As we shall see
later (Chapter 6) the distinction between
findings and laws is one of degree of generality and degree of
empirical support. In reviewing the Burleson Steiner thousand
plus propositions one finds that anywhere from five to fifty of
them are general enough to qualify as laws, depending on how
strict we make our criteria. At any rate, there is no doubt that
sociology now has a number of lawlike propositions that can be
called confirmed or trustworthy.
We have several
works in sociology that combine lawlike propositions into
systems, that is, theories. A good [p.15] example in
Hopkins' book The Exercise of Influence in Small Groups. 3 It summarizes a number of
research studies into fifteen propositions. Each is explicitly
stated and evaluated for its consistency with existing findings
and consistency with other accepted propositions. The
propositions are then ordered so that they each came to represent
a part of a process which brings into balance the ranks held and
the influences exercised by the members of any group. It is
demonstrated, both by theoretical deduction and empirical
investigations, how this process accounts for characteristics of
group leaders, opinion changes among members, and group stability
and instability. By all reasonable criteria, this is an
acceptable scientific theory.
The history of
sociology shows that we have been very eager to reach not only
such theories of modest scope but to reach for grand theories.
The last notable achievement here was made by Pitirim Sorokin.
His theory on social and cultural change 4 starts with a basic scale
of mentality, to classify the symbols that constitute social
reality. Symbols may be either 'sensate,' that is, they refer
directly to sense data, or they may be 'ideational,' that is,
their referents are many steps removed from sense data. Sorokin
reviews the flow of events in art, sci-[p.16]ence, law,
religion, and ethics from the beginnings of civilization to the
present and tabulates the symbol products of culture in terms of
the sensate or ideational mentality they exhibit. He establishes
two basic lawlike generalizations. First, he shows that in each
realm of society taken by itself, the mentality shows secular
swings from sensate to ideational and back. Second, he shows that
different realms tend to move together in their secular
fluctuations between sensate and ideational mentality. Since no
particular realm consistently seems to lead or initiate these
changes, Sorokin rejects the claims of writers who attempt to
locate the moving force of history in any particular realm of
society such as religion, politics, art, or economy. Instead, the
super rhythm, he speculates, has causes of its own operative in
any realm; adaptation and efficiency is hampered by an excessive
sensate mentality as well as by an excessive ideational
mentality, and the movement toward a consistent mentality in the
end always defeats itself.
Sorokin then
proceeds to explore, by theoretical deduction and by research,
the correlates of this super rhythm. His ambition is to show that
if we can locate the phase of a civilization on the master curve
of the super rhythm we should also be able to tell a good deal
about it: the content of its dramatic works, the topics of its
pictures, the organization of its government, the number of
inventions made, the likelihood of wars and revolutions, etc.
Most important among these correlates [p.17] are the
nature of relations between men and his fellowmen; whether they
are familistic, contractual, or compulsive. It is the large
number of correlates to sensate and ideational mentality that
makes Sorokin's theory the grandest that so far has appeared in
sociology. However, often it is not entirely clear how these
correlates are derived from the two basic generalizations, and
the empirical demonstrations of the correlatives are not always
as convincing as one might wish. Hence, the theory is still
controversial. Partly as a reaction against Sorokin's effort to
write grand theory, one of his students, Robert K. Merton,
formulated a strategy that has become widely accepted by
contemporary sociologists. Merton entered a plea for
"theories of the middle range." 5 These are miniature
theories, not grand theories; or, better expressed, partial
theories rather than inclusive theories. When we call a theory
partial (or middle range, or miniature) we admit that there are
other accepted theories which are not contradicted by, or
synonymous with, the one we call partial. Optics and
thermodynamics are examples of partial theories in physics. As
mentioned, in social psychology, anthropology, and sociology we
also have a few such theories: e.g., Festinger's theory of
cognitive dissonance, Homans' theory of elementary social
behavior, [p.18] Hopkins’ theory of influence,
Murdock's theory of kinship structures, Pareto's theory of
elites. Sociology is believed to advance most rapidly by
developing a large array of such partial theories.
At the horizon of
sociological thought looms, however, the challenging issue of
integrating these partial theories into a more inclusive whole.
The plea for theories of the middle range would be misguided if
it implied a condemnation of all efforts toward an inclusive
theory. If inclusiveness is considered a matter of degree, it
becomes again a manageable goal. (Inclusiveness is, indeed, a
matter of degree: one theory can be more inclusive than another
and yet be a partial theory in regard to a third theory). Steps
toward inclusiveness comprising integration of two or more
partial theories should be encouraged; herein lays one of the
greatest challenges of the theoretical undertaking.
Thus there are two
types of specialists in theoretical sociology: the man who
develops new partial theories out of his own or other people's
research, and the man who takes a number of partial theories
developed by others and integrates them into a more inclusive
theory.
In spite of the
fact that we have a few brilliant attempts toward grand theory
and several examples of partial theories and recognize that the
integration of the latter into a more inclusive theory is a
possibility, the dominant impression in looking at the sociology
of today is one of theoretical paucity. We have read, in [p.19]
college or privately, some more or less popular texts of
sociology. These texts contained chapter headings like "The
Family," "Social Class," "Public
Opinion," "Race Relations," etc. They dealt with
the rather interesting, but theoretically unconnected, topics
traditionally assigned to departments of sociology. They were
more like our childhood popular physics books than the systematic
physics texts of our later schooling. Present sociological
thinking has actually rather little to offer the student who
wants to go beyond this topical study to explain family
structure, social class, public opinion, race relations, and
other topics in terms of a few laws of sociology.
There is, of
course, an enormous amount of research done which gives
information about all these sociological topics. This flow of
research findings, in fact, has become so great that it is now a
losing game to try to keep abreast of all the findings.
Monographs, journal articles, research proposals, mimeographed
reports overtake man. To be a social theorist rather than a
social researcher is no refuge from the flow of this research.
The days are gone when "theory" and
"speculation" meant the same thing and the theorist did
not have to know anything except the location of the space bar on
his typewriter. Generally speaking, the modern theorist, as we
visualize him, has to know more empirical findings than the most
down to earth researcher, since he is, among other things,
concerned with the systematization [p.20] of the knowledge
researchers have acquired: one outcome of his labor is in the
form of documents that summarize the past discoveries and events
in lawlike statements. We note with gratification that we now
have obtained a few such research grounded theories in the field
of sociology. But we also note that most work in theoretical
sociology remains to be done.
The difficulty of
the sociologists' struggle with theory lies in part in the
dilemma of sociology being both a humanistic and scientific
discipline. We never escape the humanistic content of sociology
but we try to treat it scientifically. Of the elders in the
field, hardly anyone has devoted himself wholly to the task of
theoretical sociology. It is understandable that one of the first
sociological theoreticians in the modern sense, Vilfredo Pareto,
did not begin his work in theoretical sociology until he was over
fifty years old. 6 It is a sadder commentary
that, fifty years later, a representative theoretician, George C.
Homans, who received his training and has his career at one of
the world's best universities, confesses not only other pursuits
prior to his endeavor in theoretical sociology, but also to a
long process of unlearning of fettering traditional approaches to
social [p.21] thought. 7 Some fortunate members of
the generation now being trained in sociology will be the first
ever to orient themselves from the very start of their careers
toward actual theory construction.
Within the
humanistic tradition of sociology we found that "social
theory" could mean two related but different things:
classical works and criticism. Within the scientific tradition of
sociology, "social theory" also stands for two
different but related enterprises. One is represented by an
anthology such as Toward A General Theory of Action. 8 The task of the writers of
this book is to develop an orderly schema defining anything to
which sociologists (and other social scientists) should pay
attention. Names are assigned to these subjects, and the reader
is encouraged to go out and discover their concrete
manifestations in all parts of society. A more specific term than
"theory" for this system of definitions is
"sociological taxonomy." The anthology mentioned
contains mostly suggestions for a general taxonomy of the social
sciences.
[p.22] Concern
with sociological classics, sociological criticism, and
sociological taxonomy are all to the good. I, for one, enjoy
pursuing these interests in my teaching and writing. However, in
this book, I primarily want to pursue sociological theory in a
fourth sense: systematically organized, lawlike propositions
about society that can be supported by evidence: This is
"theory" in the sense this word is used in other
sciences. Taxonomy enters this enterprise only to the extent that
we sometimes need to define the terms used in our propositions.
Propositions are the central elements; definitions are auxiliary.
As a reminder that this is a different breed of animal I shall
speak of it as "theoretical sociology" rather than
"social theory." This usage is meaningless in itself,
but I believe it will serve a good purpose. Let us spell out in
some further detail how this propositional approach differs from
the taxonomical one by considering the dilemma that both want to
tackle.
An initial
difficulty for the sociological theorist is, as mentioned, the
great variety and complexity of phenomena with which his
discipline customarily deals. As we noted, the topics have a wide
range: family discord, social mobility, labor management
relations, propaganda, public opinion, crime, housing, rural
urban migration, race relations, and a series of technical
subjects related to the organizations and institutions of
government, industry, business, education, art, religion,
welfare, civic affairs, mass media, and others. It is easy to [p.23]
argue that no man knows enough or is wise enough to deal with
all these phenomena.
There have been
times when sociology was imperialistic enough to claim all
aspects of all societal phenomena as its proper realm. But the
expanding scientific knowledge about society can never be the
monopoly of any one academic discipline. It is a joint enterprise
of historians, economists, political scientists, demographers,
sociologists, anthropologists, geographers, and others. The
sociologists hold only a few of the pieces to the picture puzzle
of society. Specialization is necessary. Sociology, too, is a
specialized science.
These two
statements -- that sociology deals with just about every social
phenomenon, and that sociology is but a specialized one of the
many social sciences -- do not complement each other. How can
sociology deal with everything and yet be a specialized science
of the social world? Or how can sociology be a specialized social
science and yet deal with all societal phenomena? The diversity
of subject matter and the necessity for specialization pose a
dilemma. 9 In principle, the
resolution of the dilemma does not appear difficult. No science
seems to deal with all aspects of what common sense considers one
phenomenon. In a recent text in sociology, beginning students are
given a clear demonstration of this: [p.24]
Consider
your instructor's chair. If a specialist in the branch of
physics called mechanics were to study it, he would see it as
a combination of weights and balances; a biologist
specializing in anatomy would see it as a receptacle for the
human form and might assess its effect on the spinal column;
an economist might see it as a product of mass production, a
unit of cost and price; the psychologist might see it as a
part of the perceptual frame of the student; and the
sociologist might see in the chair a symbol of status. Like
any field of inquiry, sociology is selective in its approach. 10
Thus the
specialization of sociology lies in its concentration on certain
aspects of any social problem or any social institution, not in
an inclusive study of one or two institutions or social problems.
A first
resolution of our dilemma is to specify a small number of
definitions which delineate the few aspects of reality with which
sociology deals. These definitions, broadly speaking, tell the
sociologist what is important for him to pay attention to when he
views a human relationship, a group, or a society. The
geographer, armed with definitions such as "latitude"
and "longitude," looks upon a given area of the earth
in these terms, but can leave such problems as the age of the
crust of the earth in a given area to a geologist. Likewise, a
sociologist looking at a group in terms such as "rank"
and [p.25] "norms," which are among his key
definitions, can leave problems of the members' "personality
traits" to the psychologist, who has a series of definitions
to delineate them. Much work in sociology has concentrated on the
development of definitions of the descriptive categories that a
sociologist is to use.
This is what we
call taxonomy. The goal is an orderly schema for the
classification and description of anything social. Thus, when
faced with any subject of research, the sociologist can
immediately identify its crucial aspects or variables by using
his taxonomy as a kind of "shopping list." To
"test" his taxonomy, he takes a fresh look at subject X
and shows that the general terms defining his dimensions have
identifiable counterparts in X. For example, Parsons assigns
certain abstract attributes to a social system, and then turns to
economy, for instance, and finds that economic thinking takes
these dimensions into account. He concludes that the economy is a
social system. This is occasionally called to "derive"
X, or "explain" X -- speech habits which are
rather misleading; a better term is to "diagnose" X. To
make a sociological diagnosis of the subject matter or problem X
is to describe X in terms of a sociological taxonomy. For
example, when Parsons and Smelser suggest that the distinction
between short term supply and demand in the economy is a special
case of the distinction between performance and sanction in a [p.26]
social system 11 this is not a sociological
derivation or explanation of supply and demand; it is a
sociological diagnosis.
Taxonomies
summarize and inspire descriptive studies. Thus Parsons'
taxonomy guided Stouffer and Toby to a descriptive study which
presented the distribution of some college students on the
variable "particularism-universalism" defined by
Parsons. This variable is one among others in a set called
"the pattern variable schema" which has proven useful
in characterizing any social relation. 12 Since sociology, like
geology, botany, and geography, is largely a descriptive science,
the importance of sociological taxonomy must be taken for
granted. However, it should be emphasized again that a concern
with taxonomy and descriptive studies does not furnish any
explanations.
To know the labels
of phenomena and to know their distribution is not to explain
them. In the best case, these sets of definitions and maps of
distributions leave you where Linnaeus left biology in the
eighteenth century -- that is with denotations of species and
studies of their distribution. When Darwin formulated the
principles of the origin of any one species from others, he
pushed biological thinking toward something more wor-[p.27]thy
of being called a theory. He not only formulated definitions of
categories to investigate what cases fall into these categories;
he formulated propositions and started to verify them.
Sociological thinking, if it is to progress scientifically, is
also bound to add some propositions to the already long array of
definitions, and to let some of the effort now going into the
making of descriptive studies be allocated to the verification of
these propositions. In so doing, we should, of course, use as
many of the previously formulated definitions as we can. Darwin
was greatly aided by Linnaeus' definitions, and some -- but not
all -- of Linnaeus' definitions became definitions in Darwin's
theory.
A second
and related resolution to the dilemma between diversity of
subject matter and the need for specialization enters here. It is
represented by the program for sociology set forth by Georg
Simmel over half a century ago:
. . . we
shall discover the laws of social forms only by collecting
such societary phenomena of the most diverse contents, and by
ascertaining what is common to them in spite of their
diversity. 13
The assumption
here is that sociology will eventually discover a small number of
propositions that are valid in several diverse contexts. This
idea, that there are [p.28] sociological propositions that
hold in diverse contexts, is gradually becoming more of an
established fact and less of a wishful hope. In George Homans'
The Human Group we find a few hypotheses confirmed by such
diverse subject matter as an industrial work group, a Polynesian
kinship structure, a city street gang, and a small New England
community. This approach represents what we see as the main task
of the sociological theorist -- that is, the discovery of
general propositions.
The systematically
interrelated propositions that result from this effort are theories.
Only at this stage does it make sense to speak of "testing a
theory," "derivation," and -- most important of
all -- "explanation." To "test" a theory, we
check how well each of its propositions conforms to data and how
well several propositions in conjunction with each other account
for the outcome of a given situation. If such a
"derivation" (or prediction) is successful, we call the
outcome "explained"; that is, we claim that observed
events conform to known propositions. Thus, Homans is able to
explain the friendly feelings between brothers on the island of
Tikopia by a reference to his already established proposition
that a higher frequency of interaction results in a greater
mutual liking. 14
Theories summarize
and inspire, not descriptive studies, but verificational
studies -- studies construed to test [p.29] specific
hypotheses. The number of such studies has grown to a gratifying
extent in recent years, and every volume of the sociological
journals seems to have at least a few articles in which the
author formulates specific hypotheses and then reports data that
bear on them. If the 1gso's were particularly hospitable to
taxonomies and descriptive studies, the 1960's seems more
hospitable to theories and verificational studies. The growing
use of theory in applied sociology is helpful to this
development. 15
The following
listing of some key words may serve as a summary of the kinds of
activities we have discussed:
| |
I |
II |
| Unit |
Definition |
Proposition |
| Interrelated
units |
Taxonomy |
Theory |
Application of
unit to
new subject-matter |
Diagnosis |
Explanation |
| Research
summarized by or inspired by unit |
Descriptive
study |
Verificational
study |
We can round out
this listing by noting that some contemporary sociologists prefer
the term "frame of reference" to our
"taxonomy," and some, perhaps distressed at the
corruption of the concept of social theory, prefer the term
"model" to our "theory." The words used make
little difference so long as we remember to keep separate the two
enterprises depicted in our discussion.
1. One physicist has gone so
far as to characterize physics as a humanistic discipline. See
the postscript to T R Gerholm, Physics and Man, Totowa,
N J, The Bedminster Press, 1965
2. Bernard Berelson and Gary
Steiner, Human Behavior: An Inventory of Scientific Findings,
New York: Harcourt Brace and World, 1964.
3. Terence K. Hopkins, The
Exercise of Influence in Small Groups, Totowa, N.J. The
Bedminster Press, 1964.
4. Pitirim A. Sorokin, Social
and Cultural Dynamics, 4 vols., Totowa, N.J.: The Bedminster
Press, 1962. (Originally published 1937-1941.)
5. Robert K. Merton, Social
Theory and Social Structure, revised and enlarged edition,
New York: The Free Press of Glencoe, 1957, pp. 5-10.
6. A work by Pareto from
1901 contains a rather full blown propositional theory, and my
introduction to the English translation of the latter claims that
it is the first propositional theory in sociology. See Vilfredo
Pareto, The Rise and Fall of Elites, Totowa, N.J.: The
Bedminster Press, 1965.
7. See the autobiographical
introduction to his collection of essays entitled Sentiments
and Activities, New York: The Free Press of Glencoe, 1962,
pp. 1-49.
8.Talcott Parsons and Edward
A. Shils (eds.), Toward A General Theory of Action,
Cambridge: Harvard University Press, 1959.
9. Cf. Robert S.
Lynd, Knowledge for What, Princeton: Princeton University
Press, 1939.
10. Leonard Broom and Philip
Selznick, Sociology, second ed., Evanston: Row, Peterson
and Company, 1958, p. 3.
11. Talcott Parsons and Neil
J. Smelser, Economy and Society, New York: The Free Press
of Glencoe, 1956, p. 9.
12.Talcott Parsons, et al.,
'Values, Motives, and Systems of Action," in Parsons and
Shils, op. cit., pp. 76 ff and Samuel A. Stouffer and
Jackson Toby, "Role Conflict and Personality," ibid.,
pp. 481-496.
13. Georg Simmel, "The
Persistence of Social Groups," American Journal of
Sociology, vol. 3 ( 1898 ), pp. 829-836.
14. George C. Homans, The
Human Group, New York: Harcourt Brace & World, 1950, p.
242 ff.
15. Hans L. Zetterberg, Social
Theory and Social Practice, Totowa, N.J.: The Bedminster
Press, 1962, p. 189.
[p.30]
Definitions should
be used to facilitate communication and argument, and used only
to the extent that they make it possible to say something more
easily and clearly than would otherwise be possible. Sociologists
have spent much energy in developing technical definitions, but
to date they have not achieved a consensus about them that is
commensurate with their effort. At present there are so many
different competing definitions for key sociological notions such
as 'status' and 'social role' that these terms are no more
valuable than their counterparts, "position" and
"social relation," in everyday speech.
Yet definitions
are indispensable to the sociological enterprise. To the
annoyance of many critics, they set sociological writings apart
from historical and biographical writing. History describes and
often explains the development of a society. In their accounts
historians focus particularly on factors determining change. [p.31]
In all, the goal of the historian is very similar to that of
the sociologist. However, the historian in his accounts uses, by
and large, the language of the sources. He would say
that a certain political document stresses the individual pursuit
of happiness over the obligation to serve the king, and that a
certain theological document indicates a shift in interpretation
of the Bible's rules about taking usury so that it is taken to
refer to the Jews of ancient time and not to contemporary
Christians. The sociologist would use his own terms to describe
this; he might say, for example, that the political writer and
the theologians were both expressing 'achievement norms.' The
sociologist thus replaces the language of the sources with his
technical terms. Since historians have sifted enormous amounts of
source material to describe and explain change in a society, they
must implicitly have made many of the distinctions necessary and
crucial for the study of society. Macro-sociology translates
these distinctions from the language of the sources into a
general terminology applicable to any society. In the same way,
micro-sociology translates accounts of small groups and the
encounters reported in biographical documents into a general
vocabulary. Of course, such translations are likely to be
abstract approximations of the rich local color of the language
of the sources; this is one reason why so many artists and
humanists feel irritated at sociology.
[p.32] Discussions
of definitions often separate three different but related topics.
They may be represented by a triangle:
A term (word or
sign) is used to designate certain objects or events; the objects
or events are included in a conception; the term means this
conception.
All sciences have
their technical terms (and it follows from a sociological law
that this technical vocabulary will be called "jargon"
by outsiders). Much of the terminology of sociology is also found
in the language of educated speech, but some of it is more
specialized. Even highly educated speech is sometimes too
imprecise or cumbersome for sociological discourse. As everyone
knows, the word 'behave' might mean any activity, [p.33] but
sometimes it means 'to act with good manners'; the word 'society'
might mean the largest social system, but it might also mean the
'upper crust.' The sociologist needs some words of everyday
speech which have such an emotive or affect value in ordinary
speech that they must be re-introduced as formal definitions in
sociological discourse; 'culture' and 'bureaucracy' are examples.
Our personal
conviction is very much in favor of having a sociological
vocabulary that, in the main, is understood by most every
educated person, but in which each term has a more precise
meaning to the sociologist than to the layman. Unfortunately, the
opposite tendency prevails at present. Instead of speaking of
'equal rights,' some sociologists have learned to say
'universalistic standards,' or, instead of speaking of the
'familiarity' that prevails in some social relations,
sociologists have learned to say that 'diffuseness' characterize
the relations, and so forth, almost ad nauseum. This
makes sociology unnecessarily incomprehensible to outsiders. We
do not deny that 'diffuseness' in some contexts is a more precise
term than 'familiarity,' and that the sociologist may sometimes
need this added precision. We rather suggest that the sociologist
might still use the word 'familiarity' but give this word a
restricted and precise meaning. Among sociologists there would
then be no loss in clarity and outsiders would [p.34] still
have some understanding of sociological discourse. Sociological
terminology, in other words, should have incipient similarities
with good literary prose. 1
In sociology, as
in other sciences, conceptions may vary in content and also in
formal structure. Let us review some varieties of forms of
definitions.
An 'ostensive'
definition is "any process by which a person is taught to
understand a word otherwise than by use of other
words." 2 'Morale,' for
example, is ostensively defined by pointing to a situation and
saying, [p.35] "this is good morale." It is
entirely conceivable that the average spectator at a ball game
would be unable to give a verbal definition of the morale of the
playing teams. Yet he would not hesitate long to point out
instances of good or bad morale. In the same way, an officer
might be unable to give a satisfactory definition of morale
except by pointing to those of his units that have high morale.
In psychology and
sociology, several studies have started with ostensive
definitions. In a study of marital adjustment, a sociologist
asked a representative sample of an Indiana county to point out
to him couples among their acquaintances who were unusually happy
and couples who were unusually unhappy. In this way he obtained
two ostensively defined categories, one with high marital
adjustment and one with low. These two groups he later compared
on a variety of criteria to discover factors associated with
happiness in marriage. 3
When a term is
defined by means of other words we are, of course, dealing with
'verbal' definitions. It is of some importance to distinguish
between conceptions that assert something that can be more or
less truthful, and conceptions that merely express linguistic
conventions without assuming anything that can be proved true or
false. The rather misleading term 'real definition' usually
stands for [p.36] a truth-asserting definition: an
agreement to use in a given way certain notions which have
empirical relationships with each other. The other
"which-clause" is the heart of the matter, indicating
that these definitions are in the last analysis genuine
hypotheses which require testing before they can be accepted. An
example is furnished by a current (rather monstrous) definition
of 'social system'. It is suggested that "reduced to the
simplest possible terms, then, a social system consist of a
plurality of individual actors interacting with each other in a
situation which has at least a physical or environmental aspect,
actors who are motivated in terms of a tendency to the
optimization of gratification and whose relation to their
situation to their situation, including each other, is defined
and mediated in terms of a system culturally structured and
shared symbols". 4 This definition has
a built-in proposition about optimization of gratification.
However true a proposition like this may sound, we might hesitate
as a matter of principle to have it as an integral part of a
definition. Actually a good argument can be made that only under
special circumstances -- albeit common in our society -- do
people "optimize" their gratification; the universal
tendency seems to be that people are motivated to
"maintain" their accustomed gratifications.
There are several
varieties of these truth-asserting [p.37] definitions. As
in the above example, one criterion to be met by any case
belonging to the defined category might be that it verifies a
given proposition. When electricity is defined as anything that
satisfies Maxwell's equations we have a pure definition of this
type. Likewise, we have such a definition when a 'crowd' is
defined as anything that verifies LeBon's law of mental unity. It
is immediately realized that the value of this type of definition
rests upon the validity of of the propositions involved. If the
proposition turns out to be false, then the usefulness of the
definitions is nil. Since our knowledge of sociological events
rarely deserves unquestioned acceptance, we might suggest that
definitions in this form should be employed with caution.
A second variety
of truth-asserted definitions is found in many ideal types
developed by sociologists. Suppose we define a "primary
group" as one of small size, high morale, and of early
position in the individual's life cycle. Then we can define a
secondary group as one of large size, low morale, and late
position in the individual's life cycle. So far, all is well. But
sooner or later the writer who has defined his terms in this way
is likely to be found saying that "this group is more
primary than that group". Such an innocent remark implies
that he assumes a whole series of hypotheses to be true. He
assumes that the three variables defining a primary group are
interrelated: in other words, that a) the smaller the group is,
the higher its morale, b) the earlier it oc-[p.38]curs in
an individual's life cycle, the smaller it is, and, c) the
earlier it occurs in an individual's life cycle, the higher its
morale. None of these hypotheses is particularly well tested or
tenable. It should be clear that such assumptions revealed in the
usage of many an "ideal type" cannot be accepted in
advance of empirical testing and proof.
Often we are drawn
into truth-asserting by the use of analogous terms. In social
science it has been common to draw analogies from physical
science. An example is found in the definition of group
'cohesiveness'. Cohesiveness has been defined as the sum total or
resultant of all forces that keep a member of a group. 5 Borrowing from the
field of physics of the term 'force' might seem innocent enough
were it not for the fact that usage of the term implies at least
two propositions. In Newton's days these propositions were grand
discoveries, but since then they have become so self-evident that
we take them for granted. One of these hypotheses is that
whatever the origins of the forces -- whether from the moon
or from an apple -- they have the same consequences. Now, the
forces keeping a member in a group may vary greatly. He may stay
in the group because of the the prestige the group offers him,
because of the friends he has there, because of his need to be
punished by an authoritarian leader, and so on. To assume without
test-[p.39]ing, that all these forces have the same
consequences would indeed be presumptuous. 6 The second assumption
involved in the use of the term 'force' in the definition of
cohesiveness, is that whenever several sources of cohesiveness
are present their effects are cumulative. This principle has
proved to be immensely useful in physics: when several forces act
simultaneously, the effect is the same as if they had acted in
turn. This hypothesis is much less likely to be successfully
maintained in social science than in physical science. The
consequences of family cohesiveness deriving from both adequate
communication and adequate sexual adjustment during one year of
marriage are likely to be very different from the consequences of
a family cohesiveness based on one year of adequate sexual
adjustment and poor communication, followed by one year of
adequate communication but poor sexual adjustment. 7 Thus, we see how the
person who borrows a term from another science runs the risk of
borrowing more than a word: inadvertently he may borrow also some
propositions of this science. Clearly, definitions in the form of
analogous [p.40] terms deserve an extra careful
examination prior to their use in social theory.
In contrast to all
the above varieties of truth-asserting definitions, a 'nominal'
definition is a suggestion to name a phenomenon in a given way
without implying anything about the scientific propositions
relating to this phenomenon. Thus, nominal definitions are devoid
of hypotheses. They cannot be true or false. They can be clumsy
or elegant, appropriate or inappropriate, effective or
worthless -- but not true or false.
A common and loose
form of nominal definition is the "enumeration" of
ostensively or otherwise defined terms. Military morale is
difficult to define for research purposes. It is most readily
defined by enumeration of several factors: confidence in
officers, confidence in training, confidence in equipment,
confidence in rear echelons, identification with the war effort,
hatred of the enemy, satisfaction with the task assigned,
friendship with fellow soldiers, satisfaction with the military
system or rewards, and so forth. 8 Definitions by enumeration
give the scientist easy directions for concrete empirical
references to a concept. There seem to be, however, at least two
problems involved in the usage of definitions by enumeration. For
one, we have the risk that the factors enumerated are empirically
unrelated. As a [p.41] matter of fact and research, the
various dimensions of military morale enumerated above show
rather low intercorrelations, occasionally negative ones. To pool
them and call them 'morale' may be convenient for some purposes
but misleading for others. Secondly, the factors enumerated may
have no conceptual attribute in common. In what way is, for
example, 'identification with the war effort' conceptually
similar to 'confidence in rear echelons'? Enumerations, thus,
easily become conceptual patchworks, more confusing than
illuminating.
This last danger
is avoided in the conventional 'Aristotelian' definition. The
phenomena defined by such a definition always have two attributes
in common. One attribute -- genus proximum --
they share with a larger class; another attribute -- differentia
specifica -- is peculiar to the category defined. Many
definitions of morale follow this pattern. For example, morale
has been defined as "a disposition to act together (genus
proximum) toward a goal (differentia specifica)."
9 This type of
definition is so well known and its virtues are so obvious that
many a scholar thinks of this variety as the definition.
Lately, however,
our attention has been called to another form of definition which
is rapidly gaining ground [p.42] in many scientific
fields. Carnap calls it a 'dispositional concept.' 10 It is employed, for
example, to define the electrical resistance of a wire; the
resistance is a given number of ohms when a given number of volts
produce a current of a given number of amperes. The relation ohms
= f (volts, amperes) might be said to define resistance
dispositionally. In physiology or animal psychology, we could
likewise define hunger in terms of an equation involving hours of
starvation and number of calories in recent meals. Somewhat
simplified, the reasoning follows this pattern: if an object
(wire, animal) is subject to A (volts, hours of
starvation) and to B (ampere, calories), then we define
its X (resistance, hunger) as X = f (A,B).
In discussions of
morale -- to stay with the previous example
-- one finds statements to the effect that the test of
morale lies in the way a person or a group meets adverse
circumstances. For example, "we ascribe morale to a group to
the extent that it maintains (its) steadfastness of purpose,
maintains its solidarity, its integrity and its will to victory
even in the face of adversity." 11 This idea, that morale is
a measure of the extent a group sticks together under adverse
circumstances, is a rudimentary dispositional definition of
morale. A more precise formulation would be: if a group is placed
into a [p.43] situation of A degrees of adversity
and loses (or gains) B degrees of steadfastness of
purpose, then its morale, X, is X = f (A,B).
One is impressed
with the elegance of the dispositional concepts. However, their
use in sociology might at times be somewhat restricted. The very
formulations of dispositional concepts indicate that they require
the scientist to manipulate his object of study. For example,
previously we used the phrase "If an object (wire,
animal, group) is subject to etc." Now, our mores are such
that we can easily manipulate metal wires and laboratory animals
but not so readily human beings, human groups, and social
institutions. We are not, and do not want to be, in a position to
assign disasters and adversities to individuals and groups and
societies in order to measure their morale. Thus the practicality
of the above dispositional definition of morale is questionable.
This, however, does not mean that other dispositional concepts
might not be useful in other contexts of sociology, particularly
in micro-sociology. A person's 'attitude,' 'self-image,' 'action
repertoire,' 'commitment' might best be conceptualized as
dispositional concepts.
Max Weber has made
the most successful attempt so far to provide a taxonomy for
sociology. It appears in [p.44] the opening sections of
his posthumously published Wirtschaft und Gesellschaft. 12 There are many
reasons for the success of this particular taxonomy: it takes
into account numerous crucial distinctions made by historians; it
allows a rather easy translation of the language of the sources
into a technical vocabulary; it points out those factors we
cannot afford to overlook in a routine sociological description;
its terms have been used in the formation of some interesting
propositions.
Moreover, Weber's
taxonomy is readily seen as an orderly progression from the
simplest terms to the most complex ones. At any point we know
what any two conceptions have in common and where they differ. 13 [p.45] Consider,
for example, this string of definitions: A 'social relation' is
the presence of a probability that a social action will occur;
thus, teachers and students have a social relation as long as it
is likely that they will meet for class or conference. A 'social
order' consists of social relations guided by a set of
prescriptions; thus, education is a social order that includes
social relations between teachers, students, administrators, and
outsiders. The 'closure' of a social relation indicates the
extent to which persons are restricted from entering it; thus we
normally can say that an advanced seminar is [p.46] more
closed than a lecture course for beginners. An 'organization' is
defined as a rather closed social relation whose order is
enforced by a common leader or staff; thus a particular college
is an organization separated from other colleges because it has
its own president, deans, departmental chairmen, and so forth.
Organizations may be 'compulsive,' that is, impose their order on
anyone with given characteristics, such as an elementary school
which all children of a certain age have to attend. Or, an
organization may be 'voluntary,' that is, impose its order only
on those who have given it its personal adherance, such as a
college which we are free to leave at any time. A 'state' is a
compulsive organization that imposes its order on anyone living
in a given territory and that may legitimately use violence in
doing so. Thus Weber builds from the simplest (action) to the
most complex (state). Herein lies much of his appeal as a
taxonomist.
Let us explore
some formal aspects of such a process of taxonomy construction.
In all schemes of definitions we shall find words belonging to
the vocabulary of logic and mathematics. Words such as
"and," "or," "not,"
"imply," "equal to" belong here. They are
called 'logical terms.' 14 Unlike logical words, they
are not shared by all sciences but are specific to one or a few.
Samples are 'entrophy' in thermodynamics, 'reinforcement' in [p.47]
learning theory, 'homeostasis' in physiology, 'social role'
in sociology. For example, in a list of kinship terms, 'father,'
'mother,' 'son,' 'daughter,' 'brother,' 'sister,' 'uncle,' 'aunt'
would be extralogical terms. On the other hand, the words 'any,'
'and,' 'or,' 'of,' 'is called' are the logical words in the
definition: "Any brother of father and/or mother is called
uncle."
In an ideal theory
it should, furthermore, be possible to find a small group of
extralogical words, the 'primitive terms,' which in different
combinations with each other and with logical terms can define
all other extralogical terms of the theory, the 'derived terms.'
Any derived term, in short, is obtained by combinations of the
primitive terms and the logical words. There have been very few
efforts to systematically separate primitive and derived terms in
sociology. Talcott Parsons, in his analysis of the works by
Marshall, Pareto, Durkheim, and Weber, suggested that their
theories can be, viewed in terms of a handful of primitives, the
so-called means-end schema. Although no formal demonstration of
this has been made and the author himself eventually abandoned
the schema, 15 there can be no doubt
about its logic. A small set of building blocks, the primitive [p.48]
terms, can be combined into new units, the derived terms, to
create a sociological taxonomy.
Typologies are
constructed along similar lines but without the requirement that
the building blocks be primitive terms. The complex terms may be
obtained also by making use of derived terms previously defined.
Consider, for example the typology used in Merton's discussion of
anomie. 16 Merton sees the normative
order of society under two headings. First, there are norms
stating goals (e.g., to be successful, to rise from rags to
riches); second, there are norms setting forth the approved means
to reach these goals (e.g., to attend college, to take financial
risks). Deviance from these norms may now occur in several ways
and we obtain the following typology:
| |
Acceptance
|
| Goals |
Yes |
Yes |
No |
No |
| Means |
Yes |
No |
Yes |
No |
| |
Conformity
(e.g., Babbit) |
Innovation
(e.g., Robber Baron) |
Ritualism
(e.g., Bureau-
crat) |
Retreatism
(e.g., Hobo) |
Here 'goals,'
'means,' and 'acceptance' correspond to lower-order terms such as
the primitive ones, and 'conformity,' 'innovation,' 'ritualism,'
and 'retreatism' to complex derived terms. As we see from this
illustration, there is no logical difference between a system of
defini-[p.49]tions and a system of classification. That is
why we can call both by the same word: taxonomy.
Some primitive
terms of one science may also occur in other sciences. Geography
provides a simple example. Among its primitive terms are
"Greenwich" and "the North Pole." Bertrand
Russell has indicated that remaining primitive terms of geography
can be drawn from chemistry, physics, or geometry:
The
relation of "west of" is not really necessary, for
a parallel of latitude is a circle on the earth's surface in
a plane perpendicular to a diameter passing through the North
Pole. The remainder of the words used in, physical geography,
such as "land," and "water,"
"mountain," and "plain," can now be
defined in terms of chemistry, physics, or geometry. Thus it
would seem that it is the two words "Greenwich" and
"north pole" that are needed in order to make
geography a science concerning the surface of the earth, and
not some other spheroid. It is owing to the presence of these
two words (or two others serving the same purpose) that
geography is able to relate the other discoveries of
travelers. l7
[p.50] Thus
it appears that certain primitive terms are unique to a given
academic discipline while others are shared with other academic
disciplines. Let us call those primitive terms that are unique to
a given theory its 'minimum terms' and those that are shared with
other theories its 'borrowed terms.' 18
The distinction
between minimum and borrowed terms helps us clarify an old issue.
A number of sociologists from the time of Comte have tried to
answer questions such as "What is the proper subject matter
of sociology?" "What subject matter, if any, should be
treated in sociology but by no other science?" Much of this
discussion has been futile. However, our classification of
primitive terms provides a new and clearer criterion for
questions relating to the limits and uniqueness of the subject
matter of any theory.
It should be plain
that a theory does not have any subject matter that can be called
exclusively its own if all its primitive terms are borrowed terms
from other sciences. If so, anything that it talks about can be
exhaustively presented within the frameworks of theories from
other sciences. On the other hand, if we have one [p.51] or
more terms that can properly be called minimum terms, we also
have a unique subject matter. Any phenomenon, then, that has to
be defined in these minimum terms is our exclusive subject
matter. Therefore, instead of asking the old question,
"What, if any, is the proper subject matter of
sociology?" we instead ask, "What, if any, are the
minimum terms of a sociological theory?"
Most sociologists
have held that sociology has a minimum vocabulary of its own and
that the terminology of physics and biology is not relevant in
sociological discourse. In this vein, MacIver writes:
There is
an essential difference, from the standpoint of causation,
between a paper flying before the wind and a man flying from
a pursuing crowd. The paper knows no fear and the wind no
hate, but without fear and hate the man would not fly nor the
crowd pursue. If we try to reduce fear to its bodily
concomitants we merely substitute the concomitants for the
reality experienced as fear. 19
A few sociologists
have objected to this. For example, Lundberg:
The
principle of parsimony requires that we seek to bring into
the same framework the explanation of all flying objects....
From the latter point of view a paper flying before the wind
is interpreted as the be-[p.52]havior of
an object of specified characteristics reacting to a stimulus
of specified characteristics within a specified field of
force. Within this framework we describe the man and the
crowd, the paper and the wind. 20
The mainstream of
sociological thinking on this issue sides with MacIver. The
primitive terms we need in social theory do not consist
exclusively of terms found in physical and biological sciences.
We find it plainly impossible to say much of sociological
significance in a vocabulary based on the terms of physics and
biology.
Pareto and Weber,
as well as most contemporary social theorists, have assumed that
the building blocks of sociological definitions are terms that
denote human beings and their actions. The rationale for this
choice is found in a suggestive analogy between the position of
'actions' in all modern sociological theorizing, and the position
of 'primitive terms' in any taxonomy. The sociologists say that
all social events consist of combinations of human beings and
their actions. The logicians say that all terms of a theory can
ultimately be defined by combinations [p.53] of primitive
terms. It therefore seems useful -- at least as a
first approximation -- to assume that the primitive
terms of sociology should be words that denote human agents and
their actions.
The dangers
involved in departing from the rule that, in the last analysis,
we can only use terms representing combinations of human actions
are well illustrated
