CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . 84
REFERENCES CITED . . . . . . . . . . . . . . . . . . . . . . . 91
LIST OF ILLUSTRATIONS
Figure Page
1. Simple Communication System 0 0 0 . 0 , 0 . 0 0 9
2 0 Correction System . . . . o o . o o . . o o . * . . o . . o 33
v i
ABSTRACT
Communication theory was derived from the study o f communica
tion apparatus and i t is largely a s ta t is t ic a l and mathematical theory
which explains the transmission o f messages in e le c tr ic a l and e lec
tronic communication systems0 This theory— generally conceded to be
the product of the e ffo rts of two mathematicians» Norbert Wiener and
Claude E. Shannon— has been borrowed by the social sciences as a
framework fo r analys is , or a model, o f the processes which occur in
social and p o lit ic a l systems.
International re lations scholars, beginning with Karl W.
Deutsch, have sought operational indicators of the concepts of communi
cation theory but have met with only p a r tia l success in using the model
to describe and understand the flow of messages or information in the
in ternational system.
The communication model, as represented by the three studies
herein, has been applied to the phenomena of in ternational re la tio n s;
and i t has been a rich source of hypotheses about these phenomena; but
there are s t i l l insights to be gained from a. closer look a t the con
cepts and relationships of communication theory. The model's worth as
an analogy is great but i t cannot be regarded as a method of explana
tion of every kind of communication in in ternational re la tio n s .
CHAPTER 1
INTRODUCTION
We are immersed in a l i f e in which the world as a whole obeys the second law o f thermodynamics: confusion increasesand order decreases0 » « o ^But/ there are local and temporary islands o f decreasing entropy in a world in which the entropy as a whole tends to increase, and the existence of these islands enables some of us to assert the existence of progress,
Norbert WienerThe Human Use o f Human Beings, p. 52,
Speaking of the function of theory, Sheldon S„ Wolin has said
that i t serves "not to amass new facts but to disclose h ith erto unsus
pected relationships between them,"' His observation may be the most
cogent assessment of the value of communication theory to the study of
in ternational re la tio n s . The communication model has provided a new
way o f looking a t the subject matter and i t may have shed new lig h t
on "h ith erto unsuspected re la tio n sh ip s ,"
Purpose
The purpose of th is thesis is to describe the meaning of the
basic concepts of communication theory and to examine three studies in
in ternational re la tions which have attempted to apply some of those
1, Sheldon S, Wolin, " P o lit ic a l Theory: Trends and Goals,"International Encyclopedia o f the Social Sciences VoI, 12 (New York: McMillan Co., 1968), p. 323.
1
concepts by seeking and measuring apparent indicators of those con
cepts* Because of the necessity to p e rio d ic a lly examine the orig ins
and assumptions of any theory, 1 consider the description and exposi
tion o f the concepts themselves (Chapter 2) to be an important task*
But equally important is the necessity to examine the way in which a
theory is applied to its subject matter* Thus, three studies o f d i f
feren t systems of in ternational re lations w il l be examined (Chapter 3)
in order to assess the way in which the communication model has been
adapted* This constitutes the main problem or major question of the
whole enterprise ( i f i t must have one): How has the communication
model been adapted fo r application to the study of in ternational re
lations?
I t seems that a good place to s ta r t in seeking a standard
against which to compare the in ternational re lations applications would
be the theory and practice o f communication engineering its e lf * This
perspective, while i t may not represent the precise point of departure
from which the social science scholars embarked— the source of th e ir
knowledge and understanding of the model— i t should provide us with a
f a i r ly concrete set of meanings (d e fin itio n s and usages) of the con
cepts which are essential to the model*
I t seems obvious and important th a t, i f a model of a physical-
e le c tr ic a l phenomenon is going to be used fo r descrip tion , explanation,
or simply understanding of a subject matter which is decidedly unlike
that which the p rac titio n ers of e le c tr ic a l communication o rd in a rily
work w ith , then that model should be thoroughly understood in terms of
what the model can do, how i t operates, and what its basic conditions
and lim ita tio n s o f operation a re .
Methods
Since much of modern communication theory was predicated on the
establishment o f an adequate understanding of actual communication sys
tems (e .g . , telegraph and ra d io ), i t is essential that th is assessment
of the basic concepts be closely a l l ie d with the physical equipment o f
communication systems a lso . For, without th is empirical base, the d is
cussion o f communication theory quickly becomes mathematically and s ta
t is t ic a l ly complex and extremely d is tan t from the actual phenomena
which are sought as analogies to social phenomena. Accordingly, there
has been no attempt to reduce this investigation to a purely theoret
ical analysis of the mathematical or s ta t is t ic a l theory of communica
tio n - -both of which can represent communication phenomena as
abstracted, numerical processes.
Instead 1 have selected what appear to be the seminal works of
communication theory—works by Norbert Wiener and Claude E. Shannon—
as the theme-setting contributions to the f ie ld . ( I did not read
everything they w rote.) Working simultaneously (and sometimes in con
c e rt) on s im ila r problems of World War I I communication research,
Wiener and Shannon provided the v ita l breakthroughs in theory which
affected every phase of e le c tr ic a l and e lectron ic communication en
gineering. There were others, of course— indispensable colleagues of
Wiener and Shannon who have been less decorated fo r th e ir contribu
tions. But these two men accomplished a v e ritab le revolution (and
they published i t ) and they are the major contributors to what w i l l be
examined in th is study0 Other expositors of the theory (Pierce is
especia lly h e lp fu l) are consulted fo r c la r if ic a t io n and augmentation
of Wiener's and Shannon's ideas. Information theory (which Pierce
equates with communication theory) provides v ir tu a lly identical theo
re tic a l concepts which overlap w ith those from communication theory;
i t d iffe rs only in that its emphasis is on coding.
The studies by Deutsch, Russett, and McClelland which are the
objects of comparison and c ritic is m in th is thesis were chosen because
they represented the several major foci o f scholars u t i l iz in g the model
in in ternational re la tions studies. This choice, by no means, was a
judgment on a l l the other attempts which have been made to find con
crete indicators of the concepts from communication theory; many other
studies would probably have been equally su itab le . Rather, the choice
was made because these models (as I w i l l c a ll them from now on—e .g . ,%
Deutsch's model refers to his 1956 a r t ic le under scrutiny here) sought
or emphasized most of the basic concepts which I found to be the "core"
of communication theory (Chapter 2 ) . I
A second motive fo r picking the three models which I have in
cluded was my fee lin g that these studies represented something of a
trend in the manner of application of the general model. From the be
ginnings with Deutsch and his in terest in integration and u n ific a tio n
(as general concepts), through Russett with his emphasis on responsive
ness as a dynamic function of in teg ra tio n , to McClelland with his new
in terpreta tion of the message as an event and his techniques from
5
information theory: these three models seemed to represent va lid
attempts to measure indicators and thus they demonstrate certa in
phases in a trend of increasing sophistication o f the communication
model as applied to in ternational re la tions*
As fu rth er evidence of soph istication , i t is demonstrable that
Deutsch's indicators were crude measures of volume; Russett expanded
the number of indicators and reduced the number of actors in the com
munication system to two; f in a l ly , McClelland increased the number of
indicators to in f in i ty ( i *6 * , a l l relevant events)—with only data
gathering techniques standing in the way o f near perfect knowledge of
the flow of action in a communication network*
F in a lly , these three models were chosen because they seem to
i l lu s tra te most of the in teresting and crucia l problems of applying
the communication model to the subject matter of in ternational re la
tions,\
Sources
Considerations of space (as always) precluded th is thesis from
being an exhaustive study of the way communication theory has been
used as a framework and a guide to research (in a l l its v a r ie t ie s ) .
No attempt has been made to include critic ism s of Deutsch, Russett,
and McClelland from the f ie ld * This would have unduly burdened the
exposition and analysis; and i t would have been contrary to the p r i
mary purpose of the study—which is to c ritiq u e the three models
against the background of the orig ins of the theory from which they
have arisen .
In the same s p ir i t , the sources from which I have l i f te d the
meanings of the basic concepts were lim ited to a few o rig in a l books
and several secondary expositions* I tr ie d to use books and a rtic le s
which were published in the early and middle 1950‘ s (w ith a few excep
tions) in order to get some fee lin g fo r the point o f view (and depar
ture) which would have been ava ilab le to Deutsch and the others who
were adapting the theory fo r use in the social sciences.
CHAPTER 2
BASIC CONCEPTS OF COMMUNICATION THEORY
A s c ie n t if ic concept has meaning only because sc ien tis ts mean something by ito The meaning is s c ie n t if ic a lly va lid only i f what they intend by i t becomes ac tua l: problems are solvedand intentions are f u l f i l l e d as inquiry continues0
Abraham KaplanThe Conduct of In qu iry* p. 46.
The basic concepts of communication theory are functiona lly
in terre la ted ( l ik e the concepts of any theory) and they a re , perhaps,
best understood when they are seen as interlocking parts of an opera
ting system. In short, the concepts which in terest us, in this study
of a theory which has been applied to the analysis of in ternational
re la tio n s , are most coherently described and leg itim ate ly defined by
presenting them in a systemic form. Therefore, i t seems prudent that
th is exposition of basic concepts be carried out under the d isc ip lin e
of a framework which w i l l help the reader to o rien t himself to the
meaninq of each sp ec ific concept in the system and to the overall per
formance characteris tics of a functioning communication system. In
th is way, the meaning ( i . e . , d e fin itio n and usage) of each relevant
concept w il l become apparent; and each concept w ill be seen in its
relationship to the whole working system.
According to P ierce, one of the seminal truths o f the h istory
of science is that some of the most general and most powerful discov
eries of science have been made by sc ien tis ts who chose to study, not
natural phenomena themselves, but rather phenomena that could be re
constructed in man-made devices. By sim plifying nature, many sc i
e n t if ic models of nature have made i t possible fo r men to bring th e ir
imagination to bear on otherwise inscrutable problems. In short, some
times man's machines and his technologies provide opportunities fo r him
to theorize about the actual operations inherent in nature which he1cannot see, touch, or measure. Advancements in communication engi
neering have, since the 1830's when Morse began working w ith the te le
graph and developing his code, provided sc ien tis ts w ith devices which
they could tes t and manipulate in order to begin to understand the
process of e le c tr ic a l communication. The p ractica l applications of
communication engineering ( i . e . , "hardware11—systems and equipment)
w ill also be important and useful in th is b r ie f exposition of the
basic concepts o f communication theory.
A Visual Aid To Orientation
Since th is is an investigation of the adaptation of an idea
from one f ie ld of knowledge and practice to another f ie ld o f knowledge
and prac tice , i t seems only reasonable that the objects o f the f i r s t
f ie ld should be the leg itim ate subject matter of a study which purports
1. J . R. P ierce, Symbols. Signals and Noise (New York:Harper and Brothers, I 96I ) , p. 19.
to examine the conceptual analogies of the second. For th is reason, I
have not tr ie d to make th is into a purely theoretical discussion and
critiq u e of abstract concepts. For the most p a rt, abstract categories
and mathematically sophisticated conceptualizations of communication
processes are inappropriate; and they lack s im p lic ity ; and therefore,
the basic elements o f a communication system w il l be used as models of
the abstract categories.
To impose order on the stock of concepts which are to be ex
amined below, I have chosen Shannon's diagram of a simple communica
tion system perturbed by noise to aid the reader in orienting each
concept in the system (see Fig . 1 ). Shannon included th is flow d ia
gram in his famous essay "The Mathematical Theory of Communication,"
informationsource transm itter receiver destination
signal receivedsignal
messagemessage
noise source
Fig. 1. Simple Communication System
2published in 19^8= This diagram illu s tra te s the basic elements in a
communication system—elements which generally correspond with the
important concepts o f communication theory as i t is applied in in te r
national re la tio n s . Also, as a model„ th is diagram provides a s ta r t
ing point from which the assessment o f the three models in Chapter 3
can be launched.
But before s ta rtin g the exposition o f the concepts represented
in Shannon's simple communication system above, i t is imperative that
the technical meaning of information be explained. The modern theory
of communication is founded on th is special d e fin itio n of the ordinary
word information; and without a discussion o f th is central concept, a
study of communication theory would be incomplete.^ A discussion of
the concept of information also gives rise to several other important
related concepts which are of major s ignificance in communication
theory. These other concepts w i l l be discussed in turn before return
ing to Shannon's diagram.
A Technical D efin itio n o f Information
One of the f i r s t critic ism s usually leveled a t communication
theory, or more s p e c if ic a lly , a t th is new conception o f information
in communication theory, is the charge that semantics (language mean
ings) are not taken into consideration. I f only p ro b ab ilitie s of
2 . Claude E. Shannon, "The Mathematical Theory of Communicat io n ," in Claude E. Shannon and Warren Weaver, The Mathematical Theory of Communication (Urbana: University o f I l l in o is Press, 1949), p. 5 .
3 . Stanford Goldman, Information Theory (New York: Prentice- H a ll, In c ., 1953), p. 286.
occurrence are considered, the complaint goes» how can any meaningful
assessment be made of the value (to the rec ip ien t) o f the information
transmitted or received? This is a leg itim ate question—one which
Shannon answers emphatically and succinctly..
The fundamental problem of communication is that o f reproducing a t one point e ith e r exactly or approximately a message selected a t another point* Frequently the messages have meaning; that is they re fe r to or are correlated according to some system with certa in physical or conceptual e n t it ie s . These semantic aspects of communication are irre lev an t to the engineering problem. The s ig n ific a n t aspect is that the actual message is one selected from a set o f possible messages. The systern must be designed to operate fo r each possible selectio n , not ju s t the one which w i l l a c tu a lly be chosen since th is is unknown a t the time of design.^
Schwartz re in terp rets the essential re la tionsh ip : "The en
gineer is interested in the re la tiv e frequency of occurrence of in
dividual symbols or groups o f symbols in messages and in the fac t that
these frequencies and not the symbol meanings determine the data
/in fo rm atio n / ra te , . . In short, the semantic aspects of commu
nication are irre levan t to the engineering aspects but the engineering
aspects are not necessarily irre levan t to the semantic considerations
of what is communicated. When Shannon republished his famous essay in
book form, in 1949, he included a companion essay by his colleague,
Warren Weaver. This essay is an excellent prose in terp reta tion of
what Shannon says in his rigorous, mathematical paper. Weaver speaks
4. Shannon, "Mathematical Theory," p. 3»
5 . Leonard S. Schwartz, P rinciples of Coding, F ilte r in g , and Information Theory (Baltim ore, MdTi Spartan Books, In c ., 1963), p. 3 .
to the semantic question and helps to kn it the problems of communica
tion in general together by discussing them on three leve ls . His
three levels are quoted below;
Level A, How accurately can the symbols of communication be transmitted? (The technical problem,)
Level B, How precisely do the transmitted symbols convey the desired meaning? (The semantic problem,)
Level C, How e ffe c tiv e ly does the received meaning a ffe c t conduct ip the desired way? (The effectiveness problem ,)°
At Level A, the technical problem is concerned w ith how ac
curately sets o f symbols (w ritten English fo r example), continuously
varying electro-magnetic signals (such as telephone or radio trans
m ission), or continously varying two-dimensional patterns ( i , e , ,
te lev is ion pictures) can be transferred from the sender to the receiv
e r , At Level B, the semantic problems re la te to the id en tity or close
approximation in the receiver of the meaning intended by the sender,
of a given set of symbols transm itted. Weaver contends that these
semantic problems are extremely involved and th a t, a t best, they can
be m itigated to to lerab le levels but never elim inated e n t ire ly . At
Level C, the communication theoris t recognizes (but cannot act on) the
problem of the e ffe c t of the meaning conveyed by the transmitted mes
sage on the conduct or performance of the receiver. Weaver admits
that the semantic and effectiveness problems are c losely intertwined
6, Warren Weaver, "Recent Contributions to the Mathematical Theory o f Communication," in Claude E, Shannon and Warren Weaver, The Mathematical Theory of Communication (Urbana; U niversity of I l l in o is Press, 1949), pp, 95-96°
and overlapping and thus constitu te a d i f f ic u l t area fo r ana ly tica l
investigation ,
St should not go unmentioned that Wiener, the other great
founder of modern communication theory, conceives of the general range
of problems in communication in a remarkable s im ila r fashion— sim ilar
to Weaver's three leve ls , Wiener proposes that the human being can be
thought of as a "terminal machine" with three ch arac te ris tic aspects?
the phonetic aspect which does the technical task o f picking up the
message and transm itting i t to the brain; the semantic aspect which
evaluates the message in accordance with a complex set of already in
culcated concepts and meanings; and the behavior aspect which is the
physical expression of in ternalized experiences and takes the form of
gross actions or the use o f symbolic forms such as spoken or w ritten
languages,^ These points o f view (Weaver and Wiener) i l lu s tra te that
communication theory Is concerned with the semantic problems of com
munication; but i t conceives of them as secondary considerations to
the technical problems o f accurate transmission. The technical d e fi
n itio n of information re lates to the technical aspect or level of
communication. At th is le v e l, information must be a measurable quan
t i t y to allow problems o f accuracy to be approached and solved. In a
prelim inary and cursory way, this points to the fac t th a t, unlike the
student of in ternational re lations (who is concerned w ith the percep
tio ns , motives, and lin g u is tic differences among actors in the
7, Norbert Wiener, The Human Use of Human Beings (New York? Avon Books, I 967) , pp, 108-110,
, . 14
in ternational system)» the communication theoris t is unconcerned with
the semantic problems o f communication. In communication theory, the
special concept of information "applies not to the individual messages
(as the concept of meaning would), but rather to the s itu a tio n as a
8whole, „ „ o"
Choice and Information
The most important thing about the special d e fin it io n of in
formation in communication theory is its dependency on p ro b ab ility and
choice among equally probable a lte rn a tiv e s . Weaver provides a nearly
perfect in terpreta tion o f information embodying th is idea o f choice,
"To be sure, th is word information in communication theory re lates not
so much to what you do say, as to what you could say. That is , in fo r
mation is a measure of one's freedom o f choice when one selects a mes
s a g e , S c h w a r t z ' s d e fin it io n of information also helps to push back
the mystique of th is new conception of information as a measurable
quantity .
The greater the number of a lte rn a tiv e messages ava ilab le to the tran sm itter, the greater the in i t ia l uncertainty a t the receiver and, therefore , the greater the information transferred when the message ac tu a lly arrives at the receiver.Thus, information is a function o f in i t ia l uncerta in ty , i , e , , of message p ro b ab ility and of the number of possible select io n s ,10
8, Weaver, "Recent Contributions," p, 100,
9 . Ib id ,
10, Schwartz, Principles o f Coding, p, 7»
15
Information, Order, and Organization
The new d e fin itio n o f information stems largely from s ta tis
t ic a l mechanics and p ro b ab ility theory, where entropy is an essential
idea. Entropy must be understood as a measure of disorder; th is in
volves u n p red ic tab ility based on lack of knowledge about the p a rtic les
in a system. But, as one's knowledge about the system increases, the
more deta iled one's understanding is , the less uncertainty one has
about the system. Consequently, the entropy is less. More entropy
means more uncertainty; less entropy means less uncertainty ( i . e . ,
more inform ation). B r i1louin o ffe rs a succinct rendering of the
concept. He notes that "entropy is usually described as measuring the
amount of disorder in a physical system. A more precise statement is
that entropy measures the lack of information about the actual struc
ture of the system."*^ Khinchin, a mathematician, has generalized the
concept to include knowledge o f events as well as possible messages.
He states that we must regard "the entropy of any f in i t e space as a
measure o f the uncertainty contained in the space, and, on the other
hand, as a measure of the amount o f information given by 'removing®
th is uncerta inty , i . e . , by answering the question o f which event of
the given space ac tu a lly o c c u r r e d . T h i s rendering points to the
11. P ierce, Symbols, pp. 22-23.
12. Leon B r illo u in . Science and Information Theory (New York: Academic Press, In c ., 1956), p. 161. .
13. A. I . Khinchin, Mathematical Foundations o f Information Theory (New York: Dover Publications, In c ., 1957)s p. 83.
16
nature of entropy as the negative equivalent of information. Pierce
adds his understanding to the concept of entropy: "Entropy increases
as the number o f messages among which the source may choose increases,
St also increases as the freedom of choice (or the uncertainty to the
rec ip ien t) increases and decreases as the freedom of choice and the
uncertainty are r e s tr ic te d ," ^ F in a lly , in general terms, Wiener ex=
plains that " ju s t as the amount of information in a system is a
measure of its degree of organization, so the entropy of a system is
a measure of its degree of disorganization; and the one is simply the
negative o f the other,
Within th is discussion of information, entropy, and organiza
tio n , i t is possible to perceive the outlines of the concepts which
have led to analogies about in ternational in tegration and organiza
tio n , In the layman's ed ition of his work on cybernetics, The Human
Use of Human Beings, Wiener attempts to stretch the reader's imagina
tion and widen his perception o f his environment w ith his philosophi
cal discussions of the im plications of his new science, cybernetics.
He dwells heavily on the concepts of message and organization, c a llin g
messages "fo rm /s / of pattern and o rg a n iza tio n ,"^ Rothstein also con
siders the concept of organization and concludes that i t can be
thought of as the negative entropy of a system, ju s t as information
14, P ierce, Symbols, p, 81,
15, Norbert Wiener, Cybernetics (New York: John Wiley andSons, In c ,, 1948), p. 18,
16, Wiener, Human Use, p, 31,
tSo He shows how organization presupposes the existence of parts and
that i f there is no communication between them, no organization can
e x i s t , *7
The Source
With the technical d e fin itio n o f information completed, and
its companion concepts described, i t is now possible to begin the
setting out of the basic elements o f a communication system and to
discuss the concepts to which these elements give r is e .
Shannon describes the source in functional terms, i t produces
18messages. In almost a l l cases, the kind of source of in terest to
the communication th eo ris t and engineer is the stationary source which
puts out (emits) symbols and sequences of symbols which exh ib it a
s ta t is t ic a l re g u la r ity . This means that any segment o f a long se
quence of symbols is s ta t is t ic a l ly s im ila r to any other long segment
from the same so u rce ,^ This s ta t is t ic a l q u a lity of the source is
best understood by using the il lu s tra tio n of a w ritten language.
Shannon notes that w ritten English, fo r example, is a nearly s ta t is
t ic a l ly regular source of messages (which, in this case, are words).
Continuing, he says, of the source:
i t w i l l choose successive symbols according to certa in p ro b ab ilitie s depending, in general, on preceding choices as
17® Jerome Rothstein, Communication. Organization, and Science (Indian H i l l s , Colorado: Falcon’s Wing Press, 1958), p, 34,
18, Shannon, "Mathematical Theory," p. 4 ,
19, P ierce, Symbols, p. $8,
Well as the p a rtic u la r symbols in question. A physical system, or a mathematical model of a system which produces such a sequence of symbols governed by a set of p ro b a b ilit ie s , is known as a stochastic process.
Pierce adds, i l lu s t r a t iv e ly , that in w ritten English, the
le t te r Men occurs with a s ta t is t ic a l frequency of approximately 0.13
(13 per cent of the tim e). But, he cautions that no language is per
fe c tly regular in its s ta t is t ic a l ch arac teris tics ; and thus, "we must
exercise a reasonable caution in applying the conclusions o f the math-21ematical theory of communication to actual problems."
The major differences among sources are in the form of the
message which is em itted. Discrete sources produce messages one sym
bol a t a time (e .g . , te legraph); but continuous sources produce en
sembles or groups of functions (th in k ; bunches of lasagne noodles or
sine waves).
Measuring Information
When a source is id e n tif ie d , i t is usually done in terms of an
amount of information which St can produce. The basic language of
communication theory is the convention which has grown up a round the
un it o f measurement. Hershberger explains that " a ll conmunicable in
formation may be reduced to the basic form in which i t may be trans
mitted as a series o f yes's and no 's , or in the binary number system
20. Shannon, "Mathematical Theory,'! p. 10.
21. P ierce, Symbols, pp. 60=61.
1922
by the use of O’ s and 1*s*" Since the entropy o f communication
theory is measured in binary un its , or b its , "we may say that the en
tropy of a message source is so many b its per le t te r , or per word, or
per message. I f the source produces symbols a t a constant ra te , we
can say that the source has an entropy of so many b its per second.
The entropy o f a source, in b its per second, is its informa
tion ra te . I t is important to remember the in terchangeability of the
two concepts, entropy and information, Khinchin concludes that "the
most important c h arac te ris tic of every source is the rate a t which i t
emits information, ic e ,, the average amount of information given by
one emitted sym bol,"^
The Message
Wiener was one o f the f i r s t in the f ie ld to recognize the s ig
nificance of the message as the primary focus in communication theory.
But a single message emitted by a source conveys no information,
Wiener emphasizes that "we must have a repertory of possible messages,
and over th is repertory a measure determining the p ro b ab ility of these
messages."^5
22, W. D, Hershberger, Princip les of Communication Systems (Englewood C l i f f s , New Jersey: Prentice-H al1, In c ., 1955), p 39»
23, P ierce, Symbols. pp. 80-81.
24, Khinchin, Mathematical Foundations, p. 46,
25, Norbert Wiener, The Extrapolation, In te rp o la tio n , and Smoothing of Stationary Time Series (New York: John Wiley and Sons,In c . , 1949), p. 2. . •
20
Shannon reminds us that " i f a source can produce only one par-26tic u ia r message its entropy is zero, and no channel is required,"
This makes sense. Since, i f there is only one possible message, the
p ro b ab ility of the received message being the one sent (over a perfect
channel) must be unity (1 *0 ) , The message would not even have to be
sent. Thus, no channel is required and no information is conveyed,
even i f the message is sent, Hershberger adds that "inform ation is
conveyed by a message only to the extent that the desti nation---a
second person— is unable to pred ict which p a rtic u la r message—w o rd -
w il i be selected, , , , the essence of information is unpredictabi1-
27i t v ," Thus, in the same place, he concludes that "one does not
obtain information from a source, e ith e r by knowing in advance or by
,,28guessing,"
Wiener's g i f t o f insight which revealed the sign ificance of
the message also led to the solution o f the problem of detecting a
signal amidst background noise (the radar problem). In th is problem,
the engineer must be able to analyze a message source's past emission
in order to make some predictions (not guesses) about the probable
future characteris tics of the messages to c o m e ,
Shannon's concern w ith the message was focussed on the problem
of coding i t (the message) so that i t would match the Channel over
26, Shannon, "Mathematical Theory," p, 31®
27» Hershberger, Communication Systems, p, 4 ,
28, Ib id , , p, 6 ,
29, Wiener, Cybernetics, p, 16,
21
which St was to t ra v e l„ His greatest contribution was the proof o f
his Fundamental Theorem which shows that th e o re tic a lly perfect trans
mission can be accomplished by encoding the message (to be discussed
below)
The Transm itter
The transm itter element In Shannon's diagram is the locus of
several in teresting functions and concepts0 Shannon describes its
function as a transformative one. The transm itter must change the
symbols of the source into signals which can be propagated out from
th e ir o r ig in , through the channel, to the receiver. For example, a
telephone transforms sound pressure from the human voice (a continu
ous source) into proportional pulses of e le c tr ic current. Obviously,
the Morse Code is a way of encoding the symbols o f a d iscrete source
(the alphabet) into d iscrete signals (pulses) fo r transmission over a 31
telegraph lin e .
There are two basic processes involved a t the transm itter;
e ith e r one or the other of them, or both, may be necessary to trans
form the message into signals. These two processes are modulation and
encoding.
Modulation
The modulating function of the transm itter is accomplished by
a process which prepares the message fo r transmission. A system
30. Shannon, "Mathematical Theory," p. 72.
31» Ib id . , p. 6 .
22
designer, by choosing a p a rtic u la r type o f channel and a type of mes
sage to be conveyed, fixes the requirements to be placed on a trans
m itte r; but he may also want to code or encode his signal to fu rther
enhance the system's e ffic ie n c y . Hershberger illu s tra te s the process
of modulation by stating th a t, w ith respect to the human voice,
" l i t t l e or no information is .conveyed by an individual generating a
sustained and steady tone w ith his vocal organs. The sustained tone
in speech carries information only when i t is modulated by being
broken up into s y lla b les . " ^ 2 This il lu s tra te s the fa c t th a t, while
there are many methods o f impressing the information from a source
onto the s ig n a l, "the key process is modulation, . « A ltogether,
there are three forms of th is key process: amplitude, frequency, and
phase modulation.
To il lu s tra te that modulation is not confined to the contin
uous case, i t is demonstrable that i t also takes place in telegraphy
where the on -o ff nature o f the current is the varigation o f the other
wise steady output of the generator. Goldman o ffe rs a general def
in it io n of modulation which serves a l l cases. He says that
modulation is the deviation o f a signal from a prescribed standard p a tte rn . The pattern may be a time p a tte rn , a space patte rn , or a pattern in any set o f coordinates whatever. O rd in a rily , the prescribed standard (unmodulated) pattern of a system is considered to be known ahead of time and contains no information.grThe information in a modulated signal is in the modulation.
32. Hershberger, Communication Systems, p. 6 .
33. ib id . , p. 76.
34. Goldman, Information Theory, p. 1?4.
23
Modulation9 then is a process which prepares the message for
transmission, A certa in ch arac te ris tic is added to the form of the
message which makes i t d istinguishable against the standard pattern o f
unmodulated transmission. In short, modulation is one of the proc
esses of making signals out of messages.
Encoding
Encoding can be used to communicate even more e f f ic ie n t ly
(without regard to no ise). Pierce (who speaks of information and
communication theory synonymously) contends that one o f the chief aims
of information theory is to study the best or the most e ffe c tiv e ways
of encoding sequences of characters or symbols fo r e le c tr ic a l trans
mission, The most e ffe c tiv e way to encode the messages from a source
would be the way which would, on the average, require the least number
of binary d ig its per character or per u n it o f time than any other
35way,
Schwartz refers to the e ffic ie n c y of coding as "a strategy
that is followed in organizing a message from symbols o f specified
a p r io ri p ro b a b ility , i , e . , average entropy per symbol. Relative en-
trophy expresses the ra tio between the actual entropy per symbol and
that obtained by choosing symbols w ith equal a p r io r i p ro b ab ili
ties ,"^^ Organizing the message, according to Schwartz, means making
35® Pierce, Symbols, pp. 64, 76,
36, Schwartz, Principles of Coding, p. 22,
24
a d is tin c tio n between the sequences of symbols which are to be the
message and those which are c h arac te ris tic of the “average" production
of the source* This points out the s im ila r ity between encoding and
modulation.
D. A. Bell describes encoding in terms which ring w ith
Shannon's influence. “Coding is the name given to the process which
transforms message to s ig n a l, and may be regarded as a kind o f match
ing o f the characteris tics of message and of communication channel;
. . »“^ The way th is matching is accomplished is e ith e r by making
the redundancy of the signal d i f fe r from that o f the message, or by
a lte r in g the modulated product of the transm itter (s ign a l) s u ff ic ie n t
ly to d if fe re n tia te i t from the unmodulated product of transmission.
The la t te r , called signal coding. Is p rim arily what Shannon was
attempting to do with his coding theorems.
The coding breakthrough, in methods of combatting noise, is
not without its inconveniences, however. The engineer is faced with
the problem th a t, even in the noiseless (perfect channel) case “ in
general, ideal or nearly ideal encoding requires a long delay in the
transm itter and re c e iv e r,. . .the main function o f th is delay is to
allow reasonably good matching of p ro b ab ilitie s to corresponding38lengths of sequences. “
37e D. A. B e ll, Information Theory and Its Engineering A pplications. 3rd ed ition (London: S ir Isaab Pitman and Sons, L td ., 1962),p. 69.
38. Shannon, “Mathematical Theory," p. 31.
25
Delay Is an inherent problem in coding messages, but its main
advantage is that
by using as short a coding as possible fo r the most commonly encountered sequences, and conversely, by leaving the longer coding fo r the more rare ly encountered sequences, we have the p o s s ib ility of making the coded tex t shorter than the o r ig in a l, which obviously might constitu te a p rac tica l and economic advantage<,3?
A simple example of th is advantage is the construction of the Morse
Code, Morse knew that more information could be transmitted per un it
time i f the more frequently appearing le tte rs in the alphabet were
given the shorter codesc Therefore, he examined a p r in te r ’ s type box
to find out which le tte rs were used more in p rin tin g . His choice of
codes was so e f f ic ie n t that today, modern s ta t is t ic a l calculations
have only improved on his choice o f codes by 15 per cent, (For
example, the le t te r "e" appears most often and so Morse co rrectly gave
i t the shortest code: one d o t,)
The Receiver .
Since the function of the receiver is , th e o re tic a lly , the re
verse o f that o f the tran sm itte r, i t makes sense to consider receivers
now, a fte r discussing tran sm itters , instead of saving th is element
u n til channels have been discussed,' ,
The rece iver’s task is to turn the signal back into the mes
sage—preferably the message which was chosen by the source, Schwartz
elaborates: "the function of the receiver is to ex trac t from the
39, Khinch in . Mathematical Foundations, pp. 23-24.
26
noisy received signal a l l possible information about the transmitted hOs i g n a l T h e information which the receiver seeks is to be found by
comparing the amount of information ava ilab le to the receiver p rio r to
the transmission (a p r io r i knowledge o f the s ta t is t ic a l characteris
tics of the source) with the a po sterio ri knowledge of the message re
ceived.
Prediction and F ilte r in g
A receiver performs its function in one of two ways, depending
on the p a rtic u la r type o f communication system. The receiver may be
designed to operate on the past output of a source, which would be a
pred ic tive operation; or a receiver may be concerned with smoothing or
f i l t e r in g a desired signal from an incoming signal-noise aggrega-
4111 on,
Wiener's solution to the prediction problem in radar was based
on the premise that the receiver would be able to handle a wide range! -
of possible messages without reference to the exact nature of any one
message. Shannon's solu tion— coding messages to match channels—was
less v e rs a tile than Wiener's and i t demanded that the ideal decoder be
able to recognize and reconstruct each individual message. This com
plex decoding system is a sophisticated, expensive price to pay for42the additional e ffic ie n c y of coding information o p tim ally ,
40, Schwartz, Principles of Coding, p. 30,
41, Wiener, Stationary Time S eries, p. 10,
42, B e ll, Information Theory, pp, 149-150,
27
When a signal has been modulated fo r transmission, the task of
the receiver is to turn the modulated signal back into a message
( i . e . , demodulate i t ) . In the f i l t e r in g mode, a receiver must im itate
a transmitted signal without delay (e .g . , telephony). The f i l t e r must
screen out noise and admit s ignals. Schwartz sees the function of
signal detection in a receiver as
es se n tia lly a decision process in which decisions are made about the presence or absence o f signals. The action of decision arises from the presence of a threshold in the rece ive r. A threshold is always present in receivers because of natural biases on tubes or semiconductor devices. Thus, in the las t analys is , a receiver is a decision device. I t must decide whether a signal is present or not and i t must make th is decision with a minimum p ro b ab ility o f e rro r . 3
An increase in the s ignal-to -no ise ra tio ( i .e .# power) a t the
transm itter w i l l increase the accuracy of transmission but th is is no
guarantee of a minimum p ro b ab ility of erro r in transmission. Thus,
coding is a way to work around the problem o f noise; but i t is not an
e n tire ly sa tis fac to ry solution— and i t is im practical.
The Channel
In communication theory, a channel is a medium fo r transmis
sion of signals; but the seminal importance of the channel is that i t
represents a constraint on the rate a t which information can be trans
m itted. I f a channel represents a constraint on the information rate
of a p a rtic u la r communication System, then i t is apparent that a
second determining facto r in communication must be the fac to r o f
channel capacity.
43. Schwartz, Principles of Coding, pp. 157“ 158.
28
In the noiseless case, theoretica l channel capacity jjs the in
formation rate of the source. Again, the back-to-back relationship of
these two concepts must be kept in mind. As a ru le , a l l channels in
communication theory and engineering are 1 inear— that is , they do not
change with time. Actual lin ea r channels have no e ffe c t on the
signals transmitted over them, except to weaken them or to delay th e irM i
time of a r r iv a l .
Theoretically perfect transmission can take place i f a source
is encoded by choosing in f in ite ly long sequences o f the sourceBs emis
sions. But, paradoxically, the price fo r perfection is not in f in ite
delay. This Is not as strange as i t sounds. Schwartz8s lengthy quote
about the lim itin g case o f block encoding is illu m in atin g .
I t is important to emphasize that the capacity o f the channel can be achieved only by the use of in f in i te ly long sequences (messages). This does not mean that there is , under ideal conditions, an in f in ite delay in the reception of information. I t means rather that as an in f in i te ly long message arrives a t the rece iver, the la t te r can examine any portion of the message already received and, by means o f coding ( i . e . , in ternal constraints among the symbo ls ), correct errors among the received symbols. With f in ite - le n g th messages only some of the errors may be corrected. In the lim it of in f in ite message length a l l errors may be corrected, provided that the rate a t which information is transmitted does not exceed the capacity of the channel.^5
This process, involving in f in i te sequences, may be compared to
the way one might view time and h isto ry: the message is of in f in ite
k k o P ierce, Symbols., p. 32.
4$. Schwartz, Principles of Coding, p. 4?.
29
length but we have decoded a t least part o f the sequence already
rece i ved.
In the continuous case (e 0g6$ te le v is io n )p a l l errors cannot
be corrected; but continuous systems can to le ra te a certa in amount of
discrepancy between the transmitted message and the received message.
The problem is solved— the discrepancy is to lerated-"by the process of
samp1ing the function (curve) enough times so that the value of the
ordinates o f the curve can be reconstructed a t the receiver with
enough d e ta il to insure an accurate reception of the information in
the message.
Noise w i l l be treated la te r ; but the noisy channel needs
c la r if ic a t io n here. In the noiseless case, the channel does not re
duce the information rate of the source. But, i f noise is introduced
into a channel, the signal which emerges from the channel (a t the re
ceiver) is corrupted. The noise in the channel has introduced an
additional ambiguity or uncertainty into the system which necessarily
reduces the actual information rate because the signal received may
not be the signal which was sent.
Shannon's basic mathematical work on channels provided proof
that the noisy, channel adds ambiguity or entropy to the a p r io ri en
tropy of the source. Thus, f in a l information rate and channel
46, Claude E. Shannon, "Communication in the Presence of Noise," Proceedings of the In s titu te o f Radio Engineers, v o l. 37» no. 1 (January 1949) 8 p. 21.
30
capacity are defined by the sum of the in i t ia l information rate and
the negative uncertainty facto r of the noisy channel» What comes out
of the channel is equal to what goes in , minus the net e ffe c t of
. 4?noise.
The Destination
Shannon's ana ly tica l diagram distinguishes between the re
ceiver and the destination in a way which is foreign to most common-
sense notions about in ternational p o lit ic a l communication (o r, fo r
that m atter, to any form of human communication). But, because he
does discuss i t separately, i t should be afforded a t least a mention
here. The destination , fo r Shannon, is the person or thing fo r whom
or fo r which the message is intended. Once the receiver has done its
work, the human receiver or destination can apprehend the message and
in terp re t i t in lig h t o f its semantic meaning. The semantic-
in terp re tive function o f the destination is e ssen tia lly s e lf -
explanatory; but i t does, needless to say, constitu te the focus of
most in teres t in in ternational p o lit ic a l re lations ( i . e . , foreign. „ x 48 p o lic y ).
Noise
Noise may enter the system a t any point between the trans
m itte r and the receiver, adding uncertainty to the message and placing
47= Shannon, "Mathematical Theory," pp. 35-39=
48. Ib id . , p. 6 .
/
a lim it on the system's performance. But, what Shannon's simple d i
agram does not show is the p o s s ib ility of noise jm transm itters and
receivers®
Man-Made Noise and Inherent Noise
Man-made noise usually impinges on communication channels—
that is , a f te r the signal has been propagated® This type of noise can
be controlled to almost any degree of e ffic ie n c y desired since i t is
generated by other force f ie ld generators and broadcasting systems
(antennas, fo r example, can be relocated or shielded)® But, inherent
noise orig inates in the tubes, diodes, and other apparatus of trans
m itters and receivers® This type o f noise, called thermal noise, is
not eradicable but i t can be controlled w ith other e le c tr ic and£iQ
electron ic devices®
Thermal noise occurs in any e le c tr ic a l device because of the
ag ita tio n of the molecules of the substance (Brownian motion)® But
there is a thermal noise, sometimes called Johnson noise, which o rig
inates outside the communication system— usually encountered as back
ground noise from the thermal radiation of the s ta rs . This noise has
p erfec tly random s ta t is t ic a l properties of which communication theory
takes advantage ( i . e . , i t is treated as a constant)
49® Hershberger, Communication Systems® pp. 56-57®
50. P ierce, Symbols. pp. 188-189.
Combatting Noise
Shannon's contribution to communication theory appears even
more s ig n ific a n t when one considers what the engineer could do about
noise before 1945. The engineer could (1) increase transm itter power,
(2) make the receiver less noisy, (3) insulate the receiver against
various frequencies, (4) modulate the frequencies to obtain very large
bandwidths, or (5) make transmission o f messages r e d u n d a n t . The
revolutionary thing which Shannon did was to avoid the noise problem
altogether by coding around i t . He exchanged message bulk fo r trans-52mission delay. Message compression results from coding.
Up to th is point (fo llow ing Shannon's diagram to the r ig h t-
hand side) the discussion has been on un i-d irec tio n a l systems. These
systems are characterized by one-way flow . But now two-way flow must
be considered. Feedback jo ins the p ic tu re .
Feedback
In the same revolutionary essay, in 1948, in which he put
fo rth his fundamental theorems on coding. Shannon included a discussion
of a correction system involving a feedback loop. The schematic draw-
53ing which he provided of th is process is reproduced below.
51. ib id . , p. 146.
52. Brockway McMillan, "The Basic Theorems of Information Theory." Annals of Mathematical S ta t is t ic s , v o l. 24, np. 2 (June 1953), p . 207.
53. Shannon, "Mathematical Theory," p. 37.
33
correction data
observer
signal
source transm itter receiver correction
Fig. 2. Correction System
In this correction system, the observer monitors the system
before and a fte r transmission has taken place in order to assess the
f id e l i t y of the transmitted message. "The observer notes the errors
in the recovered message and transmits data to the receiver to correct
the e r r o r s . T h e input message, M, becomes d istorted somehow in
transmission and emerges a t the receiver as a corrupted version of the
o rig ina l message: M becomes M*. M d iffe rs from M1 by the amount of
ambiguity in the channel. Thus, th is same amount of additional in fo r
mation must be supplied, per second, a t the receiving point to correct
the received message.'*'* A correction channel, then, keeps the
54. Ib id . , p. 36.
55. Ib id . , p. 37.
34
transm itter informed of the additional amount of information which is
needed to make the messages already received match the messages which
were sent®
Schwartz emphasizes the point made above about one-way in fo r
mation flow® He notes that the basic subject matter of information
theory is the u n i-d irec tion a l communication system; but he goes on to
point out that feedback, or a correctional c a p a b ility , "gives a commu
nication channel an a b i l i t y to adjust its transmitted information rate
to the signal and noise conditions on the channel, speeding up when the
noise level is low and slowing down when i t is high® Thus systems
equipped with a feedback lin k are adaptive communication systems®^
Figure 2 is an il lu s tra tio n o f a no ise-free feedback channel.
That means that i t has been assumed that the information conveyed along
th is channel is not subject to corruption® This is a reasonable assump
tion (in some cases) according to Schwartz, since the operation of the
correction device only demands a simple decision about the accept
a b il i ty of the received message® For example, the message on the main
or forward channel may require ten b its of information to transmit i t
to the receiver; but the feedback message—a yes-no or on -o ff message
about the accep tab ility o f the sequence— is a simple decision requiring
only one b it o f information to be sent back to the tran sm itte r. The
importance of the feedback loop is that the en tire message can
56® Schwartz, Principles of Coding, p. 5®
35
eventually be received a t any desired erro r p ro b a b ility , depending on
57how involved a system one wishes to design,,
Schwartz points out that feedback can be of two general types
(not mutually exclusive): Decision-Feedback and lnformation-Feedback@
Decis ion-Feedback
The simplest type of feedback system is one in which the cor
rection device e ith e r accepts the message or re jects i t : "the receiv
e r , as a resu lt o f a decision made on each received message, e ith e r
accepts the message and records the corresponding symbol or rejects i t
as ambiguous and reports the re jection to the transm itter* The in
formation in the ambiguous message is discarded, and the transm itter
subsequently repeats the message*"^
Information-Feedback
Schwartz id e n tifie s a second type of feedback system—more
complicated than the simple decision system explained above* The sec
ond type of feedback is called inform ation-feedback because "the feed
back channel is employed to report information about the received
message to the tran sm itte r, w ith the decision to accept, or to re jec t
and co rrec t, being made subsequently a t the tra n s m itte r*^
The report which the receiver makes back to the transm itter is ,
in an operating system, subject to d is to rtio n and fading, ju s t as the
57* Ib id * , p* 196,
58. ib id .
59* ib id . , p. 193.
36
signal in the main forward channel is subject to noise. Thus» Schwartz
contends that the optimum system is one in which coding and decision-
feedback are combined.
In this system, the feedback channel reports accept-re ject decisions to the tran sm itter, thus minimizing the e ffe c t of errors due to fading and disturbances in the feedback channel , a resu lt that may be accomplished by u t i l iz in g a suff ic ie n t ly asymmetrical decision mechanism a t the transm itter to in terp re t the feedback information, so that re je c t-to - accept (R-»A) errors occur with a minimal frequency.®0
Schwartz c a lls th is system " fa i l-s a fe ” because of its r e l ia b i l i t y and
possible range of performance. .
Without some type of feedback information reaching the trans
m itte r , coding protected systems (non-feedback) are subject to "burst-
type noise, signal fading, and sudden changes in channel conditions,
since the transm itter cannot determine when such changes occur.
These shortcomings are avoided in Schwartz's suggested " fa il-s a fe "
system. He provides several other feedback systems which are combina
tions of the various types already mentioned. Each has its advantages
and disadvantages fo r various forms o f communication; but a l l are de
signed to insure r e l ia b i l i t y and increase the system’ s performance of
its technical function: accurate transmission o f information.
Goal-Oriented Feedback
Feedback, as i t has been discussed so fa r , has been concerned
with the performance of a communication system. Feedback has been
60. Ib id . , p. 221.
61. Ib id . , p. 192.
37
described as another method of improving the transmission of informa
tion from a source to a destination . No other purpose has entered
into the discussion of communication systems except the s e lf is h ,
internal purpose o f rapid and re lia b le transmission of information.
But, where the communication th eo ris t and engineer drop the problem
of message transmission, the systems engineer and the cybernetician
pick i t up.
The system which has a behavioral goal beyond e f f ic ie n t commu
nication of information is a system which uses information fo r a pur
pose. This is why the cybernetician, Norbert Wiener, chose the
extremely logical phrase, "communication and co n tro l," to characterize
his understanding of the new science. The goal-oriented, purposive
system which uses information to maintain i t s e l f in an environment is
communicating with that environment in order to control its behavior
and maintain its id e n tity .
The way in which a purposive system achieves its goals is by
receiving information from its environment and adjusting its internal
mechanisms or its external behavior to cope with the environmental de
mands. This response to the environment is accomplished by the process
of feedback.
Wiener's goal-oriented system is s im ilar to Shannon's adaptive '
communication system (F ig . 2 ) . Thus, re ferring to Shannon's correc
tion system, i t can be seen that the fundamental processes in i t are
s im ilar to those described in th is sumrnary (below) of communication
and control by Wiener.
This control o f a machine on the basis of its actual performance rather than its expected performance is known as feedback, and involves sensory members which are actuated by motor members and perform the function of te l 1-ta le s or monitors— that is , of elements which indicate a performance. I t is the function of these mechanisms to control the mechanical tendency toward disorganization; in other words, to produce a temporary and local reversal o f the normal d irection of entropy,
Wiener was not the f i r s t to ta lk about the concept of feedback;
but he was one of the f i r s t to perceive the process of communication
as a flow of messages— "forms and patterns o f organization;" and
therefore, he was p rim arily responsible fo r link ing up the phenomenon
of communication w ith the phenomenon of co n tro l. Feedback was already
known to e x is t in systems which were characterized by a homeostatic
mechanism of internal equilibrium . But there are s lig h t differences
in the processes. Pask explains: "Whereas re fle x is preserves the or
ganism against the flu x o f its environment, homeostasis counters the
in te rn a lly generated changes which are prone to d isrupt the proper
structure and disposition of parts in the organism.
Wiener makes a d is tin c tio n between feedback in postural s itu a
tions o f the human body and internal homeostatic s itu a tio n s: "Our
homeostatic feed-backs have one general d ifference from our voluntary
and our postural feed-backs: they tend to be slower;" and he contin
ues: "The typ ical e ffectors of homeostasis— smooth muscles and
62. Wiener, Human Use, p. 36.
63. Gordon Pask, An Approach to Cybernetics, Science Today Series (New York: Harper and Brothers, I 96I ) , p. 12.
39
glands—are likewise slow in th e ir action compared with striped64muscles» the typical effectors of voluntary and postural a c t iv ity *"
Goal-oriented feedback may be positive or negative. Positive
feedback stimulates those responses (encourages them) in the system
which tend to f u l f i l l the purpose or goal sought® Negative feedback
stops (discourages) the system from pursuing one tack when i t becomes
disfunctional to the goal o f the system (e 0g .» the steam engine65
When the system responds slowly to the con tro llin g feedback
messages coming in on its performance» i t is said to be exh ib iting lag®
Since no feedback channel is p e rfe c t, a l l systems lag in goal achieve
ment. To compensate fo r th is and other environmental parameters» sys
tems sometimes have lead b u ilt into them so that the system can perform
e ffe c tiv e ly in its normal environment (e.g®, a radar guided a n t ia ir
c ra ft gun is programmed to aim in fro n t o f the spot a plane ac tu a lly
occupies in the sky)
At th is point in the discussion we have come fu l l c irc le : From
a d e fin itio n of information and organization, through un id irectional
and two-way communication, to feedback, we have returned to a
consideration of the problem of furnishing the system w ith adequate
information fo r its proper functioning- and su rv iva l. Feedback means
64® Wiener, Cybernetics* p. 135®
65® Wiener, Human Use, p. 206®
66® ibid®, pp® 36-37®
40
co n tro l; and Wiener, better than anyone (and sooner), perceived the
profound importance o f its meaning. For him,
information is a name fo r the content o f what is exchanged with the outer world as we adjust to i t , and make our adjustments f e l t upon i t . . . .To liv e e ffe c tiv e ly is to liv e with adequate information. Thus, communication and control belong to the essence of manjs inner l i f e , even as they belong to his l i f e in so c ie ty^ .'
This b r ie f exposition of communication theory is complete.
The basic elements and concepts, ideas and p rin c ip le s , have been
arranged and presented in the Coherency of a general communication
system. The emphasis which I have given each concept may not be pre
c ise ly that which a communication engineer or th eo ris t would consider
adequate or proper; but I fee l somewhat ju s t i f ie d in the form and sub
stance of the presentation because of the use of the diagrams which
Shannon himself used.
In the next chapter, the three models of in ternational re la
tions w il l be presented and evaluated in terms of the meanings ascribed
to communication systems and th e ir concepts in th is chapter. Hopeful
ly , the follow ing evaluation w il l be taken as constructive c ritic is m
and not simply judgment.
67. Ib id . , pp. 26-27.
CHAPTER 3
THE COMMUNICATION MODEL:THREE APPLICATIONS IN INTERNATIONAL RELATIONS
When you can measure what you are speaking of and express i t in numbers you know that on which you are discoursing,. But i f you cannot measure i t and express i t in numbers, your knowledge is o f a very meagre and unsatisfactory kind.
Lord KelvinQuoted in Arthur PorterCybernetics S im p lified , p. 34.
In the f ie ld o f in ternational re la tio n s , Karl W. Deutsch was
one scholar who took Lord Kelvin a t his word. Throughout the 1950*5
and early 1960, s , he dominated the f ie ld with his in terpreta tion of
communication models and th e ir place in the study of in ternational
in teg ratio n , u n ific a tio n , and organization. He tr ie d to measure
things.
Apparently c a p ita liz in g on his personal acquaintance with
Norbert Wiener a t the Massachusetts In s titu te of Technology, Deutsch
got into the area of communication qu ite ea rly —-perhaps he was the
f i r s t —-with the publication of his in terpreta tion o f the use of com
munication models in the social sciences.* In 1950, two years a fte r
his book Cybernetics was published, Wiener came out with a layman's
1. Karl Wo Deutsch, "On Communication Models in the Social Sciences." Public Opinion Q uarterly , v o l. 15, no. 3 (F a ll 1952).
41
ed ition of his ideas on the new sciences The Human Use of Human Beings
became the stimulus fo r much thinking in the social and behavioral
sciences concerning new approaches and new questions to ask. Deutsche
present a t the creation (in a manner of speaking), was greatly in=\
fluenced by the rigor and apparent fru itfu ln e s s of the new model o f
systematic processes. The revolutionary new p rin c ip le was feedback.
The most extensive use o f the communication model in in te r
national re la tions has been in the area o f theoretica l frameworks fo r
analysis--as a ta k e -o ff point fo r m odel-building. The number of
studies which ac tu a lly attempt to operationalize the concepts of com
munication and in teg ra tio n , communication flow , or cybernetic steering
has been sm all. However, there is an enormous lite ra tu re on in te r
national p o lit ic a l communication dealing with topics such as public
opinion, propaganda, national values, national character, and cross-
national perceptions.^ In most of these studies communication is
considered from the semantic perspective. Davison and George provide
a d e fin itio n o f in ternational p o lit ic a l communication which seems to
characterize the perspective of these "semantic" studies.
By 8 In ternational P o lit ic a l Communication8 we re fe r to the use by national states of communications to influence the p o li t ic a l ly relevant behavior of people in other national s ta tes . Thus we include the propaganda and information a c t iv it ie s of most government agencies—especia lly the State and Defense Departments—and certa in aspects o f
2. Bruce Lannes Smith, "Trends in Research on In ternational Communication and Opinion," Public Opinion Q uarterly , v o l. 20, no. 1 (Spring 1956) i see also W. P h illip s Davison, In ternational P o lit ic a l Communication (New York: Frederick A. Praeger, 1965).
43
diplomatic communication, but we exclude the a c t iv it ie s of the press associations and bodies which are interested p rin c ip a lly in in ternational education in re lig ious missionary a c tiv i t ie s ,5
The important thing to remember about a l l o f these kinds of
studies is that they " re fe r to the tran sfer o f meaning, whether by
w ritte n , spoken or p ic to r ia l symbols, or by various types of action,"**
The variables usually associated with these studies of in ternational
communication are: "Who says what to whom through what mediurn fo r
what purpose under what c?rcumstances and with what e ffects?"** Host
scholars have ta c it ly agreed that th is is a convenient way of ta lk ing
about the main elements o f any communication process. But, Davison
and George warn that "the variables in question are in tr ic a te ly re
lated and most e ffo rts to study and s ta te th e ir in terre lationsh ips
have been seriously over-s im p lified ."^
One study of in ternational re lations which could never be
labeled as over-s im p lified was Richardson's mathematical study o f past
warSo^ This l i f e work was an attempt to lin k certa in indicators of
h o s t i l ity w ith the eventual outbreak of war using in tr ic a te
3. W, P h illip s Davison and Alexander C, George, "An Outline fo r the Study of In ternational P o lit ic a l Communication." Public Opinion Q uarterly , (Winter 1952=1953), p. 501.
4 . Ib id .
5 . Ib id . , p . 502.
6 . Ib id .
7. Lewis F. Richardson, Arms and Insecurity (P ittsbu rg , Pa.: Boxwood Press, I960 ); see also Anatol Rapoport, "Lewis F. Richardson's Mathematical Theory of War," Journal of C onflic t Resolution, vo l. 1, no. 3 (September 1957)»
44
mathematical formulas and techniques. His study was s ig n ific a n t
enough to be among the s ix minor approaches to the study of communica
tion which BereIson id en tifie d in a v ir tu a l obituary on "The State o f8Communication Research," in 1959.
Richardson's super-quantitative study is s ituated a t the fa r
end of the spectrum—opposite the "semantic" studies, in between the
two extremes, u t i l iz in g the communication model as a theoretical frame
work, yet attempting to find concrete indicators o f communication flow ,
there has been a number o f studies which have attempted to quantify
certa in conditions of in tegration and community. Deutsch, of course,
did much of the pioneering work in th is area of in ternational re la
tions.^ But Russett and McClelland have also contributed useful
studies of systemic communication between and among actors in an in te r
national system.
For reasons which have already been expounded in Chapter 1, 1
have chosen studies by Deutsch, Russett, and McClelland as useful
examples of the application of the concepts of communication theory to
the study of in ternational re la tio n s . Each of these authors has chosen
certa in perspectives from the theory and attempted to study selected
hypotheses from these vantage points. In this chapter, I w il l attempt
to present enough of each of th e ir studies to characterize each
author's approach to his subject m atter; and the analysis which follows
8 . Bernard BereIson, "The State of Communication Research," Public Opinion Q uarterly , v o l. 23, no. 1 (Spring 1959)•
9 . Karl W. Deutsch, P o lit ic a l Community a t the In ternational Level (Garden C ity , New York: Doubleday and Co., 1954).
w ill be an attempt to evaluate the use o f the concepts which each
author chooses fo r his focus. No attempt has been made to force each
of the following studies into a p a rtic u la r framework— such as the
communication system used in Chapter 2 . Instead, I have chosen to
approach each study on its own ground—as i t were— in an attempt to
understand what the author means when he uses the various concepts
from communication theory. At times, one o f the author"s concepts may
appear to be a varia tio n of one of the ideal types discussed in Chap
te r 2 . In these cases, generosity w i l l prevail and approximations
w ill be evaluated fo r what they are—and how useful they appear. I t
should be stated again that th is analysis is not designed or intended
to be merely judgmental: i t is not a matter o f whether or not a
s c ie n tif ic model may be used to study in ternational re la tio n s ; ra th er,
the question is , how and how well?
Peutsch: In tegration
Since the early 1950's, Peutsch had worked extensively on the
problems of integration and community a t the in ternational le v e l. But
most of his work in that period had been mainly theoretical or hypoth
esis generating (both endeavors are va lid and valuable en terprises).
However, in 1956, Peutsch contributed a "flow" study to a10special communication issue of the Public Opinion Q uarterly . This
is the study which w il l be analyzed in the following pages. I t is the
10. Karl W. Peutsch, "S h ifts in the Balance o f Communication Flows," Public Opinion Q uarterly , v o l. 20, no. 1 (Spring 1956).
4 6
f i r s t re a lly experimental study which Deutsch undertook; i t was his
attempt to operationally define his concepts by measuring them0 Kaplan
recounts that “operationism” has been a persistent school of thought
within a larger context o f semantic empiricism ( io e ,» what cannot be
known by experiences9 through the senses» cannot be sa id ): ”To each
concept there corresponds a set of operations involved in its s c i
e n t if ic use* To know these operations is to understand the concept as
fu l ly as science requires; without knowing them, we do not know what
the s c ie n tif ic meaning of the concept is , nor even whether i t has a
s c ie n tif ic meaning*” * * While there are numerous critic ism s of th is
approach to concept d e fin it io n , they are of no concern here, since i t
is merely a fac t that Deutsch attempted to measure communication
flows--and thereby define his concepts*
A fte r a good theoretica l statement concerning the princip les
of communication theory which would guide the presentation which was
to fo llow , Deutsch launches into a ju s t if ic a t io n fo r attempting to
measure something— communication flow—which had previously only been
alluded to or had been taken fo r granted while being treated q u a lita
t iv e ly * The purpose o f th is study was to get a “be tte r understanding
of some aspects of the integration or consolidation o f communities,12countries, or other kinds of organizations*” To achieve this pur
pose, Deutsch defines integration (in a very haphazard way) as the
11* Abraham Kaplan, The Conduct of Inquiry (San Francisco: Chandler Publishing Co*, 1964), p* 40*
12* Deutsch, "Communication Flows,” p. 160*
47
extent of an organization 's in teraction with organizations outside its
boundaries. Internal cohesion is also an important concept in th is
study since i t represents the extent o f the organization's internal
communication with i t s e l f . In e f fe c t , the two concepts are construc
tions representing certa in ratios o f in-country and out-country commu
nication flows: "The ra tio of intra-boundary processes w ith in any
organization to cross-boundary processes o rig ina tin g or terminating
in the same organization o ffers one basic operational measure fo r the
cohesion of any such organization, as well as of the extent of its13integration with others outside i t . "
The hypothesis upon which he bases his selection of these
ra tio s — these indicators of integration and cohesion—while in s ig h tfu l,
is nonetheless of subtle , almost strange parentage. Deutsch has i t on
good authority from the mouth of the master, Wiener h im self, that in
the case of a simple goal-seeking device, the intake-output ra tio (in
terms of b its of information necessary fo r e ffe c tiv e system maintenance
and survival in an environment) is on the order of one-to-one ( 1/ 1) .
This means that fo r e ffe c tiv e operation, the organism or system must
14get a b it of information fo r every b it of information i t gives.
Deutsch then takes th is ideal ra tio as a standard against which to
measure the performance of some typical systems.
He points out the fundamental p rin c ip le of communication
theory: the individual message, by i t s e l f , contains no useful
13. Ib id . , p. 149.
14. Ib id . , p. 146.
information regarding the message source; only a great number of mes
sages together can convey information by creating a pattern of s ta t is
t ic a l s ignificance* He asserts that to be successful» a study of
communication must be able to measure the flow of messages,^ Deutsch
admits that there are other important aspects of communication flow:
speed; f id e l i t y ; e ffe c ts ; and number o f in i t ia l choices— a ll are v ita l
characteris tics of message flow* But, he concludes—cap itu la ting to
ease-of-col lection pressures— that "volume— that is to say, f r e
quency—of communication is the f i r s t dimension of in ternational and
domestic communication flows that is l ik e ly to be measured with any„16
degree of success0
I t is important to remember that the whole study is predicated
on the hypothesis which Deutsch got from Wiener in an oral conversation.
17a t the Massachusetts In s titu te o f Technology in the spring of ?955o
This hypothesis states that an optimum intake-output ra tio of a func
tioning system would be in the general neighborhood o f u n ity . With
th is ideal ra tio in hand, Deutsch sets about to measure his chosen
indicators of internal and external flow . From these measurements, he
constructs certa in Received/Sent (R /S ), Local/Non-Local (L /N ), and
Domestic/Foreign (D/F) ra tios of mail flow .
Using data from the s ta tis t ic s of the Universal Postal Union,
Bern, Switzerland, fo r the relevant years, Deutsch develops ratios of
15. Ib id , , p. 144.
16. Ib id . , p. 145.
17. Ib id , , note 5 , p. 146,
49
received-mai1 to sent-mai1 fo r fo r ty -s ix countries and finds that more
than three-fourths of them receive more than they send— that is , th e ir
R/S ratios are more than one ( I o0 ) . ^ In most cases, i t seems, these
ratios turn out to be very gross indicators of the degree of indus
tr ia l iz a t io n and d iv e rs ific a tio n . The countries w ith R/S ratios con
siderably below 1.0 sent more mail than they received fo r the years
recorded (e .g . , France 0.52; Germany 0.65; United States 0 .8 1 ). The
countries a t the other end o f the rank order scale , those with R/S
ratios considerably above one, received more mail than they sent (e .g . ,
Mexico 3 .12; B razil 2 .78; Indo-China 1.71)» A middle range of coun
tr ie s clustered around 1. 0 ; these appeared to represent the ideal or
optimum ra tio (e .g . , Union of Soviet S o c ia lis t Republics 1.02; Greece
1.18; Sweden 1 .4 8 ) .^ These ratios are crude indicators of one type
of communication flow . They constitu te only a prelim inary hypothesis
about a country’ s condition of in tegration with the world.
Seeking another type o f indicator of in tegration and its cor
responding concept, cohesion, Deutsch turns next to ratios which ex
press the d if fe re n t ia l in local communication w ith respect to non-local
communication. Using s ta tis t ic s from the United States Post O ffice
Department, he reverts to the discussion o f mail flow w ith in the United
States to make a case fo r Domestic-Foreign (D/F) flows la te r in the
study. What he is building up to is the proposition that a country’ s
18. Ib id . , p. 147.
19. Ib id .
502D/F ra tio , coupled with an aggregate s ta t is t ic , AL /P , of geographical
area (A ), population (P ), to ta l le tte rs sent (L ) , and le tte rs per
capita (L /P ) , is an indicator of that country's in tegration with or20isolation from the rest o f the world,
) 2 When a l l the countries had been assigned D/F ratios and AL /P
measurements, they were p lotted on a two-dimensional a x is . This was
done with the measurements from four time periods. What resulted was
a "regression lin e ," The plots tended to f a l l along a sloping lin e
which could be represented by a certa in calculable formula. Deutsch •
found that the slope of the regression lin e fluctuated from period to
period in such a manner that one general conclusion could be drawn from
the data display: as countries got larger and more in d u s tr ia lly ad
vanced , they tended to begin to send more mail than they received.
Deutsch was impressed by the way in which the regression lines had
shifted fo r each of the four periods examined. But, owing to the
sketchy nature of the data fo r the year 1880, i t would have to be
assumed that the s h if t in the lin e from th is time period to the next
would not be too dependable in any close comparison with the other re
gression lin es .
Deutsch suggests other indicators of communication flow such
as m igration, exchange students, and s c ie n tif ic notation in scholarly
jo u rn a ls ; but his prelim inary tests of these indicators are on flows
w ithin the United States, fo r the most p a rt. At any ra te , they appear
20. Ib id . , p. 152.
51
almost as afterthoughts, and they are , a t best, inconclusive. The
primary data in the study is the D/F mail flow . Secondarily, a
country's cohesion as a function of an increasing volume of mail flow
w ith in its boundaries is important to the main purpose o f the a r t ic le .
The measure of iso lation (converse o f in tegration) o f a country with
respect to the relevant environment is defined by the magnitude of the
D/F ra tio in re la tio n to a l l the rest of the countries arrayed. An
example of th is is the extreme case of the Union of Soviet S o c ia lis t
Republics which had a D/F ra tio in 1936 of 96.0~=which Deutsch thought
seemed "to represent the world's record to date in postal iso1a=. . ..21 tionism ."
Deutsch's search fo r the secret o f integration (or a t least fo r
a clue to i t ) , or community, or cohesion was attempted using methods
which could only measure crude volumes o f flow from one time period to
the next. His conclusion a t th is point in his experience with the
communication model ( 1956) was that "most of the work on measurement
22of in ternational communication flows is yet to be done."
Having discussed his main theoretical in te res ts , his hopethesis
about in teg ratio n , and his operational d e fin itio n s o f communication
flow , i t is now possible to make some sort o f an assessment of
Deutsch's use of the communication model in th is study.
21. Ib id . , p. 154.
22. Ib id . , p. 160.
Deutsch: Ana lys is
Deutsch's hypothesis about in tegration as a function o f cross
boundary mail flow is taken from the two-way communication system
model« Since his basic premise is that an organism must give about as
much information to Sts environment as i t gets from its environment,
his assumption must be, i t seems, that the organism, to survive, i oee,
to maintain i ts e lf and perform adequately, must exchange information
with its environment, or perish. Based on what we have seen from
communication theory and the idea of organization as the input of in
formation, th is is a sound assumption. But common sense and a l i t t l e
re flec tio n on the h istory of some important countries indicate that
many nations can and have successfully existed in v ir tu a lly complete
Iso lation from the rest o f the world (e .g . , China before P erry ). This
must be the case—a t least in the measurement of the indicators chosen
by Deutsch. In other words, even i f a country sent no m a il, students,
or immigrants abroad or its scholars fa ile d to c ite any foreign schol
ars in th e ir work, the country would be able to survive, maintain
i t s e lf as an organized s ta te , and be e ffe c tiv e in other areas of trans
action such as trade, foreign conquest, or foreign investment. Thus,
by themselves these indicators (but mainly mail flow) of integration
cannot be a t a l l conclusive. But they can, as Deutsch admits, be used
to substantiate "q u a lita tiv e " information about a country's degree of
in tegration or is o la tio n ; these indices can also help to lend an idea
of the magnitude of some phenomena which might otherwise seem
inconsequential.
53
From another perspective, Deutsch's model could be considered
as a form of the un id irectional system. I t could be viewed as simply
a description of certa in message flows a t certa in points in time. In
th is mode, the model would not have the feedback c a p a b ility ; although
i t would be erroneous to^suggest that there was not a continuous pro
cess going on ( i . e . , a constantly s h iftin g D/F r a t io ) . But, Deutsch
never supplies the structura l evidence or conceptualization fo r the
type of feedback process which would, a t le a s t, describe the exchange
of information taking place. -
A second area of analysis must focus on Deutsch's use of the
concept of messages. He c le a rly recognizes the p rin c ip le that the
individual message conveys no information. Messages must be received
in quantity , as aggregates (functions) or sequences of symbols, before
they create patterns a t the receiving end of the transmission. Though
Deutsch has been c r it ic is e d fo r weighting a l l of his "messages"
equally— i . e . , not distinguishing among the contents o f certa in mes
sages such as post cards and diplomatic le t te rs — he is vindicated by
the communication theory dictum concerning patterns as opposed to theS
content of individual messages.1* Deutsch is r ig h tly concerned only with
the transmission and reception (flow ) of long sequences o f messages.
Thus, his measurement of gross volume is a generally acceptable indica
to r of communication flow and can be re lie d upon to t e l l the observer
a t least something about the sta te of the system's information-
organization condition (in e ith e r the one-way or the two-way communi
cation system). But, in the f in a l analysis, Deutsch's focus on the
volume of the flow alone fa l ls severely short o f the understanding
which has already been established in Chapter 2 about the nature o f
the transmission o f messages as sequences of symbols with d is tin c tiv e
s ta t is t ic a l ch arac te ris tics ,
Deutsch acknowledges that there are other dimensions of commu
nication flow besides volume ( 1» e ., speed, f id e l i t y , number of o rig
inal choices, and e ffec tiven es s ), but he realizes that data on volume
w ill be the most read ily availab le and the most descriptive of the con
cepts he has operationalized as flows—with the least amount of work.
The most that could be said fo r the flows which Deutsch succeeds in
quantifying is that these flows do infuse (when they are coming into)
a country with a certa in type of information which tends to "organize"
the e n tity in spec ific ways. For example, immigrants into a country,
as Deutsch suggests, may be thought of as messages. They can change
the demographic structure of th e ir adopted country (or portions of i t ) .
A large contingent o f immigrants from Cuba, fo r examp1e , has had an
impact on the State of Florida by th e ir very presence. In addition ,
th e ir m inority group status, the p o lit ic a l connotations of th e ir migra
tion , and th e ir impact o f the job market, schools, and housing capabil
it ie s of the State are a l l factors associated with the fa c t that they
have come.
At other times in American h is to ry , the Negro slave, the Ir is h
immigrant and the Polish immigrant have a ll been inputs to the United
States, measurable not only as volume (as Deutsch recognizes) but also
by th e ir peculiar and sometimes common characteris tics as communication
55
flow 0 The conditions o f th e ir emigration and the conditions o f th e ir
reception in the new country constitu te transmission characteris tics
which might be e ffe c tiv e ly probed to find out about the characteris tics
of the source (abandoned country) as well as the s ta te o f the overall
system in terms of how i t handles the flow— ioe», performance thresh
old*
To carry the example a l i t t l e fu rth e r , I suggest that a p ro f ile
on an immigration flow could be developed to any degree o f d e ta il de
sired* This p ro file could be construed as the s ta t is t ic a l character
of the sequence. Various received sequences could be analyzed to re
veal th e ir p ro file s and an increasingly more d e ta ile d , complete d e f in i
tion or "p icture" of the e n tire sequence could be constructed. At any
ra te 9 perhaps each of the indices used by Deutsch could be used in
th is more sophisticated way in order to more closely approximate the
theoretical and engineering ideal and a t the same time get more out of
the data. But, always, i t should be remembered that studies of th is
type can only be corroboration fo r the q u a lita tiv e investigations which
tru ly investigate the human interaction which is , in the end, the only
level a t which real understanding and subsequent in tegration can take
place.
Cross-boundary output could be treated in much the same way as
input was treated above; but, in ad d itio n , i t should be noted that
C ybernetica lly , what a system or an organism "puts out" is action— fo r
the purpose of maintaining i t s e lf by adjusting to its environment.
There is an exchange of information w ith the environment, to be sure.
but i t has th is manipulative ch ara c te ris tic —action* An organism
perceives its environment by taking in information; th is i t does by
sensing its position or s itu a tio n w ith in the environment and persuing
its goals by a lte r in g its behavior in ways which w i l l lead to goal
achievement. Again, its outputs are a c t iv it ie s (motor, glandular, or
nervous) which are themselves programmed to help the system achieve
its goals.
Viewed from th is feedback perspective, Deutsch's ratios of
flows do not make too much sense. He equates out-flow with in -flow
and hypothesizes that in -flow divided by out-flow (D/F) w i l l y ie ld a
measure or an index of in teg ra tio n . But the receip t of a small amount
of information from the environment— say in the form of a perception of
heat a t the fin ger t ip s —can tr ig g er a larger motor response in the arm
and body of many times the amount of information taken in by the sen
sory receptors. Thus, without c la r if ic a t io n from some quarter, i t
remains, a t best, an “oral communication" from Wiener that an input-
output ra tio o f one to one ( 1/ 1) is the ideal against which to measure
any system's performance. One might be tempted to think that a ra tio
o f 10 to 1 would be more “normal" fo r Deutsch's data (and he does
speculate that la rg e r, more advanced countries might show a consistent
p ro c liv ity to stay up a t a high D/F ra tio o f 15 or more); but th is
would be an erroneous concession to what appears normal, as well as an.
admission that his o rig in a l premise ( 1/ 1) was fa u lty .
But, in id en tify ing these ratios of Domestic to Foreign mail
flow , Deutsch may have uncovered something which he did not an tic ip a te
and which could not be dea lt with under his concept of in teg ratio n .
His concept of cohesion was less developed than his concept o f integra=
tio n . A country's in ternal mail flow (or communications of a l l kinds)
may be construed as an indicator of its s ta te of organization and
cohesion. Perhaps Deutsch's concentration on the ra tio masked the
re la tiv e importance of the domestic flow by i t s e l f . This flow might
represent what Deutsch called "secondary" message flow w ith in an
organism by which i t informs a l l the sub-systems of the s ta te of the
whole system— keeping i t ready to receive and process "primary" mes
sages from the environment, Deutsch had stated th is d ifference in
types of messages in his e a r l ie r a r t ic le on communication models; but,/ 23
apparently he did not consider i t germane to the 1956 study.
F in a lly , i t is important to remind ourselves that the quan
t i f ia b le flows of in ternational communication a re , and must be, pat
ently d iffe re n t from the q u an tifiab le flows of communication
engineering. When the communication engineer speaks of coding and
modulating a source's output, he is ta lk ing about sp ec ific operations
which can be carried out on the signal to make i t transm ittable or to
match i t to the channel. He can make predictions about a future trans
mission based on his knowledge of the source's past performance. In
short, his knowledge of the characteris tics of the message, o f the
channel, of the transm itter and rece iver, and of the requirements o f
the operator or the d es tin a tio n --th is knowledge— is nearly complete.
On the contrary, the scholar who wants to know something about the
23. Deutsch, "On Communication Models," p. 369.
performance of an in ternational system or sub-system is confronted on
a l l sides w ith imprecision, ambiguity, and the semantic fac to r: a l l
which make his task of flow perception, id e n tif ic a tio n , and measure
ment exceedingly d i f f i c u l t , i f not impossible. And in the end, he can
never be sure that what he has observed is the resu lt of in tegration or
the cause of i t .
To round out the c ritic is m is to note that Deutsch does not
even discuss channels. Evidently, he takes them fo r granted because
they are e ith e r not important or because they defy id e n tific a tio n and
description. But channels are a v ita l element in communication en
gineering and theory; and fo r th is reason, i f no o ther, i t would seem
imperative that a model builder in in ternational re la tio n s , or in any
d is c ip lin e , would a t least account fo r th e ir insign ificance, i f he
cannot deal with them as variab les .
No c ritiq u e of Deutsch's use of the model could leg itim ate ly
accuse him of misusing the concepts because he used what he could—
what he perceived as relevant and applicab le . He had to match his con
cepts to his subject m atter. This is how he defined them. The indices
which he id en tifie d are not bereft o f s ign ificance; but they are
severely lim ited in th e ir correspondence with the ideal concepts d is
cussed e a r l ie r ; and they lack the precision of logical re lationship
which characterizes the communication model o f Chapter 2 . This c r i t
icism does not invalidate Deutsch's work. I t only q u a lifie s its
veracity as an operational manifestation of communication modelling.
59
Innovation is welcome, so long as i t is fu l ly explained and
deviance is accounted fo r 0 But serious epistemological and method
ological problems can arise when a researcher changes some of his
premises and not others; and when he attempts to make his whole e f fo r t
hang together on a common-sense in terp re ta tion of the subject m atter.
Russett; Responsiveness
Russett's study of Anglo-American relations is an extension of24
Deutsch's search fo r indicators of in teg ratio n . But Russett has
chosen a more spec ific focus than Deutsch's general in teres t in
in tegration . Russett chose to investigate what he called responsive
ness. Responsiveness can be understood as a cybernetic concept ( i . e . ,
a process involving communication and contro l) . I t connotes an or
ganism's or a system's sensory awareness of its environment, its commu
nication with that environment, and its attempt to control its own be
havior based on the sensory information from its environment. In other
words, by a process of feedback, the system organizes and reorganizes
i t s e l f in order to maintain its id en tity and survive in its environ
ment.
What Russett has done is to id en tify a p a rtic u la r instrumental
process— responsiveness—which characterizes the state of integration
a t a p a rtic u la r time. Deutsch was attempting to investigate the s tate
of a country's integratedness with its environment by certa in
24. Bruce M. Russett, Community and Contention (Cambridge, Mass.: Massachusetts In s titu te of Technology Press, 1963).
6 0
measurements of flows which he f e l t were indicators of progress toward
or away from th is state of in tegration .
Taking up where Deutsch le f t o f f 9 Russett narrowed the scope
of his inquiry to two countries and chose to investigate one of the
primary processes (responsiveness) by which a country matches its
caoabi1it ie s to its loads. I t is s ig n ific a n t, even c ru c ia l, that
Russett redefines the concept of integration fo r th is study. In
Deutsch*s terms, the United States and B rita in would already be in te
grated. Deutsch*s d e fin itio n of in tegration asserts th a t, w ith in a
te r r i to ry , a "sense of community" is attained and certa in in s titu tio n s
and practices become '(strong enough and widespread enough to assure,
fo r a 'long* time, dependable expectations of ‘peaceful change* among
its po p u la tio n ."^
Russett's d e fin itio n asserts no f in a l i t y to the process of
in tegration; but rather i t turns on the dynamic ongoing nature of a
relationship between a country's c a p a b ilitie s fo r responsiveness and
the loads placed on those cap a b ilitie s by another country. Russett
explains: " I f we say two countries are becoming more integrated, we
mean that the growth of c ap ab ilities is outpacing the increase in
loads. . . .Thus integration is a process, and we shall re fe r to the
ra tio of c ap a b ilitie s to loads a t any time as the s ta te of Integra-26tio n , of which responsiveness is largely a function."
25. Karl W. Deutsch, e t a l . P o lit ic a l Community and the North A tla n tic Area (Princeton: Princeton University Press, 1957)» P. 5 .
26. Russett, Community and Contention, p. 39.
61
In Russett8s theoretical framework, responsiveness is a
general term— d is tin c t from a spec ific response. Responsiveness is
"the probabi1jtv that the demands of one party w il l be met with indul-27gence rather than deprivation by the other p a rty ," This frees him
from the deductive imperative of having to show which request evoked
each response; besides, the flow of responses— the trend or pattern o f
responsiveness— is what is sought, in keeping with the princip les of
communication theory. But, the in a b ility to locate the feedback
channel and id en tify the nature of its process is a serious short
coming, i f the researcher is concerned with th is process s p e c if ic a lly ,
and not ju s t w ith a description of a communication network in general
terms,
Russett id en tifie s two types o f actors who can exh ib it re -
sponsiveness: (1) Governments acting o f f ic ia l ly ; and (2) individuals
acting u n o ff ic ia lly (e ,g 0, government o f f ic ia ls or members of in terest
groups), A spec ific demand placed on one party may be met in one of
three ways: (1) The demand may be given atten tion and indulgence; or
( 2) i t may be met w ith a tten tion and p a rtia l indulgence; or f in a l ly
(3) a demand may be d e lib e ra te ly ignored or inadvertently unnoticed and
the resu lt w il l be deprivation. Russett admits that while i t is pos-
, s ib le to make ana ly tica l d is tin c tio n between inadvertent and deliberate28deprivation, i t is v ir tu a lly impossible to do so in p rac tice . But in
the long run, de liberate inattention to a series of demands or request
27. Ib id . , p. 30.
28. Ib id .
' ' ' ■ )
62
would probably become apparent and thus no real problem arises
( i . e . , when the technical level o f communication has been assured,
tested and proven fo r certa in types of messages, i t would probably
become apparent that the problem o f non-responsiveness was, in fa c t,
a semantic problem—or perhaps, the effectiveness problem).
Having discussed responsiveness in terms of Russett’ s general
theoretical framework, i t is now possible to explore his d e fin itio n o f
responsiveness in greater d e ta il . He asserts that p o lit ic a l units
(countries) have certa in ca p a b ilitie s which are: (1) The capacity to% '
act (e ,g , , m ilita ry strength, wealth, s ize — "power"), (2) the capacity
to perceive and communicate ( i , e , , c a p a b ilitie s which he ca lls a tten
tion and communication such as perception o f one's v ita l interests and
communication of those perceptions to another p o lit ic a l u n it's decision
making centers); ( 3) the capacity of mutual id e n tific a tio n (th is cannot
e x is t without exercise of the attention-communication capacities but
i t need not always fo llow these c a p a b ilit ie s ) . C ap ab ilities (2) and
(3) are of prime in teres t to Russett: "A ll these f a c i l i t i e s —atten
t io n , communication, and mutual id e n tif ic a tio n —we shall c a ll 'capab il
it ie s fo r responsiveness,' They must carry the demands which each29
country makes on the o th er," While c ap a b ilitie s fo r action (number 1
above) sometimes overlap with c a p a b ilit ie s fo r responsiveness, Russett
contends that i t is useful to conceptually distinguish the two because
action presupposes a cap ab ility to act while responsiveness only
presupposes an intention to ac t,
29. 1 b id , , p, 27.
63
So f a r , Russett has said that a country has certa in c ap ab ili
ties fo r responsiveness which are a tte n tio n , communication and mutual
id e n tific a tio n * Now i t is necessary to describe what he called the
loads or demands which ac tiva te the process o f response. Demands are
not only formal requests or demands, but also "informal requests, pleas
made by p riva te individuals or agencies e ith e r to the government or to
other private ind iv iduals , and merely ’ s itu a tio n s ” that need a tte n tio n ,
whether or not an e x p lic it request is ever m ade,"^ Loads or demands
have three characteris tics which distinguish them: (1) D irection—
degree of complementarity or c o n tra rity with other demands being made
on the decision making center; (2) number; (3) weight--importance
attached to the demand by the requester.
What occurs, then, is that a country has loads placed on i t
and attempts to deal w ith those loads by exercising its cap ab ilities
fo r responsiveness, Russett’ s theoretica l framework constitutes a
model of a process which is represented by a ra tio of cap a b ilitie s to
loads. As he says, "the responsiveness of A to B is thus largely a
function of the ra tio o f the c a p a b ilitie s fo r responsiveness which A
directs toward B over the loads coming from B to A a t any given
tim e," But Russett recognizes that no s ituatio n is ever so simple.
He admits that c a p a b ilitie s can never be precisely specified or
30, Ib id , , p, 29,
31. Ib id ,
64
measured and that "the state of a f fa irs In the in ternational arena is
also an extremely important v a ria b le .
C apab ilities fo r responsiveness— a tten tio n , communication, and<
mutual id e n tif ic a tio n — require certa in " fa c i l i t ie s " fo r th e ir re a liza
tio n . These fa c i l i t ie s are e ith e r formal in ternational and supra
national in s titu tio n s fo r channeling a tten tion and communication; in
formal practices which do the same thing; or merely habits and memories
of a tte n tio n . These fa c i l i t ie s depend upon certa in physical channels
fo r the transmission of perceptions such as te letype machines and
telegraph lines; but Russett is not concerned with th e ir existence.
He recognizes that some delay is b u ilt into these channels but he con
siders i t normal; and he states that the importance o f the channels
lies in the frequency of th e ir use. "This use, the contacts ac tua lly
made, is what we mean by the fa c i l i t ie s fo r communication and atten tion
between nations
Given a two-country system in which demands and responses are
exchanged, how does Russett go about measuring the extent of the re
sponsiveness which would characterize the state of in tegration ex isting
between the two countries? How does he a c tiva te the model?
Russett chose what he determined were the f iv e most normal
years of Anglo-American re lations as the periods o f time which would
y ie ld the most va lid and revealing information about Anglo-American
32. Ib id .
33. ib id . , p. 27o
65
re la tio n s . The years 1890, 1913$ 1928, 1938, and 1954 were years, he
concluded, which were not beset by any v io len t upheavals or extraor
dinary events (wars, fin an c ia l crises) and would probably not y ie ld
data skewed in e ith e r d irec tio n — toward or away from a normal state o f
in tegration and the subsequent level of responsiveness. What kinds
of data would be most appropriate and ava ilab le fo r tracing the trends
in each country's c a p a b ilitie s fo r responsiveness?
As indicators of the a tten tion c a p a b ility , Russett measured
"the proportion o f space devoted to the other country in e l i t e news
papers, the proportion of references in scholarly research to work done
in the other country, and the amount of a tten tion given another nation
34in a s ta te 's education system." The indicators he chose in order to
characterize and measure the communication cap ab ility were "trends in
mail flows, trade in goods and services, telephone and telegraph commu
n ica tio n , student exchange, travel fo r business and pleasure, various
kinds of contacts among members of the e l i te s , magazine and motion
35p ictu re exchange, and m igration ," F in a lly , as indicators of mutual
id e n tif ic a t io n , he notes that many of the same measures of communica
tio n - -p a rtic u la r ly movies and m igration --are also useful as measures of
th is las t c a p a b ility . "Other indicators of mutual id e n tific a tio n are
the amount of approval expressed in the e l i t e papers and in mass36
opinion as measured by survey data ." _____________
34, Ib id . , p. 31.
35, Ib id .
36, Ib id .
66
A fte r each of these indicators had been measured fo r each
yearly period in question* a statement about the level o f responsive
ness of A to B and B to A could be made (conveniently* A and B can be
thought of as America and B r ita in ) . But i t is important to remember
that Russett insists on a combinational resu lt based on a l l three of
the measured c a p a b ilit ie s . In short* a l l three indices must point in
the same d irec tio n — eith e r toward indulgence or toward d e p r iv a tio n -
before a judgment can be made about the conclusiveness of the data.
This approach, he fe e ls , avoids the need to weight the indices. For
him, "a tte n tio n , communication* and mutual id e n tific a tio n are a l l
37essential to responsiveness; there can be no weighting of the th ree ."
Russett recognizes that there might be a question about
whether an indicator was a cause of community or an e ffe c t of an a l
ready existing sense of community. He answers th is question by invok
ing the pervasive influence of feedback. He concludes that "most of
the aspects of cap ab ility that we shall examine are in fa c t both cause
and e ffe c t . The messages not only contribute to the development of
community but also are themselves promoted by the community's e x is t
ence, as the two elements are mutually re in fo rc in g ." ^
Since.he is concerned, not with the factors which promote a
trend in a p a rtic u la r c a p a b ility , but with what the overall trend is ,
and its e ffe c t on in teg ra tio n , he fee ls absolved of the alleged sin of
conceptual ambiguity in th is area of mutual reinforcement. In
37. Ib id . , p. 32.
38. Ib id . , p . 33=
addition , he declares the role of force in in tegration as outside the
scope of his study, since he is concerned with two nations which have
b u ilt th e ir integration on free and cooperative associations fo r
problem solving,,
Russett: Analysis
One of the f i r s t things which should be apparent about
Russett's model is its great v a rie ty o f indicators (grouped under the
three cap ab ilitie s ),, While th is could conceivably be a sign of the
greater a v a i la b il i ty o f data ( fo r Russett as compared with Deutsch),
i t is , more probably, a function of Russett's more ambitious design
(and i t should be noted that Russett's study was o r ig in a lly done as
his doctoral d issertation under Deutsch) 0 By iden tify ing many more
types o f flows as "message1* flows, Russett is acknowledging the fa c t
that nearly every public event— i ee 0, every event that becomes public
knowledge— is ac tu a lly a message of sorts which conveys information
about the conditions of the system in which i t takes place. Thus, by
attempting to expand his l i s t of indicators of a tte n tio n , communica
tio n , and mutual id e n tific a tio n to include some q u an tifiab le flows of
events, Russett is follow ing the p rin c ip le of examining more and more
of the relevant sequences of symbols or events which occur in the fo r
mal and informal relationships between two countries. Id e a lly , i f one
could examine a l l o f the sequences o f events of human existence, one
could decode the sequences and "get the message" of h is to ry . But,
6 8
th e o re tic a lly , one does not even have to decode a ll the sequences
in order to learn a t least something about the character o f the
transmission,
Russett is rather careless in his use of the concept of re
sponsiveness, A n a ly tic a lly , he divides the concept into three compo
nents (a tten tio n , communication, mutual id e n t ity ) | but he then admits
that these components are sometimes indistinguishable from one another
in p rac tice . In the data co llection and c la s s ific a tio n stage, then,
i t must be necessary to make some rather a rb itra ry decisions about
whether an event should be lis te d as a tte n tio n , communication, or
mutual id e n tif ic a tio n . Furthermore, when there is an overlap, should
a c ap ab ility be counted under a l l three categories? This would destroy
the distinctiveness of the component aspect of responsiveness by admit
tin g , perhaps, that everything is re lated to everything (which may be
tru e , but i t is not a very useful revelation fo r purposes of descrip
tio n , le t alone explanation),
Russett1s d is tin c tio n between a sp ec ific response and respon
siveness, as a general condition, is consistent with communication
theory's principal in terest in the pattern or sequence of messages
which conveys information. Each response is a yes-no or on-off prop
o s itio n , i , e , , e ith e r indulgence or deprivation is the nature of the
specific response. Together, a series of deprivations and indulgences
te l ls the observer something about the source ( i , e , , the sender),
In a sense, Russett has constructed his responsiveness concept
in a way that seems to p a ra lle l the engineering d is tin c tio n between
69
the source and the transm itter* The ca p a b ility of mutual id e n tif ic a
tion could be thought of as the source of the messages. In other
words, a l l the previous experiences, memories, lo y a ltie s , and a ttitu d es
o f a nation (w ith , toward, about another nation) could be thought o f
as a fund or source of possible choices from which the transm itter
selects when a message is ac tu a lly communicated. The actor (govern
mental or ind iv idua l) could be thought of as the transm itter of the
message, since he inputs the response (pos itive or negative) into the
channel„ The transm itter could be viewed as a modulating and coding
device which prepares the message fo r transmission,
Russett chooses to disregard the influence of physical channels
on responsiveness—accepting the technical problems o f delay and assum
ing general adequacy. This is not a c tu a lly an unreasonable assumption
because, in in ternational re la tio n s , a l l the actors would be faced with
ostensibly s im ila r parameters of channel capacity. On the other hand,
i t could be argued that channel capacity is a lim itin g fac to r and care
fu l a tten tion ought to be paid to i t ,
Russett does acknowledge the p o s s ib ility of overload a t the
receiving end of a communication c h a n n e l H e admits that the deci
sion making centers of government, which must be a tte n tiv e to requests
from the other countries, do become overloaded by the in flux of demands
and requests; and th is overloaded condition forces certa in internal
reorganizations in manpower (and sometimes structure) to allow the
decision centers to handle the extra load. During these overloaded
40* Ib id , , p , 28,
periodss and because of upheavals in the larger system (outside the
two-country system), decision centers may become overloaded to the
point of breakdown. When th is receiving function has deteriorated too
fa r , i t is l ik e ly that non-responsiveness (between A and B) w il l in
crease due to in atten tio n .
Russett does not label this overload (which results in inatten
tion ) noise, but i t could be construed in th is way. His model has no
way of managing the noise problem in periods o f great in ternational
stress; and this is why he had to choose "normal" years fo r the study.
In any event, these normal years were chosen in an attempt to discover
what the normal flow o f requests and responses might look lik e ,
Russett has used the idea of decision making centers in surprisingly
close conformity to the meaning and function of receivers (discussed
in Chapter 2 ) .
The discussion of overload above leads to the discussion of
another problem in Russett11 s model: the general d e fin itio n and use of
the concept of loads or demands. In i t i a l l y , he acknowledges three
components of loads: d ire c tio n , weight, and number. The theoretica l
importance of these loads is placed on par with that o f the concept of
responsiveness. But i t turns out that loads and demands are studied
as much more amorphous "situations" and threatening trends or a t
mospheres of " f r ic t io n ," than the re la t iv e ly more precise delineation
of ca p a b ilitie s which he provides. Trade discrim ination and investment
r iv a lry , fo r example, are considered as indicators of loads. The im
portant element which is central to communication theory is preserved.
however, in his treatment of loads0 He adheres to the p rin c ip le that
i t is the probabi1itv generated by the sequence of loads which is im
portant* Thus loads-as-a-general-condition of tension is a good way
to conceptualize the p ro b a b ilis tic s itu a tio n in which responsiveness
reveals the state of in tegration*
The most serious flaw in the model is the exclusion of events
which occur outside the immediate system of Anglo-American concerns®
But, given the fa c t that l i t t l e happens a round the world which the
United States fee ls is not in its in te re s t, i t may be a se lf-co rrec tin g
deficiency (so long as the United States is one o f the countries in
the study)®) By choosing "normal*1 years, and by trea tin g the outer
in ternational system as irre le van t to the Anglo-American sub-system,
Russett has severely lim ited his model to considerations which, in the
long run, may not be very accurate or in teresting e ith e r to the student
or the p ra c titio n er of in ternational relations®
I t is in teresting to note that Russett ac tu a lly develops sev
eral measures of responsiveness along the conceptual lines of response
of individual and response of government® By measuring trade, invest
ment, and a llian ce trends, for example, he is investigating governmen
ta l or systemic trends; while his a tten tion to e l i t e a ttitu d e s ,
perceptions, and so c ia liza tio n points to his in teres t in individual
responsiveness® This cleavage is a most important fac to r in assessing
the characteris tics of the messages a spec ific type of source is lik e ly
72
to put outo For th is reason, i t is puzzling why Russett did not
attempt to tabulate separately his trends in governmental and ind iv id
ual responsiveness.
F in a lly , the place of the concept o f feedback in Russett's
model must be analyzed. Aside from several apologies about the
"mutually re inforcing" interdependency ex is ting between the three
ca p a b ilitie s and the sta te of in teg ra tio n , Russett makes no overt
attempt to f i t the concept of feedback into his model. He acknowledges
the fa c t that i t Is d i f f ic u l t to d istinguish between the cap ab ilitie s
as causes of integration and the c a p a b ilitie s as e ffec ts o f integration
already achieved; but he w rites th is ambiguity o ff w ith a reference to
Wiener's development of the concept. This is not adequate treatment
fo r a concept which is essential in a two-way communication system. In
short, he provides a f a i r ly adequate v is ib le s tructu ra l p icture of the
forward channel linking each country with the o ther, but he leaves
the feedback channel poorly defined and v ir tu a lly without p ic to r ia l
id e n tity .
In summary, th is analysis finds Russett's model to be a f a i r ly
close approximation o f a communication system as described in Chap
te r 2 . His use of the basic concepts is generally sound; however he
does f a i l to make the process of feedback very c le a r. He has made an
ambitious attempt to demonstrate the factors involved in integration by
dealing with an instrumental process of in tegration— responsiveness.
His empirical referents fo r communication flow are re la t iv e ly numerous
and varied; they promise greater breadth o f sequence id e n tific a tio n
73
and decoding,. The reduction of uncertainty by information in flux is
apparent, though undiscussed, in the re d e fin itio n o f integration as a
continual, building process.
McClelland: Event In teraction
I f Russett succeeded in increasing the number of indicators of
in teg ra tio n -o v e r the number employed by Deutsch=”McClelland has
fashioned, out of the communication model, an approach to in ternational
re lations which has the po tentia l o f trea tin g a l l events as relevant
•'messages'* in a communication system,,^ McClelland's approach, gather
ing newspaper accounts of events, c lass ify ing them into categories, and
computing the frequency o f each type o f event, considers flows of
events as the flows which reveal patterns of in ternational behavioro
McClelland views the in ternational configuration o f nations as
a system (made up of sub-systems) which is characterized, a t any given
time, by the flow of events taking place* As they occur, these events
form patterns and structures of in teraction among the actors in the
system* He recognizes, with communication theorists and engineers,
that in a given sub-system consisting of two countries "one of the
things that would be p ro fita b le to discern would be the patterns and
sequences and successions of actions of one side and the responses of
the other s ide;" but, admitting the complexity of th is type of dynamic
41 * Charles A* McClelland, "Access to B erlin ; The Quantity and Variety of Events, 1948-1963," Jo David Singer, ed* Q uantitative In ternational P o litic s (New York: Free Press, 1967)0
lh
feedback analysis , he concludes that "how to do th is most e ffe c tiv e lyI fO
s t i l l remains an unsolved problem,.
Not being able to say precisely which action or event (by one
actor) produced a subsequent event or response by another ac to r,
McClelland contends that an "a ttra c tiv e a lte rn a tiv e is to raise the
question of how the opposing actors ‘mix’ th e ir behaviors over a speci
fie d length o f t im e ," ^ This is a good hypothetical question which
comes d ire c tly out of communication theory’ s understanding of the
s ta t is t ic a l characteris tics of a source0 In other words, an examina
tion of the s ta t is t ic a l characteris tics of the message sequences pro
duced by a source may, in time, y ie ld (1 ) a knowledge of the kinds of
message patterns to expect from the source and thus (2) an understand
ing of the performance of the communication system* From Chapter 2 ,
we know that by examining the past sequences of messages from a source,
i t is possible to say something about the probable fu ture production
of sequences of messages* But, as always, i t is the flow or the tim e-
series sequence which yields an increasingly complete p ictu re of the
ch aracteris tic performance of a communication system*
McClelland's focus in th is study is what he c a lls the problem
of the "acute in ternational c r is is *" Because crises are dangerous to
peace and create untold tension in in ternational re la tio n s—p a rtic
u la r ly when the superpowers are involved—McClelland feels that an
h is to rica l examination of the events leading up to and following a
42* Ib id * , p* 169*
43* Ib id *
75
c r is is may y ie ld important insight into the "change of s ta te" of the
p a rtic u la r in ternational system. This change of s ta te would be re fle c
ted in the type of events which occurred and th e ir "m ix," i 0ee, the
re la tiv e frequency of certa in types of events a t d iffe re n t times. The
threshold of event in tensity and v a rie ty , beyond which the system w ill
change from a non-crisis to a c r is is s ta te , is the conceptual objective/i/.
of this examination o f past events,
McClelland's preoccupation w ith crises and th e ir characteris
tic s is epitomized by his b e lie f that
the s ig n ific a n t varia tio n in the flow of action in a system is the central in terest o f in ternational system analys is , , , oThus, the long-term objective of both theory and research on the in ternational system is not to find out everything about the system; i t is only to construct an appropriately complex and s u ff ic ie n tly accurate set ofperformance indicators so that 'readings' o f the s ta te ofthe system can be taken a t any time, ^
To apply these theoretical in terests to a concrete system, and
to test his propositions about acute in ternational crises in general,
McClelland studied the sub-systemic c r is is defined by the problem of
access to B erlin during the years 1948 through I 963, He considers the
B erlin access s ituation to be a sub-system of the larger East-West sys
tem of p o lit ic a l events. Narrowing the focus in th is way, he is able
to consider only that certa in class of " lo ca l" events which is relevant
to the sub-system. These local actions are defined as events which
"have (1) challenged d ire c tly the Western position in B erlin and on the
44, Ib id , , p, 160,
45, Ib id . , p, 162.
76
access routes, or (2) have been defensive measures mounted against
these ch a llen g es ,"^ He s p e c ific a lly excludes the actions of o f f ic ia l
government agencies— high diplomacy— because of his desire to avoid
any consideration of motivation or perception.
Iden tify ing and c lass ify ing the relevant local events con
s titu te s the major preoccupation of the study, i t seems, The New York
Times Index was used as a source lis t in g of events pertaining to the
B erlin s itu a tio n . When co llected , th is chronology o f events consti
tutes , in a very crude way, the flow o f action in the sub-system. The
tes t of sorting the "re levant11 events from the irre levan t ones was ,
the simple question; "'Does this item re la te d ire c tly to moves made in
the East-West action fo r position and access e ith e r to and from B erlin
or w ith in B e r lin ? '" ^
When the relevant items had been selected (1791 of them), each
of them was coded or translated to f i t into one of the most appropriate
of 18 d iffe re n t categories of action . The categories were chosen to
conform as closely as possible to the language used in the Index; and
a set of synonyms was developed fo r each category. The categories
settled on were as follow s: Accede, Withdraw, Request, Propose, Bar
gain, Convey, Abstain, Protest, Reject, Deny, Accuse, Demand, Warn,48Threaten, Decree, Demonstrate, Force A ttack, I t should be apparent
46, Ib id , , p, 160.
47, ib id , , p. 167,
48, Ib id , , p, 168,
77
that the l i s t Is rank-ordered to denote a range of actions running
from cooperation, through bargaining, to extreme c o n flic t; but there
is no assumption o f in ternal distance between types„
The next operation on the data was a calculation of the f r e
quency of the occurrence of each type of action , fo r each of the two
sides (East-W est), fo r several d iffe re n t time periods (e .g . , each month
and y ear). McClelland could have stopped here, as Deutsch and Russett
did , s e ttlin g fo r the gross measurements o f volume and v a rie ty of
events. But he chose to subject his frequency data to a fu rther set
of manipulations based on techniques from information theory.
Viewing his 18 types of action as possible choices or a lte rn a
tives which each side had an equivalent chance of selecting or "emit
t in g ," McClelland proceeded to quantify the amount of information in
each time period— information defined and lim ited by the number of
choices involved. As outlined in Chapter 2 , the a p r io r i uncertainty
in a s ituatio n involving the selection of a number of equiprobable
a lte rnatives can be measured. McClelland did th is fo r each actor and
time period under study, c a llin g th is measurement the maximum uncer
ta in ty of the s itu a tio n . Then he figured out the absolute uncertainty
fo r each actor and time period— based on the actual number of events
occurring in each category. F in a lly , using these two measurements of
uncertainty, he constructed the ra tio o f absolute to maximum uncertain
ty , c a llin g i t the re la tiv e uncertainty o f the s itu a tio n ( i . e . , each
p erio d ). And th is measurement represented the "va rie ty " of events
In it ia te d by each side in the East-West c o n flic t over access to
Berl in
McClelland concluded th a t, when the re la tiv e uncertainty
measurements fo r each actor were arrayed in monthly or yearly succes
sion, the highest measures of v a rie ty coincided with the "known" c r is is
periods of 1948 and I 96 I ; that is , the mix of the number of actions
taken by each side increased during a c r is is period and th is was re -eg
flee ted in the re la tiv e uncertainty index fo r each period,
McClelland: Analysis
I t has already been noted that McClelland's use o f information
theory techniques is in conformance with the princip les of information
and uncertainty in communication theory. His concept of events as
messages is also a leg itim ate use of the idea that the number of prob
able events in an a p r io r i s itu a tio n is ind icative of the amount of
information which can flow when the choice is made among the a lte rn a
tiv e s . Wiener provides a spec ific support fo r th is conception of mes
sages as events. He speaks of a message transmitted in a l l directions
a t once--as a sort o f "to-whom-it-may-concern" message—which radiates
out from its source u n til i t is received by some concerned receptor.
R ightly , McClelland attempts to iso la te his sub-system an a ly t
ic a lly from the larger system. This makes sense because of the
— 1 ------49o Ib id . , p. 172.
50. Ib id . , p. 184.
51. Wiener, Human Use, pp. 96-97=
79
necessity to reduce the scope of the problem to manageable proportions
and in terms of the inherent problem of constructing a simple yet
convincing model„
He has to concern himself w ith the " lo c a l11 s itu a tio n . Other
message-events may be related to the problem (e .g . , a Korean war may
d is tra c t the a tten tion of the actors from the local s itu a tio n ) butj
these remote events are too d i f f ic u l t to id en tify w ith the sub-system
and are thus le f t outside its boundaries. Th is , p ra c tic a lly , is
McClelland's only way of "getting on" w ith the study. His attempt to
show that perturbations in the larger system cause perturbations in the
sub-system seems inconclusive since his data array fa ile d to show one
out of the three major crises which ac tu a lly did occur during the 16
year period .^2
McClelland's focus on a threshold as a graphic indicator of a
c r is is s ituatio n is a leg itim ate , i f awkward, borrowing of the p r in c i
ples involved in receivers. Jumping an a ly tica l levels fo r a moment, i t
is possible to view the threshold idea and the system performance idea
as in terre la ted parts o f a communication system. Since McClelland's
goal is to develop a method of "reading" the system to determine its
state a t any given tim e, the system's performance could be viewed as
"on" when the threshold had been reached and "o ff" when the performance
level fa ile d to reach the c r is is threshold. But th is makes the whole
system into a c r is is system—with undue emphasis on a c r is is as a good
performance. Further, th is approach (threshold) screens out a l l the
52. McClelland, "Access to B e r lin ," p. 184.
8 0
in teresting and important phenomena which occur below the threshold
1 ine»
Another problem with the model is that McClelland fa i ls to
account fo r the “quiet periods"— except to say that the system is non
existent a t these less intensive times. His quiet periods were a r
b i t r a r i ly defined as periods of time in which fewer than f if te e n events
of a l l types occurred. Below th is le v e l, he f e l t that the system was
not in existence because nothing much was happening. This is not
very h e lp fu l, even though i t does re lieve McClelland from the burden
of his problematic data. To rid himself o f th is lackluster data, he
reduces the number of relevant months fo r scrutiny from the orig ina l
192 (16 years) to 41 This amounts to nearly an 80 percent reduction
in the "active" or functioning "on time" of the sub-system. By any
standards, th is appears to be a wholesale decimation of the data base.
Only 41 months are used to c o lle c t, on the average, about f if te e n
events per month—and that is counting East and West together. In
addition to the small s ize of the relevant sample (in months), i t is
c le a rly inaccurate to postulate that the system, the communication
network, goes out of existence simply because no messages are sent over
i t .
McClelland admits that his method of data gathering (scouring
the Times Index) is wrought with severe de fic ien c ies . He also acknowl
edges the fac t that his resu lts , using information theory techniques,
53. Ib id . , p. 173.
54. Ib id . , p. 182.
81
are severely lim ited by the paucity o f events in most months of the
test period. What th is leads one to suspect is that McClelland's
fascination with the p o s s ib ilit ie s of information analysis (in the
technical sense) led him to seriously overestimate the strength of his
data, and thus, the conclusiveness of his findings.
The major fa u lt which must be pointed out in McClelland's
study is his reliance on an extremely small number o f events to repre
sent his "flow" or sequence of messages. Event analysis is possible
and consistent with communication theory; but. i f too few messages are
considered, i t v io la tes the sense of the message as i t is known in
communication theory.
Another in teresting problem with th is model is one which
McClelland mentions but disregards—out of necessity. He admits that
there is a school of thought that would c r it ic is e his behavioral
approach to the problem of c r is is . This c r itic is m focuses on the psy
chological component of in ternational c rises . McClelland notes that
th is psychological approach "might reg is te r to a lim ited extent the
perceptions of national a c to rs ,” by proceeding with "some type of con-CC
tent analysis of diplomatic messages and statements. . . ." But he
dismisses th is approach in favo r.o f his own. Perhaps he does this be
cause of the greater ease of co llectio n or the supposed advantage of
escaping from the in terp re tiv e problems of semantics.
But, th is type of analysis, o f diplomatic messages and s ta te
ments, might demonstrate that the sub-system was very much a liv e during
5 5 ° Ib id . , p . 165.
82
those alleged "quiet periods." At any ra te , th is type of communication
between actors cannot be a r b it r a r i ly dismissed as unimportant simply
because other "hard" data provide an a ttra c tiv e or easy a lte rn a tiv e .
A f in a l general assessment of"McClelland's adaptation of the
communication model must include a comment on his choice o f "quantity
and varie ty" of events as indicators of some sort o f c r is is in the
system. He is forced to admit that his measurements of va rie ty (using
information analysis) are not e n tire ly independent of his measures of
quantity or volume. Given the fa c t that he must discard much of his
data when i t does not y ie ld conclusive evidence of a flow , one is
forced to ask: Why not s tic k with volume or quantity i f i t is almost
as i l lu s tra t iv e o f the c r is is he seeks to identify? The answer would
be, obviously, that his methods and his techniques got the better of
him. He was fascinated with the idea of a threshold which might be
v isu a lly apprehended and with the pred ic tive p o s s ib ilit ie s which that
threshold e n ta ile d . But without a great deal more volume in his com
p ila tio n of relevant events, i t appears to be only an in teresting
hypothesis (threshold) which defies tes ting—a t least fo r the present
techniques of data gathering.
McClelland's apology fo r his focus on volume is hauntingly
sim ilar to Deutsch's, made some eleven years e a r l ie r . Deutsch admits
that "where such q u an tita tive models are attempted, . . .they raise the .
question of operational d e fin itio n s and techniques of measurement fo r
the variables they e m p l o y . A n d he goes on to concede that volume
56. Deutsch, "Communication Flows," p. 144.
83
w ill be the most eas ily measured variab le or dimension of communication
flow . Again, he realizes that "most o f the work on the measurement o f
in ternational communication flows is yet to be done,"
Eleven years la te r , McClelland must also admit that "the
changes in quantities in some time periods stand out and can be d is t in
guished re ad ily , but often the meaning of the numbers tend to be ob-
scured," He, too, makes the standard disclaim er and the a ffirm ation
of tomorrow:
At some future date, the rearranging and testing of d iffe re n t combinations o f indicators probably w il l become a major research preoccupation. Meanwhile, the task is to begin the construction, one a t a tim e, of possible indicators in the expectation th a t, sooner or la te r , they w il l be assembled and used in an e ffe c tiv e composite form of measurement,^
Thus, in a s p ir i t of honest, i f irreve ren t, scepticism we can
ask: How fa r have in ternational re la tions scholars come in the
application of the communication model to the study of in ternational
relations?
57» McClelland, "Access to B e r lin ," p, 169,
58, Ib id , , p, 162,
CHAPTER 4
CONCLUSION
Whether our investigations in the social sciences be s ta t is t ic a l or dynamic—and they should p a rtic ip a te in the nature of both— they can never be good to more than a very few decimal places, and, in short, can never furnish us with a quantity of v e r if ia b le , s ig n ific a n t information which begins to compare with that which we have learned to expect In the natural sciences.
Norbert Wiener Cybernetics. p. 191.
8n the beginning, the task undertaken was to probe the ques
tion ; How has the communication model been adapted to the study of
in ternational relations? To answer that question in fu l l would have
been a monumental task. But what was learned here was the manner in
which three d iffe re n t authors attempted to operationalize some of the
concepts basic to the theory. The adaptation, however, has not been
a complete success— nor should we have expected i t to be so.
At the end of Chapter 3 , i t was noted that McClelland had ex
perienced substantia lly the same problems and expressed essen tia lly
the same reservations as Deutsch had in his early use of the model.
This does not mean that the trend postulated e a r lie r is non-existent.
That old problems of finding va lid and measurable indicators of flow
s t i l l haunt present day applications of communication theory is not
evidence that the trend is non-existent. Rather, the recurrent
84
85
problems of finding and measuring, in large q u an tities , su itab le in
dicators of flo w --w h ile trying to adhere to the meanings of the basic
concepts of communication theory and the princip les of th e ir in te r
re lationships— has, simply, been an extremely d i f f ic u l t enterprise®
McClelland's new work in event analysis may prove to be the most useful
adaptation o f the theory which w i l l be able to help scholars in th e ir
study of the in ternational system. But, as has been pointed out, there
are serious problems of data gathering and manipulation which s t i l l
need study and refinement.
Returning to Wol in 's understanding of the function of th e o ry -
mentioned in the beginning of this thesis— i t can be modestly claimed
that th is investigation of the basic concepts of communication theory
has helped to c la r ify the concepts and has led to a be tte r understand
ing of th e ir in terre la tio n sh ip s . But, is i t necessary that these
rigorous d e fin itio n s be imposed on the subsequent applications which
have been made of the theory?
1 th ink that I am inclined to agree with Kaplan who refers to■)
a model as "the embodiment of a s tructura l analogy," And, he con
tinues, "no theory is to be condemned as 'merely an analogy' ju s t
because i t makes use of one," The conclusion which I reach, then,
1, Kaplan, The Conduct of In qu iry , p. 266,
2 , ib id .
86
concerning the application of communication theory to subject matter
foreign to i t is , w ith Kaplan, that
no two things in the world are wholly a lik e , so that every analogy, however close, can be pushed too fa r ; on the other hand, no two things are wholly d is s im ila r, so that there is always an analogy to be drawn, i f we choose to do so,, The question to be considered in every case is whether or not there is something else to be learned from the analogy i f we do choose to draw i t . 3
This "something else to be learned" is precisely what is gained by the
application of the model to the in ternational system. I f the In te r
national re lations scholar does not try to push the analogy too fa r ,
i f he avoids the fa lla c y of what Kaplan c a lls p ic to r ia l realism
("The mistake of p ic to r ia l realism is fo rgetting that the s im ila r ity
exists only in a given perspective, that i t depends upon a p a rtic u la rbmode of representation." ) , then he can use the model to describe and
even to understand the in ternational system --if not to explain its
processes.
The conclusion which may be apparent from the above observa
tions is that the communication model is an analogy. I t is a conven
ient way of ta lk ing about certa in processes in in ternational re la tio n s .
I t can disclose relationships in a new l ig h t . I t can help the student
and the scholar to a better understanding of events by providing him
with a perspective—a framework fo r analys is . But, in the end, the
analog— the model— cannot re a lly begin to explain the processes or
3. Ib id .
4 . Ib id . , p . 287.
87
the performance of a system cast in communication terminology. Speak
ing of the Incautious philosopher, but using words of no less impor
tance to the social s c ie n tis t, Hans Reichenbach comments on the nature
of explanation.
S c ie n tific explanation demands ample observation and c r it ic a l thought; the higher the genera lity aspired to , the greater must be the mass of observational m a te ria l, and the more c r it ic a l the thought. Where s c ie n tif ic explanation fa ile d because knowledge of the time was in s u ffic ie n t to provide the righ t genera liza tion , imagination took its place and supplied a kind of explanation which appealed to the urge fo r genera lity by satis fy ing i t with naive para lle lism s.Superfic ial analogies, p a rtic u la r ly with human experiences, were confused with generalizations and taken to be explanatio ns . The search fo r genera lity was appeased by the pseudo explanation. 5
He adds that "a l i t e r a l in terp re ta tion of analogy thus supplies a
pseudo explanation, which by the u n c ritica l use of a p ictu re brings
many d iffe re n t phenomena together under one label
These are the dangers and the fru s tra tio n s of attempting to
adapt a theory from the re la tiv e ly precise s c ie n tif ic f ie ld of commu
nication theory and engineering fo r use in the social sciences—
p a rtic u la r ly in ternational re la tio n s . The adaptation must not be so
simple that i t is of no use fo r understanding and description; but i t
must not attempt to be so complicated and comprehensive that i t becomes
a pseudo explanation.
Even i f we decide to use the model in moderation—expecting
neither too l i t t l e nor too much— there is a danger, Kaplan says,
5. Hans Reichenbach, The Rise of S c ie n tific Philosophy (Berkeley; University of C a lifo rn ia Press, 1963), p. 8 .
6 . Ib id . , p. 13.
’’that the model lim its our awareness o f unexplored p o s s ib ilit ie s of
conceptualization. We tin ker with the model when we might be better
occupied with the subject matter i ts e lf ." ^ This danger bothered
Norbert Wiener too. He was firm ly convinced "th a t society /coul<l7
only be understood through a study of the messages and the communica-O
tion fa c i l i t ie s which belong to i t ; . . But he was reluctant to
place his blessing on social science attempts to usurp the communica
tion model— the study of messages— and he resisted , apparently, certa in
pleas addressed to him which had urged him to apply his ta lents to
social problems. He was hesitant to ta lk about cybernetic explanation
in society, noting that
to begin w ith , the main quantities a ffec tin g society are not only s ta t is t ic a l , but the runs of s ta tis t ic s on which they are based are excessively short. . . .For a good s ta t is t ic of society, we need long runs under essen tia lly constant cond itions . ju s t as fo r a good resolution of lig h t we need a lens with a large aperture. . . .Thus the human sciences are very poor testing-grounds fo r a new mathematical technique.9
Perhaps these words» from the master from whom he had learned the
language of the new science, had f in a l ly caught up with Deutsch; be
cause, in 1966, he indicates his repudiation of the s in g le , a l l -
encompassing point of view provided by the communication model and
writes what amounts to an epitaph on the passing of the period during
which i t dominated his thinking:
7. Kaplan, The Conduct of In qu iry , p. 279e
8 . Wiener, Human Use, p. 25.
9 . Wiener, Cybernetics. p. 34.
8 9
Cybernetic models as well as information and communication theory can be applied in p rin c ip le very w idely, but they cannot cover comprehensively a l l o f our knowledge o f socie ty , There are important processes in social development which are not p rim arily communication processes or control processes and where, therefore , cybernetic models are not h e lp fu l,10
Yet, even though the communication model has fa lle n out of
fashion—whether because of the d i f f ic u lty of applying i t or because
of the fu s illa d e of “slings and arrows of outrageous p ro b ab ility"— i t
s t i l l contains some useful analogies o f processes in the in ternational
system which have not been explo ited.
This thesis has investigated communication theory through the
meanings of its basic concepts. In add ition , fo r c la r i t y , i t has pro
vided the reader with the visual and conceptual aid of illu s tra tio n s
u t i l iz in g communication engineering apparatus. That these conceptuali
zations and common-sense understandings of communication processes have
so thoroughly permeated our everyday language is abundantly and r ic h ly
obvious. I f i t were not so, i t would have been exceedingly hard to
describe the concepts a t a l l . Therefore, the systems perspective and
the language o f communication and cybernetic theory have become part
of our vocabulary and our way of th inking . To have had such a great
general impact is no mean accomplishment fo r a theory. Whether the
technical aspects of its adaptation to the study o f in ternational re la
tions (or any other d isc ip lin e ) are ever worked out in fu l l d e ta il , i t
10. Karl W. Deutsch, "The Theoretical Basis of Data Programs," Richard L. M e rr itt and Stein Rokkan, e d s ., Comparing Nations (New Haven, Conn.: Yale U niversity Press, 1966), p. 31.
can be unequivocally stated that communication theory has enjoyed
a wide and extended application in the social sciences.
REFERENCES CITED
B e llp Do A0 Information Theory and its Engineering A pplications*3rd ed itio n . London: S ir Isaac Pitman and Sons L td ., 19&2.
Berelson, Bernard, “The State of Communication Research,1* Public Opinion Q uarterly , vo l. 23, no. 1 (Spring 1959), pp. 1-5.
B rillo u in , Leon. Science and Information Theory. New York: AcademicPress, In c ., 1956.
Davison, W. P h illip s . International P o lit ic a l Communication.New York: Frederick A. Praeger, 1965.
, and George, Alexander C. "An Outline fo r the Study of International P o lit ic a l Communication." Public Opinion Q uarterly ,Special Issue on Communication Research (Winter 1952-1953), pp. 501-511.
Deutsch, Karl W. "On Communication Models in the Social Sciences." Public Opinion Q uarterly , v o l. 15, no. 3 (F a ll 1952), pp. 356-380.
___ , P o lit ic a l Community a t the In ternational Level: Problems ofD efin ition and Measurement. Garden C ity , New York: Doubledayand Co., 1954.
"Shifts in the Balance of Communication Flows: A Problem ofMeasurement in In ternational R elations." Public Opinion Q uarterly , v o l. 20, no. 1 (Spring 1956), pp. 143-160.
, "The Theoretical Basis of Data Programs." M e rr it t , Richard L.and Rokkan, S te in , eds. Comparing Nations: The Use of Quantitativ e Data in Cross-National Research. New Haven, Conn.: YaleUniversity Press, 1966.
Deutsch, Karl W ., B u rre ll, Sidney A ., Kann, Robert A ., Lee,Maurice J r . , Lichterman, M artin , Lindgren, Raymond E .,Loewenheim, Francis L . , Van Wagenen, Richard W. P o lit ic a l Community and the North A tla n tic Area: In ternational Organization in the Light of H is torica l Experience. Princeton, New Jersey: Princeton U niversity Press, 1957.
Goldman, Stanford. Information Theory. New York: P re n tic e -H a lj, In c .,1953.
91
92
Hershberger, W. D. Principles o f Communication Systemso Englewood C lif fs , New Jersey: Prentice-H al1, In c ., 1955®
Kaplan, Abraham. The Conduct of Inquiry: Methodology fo r BehavioralScience. San Francisco: Chandler Publishing Co., 1964.
Kh inch in , A. I . Mathematical Foundations of Information Theory. New York: Dover Publications, In c ., 1957®
McClelland, Charles A. "Access to B erlin : The Quantity and Varietyof Events, 1948-1963®" Singer, J . David, ed. Q uantitative In ternational P o lit ic s : Insights and Evidence. New York: FreePress, 1967, PP® 159-186.
McMi1lan, Brockway. "The Basic Theorems of Information Theory."Annals of Mathematical S ta t is t ic s , v o l. 24, no. 2 (June 1953), pp. 196-219®
Pask, Gordon. An Approach to Cybernetics, Science Today Series.New York: Harper and Brothers, 1961.
P ierce, J. R. Symbols. Signals and Noise: The Nature and Process ofCommunication. New York: Harper and Brothers, I 96I .
Porter, Arthur. Cybernetics S im plified . London: English UniversityPress, L td ., 1969®
Rapoport, Anato!. "Lewis F. Richardson's Mathematical Theory of War." Journal of C onflic t Resolution, v o l. 1, no. 3 (September 1957), pp. 249-299®
Reichenbach, Hans. The Rise of S c ie n tific Philosophy. Berkeley: U niversity of C a lifo rn ia Press, I 963.
Richardson, Lewis F. Arms and Insecurity: A Mathematical Study ofthe Causes and Origins of War. P ittsburg: Boxwood Press, i 960.
Rothstein, Jerome. Communication. Organization and Science. Indian H il ls , Colorado: Falcon's Wing Press, 1958.
Russett, Bruce M. Community and Contention: B rita in and America inthe Twentieth Century. Cambridge, Mass.: Massachusetts In s t itu te o f Technology Press, 1963®
Schwartz, Leonard S. Principles of Coding. F ilte r in g , and Information Theory. Baltim ore, Md.: Spartan Books, In c ., 1963.
93
Shannon, Claude E. "The Mathematical Theory of Communication."Shannon, Claude E. and Weaver, Warren. The Mathematical Theory of Communication. Urbana: U niversity of I l l in o is Press, 1949.
, "Communication in the Presence of Noise." Proceedings of theIn s titu te of Radio Engineers, vo l. 37, no. 1 (January 1949),pp. 10-21.
Smith, Bruce Cannes. "Trends in Research on In ternational Communication and Opinion, 1945-55.11 Public Opinion Q u arterly , vo l. 20, no. 1 (Spring 1956), pp. 182-195.
Weaver, Warren. "Recent Contributions to the Mathematical Theory of Communication." Shannon, Claude E. and Weaver, Warren. The Mathematical Theory of Communication. Urbana: U niversity of111inois Press, 1949.
Wiener, Norbert. Cybernetics: Or Control and Communication in theAnimal and the Machine. New York: John Wiley and Sons, In c .,1948.
, The Extrapolation, In terp o la tio n , and Smoothing of StationaryTime Series: With Engineering A pplications. New York:John Wiley and Sons, In c ., 1949.
, The Human Use of Human Beings: Cybernetics and Society.New York: Avon Books, 1967.
Wolin, Sheldon S. "P o lit ic a l Theory: Trends and Goals." In ternational Encyclopedia of the Social Sciences, vo l. 12. New York: McMillan Co., 1968, pp. 318-330.