An Approach to a Science of Administration

Author(s): Edwin O. Stene

Source: The American Political Science Review , Dec., 1940, Vol. 34, No. 6 (Dec., 1940), pp.
1124-1137
Published by: American Political Science Association

Stable URL: https://www.jstor.org/stable/1948193

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Political Science Association is collaborating with JSTOR to digitize, preserve and
extend access to The American Political Science Review

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 PUBLIC ADMINISTRATION

An Approach to a Science of Administration. Writers on public ad-

ministration place much emphasis upon the possibilities and importance
of discovering and applying scientific principles in their field of study. But
very few of them have ventured to state the basic premises upon which
they seek to build that science. Many of those whose writings imply that
major principles have been discovered announce, not premises, but con-
clusions, which, regardless of their practical merits, can hardly be called
anything but opinions.1 On the other hand, several scholars seek to escape
from errors of commission by avoiding the use of such scientific terms as
"principles" or "efficiency." If they go beyond descriptive analyses to
advocate particular plans of organization or methods of procedure, they
use terms which denote value judgments, thereby admitting by implica-
tion that they are expressing mere opinions.2 In a few published discourses,
basic premises are stated and reasoning is developed therefrom. However,
most of those premises-or "principles"-are referred to by name only,
such as "the principle of leadership"; they are not stated in terms of pre-

1 For a distinction between "principles" and "opinions," see L. D. White, "The

Meaning of Principles of Public Administration," in White, Gaus, and Dimock,
Frontiers of Public Administration (1936), Chap. 2. For statements announcing im-
portant "principles" of public administration, see Report of the President's Com-
mittee on Administrative Management (1937), p. 3; Harvey Walker, Public Adminis-
tration (1937), pp. 61-62; A. E. Buck, Reorganization of State Administration in the
United States (1938), p. 14. Buck uses the term "standards," but his standards of
administration are apparently derived from what he and others regard as scientific
tests of experience. In a criticism of the "prevailing theories" advocated by Buck,
Willoughby, and others, Charles S. Hyneman cites several other writings which
support the view that those theories are mere opinions-or worse. Charles Hyneman,
"Administrative Reorganization," Journal of Politics, Vol. 1, pp. 62-75 (Feb.,
1939).
2 In his recently published Introduction to the Study of Public Administration
(1939), L. D. White carefully avoids the use of the term "principles" and of other
terms which might suggest that he is stating "scientific laws"; but he makes frequent
use of another type of positive statement, such as "good administration requires"
(p. 46), "it is also desirable" (p. 79), and "the interests of good administration
suggest" (p. 570). Of course the social scientist cannot disregard human concepts of
value, because the very purpose of social organizations is to satisfy human wants,
and no organization can function long if its activities produce general dissatisfaction.
But the student must seek to measure public value judgments objectively, by
taking his standards from official statements of policy and by observing human be-
havior which may serve as evidence of general and public concepts of value. The
"social efficiency" to which Millspaugh and Hyneman refer (see Hyneman, op. cit.,
pp. 66-67), like Dr. White's "good," may be used scientifically if it can be observed
and measured objectively. But it is doubtful whether political scientists are fully
agreed on the values which they set forth, and therefore the conclusions drawn from
such general concepts of value must be classed as opinions.

1124

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 PUBLIC ADMINISTRATION 1125

cise causal relations which can be verified or which can serve adequately
as bases for further reasoning
It may be regarded as unwise to venture a statement of what one con-
siders the basic premises upon which a science of administration may be
built. But every body of theory is built upon fundamental assumptions,
either expressed or implied. Moreover, a body of theory is complete, and
has scientific value, only when the premises are sufficiently clear to permit
objective scrutiny and verification. Erroneous hypotheses, stated pre-
cisely, may be more scientific than vague or unexpressed assumptions;
for only the former will lend themselves to verification. In other words,
trial and error is an essential part of scientific method.
What, then, are the first steps in an approach to a science of administra-
tion? Dr. Luther Gulick recently set forth what he considered the require-
ments of a scientific approach to the social studies at the present time.
These requirements were: "(1) analysis of phenomena from which we may
derive standard nomenclature, measurable elements, and rational con-
cepts; (2) the development of extensive scientific documentation based
upon these analyses, and (3) the encouragement of imaginative approach
to social phenomena, and the publication and circulation of hypotheses
so that they may be scrutinized by others in the light of experience, now
and in future years."4
A number of essentials are suggested by the foregoing statements, but
there is involved a clear recognition of two basic methods of approach
to scientific study. Statement number "2" calls for the empirical method,
by which we collect and describe appearances in order that general con-
clusions or laws may be derived therefrom. Statement "3" suggests the
rational or deductive method, whereby we seek to determine causal rela-
tionships among the appearances which we observe. This latter method
involves the adoption of definitions, axioms, and propositions, which are
derived from empirical observations, and which are capable of serving as
premises to reasoning. Conclusions reached by the rational method must
be consistent with the appearances gathered; but, being consistent, those
conclusions, and the premises from which they are derived, may serve
as bases for other conclusions beyond the immediate reaches of em-
pirical observation. The two methods are not independent and alternative
processes of scientific study; they are mutually dependent phases of in-
vestigation, and one cannot be separated from the other without impairin

I Papers on the Science of Administration (1937), edited by Luther Gulick and

L. Urwick, perhaps contains the best collection of articles representing an attempt
to develop a rational scientific approach to the study. But one searches almost in
vain for the statement of a "principle" in precise terms or in terms of causal relation-
ship.
4 Gulick and Urwick (eds.), Papers on the Science of Administration, p. 194.

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 1126 THE AMERICAN POLITICAL SCIENCE REVIEW

the value of both. However, it is possible for individuals, according to

their own temperaments or according to the facilities which they have
available, to specialize in one or the other of these phases.
There may be some dispute as to whether the rational method can be
depended upon until the empirical studies are well advanced. Political
scientists who give advice regarding fields and methods of possible re-
search seem to emphasize the need for empirical study.5 There is danger,
however, that the empirical studies will be lacking in direction or meaning
until they are capable of being interpreted in the light of propositions
brought forth by the rational or theoretical approach. Pithecanthropus was
not discovered until after Darwin had expounded his theory of evolution,
and the discovery probably would have had little significance prior to that
time. Principles of economics which were originally derived from rela-
tively superficial observations have served as guides to extensive empirical
*tudies in recent years, but thus far the major conclusions derived from
the rational analyses have been changed very little.
Without disparaging the importance of empirical research, therefore,
one may be justified in taking the view that the early development of a
rational theory is indispensable to the advancement of scientific method
in the study of administration. Since the development of a rational theory
is the concern of this article, the problem of selecting basic definitions,
axioms, and propositions becomes fundamental. For that reason, the
premises which are stated in the following pages are classified as "axioms"
and "propositions," although they would more properly be termed
"hypotheses" which are set forth in order that they may be tested and
scrutinized in the light of experience.

Administration is a type of cooperative activity; it is concerned with

the performance of tasks that require collective efforts on the part of
two or more persons. It follows, therefore, that the first term which re-
quires definition is organization.

6 Cf. Oliver P. Field, Research in Administrative Law (1937); L. D. White, Re-

search in Public Personnel Administration (1939). All of the publications of the
Committee on Public Administration, Social Science Research Council, are reports
of empirical studies. Perhaps the emphasis is justified by the definition of "research";
but see W. F. Willoughby, "A Program for Research in Political Science," in this
REVIEW, Vol. 27, pp. 1-23 (1933).
6 Most textbooks, and many other treatises on the subject, start by defining
administration. However, the writer has found no definition which presents a clear
demarcation between administration and other types of cooperative activity, and
he doubts whether the concept will lend itself to precise definition. Since adminis-
tration is a type of organization, the elements of the latter are also elements of the
former.

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 PUBLIC ADMINISTRATION 1127

"Organization" is defined in the Standard Dictionary as "a number of

individuals systematically united for some end or work." James Mooney
says that it "means concerted human effort" ;7 while Chester I. Barnard
defines "formal organization" as a "system of consciously coordinated
activities or forces of two or more persons."8 From these definitions and
others we may conclude that the primary elements of every social organi-
zation are (1) persons, (2) combined efforts, and (3) a common purpose,
or a common task to be performed. Therefore, by using the term "formal
organization," in order to exclude subconscious or accidental cooperation,
the following definition may be set forth:

Definition: A formal organization is a number of persons who system-

atically and consciously combine their individual efforts for the ac-
complishment of a common task.9

It follows from the above definition that the effectiveness of an organi-

zation is measured by the extent to which it carries out its purpose. To be
sure, that purpose is not always clear in the minds of contributing mem-
bers, and individuals frequently cooperate for different and personal
reasons. In the case of an administrative organization, however, objective
measurement of effectiveness can be made only in terms of the legal or
official statement of ends to be sought.
But in order to carry out its prescribed tasks over a long period, an
organization must survive; and to survive, it must avoid undue and un-
necessary consumption of human energy and of other human sacrifices.
Every individual belongs to several organizations, each of which requires
effort or some other sacrifice on his part. To some extent, these organiza-
tions must compete for his contributions. We may assume, therefore, that
the less human effort and sacrifice required by an organization to perform
a given task under given circumstances, the greater the assurance that the
organization will survive and continue to fulfill its purposes."0

7Gulick and Urwick (eds.), Papers on the Science of Administration, pp. 91-92.
8 Chester I. Barnard, Functions of the Executive (Harvard Univ. Press, 1938),
p. 73. The presence or absence of a joint purpose is the principal basis on which
Barnard distinguishes between formal and informal organization. Op. cit., Chap. 9.
Some authorities define organization as the arrangement of personnel for the
accomplishment of a purpose. Cf. Gaus, White, and Dimock, The Frontiers of Public
Administration (1936), pp. 66-67, and L. D. White, Introduction to the Study of
Administration (1939), p. 37. While the arrangement of personnel is essential, and
while that is the center of attention in organization charts and plans, the definition
is suggestive of preparation rather than of action. No organization actually exists
except in so far as there is concerted activity.
10 This is apparently what Mr. Gulick has in mind when he says that "in the
science of administration, whether public or private, the basic 'good' is efficiency."
Gulick and Urwick, op. cit., p. 192. The idea is basic in all discourses on public
administration-even among those who object to the emphasis placed upon "effi-

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
1128 THE AMERICAN POLITICAL SCIENCE REVIEW

The degree to which an organization accomplishes its task depends upon

three primary factors; namely, (1) the accomplishments of individual
members, (2) the number of members, and (3) the coordination of in-
dividual efforts. The first two of these factors give potential power or
effectiveness to the organization because they represent the energy con-
sumed by it. The third, with which this article is particularly concerned,
is the factor which determines the efficiency of the organization.'
Coordination involves several essentials, among which are (1) the at-
tachment of individual accomplishments to the common task, (2) the
avoidance of individual activities which interfere with or nullify one
another, (3) the performance of all activities necessary to make any given
individual accomplishment contribute to the common purpose, and (4)
the timing and placement of individual activities in such relationship to
each other that the common purpose will be accomplished.'2
James Mooney lists coordination as "the first principle of organiza-
tion."'3 The propriety of that designation becomes obvious when one
realizes that all other phenomena which serve to promote the attainment
of organizational objectives operate through coordination. Therefore, the
principle of coordination, which I state here in terms of causal relation-
ship, may be classed as the first axiom in a rational theory of administra-
tion:

Axiom I: The degree to which any given organization approaches the

full realization, of its objectives tends to vary14 directly with the co-
ordination of individual efforts within that organization.

II

Having stated the basic axiom, one faces the necessity of explaining the
phenomena which serve to promote or to impair coordination. An exami-

ciency in operation." Difficulties and disagreements arise in the attempt to deter-

mine the relative importance of different kinds of human sacrifice.
11 These factors cannot be isolated. As the later discussion will indicate, coordi-
nation is largely a consequence of individual efficiency, and individual accomplish-
ment is also dependent upon coordination. The factors may be considered separately,
however, in the sense that price, demand, and supply are considered separately by
the economists.
12 The result of coordination is efficiency of the organization. The result of man-
power, individual accomplishment, and coordination is effectiveness of the organ-
ization.
13 Mooney defines "principle" as a characteristic which is universal. By his
definition, coordination is a principle of organization because it is characteristic of
all organizations. Having adopted that definition, he is content merely to name
"principles" without explaining them in terms of causal relationship. The present
article is based on the proposition that coordination, while essential to all organiza-
tion, is a variable which tends to produce like variations in the effectiveness of an
organization.

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 PUBLIC ADMINISTRATION 1129

nation of published materials in the field leaves one with the impression
that most students of administration agree with Mooney, who places "the
principle of leadership and command" next in importance to his principle
of coordination. Mooney illustrates the principle by pointing out that if
two men combine their efforts to lift and move a heavy stone, one must
give a signal, such as "heave ho," in order that the two men can lift in
unison.'5 But the illustration involves the start of a new operation rather
than the continuous functioning of an organization, and therefore it seems
that Mooney has overlooked another vital factor in administrative co-
ordination.
If, instead of the two men moving a stone, one observes an amateur
tennis or volley-ball team at play, or even a store or office force at work,
one may be impressed by the fact that either work or play is coordinated
for long periods of time without any evidence of leadership. While leader-
ship is essential and important, it does not operate continuously.
But coordination is continuous in any but the most casual organizations.
This continuity of cooperative interaction is possible only when each mem-
ber of an organization has learned his task and certain vital relationships
between his own activity and the activities of those about him. In other
words, a routine of interactions has been established; the relative positions
of workers, the individual tasks, and the sequence of operations are known
because of a series of past decisions and experiences. The parts of the
organization are coordinated much the same as the muscular movements
in one's body are coordinated when one walks. No continuous conscious
effort is necessary; the organization has acquired a habit of coordination.
Definition: Organization routine is that part of any organization's
activities which has become habitual because of repetition and which is
followed regularly without specific directions or detailed supervision by
any member of the organization.

Proposition I: Co6rdination of activities within an organization tends

to vary directly with the degree to which essential and recurring func-
tions have become part of the organization routine.'

14 In all axioms and propositions, the phrase "tends to vary" is used-instead

of the unrealistic "other things being equal"-to allow for factors that would
produce contrary tendencies.
1I Gulick and Urwick, op. cit., p. 91.
16 To many persons the term "routine" implies the absence of individual initia-
tive and imagination. However, an analogy may serve to illustrate the difference.
A violinist who devotes his entire attention to matters of technical proficiency (the
routine of playing his instrument) never becomes a great artist; but neither is it
possible for the musician to give his attention to the creative phases of his art until
he has mastered the technical problems so thoroughly that he need not think about
the details of his finger or arm movements. The routine must be mastered before
the violinist becomes a musician. In an organization, the conscious attention is the

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
1130 THE AMERICAN POLITICAL SCIENCE REVIEW

The word "essential" is included in Proposition I because routinized

activities do not always promote coordination. Every individual who has
observed his own behavior knows that while habit makes coordination
possible, it may also serve to interfere with coordination. For some un-
known reason, perhaps an unexpected situation, an athlete may shift
certain muscles unconsciously. The movement becomes habitual and the
athlete finds himself "in a slump" until more effective coordination can be
restored. In like manner, an organization may adjust itself to certain ir-
regularities, with the result that a new sequence of interaction becomes
routine.'7 And the new habits of an organization, like those of an indi-
vidual, may be "good," or "bad," i.e., they may serve to promote, or they
may impair, the accomplishment of an organization's objectives.

ItI

Organized routine has another important limitation, namely, that it is

restricted to regular and recurring activities. The individual knows that
situations arise in which he cannot depend upon habit. He must exert
conscious mental effort to coordinate his actions for a desired purpose.
Again, the analogy is to be found in the organization. Action of a non-
routine character, that is, action in which coordination has not been se-
cured through repetition, must be preceded by conscious consideration
and decision. Decision is necessary whenever an organization is formed,
whenever routine interactivity is deliberately changed, and whenever
action is called for which has not become routine in character. While these
cases occur only at intervals, the existence of methods of making decisions
for an organization is essential to coordination of efforts.

Definition: Decision is the conscious consideration and conclusion re-

garding a course of action. The decision of an individual may be known
only to himself; but the decision of an organization must be made
known to more than one person and must be made by a person or per-
sons who are recognized as having authority to speak for the organiza-
tion.

When action is coordinated by means of decision, it must be because

the decision is consistent within itself, with existing routine (except

attention of leaders, and the routine has been established when the organization
carries on without conscious attention of leaders. The "dead level of mediocrity and
routine" of which P. W. Melton speaks ("Administration in a Federal Government
Bureau," in this REvIEW, Vol. 33, pp. 835, 840) apparently arises because the
executive devotes his attention to matters that should be routine and habitual;
not because recurring functions have become routine.
17 L. Urwick apparently has somewhat the same idea in mind when he points
out that "in organization there is a tendency to compensation, comparable to the
tendency found in the human organism." Gulick and Urwick, op. cit., p. 84.

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 PUBLIC ADMINISTRATION 1131

where changes are intended), and with other decisions affecting the same
set of interrelations. Coordination in action is a function of coordination
in decision.

Axiom II: Co6rdination of organized activities tends to vary directly

with the coordination of decisions by which the course of activities has
been determined.

Here lies the virtue of leadership. The problem of coordinating activi-

ties becomes more difficult as an organization becomes larger and more
complex. But, on the other hand, the number of persons who participate
in making decisions can be much smaller than the number required to
carry out those decisions. In Mooney's illustration referred to above, two
men were required to lift and move the stone, but one could decide when
to lift. In fact, when a simple decision is called for, such as fixing the exact
time to act, one may decide for thousands.
The essence of leadership, as illustrated by Mooney, is that action is
coordinated by means of concentrating the function of decision in one or
a few persons.'8 Because performance may be coordinated through de-
cision, this concentration of the authority to decide simplifies the problem
of coordinating performance as well as decision. Thus we have a rational
basis for the oft-repeated "principle of unity of command.'9

Definition: Leadership, as an element of organization, is a system

of coordinating action by concentrating in a small number of persons
the function of making decisions for the organization.

As will be pointed out later, other factors have vital influence upon the
coordination of decisions, as well as upon the rate at which coordinated
performance follows as the result of coordinated decision. However, these
factors do not invalidate Axiom II or the following axiom:

Axiom III: In any particular type of organized activity, coordi-

nation tends to vary inversely with the number of persons directly
participating.20

18 The term "leadership" is used here in a limited and non-personal sense. Types
of leadership or personal characteristics of leaders are not considered in this article.
19 Gulick and Urwick, op. cit., p. 9. Marshall Dimock, Modern Politics and Ad-
ministration (1937), p. 270. For views approving the idea, see also L. D. White,
Introduction to the Study of Public Administration (1939), p. 51; A. E. Buck, The
Reorganization of State Governments in the United States (1938), pp. 17-19.
20 The reader should recall that while coordination promotes efficiency, it is
only one of the factors which determine the effectiveness of an organization. (Cf.
note 12, supra.) This axiom does not lead to the conclusion that several small and
independent organizations are more efficient than one large organization. If the
same task is to be performed, integration does not increase the number of persons
involved.

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 1132 THE AMERICAN POLITICAL SCIENCE REVIEW

Axiom III is of particular importance with respect to leadership, since

the acts of decision lend themselves most easily to concentration. With
this kind of application in mind, we may combine Axiom III with Axiom
II and derive the following proposition:

Proposition II: Co6rdination of organized activity tends to vary

directly with the degree to which the function of making decisions for
the organization is concentrated.

Particular caution is required to avoid possible misinterpretation of

Proposition II. In the first place, we must remember that to concentrate
the power to decide is not necessarily the same as to concentrate the func-
tion. Over-concentration of the power may result in a dispersing of the
function. Moreover, the lack of communication and knowledge, which
factors are discussed below, may serve to nullify an attempt to promote
coordination by concentrating the function of deciding.

IV

In considering the relationship between decision and performance, we

must not overlook the important element of communication. Obviously,
no decision can affect action if that decision is not communicated to the
points at which the relevant action takes place. In like manner, decisions
cannot be coordinated with the routine activity unless facilities are pro-
vided to communicate appeals for decision from the points where the
need has been observed.
The importance of communication is a factor to which many advocates
of integration and "unity of command" seem to have given inadequate
attention. Too often their arguments suggest that if the authority to make
decisions is properly concentrated, communication will take care of itself.
But anyone who carefully observes an administrative agency in operation,
even for a short period, is likely to be impressed by the difficulties of com-
munication. The individual worker frequently does not know what ques-
tions he should decide for himself, or to whom he should appeal for de-
cisions when an appeal is called for. Many times he finds that the person
to whom he goes is too busy to help him, and time is lost while he awaits a
decision. Such cases become truly serious when intermediate supervisers
do not dare to make decisions, or when higher officials insist upon personal
control of more decisions than they have time to handle expeditiously. Add
to this the cases where lines are crossed and subordinate officers go over
the heads of their immediate superiors, and possibly outside of the official
organization to political party headquarters, for decisions, and we have a
picture of the disorganization in many of our governmental agencies.
The reader may have noted that the "lines of communication" ems
phasized here appear to be synonymous with what is generally known as

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 PUBLIC ADMINISTRATION 1133

."lines of authority." In fact, the form

in operation; they function only when communication can be maintained
between points of decision and points of action. Legal, or "official," au-
thority is worthless, or worse than worthless, when the person who pos-
sesses authority does not have, or does not see fit to use, facilities neces-
sary to make and communicate to points of action the decisions which he
is required to make, or which he insists upon making himself.
The existence of lines of communication between points of decision and
points of action depends upon two important elements, namely, (a) free-
dom from interference by other lines of communication which may
monopolize the time of persons in authority, and (b) recognition of the
existence of authority to make decisions.
(a) The first element may be described as the readiness to communicate
at the upper level, or point of decision. Mention has already been made
of the fact that decision cannot serve to coordinate action if the person
in authority cannot be reached for a decision when called upon. A con-
gested waiting room often tells only part of the story about the difficulties
of reaching the administrator in the inner office; subordinates who are
conscious of the situation may postpone vital action for days because the
superior is "obviously too busy." Effective coordination is dependent
upon the ability of subordinates to appeal to the points of decision when
necessary, and to get decisions which are communicated to points of action
without undue delay.21
(b) The second element mentioned above, namely, recognition of the
existence of authority, may be described as the readiness to communicate
at the lower level-the points of action or of subordinate decision. An of-
ficer cannot make decisions which serve to coordinate action if his order,
or communication, is not recognized by subordinates. Such a situation
often arises when no definite lines of authority have been agreed upon,
when a superior officer makes a practice of giving specific orders to
workers without considering the importance of communicating through
intermediate authorities, and when a person at an intermediate level is
known not to be in good standing with his superiors.22

21 Not every appeal requires immediate decision; but an administrator must

have time to make decisions which are called for immediately, and to recognize the
order in which questions ought to be decided and the information to be obtained
and persons to be consulted before an order is issued.
22 Under a spoils system, it is often necessary to retain an intermediate officer
for some time after a change of party control, because his technical knowledge is
indispensable pending the time when the new employees learn their tasks. When
such an officer is flanked above and below by members of the new party, he usuall
exercises no authority; and the practice of passing over his office in departmental
communication may become so habitual that his successor also exercises very little
influence. The practice of jumping lines of authority (breaking the official channels

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
1134 THE AMERICAN POLITICAL SCIENCE REVIEW

To a large extent, the maintenance of efficient channels of communica-

tion depends upon the establishment of a routine. If the worker knows
whom he should consult when in doubt, if the subordinate officer knows
what kinds of questions he may decide and what kinds he should pass on
to his superior, and if the superior likewise knows when he should not
interfere with his subordinates, the system of communication usually
operates efficiently. On the other hand, when the superior follows no con-
sistent plan, or when he refuses to delegate authority to decide detailed
or technical matters, the entire system of communication may be con-
fused and inefficient.
Space does not allow further consideration in this article of the nature
of administrative communication or the conclusions that may be drawn
with respect thereto. If the following proposition is accepted as valid,
further conclusions can be drawn by others who may continue the study,
and systems of organization may be tested by those conclusions. Suffice it
to state here one apparent conclusion, which is that coordinated adminis-
trative performance is most probable when the point of decision is near
to, or within easy reach for communication from, the points of action-
provided, of course, that the former is sufficiently high in the hierarchy
of organization to avail itself of communication lines to all points of
action to be governed directly thereby.23
Proposition III: Co6rdination of performance tends to vary directly
with coordination of decision, but the changes in the former follow
changes in the latter at a rate which varies directly with the efficiency
of the lines of communication between points of decision and points
of action.
V

Propositions II and III and Axioms II and III have been set forth as
possible principles of authority, or, in the highly descriptive terminology
used by Mooney, principles of "perpendicular coordination."24 But
Mooney points out that the "principle of horizontal coordination" is
equally important. He explains that this principle operates through the
"universal service of knowledge" rather than through authority and
command.
Basically, "horizontal coordination" is little more than another phase
of the "system of communication" already referred to. It takes on special

of communication) is especially common when a busy director functions under a

not-so-busy permanent board or commission.
23 Further clarification of the nature and causes of red tape, the merits of var
systems of field organization, and the merits and demerits of the board type of ad-
ministrative control are only a few of the problems that might be analyzed on the
basis of the proposition drawn from this discussion of "lines of communication."
24 Gulick and Urwick, op. cit., pp. 93-94.

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 PUBLIC ADMINISTRATION 1135

characteristics because it operates principally before rather than after

decision, and because it brings to attention another factor in coordination,
i.e., knowledge.
Proposition I indicates that organization routine is a major and con-
tinuous factor in coordinating organized activity. Since the process of
decision operates intermittently, it can ordinarily promote coordination of
performance only when the decisions are adapted to, or coordinated with,
the existing routine. In order that a decision may be adapted to the exist-
ing routine, it is necessary that it be based upon knowledge about that
routine. An officer who is far removed from the points of action can make
decisions consistent with existing routines only by communicating with
operating agencies, or with other agencies which have, or can obtain, the
essential information.
Another illustration of the importance of a service of knowledge arises
from the fact that administrative decisions affect many activities which
are far beyond the recognized sphere of authority of the officers who make
the decisions. For example, a state director of revenue may make a de-
cision designed to insure more effective collection of sales taxes. In so far
as this decision affects the administration of income taxes or other special
taxes, the director may issue orders to promote coordination. But the re-
sulting activity of governmental agents and employees may have to be
coordinated with the methods by which business is carried on in the state,
or with the constitutional principles of our federal system of government,
either of which is beyond the control of the director of revenue. Knowledge
of certain business methods and of constitutional principles, therefore,
becomes fundamental to securing administrative coordination.
A single decision may serve to govern activities within several depart-
ments of an administrative system, and, at times, even to govern action
within several units of government. But no decision or set of decisions can
cover all governmental activities-not to mention related economic and
social activities. Furthermore, if decisions affecting only the internal
operations of administration were made at an administrative level high
enough to cover all cases related thereto, the probabilities are that lines of
communication would be clogged at the upper levels, and that co6rdina-
tion would be impaired rather than improved by the attempt to co6rdi-
nate through control instead of through adaptation.
Because coordination through control is possible only within narrow
limits, it follows that almost every decision is made with the idea of
changing action within definite and narrow limits and of coordinating that
action with the related activities not governed by the decision. The
limited sphere of the decision may be called the "governed" or "strategic"
factor, whose control, within the sphere of coordination with ungoverned
factors, may produce new action designed to carry out the purpose of the

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 1136 THE AMERICAN POLITICAL SCIENCE REVIEW

decision.25 An objective of this character necessarily r

cision be based upon knowledge about the related factor
is no assurance that the decision will promote coordination. We may con-
clude, therefore, that a decision promotes coordination between "gov-
erned" activity and "ungoverned" activity only to the extent that the
decision is based upon knowledge about the related, "ungoverned" ac-
tivity. In the same manner, coordination between activity which results
from a decision and other activities within the same administrative agency
is assured only if the decision is based upon knowledge about existing
routine. This idea may be expressed in another way by saying that effec-
tive coordination requires that, over a limited period at least, decisions
shall be governed by action to a greater degree than action is governed by
decisions. For only the "strategic factors" can be governed by decisions,
while the latter, in turn, must be guided by all related activities if co6rdi-
nation is to be promoted.
The causal relationship between knowledge and coordination is ex-
pressed in the following proposition:

Proposition IV: Co6rdination between activity resulting from a set

of decisions and related activities not governed directly by those de-
cisions tends to vary directly with the extent to which the decisions are
based upon knowledge about the related activities.
But the persons who make decisions cannot collect and analyze essential
information without impairing lines of communication between points of
decision and points of action. Furthermore, the less important decisions
must be made at points where the records at hand do not provide adequate
information about other administrative units. Therefore it becomes neces-
sary to transmit information from one agency to another through channels
of communication which will not interfere unduly with the channels used
for purposes of communicating decisions. And it becomes necessary, also,
that in any large administrative organization certain agencies be main-
tained for the sole purpose of collecting, analyzing. and transmitting to
points of decision such information as may be useful in coordinating
strategic factors of organized activity with related activities not governed
directly by administrative decisions. Such information can be interpreted
and transmitted to points of decision when called for, when it appears
that the information itself may lead to decisions, or when the knowledge
may influence decisions that are made immediately upon appeal to the
respective positions of authority in an organization.
Again it becomes important that facilities be made available for an
effective system of communication. The following proposition expresses

25 Cf. Chester I. Barnard, Functions of the Executive, Chap. 14, for further defini-
tion and illustrations of the "strategic factor."

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms
 PUBLIC ADMINISTRATION 1137

a relationship between knowledge and coordination, similar to that ex-

pressed in Proposition III with respect to leadership:
Proposition V: The extent to which decisions are based upon knowl-
edge about related activities tends to vary directly with the efficiency of
lines of communication between points of decision and points at which
facts are collected and analyzed.
Propositions relating to "horizontal coordination" and the importance
of knowledge as a basis of decision emphasize a phase of administration
which deserves particular attention. Information is basic to any improve-
ment in organization routine; it is essential in order to adapt decisions to
existing routines; and it is the key to development of an efficient system
of communication within an organization. The principles involved suggest
the importance of certain democratic processes within administration,
such as consultation with the minor officials in charge of activities which
will be affected by a contemplated decision, and the use of conferences
with interested parties outside of the administration when decisions may
affect personal rights of citizens.
But it is not the purpose of this article to apply the hypotheses which
have been set forth. A few "axioms" and "propositions" have been pre-
sented in the hope that they will be scrutinized and tested in the light of
experience, that those parts which are found good may be retained while
other parts may be modified or discarded, and that many other hypotheses
may be added thereto. Finally, the writer hopes that, by extension and
improvement of the above method of approach, a body of theory may be
found adequate to support and guide the empirical studies necessary to
the development of what may be called, without apologies, a science of
administration.
EDWIN 0. STENE.
University of Kansas.

The Lack of a Budgetary Theory. On the most significant aspect of

public budgeting, i.e., the allocation of expenditures among different pur-
poses so as to achieve the greatest return, American budgetary literature
is singularly arid. Toilers in the budgetary field have busied themselves
primarily with the organization and procedure for budget preparation, the
forms for the submission of requests for funds, the form of the budget
document itself, and like questions.' That these things have deserved the
consideration given them cannot be denied when the unbelievable resist-

I See A. E. Buck, Public Budgeting (New York, 1929); J. Wilner Sundelson,

Budgetary Methods in National and State Governments (New York State Tax Com-
mission, Special Report No. 14, 1938); ibid., "Budgetary Principles," Political
Science Quarterly, Vol. 50, pp. 236-263 (1935).

This content downloaded from

191.17.158.70 on Sat, 26 Jun 2021 15:07:40 UTC
All use subject to https://about.jstor.org/terms