Está en la página 1de 52

CAUSAL MECHANISMS, CORRELATIONS,

AND A POWER THEORY OF SOCIETY








James Mahoney
Department of Sociology
Maxcy Hall, Box 1916
Brown University
Providence, RI 02912
James_Mahoney@brown.edu







1



ABSTRACT


This paper explores how theories of causal mechanisms can help empirical analysts
improve the quality of their research. Drawing on philosophical realism, the paper conceives
causal mechanisms as unobserved entities and processes that act as ultimate causal forces in
the world. Defined as such, causal mechanisms can help empirical researchers better explain
particular outcomes, derive new testable hypotheses, and integrate existing correlational
knowledge. To illustrate these applications, the paper examines the two social science traditions
with the most developed causal mechanisms: functionalist theory and rational choice theory. In
addition, the paper discusses the applications of a new causal mechanism associated with a
power theory of society. This power mechanism suggests that many important social
outcomes are ultimately generated by the physical properties (i.e., mass, force, and direction) of
actors. The paper concludes by considering the normative entailments of treating functionalist,
rational choice, and power mechanisms as ultimate causes in the social world.




2
CAUSAL MECHANISMS, CORRELATIONS,
AND A POWER THEORY OF SOCIETY

Recently, several analysts have called for new theories of causal mechanisms in the
social sciences (e.g., Elster 1989, 1998; Goldthorpe 2000; Hedstrm and Swedberg 1998;
Stinchcombe 1993; Tilly forthcoming). Despite differences in their understanding of causal
mechanism, these scholars all argue that the identification of such mechanisms can substantially
improve empirical work. For some, this empirical payoff entails using theories of causal
mechanisms to explain why certain variables are regularly conjoined with one another; that is,
identifying the processes through which one variable affects another variable. For others, the
payoff involves using causal mechanisms to explain diverse outcomes from a common
theoretical approach; that is, subsuming multiple outcomes under one basic theory.
Although this agenda promises a great deal, its contribution depends on analysts actually
using theories of causal mechanisms in empirical work. To this point, however, writings about
causal mechanisms do not appear to have had much of an impact. Part of the problem may be
that, for empirical researchers, the concept of causal mechanism is still too unclear and its
specific application too vague to be readily employed in analysis. Indeed, a bewildering number
of definitions of causal mechanism can be found in the literature. Likewise, the actual steps
involved in using causal mechanisms for empirical research are rarely specified in the literature.
As a result, despite pleas for incorporating theories of causal mechanisms in substantive
research, this strand of social theory remains largely divorced from empirical analysis.
This paper attempts to build a bridge between social theory focused on causal mechanisms
and analyses centered around the testing of empirical hypotheses. It does so by introducing and
defending a particular conceptualization of causal mechanism grounded in philosophical realism,
3
and showing how this conceptualization can inform substantive research. Causal mechanisms
are defined here as unobserved entities and processes that act as ultimate causal forces in the
social world. By assuming the existence of such ultimate causes and employing certain bridging
assumptions, analysts can enhance empirical research in several specific ways. First, causal
mechanisms can help scholars address questions about the black box linking independent
variables with dependent variables. Second, by subsuming independent variables under a
common theoretical framework, causal mechanisms can help scholars understand why diverse
explanatory factors are all empirically associated with a particular dependent variable. Finally,
scholars can use causal mechanisms to logically deduce new and perhaps counterintuitive
hypotheses that can subsequently be tested in empirical research.
Currently, the two most developed causal mechanisms in the social sciences are found in
the core assumptions of rational choice theory and functionalist theory. The basic mechanism of
rational choice theory entails the claim that individuals will select the option that best enables
them to realize their private goals when faced with a range of behavioral alternatives.
1
For
functionalist theory, this mechanism entails the claim that enduring features of a social system
exist because they are functional for that system.
2
Below I explore how these mechanisms can
help empirical researchers subsume explanations under a single theoretical framework and better
understand why certain variables are associated with other variables. Yet, I am aware that
rational choice theory has come under attack for its limited empirical application (e.g., Green and
Shapiro 1994; Gould 2000), and that functionalism is currently a stagnant theoretical tradition
(e.g., Turner and Maryanski 1979). In response, I argue that these traditions could be revitalized
by recognizing that their core claims are not empirical hypotheses to be evaluated as either true
4
or false, but rather causal mechanisms to be evaluated according to their usefulness for empirical
work.
In addition to defending the causal mechanisms used in rational choice and functionalist
theories, I seek to show that a good deal of social science analysis assumes the existence of an
alternative mechanism. This mechanism is built around the notion that social outcomes are the
product of the physical properties (i.e., mass, force, and direction) of actors. I see this
mechanism as constituting the baseline assumptions of what can be called a power theory of
society, and I argue that such a theory can make substantial contributions to empirical research. I
develop this argument by considering how the mechanism underpins deductive theorizing in the
field of comparative-historical analysis, a research area that is sometimes considered purely
inductive.

LINKING SOCIAL THEORY AND EMPIRICAL RESEARCH

Unlike some areas of social theory, recent work on causal mechanisms has been concerned
with making links to empirical analyses, including quantitative sociology. This agenda is
especially important now because social theory and empirical sociology are increasingly distant
from one another.
3
In this section, I seek to illustrate how this division is harmful to the advance
of social research. I argue that the findings of empirical researchers are incomplete and not fully
intelligible without theories of causal mechanisms; by contrast, theories of causal mechanisms
are entirely speculative until their usefulness is revealed through empirical correlations. A
complete science must therefore strive to identify both empirical regularities and causal
mechanisms (see Dessler 1991; Jasso 1989).

5
Moving Beyond Correlations
The concern with causal mechanisms grows out of a distinction sometimes made in the
philosophy of science between correlational analysis and causal analysis. Whereas correlational
analysis involves identifying antecedents regularly conjoined with outcomes, causal analysis
consists of specifying the mechanism that underlies and generates empirical regularities and
outcomes. In contrast to correlational research, the challenge of causal analysis involves
postulating entities, properties, processes, relations . . . that are held to be causally responsible
for the empirical regularities to be explained (McMullin 1984a, p. 210; see also Blalock 1961,
pp. 11-13; Keat and Urry 1982, chap. 2). These mechanisms explain why social regularities
exist in the first place; knowledge of their operation allows researchers to go beyond
correlations.
At the base of this argument is a rejection of the Humean model of constant conjunction for
understanding causation. According to Hume (1748/1988), we infer causation when we
repeatedly observe putative causes followed effects; that is, two events are constantly
conjoined in our experience. However, we can never actually confirm the existence of
causation because the imagined necessary connections linking the two events cannot
themselves be observed or known. In contrast to this Humean model, scholars concerned with
causal mechanisms argue that exploration of the black box connecting independent and
dependent variables is essential to good causal research. On this view, we are not satisfied with
merely establishing systematic covariation between variables or events; a satisfactory
explanation requires that we are also able to specify the social cogs and wheels that have
brought the relationship into existence (Hedstrm and Swedberg 1998, p. 7; see also Elster
1989, p. 3; Glennan 1996; Harr 1970, p. 230; Harr 1972, p. 170).
6
On a substantive level, the identification of causal mechanisms can help empirical analysts
overcome two basic problems associated with their research. First, correlational findings are
almost inevitably vulnerable to charges of selectivity and omitted variable bias (see Lieberson
1985). As a consequence, scholars are often left uncertain whether a given association reflects
causation, or whether the association is simply the spurious product of an unknown antecedent
variable. Second, empirical researchers typically discover that a highly heterogeneous group of
independent variables are associated with an outcome, and they lack tools for understanding why
such diverse factors are related to the phenomenon of interest. The result can be an absence of
theoretical integration, which in turn contributes to the fragmentation of knowledge that
currently characterizes the social sciences (Dessler 1991; Srensen 1998).
The identification of causal mechanisms can help researchers overcome these problems by
providing the theoretical basis for understanding why an association between empirical variables
exists (Hedstrm and Swedberg 1998, pp. 8-9). For example, by logically deriving a correlation
from assumptions about a causal mechanism, analysts can explain why the independent variable
of the correlation affects the dependent variable. Likewise, by showing how diverse independent
variables all affect a given dependent variable by virtue of a common causal mechanism,
scholars can integrate existing correlational findings and put social science analysis on a better
foundation for accumulating knowledge.

Defining Causal Mechanisms
Unfortunately, a good deal of confusion currently surrounds the precise meaning of causal
mechanism. This terminological confusion is one reason why theoretical work on causal
mechanisms has not attracted wide attention among empirical analysts. Here I attempt to make
7
sense of existing definitions by arranging them into four groups, each of which contributes
insight toward a synthetic conception of the term (see Table 1).
---------------------------
Table 1 about here
----------------------------
First, the simplest definitions treat causal mechanisms as synonymous with independent
variables or causal factors that help explain outcomes (see Table 1). For example, Boudon
defines a mechanism as the well-articulated set of causes responsible for a given social
phenomenon (1998, p. 172). This definition usefully suggests that causal mechanisms are, in
fact, causes, and that they help produce outcomes. However, it does not distinguish the idea of
mechanism from the standard notion of an independent variable, raising the question of what the
study of mechanisms adds to correlational research.
Second, other definitions see mechanisms as intervening variables, events, or processes that
explain how one variable influences another (see Table 1). These definitions are useful in that
knowledge of intervening processes can help researchers understand why a given independent
variable exerts a causal effect on a given dependent variable. In addition, knowledge of
intervening processes can increase ones confidence that a statistical association is not spurious.
However, this kind of definition leaves open the possibility that causal mechanisms must be
identified and analyzed as correlations. That is, to locate a causal mechanism, the analyst must
show how intervening processes and events are themselves correlated with both independent and
dependent variables.
4
In this sense, the analyst explains a correlation by appealing to another
correlation, begging the question of why the new association exists (King, Keohane, and Verba
1994, p. 86; McMullin 1984a, p. 206). To explain the new correlation, the analyst must identify
an additional mechanism, which itself will contain a black box and require explanation. Under
8
this definition, in short, the distinction between an independent variable and a mechanism
becomes somewhat arbitrary, and the analyst may be forced into an infinite regress in search of
deeper and deeper mechanisms.
Third, other scholars view causal mechanisms as underspecified causal propositions that
can be applied to a fairly wide range of cases (see Table 1). For example, according to Elsters
(1998) influential definition, mechanisms are frequently occurring and easily recognizable
causal patterns that are triggered under generally unknown conditions or with indeterminate
consequences (p. 45). This definition assumes that a mechanism identifies a cause-effect
relationship that is applicable in many social situations. Such a cause-effect relationship differs
from empirical work on correlations in that it is underspecified; that is, it makes reference to
analytical constructs that are not actually observed (Rueschemeyer 2001). For example, one of
Elsters mechanisms is the spillover effect, defined as follows: if a person follows a certain
pattern of behavior P in one sphere of life, X, he will also tend to follow P in sphere Y (p. 54).
This mechanism identifies a probabilistic relationship between an underspecified independent
variable (a type of behavior in one sphere) and an underspecified dependent variable (a type of
behavior in another sphere).
Although these underspecified causal propositions do not have empirical content, they can
be used to derive empirical hypotheses. For example, the spillover effect could be used to
hypothesize that alienation at work will produce alienation at home, or that greater participation
at work will produce greater participation in politics (Elster 1998, pp. 54-55). While clearly
useful, this definition of causal mechanism nonetheless fails to explain why the underspecified
cause-effect relationship embodied in mechanism itself obtains. For example, with respect to
Elsters spillover effect, the mechanism itself does not explain the process through which a given
9
type of behavior might spillover into a new sphere; that is, it does not explain why the
underspecified independent variable affects the underspecified dependent variable. As a result,
scholars have no way of knowing when the spillover effect will be in operation as opposed to
some other mechanism, including a mechanism that predicts exactly the opposite of the spillover
effect, such as the crowding-out effect (i.e., if a person follows a certain pattern of behavior P
in one sphere of life, X, she will not tend to follow P in sphere Y). In short, this definition of
mechanism itself contains a black box, not identifying the reasons why the underspecified
association holds and thereby posing problems for empirical research.
Scholars associated with the realist school in the philosophy of science offer a fourth
definition that, I believe, synthesizes insights from the above definitions while avoiding their
shortcomings (see Table 1). This synthetic definition understands a causal mechanism as an
unobserved entity, process, or structure that acts as an ultimate force in generating outcomes. An
ultimate force is a final cause that itself requires no further explanation (Harr 1970, pp. 101-4;
see also Jasso 1998, pp. 5-6; Steinmetz 1998, pp. 24-25). When causal mechanisms are
understood to be ultimate forces, they cannot be studied through direct empirical means.
However, by appealing to their existence, scholars can explain diverse empirical outcomes from
a single theoretical perspective and offer an account of why empirical correlations exist in the
first place.
In the natural sciences, it is common for researchers to posit causal mechanisms in this
sense of unobserved ultimate forces. For example, scientists assume the existence of strings in
superstring theory, quarks in particle theory, and gravitons in gravitational theory (Hacking
1983; McMullin 1984b; Musgrave 1985). The utility of these entities is not as independent
variables that explain variation on an outcome; rather, they are believed to be the real
10
mechanisms that physically constitute or otherwise generate phenomena (including empirical
correlations) in the natural world.
When a scholar asserts that a mechanism causes some outcome, the scholar is typically
presenting a claim that is, by itself, tautological or otherwise not directly testable. For example,
it is tautological to argue that gravity causes objects to fall, given that gravity is defined as a
hypothetical force that will have this effect. Likewise, in rational choice theory, it is tautological
to argue that actors carry out particular actions because these actions are the most appropriate for
achieving their goals, given that one cannot falsify the proposition with empirical evidence. The
nonfalsifiable nature of explanations based on causal mechanisms is closely related to their status
as ultimate causes. Once an explanation based on a mechanism has been proposed, it is no
longer relevant to ask why the mechanism produces the outcome in question, because the
mechanism stands in a necessary relationship to the outcome. If the mechanism really exists and
is operating, it will produce the outcome of interest. In this sense, explanations that appeal to
causal mechanisms are explanations without black boxes.

Using Causal Mechanisms in Empirical Research
Because they lead to tautological propositions, causal mechanisms have little value unless
they are linked to correlational research through bridging assumptions (see Kelle and
Ludemann 1998; Morton 1999). The process of formulating bridging assumptions is in fact one
of the main activities that practicing theorists carry out when building their models. Yet, it is
surprising how little has been said about the use of such assumptions in the social sciences. This
silence has stood in the way of better linking social theory and empirical research. As a remedy,
I discuss three ways in which bridging assumptions can be used in conjunction with causal
mechanisms to generate empirical propositions (see Table 2).
11
---------------------------
Table 2 about here
---------------------------
First, researchers can use bridging assumptions to test whether specific outcomes are
produced by a given causal mechanism. With this strategy, the researcher begins by asserting
that a specific outcome of interest is the product of a particular mechanism. For example, a
biologist asserts that a given trait of a species (e.g., bipedalism among humans) exists because it
is (or was) conducive to survival and reproduction, thereby using the natural
selection/functionalist mechanism of biology. Next, the analyst introduces a set of assumptions
from which the initial assertion is logically derived. For example, the biologist makes certain
assumptions about the ancestral environment in which bipedalism first developed (e.g.,
availability of food, type of predators) to logically deduce the assertion that bipedalism conferred
reproductive advantages. In this sense, bridging assumptions are used as premises and the
proposition that a natural selection mechanism causes bipedalism is treated as a conclusion that
necessarily follows from these premises. The analyst then empirically assesses the premises of
the argument by formulating them as testable hypotheses. For the biologist studying bipedalism,
this entails using the fossil record to support assumptions about the nature of the ancestral
environment. Testing premises is relevant because if one establishes that they are likely true,
one has reason to believe the conclusion is also likely true, given that the conclusion follows
logically from the premises.
In a second strategy, the analyst derives testable hypotheses through a logical proof that
combines a causal mechanism with other bridging assumptions. Here the goal is to work
deductively from a set of premises that include both a causal mechanism and additional bridging
assumptions to new propositions. For instance, the natural selection mechanism of biology can
12
be combined with assumptions about parental investment to arrive at the prediction that the
relative degree to which females versus males of different species are eager to initiate courtship
and mating will covary inversely with the amount of resources and time they must invest to
produce surviving offspring. Propositions derived in this fashion are of particular relevance to
empirical researchers because they contain no black boxes; that is, the logical connection
between the independent variables and the dependent variables of the propositions is explained
by the theoretical proof itself. Unlike the first strategy, the premises used to derive conclusions
in this strategy are not tested themselves (Cohen 1989; Jasso 1988; Stinchcombe 1993;
Kanazawa 1998). Rather, it is the propositions logically derived from the premises that represent
the hypotheses to be empirically evaluated. As a general rule, scholars strive to use as few
bridging assumptions as possible to generate as many testable hypotheses as possible.
5

A final strategy involves working backward from one or more existing correlations to a
causal mechanism. With this strategy, the analyst attempts to show how already established
correlations can be logically derived from a set of premises that includes a causal mechanism
plus additional bridging assumptions. For example, by assuming that natural selection drives
evolution and that the ancestral environment featured individuals living in small villages near
close kin, sociobiologists can offer a theoretically grounded explanation for the association
between todays large, impersonal urban settings and crime and uncivil behavior. This strategy
parallels the second strategy above, except that the analyst begins with correlational findings that
have already been supported and works back to the premises from which one can logically
deduce these correlations.
6
The analyst does not actually test any correlations, since this work
has already been completed. Instead, the analyst integrates existing knowledge by showing how
established associations can all be explained in terms of a single mechanism.
13

THEORIES OF CAUSAL MECHANISMS

In this section, I discuss the two social science theories with the most developed causal
mechanisms: functionalism and rational choice analysis. This discussion attempts to make
concrete some of the points raised above and demonstrate how research based on these
mechanisms can best serve sociology.

Functionalism
By the late 1970s, functionalism as a theoretical approach was pronounced dead (Turner
and Maryanski 1979, p. 141) and an embarrassment to contemporary theoretical sociology
(Moore 1978, p. 321). Since then, it saw a brief but aborted revival though neo-functionalism
(Alexander 1998); today almost no social theorist would identify himself or herself as a
functionalist. However, I believe it is possible and desirable to again try to revive this theoretical
program. When conceived from the perspective of causal mechanisms, as opposed to
correlational analysis, functionalism has yet to be proven inherently defective.
7

Adversaries of functionalism tended to evaluate the theory using criteria appropriate for
correlational analysis rather than causal mechanisms. Thus, functionalism was criticized for
offering non-verifiable, tautological, and overly abstract claims about the nature of society (e.g.,
Buckley 1957; Dore 1961; Hempel 1959; Merton 1949; Nagel 1957).

Yet, from the perspective
of causal mechanisms, these criticisms do not carry much weight. A more important concern is
that functionalists failed to establish links with correlational research, even as they held up their
approach as a universalistic theory of society (e.g., Davis 1959). Too often, functionalists simply
asserted that a particular societal trait (e.g., religion) exists because it fills a requisite need for a
14
social system (e.g., survival). However, a claim like this cannot be directly tested and is of little
value in itself. To be of use, the claim must be logically derived from a set of testable premises
or used with bridging assumptions to derive new propositions.
8
Lacking a strong connection to
empirical research, functionalist theory eventually degenerated into abstract exercises in
conceptual classification, exemplified by the later writings of Talcott Parsons.
9

If it is to be revived, functionalism must be understood as embracing a causal mechanism
not a set of directly testable hypotheses that can be applied to select research questions in
sociology. The basic functionalist mechanism treats social aggregates as systems with internal
parts that exist because they serve (or previously served) systemic functions; it assumes that
societal structures would not exist in their present form if they did not currently or previously
serve a function for the social system within which they are embedded. Beyond this claim,
functionalist theory says little about the workings of societies.
As many observers have noted, studies that draw on the functionalist mechanism make
teleological claims. That is, they assume outcomes occur because they help realize some future
end or goal; the future end or goal acts as a cause of the outcome. However, not all teleological
explanations are illegitimate (Stinchcombe 1968, pp. 80-101; Turner and Maryanski 1979, p.
118-19; Elster 1982, p. 455). They are illegitimate only when the analyst fails to specify the
reasons why a future end or goal leads to the occurrence of a given outcome. Traditionally,
functionalists tried to identify such reasons by specifying the needs of society that must be met
to ensure its survival or stability (e.g., Aberle et al. 1950; Levy 1952, p. 151). These functional
requisites, however, were often stated so vaguely that it was impossible to decide empirically
when a given need was being met. For example, it was difficult to know when a systems need
for normative regulation or socialization was adequately met.
15
Insights about natural selection in Darwinian theory can help functionalists better specify
system needs. Darwinian theory assumes that biological traits are initially introduced into a
population through an essentially random process (i.e., mutation),
10
and that the ability of a trait
to persist in the population depends on the reproductive success of the organisms who carry it. If
a trait confers a reproductive advantage to its carriers in a given environment, the trait will
proliferate. In many cases, in fact, the trait will continue to exist even when, in a subsequent
environment, it no longer confers a reproductive advantage.
Functionalists could fruitfully extend these ideas to explain the existence of social
structures found widely across societies (e.g., social stratification, religion, or government) (see
Merton 1949, p. 50). Like biologists, functionalists need not explain why a given social structure
was first introduced into a society; its initial appearance can be treated as exogenous to the
theory.
11
However, once the trait appears in at least one member of the population, functionalists
must offer an account of the winnowing process through which it spreads to other members.
Here is where the natural selection argument of biology can be put to good work.
Many of the societal traits that interest functionalists developed in an early environment
when human society was organized as hunter-gatherer bands or primitive agricultural
settlements. These early societies were vulnerable to elimination through two primary sources:
warfare and/or starvation. For example, there is strong evidence that warfare was extremely
common in early agricultural societies, and that societies that were better adapted to warfare
extended their control of territory at the expense of societies that failed to meet the exigencies of
warfare (e.g., Mann 1986). Hence, traits that favored success in warfare should have been
selected for in this early environment (Wendt 1999, pp. 321-23).
12
Likewise, given that
starvation could and did eliminate hunter-gatherer societies, one would expect hunter-gatherer
16
bands that developed features that enabled them to avoid this outcome to proliferate at the
expense of those that did not. Especially given the possibility of cultural selection, in which
units acquire traits through imitation, it is hardly unrealistic to believe that societies would
eventually converge on certain social structures that provided an early adaptive advantage. Once
these traits were institutionalized, they could continue to exist through a path-dependent process,
even long after they had outlived their original useful purpose (Mahoney 2000).
Although some observers might see this call for employing ideas about natural selection as
an unwelcome return to previous failed efforts to use Darwinian theory, I believe that
contemporary sociologists can profitability revisit some of the classic works carried out by
functionalists in the 1950s and 1960s. Of these works, perhaps the best known and most
controversial is Davis and Moores (1945) analysis of social stratification. These authors argued
that stratification along the lines of status and income exists because it is functional for society.
Table 3 summarizes four premises that constitute a modified version of this argument. The first
premise asserts that individuals inherently desire status and/or income. The second premise adds
the assumption that individuals, ceteris paribus, do not desire social positions that require
training. A third premise explains why individuals sometimes fill social positions that require
training: status and/or income compensate for the costs of training. A final premise then paves
the way for the authors conclusion about the functionality of stratification: in order for society
to survive, it is necessary that individuals occupy certain social positions that require training.
---------------------------
Table 3 about here
---------------------------
To assess this argument, the three strategies discussed above are available to researchers.
First, the premises of the Davis-Moore argument can be directly tested. In the past, critics
17
argued that certain of these premises are not true, especially the notion that training is inherently
undesirable (Tumin 1953). In addition, some observers wondered why training is or was
necessary for societal survival.
13
Nevertheless, the most rigorous tests of these hypotheses were
generally supportive (e.g., Cullen and Novick 1979).
Second, the argument could be assessed by asking about additional hypotheses that should
be true if we assume the premises of the argument are true. To derive such hypotheses, it would
be necessary to decide whether selection pressures in favor of social stratification exist in
contemporary societies, or whether these pressures operated only in the past. Davis and Moore,
for their part, believe that selection pressures are still at work in the contemporary world and that
failure to conform to such pressures produces social instability. Assuming this is true,
14
testable
hypotheses about todays societies can be formed. For example, the Davis-Moore argument
suggests that the level of social stability in society should be, in part, a product of the degree to
which there is a positive linear relationship between occupational training and occupational
income/status. Likewise, the argument suggests that governments that attempt to eliminate the
status differentials across occupations marked by different levels of training should not be able to
survive over the long run. These hypotheses can, in principle, be tested with empirical measures
of social stability, occupational training, and occupational income and status.
Third, the Davis-Moore argument could be used to make sense of existing correlations that
already have received significant empirical support. One such correlation concerns the
relationship between the level of economic development of a society and its level of economic
inequality. Substantial research on nation-states has confirmed the Kuznets inverted-U
hypothesis, which holds that inequality first increases during the process of economic growth,
but subsequently decreases as economic growth continues (see Lindert and Williamson 1985;
18
Nielson and Alderson 1995). Despite agreement concerning this association, scholars have not
reached any consensus regarding why the association exists. The Davis-Moore theory could be
used to explain the correlation through the following argument. During the early phases of
economic growth, new occupations emerge that require some individuals to receive dramatically
more training and thus income and status than others. This change in occupational structure
therefore increases the level of income inequality. However, with sustained economic
development, the pronounced differences in the occupational training required for individuals
begin to decline as more and more individuals are trained to work in a modern economy. Income
and status inequality follow suit, leveling off and then declining in the course of economic
development.
To fully assess the validity of a functionalist explanation of social stratification, one would
have carry out empirical tests, which is not my purpose here. Nevertheless, the example does
illustrate how functionalism could be revived in sociology. It is remarkable that virtually no
active sociologist identifies himself or herself as a functionalist, given that the theory seems
appropriate for understanding how certain social institutions (e.g., governmental, educational,
religious, health, familial institutions) proliferated to the point that they are now found across
most or all modern societies. In fact, while I have focused on early selection pressures as a
means of reviving functionalism, the theory could be used to study how certain modern social
institutions (e.g., the family) may persist over time because conscious learning and design have
led actors to believe that these institutions are functional for society. The potential for research
along all of these lines is there, but it is up to scholars to actually seize the opportunities.


19

Rational Choice Theory
In the 1950s, it was common for analysts to assert that functionalism was the only viable
theory for social analysis. Fifty years later, one hears remarkably similar claims about rational
choice theory (e.g., Geddes 1991; Kiser and Hechter 1991, 1999; Wallerstein 2001). Despite
these claims, there are good reasons to believe that rational choice theory, at least in sociology,
could meet a fate similar to functionalism in the not too distant future. This is true because the
very developments that led to the swift abandonment of functionalism increasingly apply to
rational choice theory.
First, one hears mounting criticisms that rational choice theory is abstract, tautological, and
circular (e.g., Almond 1991; Green and Shapiro 1994). Again, these concerns stem from the
tendency of critics to evaluate the program from the standpoint of correlational analysis rather
than a theory of a causal mechanism. Second, rational choice theorists have been slow to link
their work to correlational research. Like functionalists before them, these theorists often spend
too much time specifying elaborate models at the expense of generating testable propositions
(Levi 1999, pp. 154-55; Munck 2001, p. 200). Finally, the universalistic aspirations of some
rational choice theorists have become a focal point of criticism. Skeptics and even pragmatic
rational choice theorists themselves contend that it is unrealistic to think that rational choice
theory can be applied to all domains of social analysis (e.g., Green and Shapiro 1994, pp. 27-28;
Munck 2001, p. 188; Tsebelis 1990, pp. 42-43).
As others have noted, the core of rational choice theory has no empirical content (e.g.,
Denzin 1990; Kelle and Ludemann 1998, p. 112). This is true because the theory is nothing
more than the basic mechanism discussed above: given a set of alternatives, individuals will
choose the option that best enables them to realize their goals. To be of use, this mechanism
20
must be empirically specified through bridging assumptions. In rational choice theory, bridging
assumptions typically identify actors, preferences, available choices, opportunity costs, and the
means through which individual choices combine to produce social outcomes (Friedman and
Hecther 1988). In conjunction with such assumptions, rational choice theorists can pursue the
three strategies discussed above and thereby build a bridge to correlational analysis.
The first strategy involves using testable assumptions to logically derive the conclusion that
the rational choice mechanism produces an outcome of interest. For example, to explain why a
collection of individuals fails to produce a public good from which all would benefit, rational
choice theorists draw on bridging assumptions about egoism, non-excludable goods, and the
relative contribution of a single individuals participation to show how this seemingly illogical
outcome is the result of rational individual choices (see Olson 1965). Theoretical proofs that
take this form are internally coherent in the sense that, if the premises are true, the conclusion
about the causal role of the mechanism in producing the outcome necessarily follows from these
premises. With this strategy of explaining particular outcomes, the challenge is to substantively
demonstrate the validity of the premises. For example, when using Olsons theory of collective
action to explain the absence of some specific public good, theorists must be held accountable
for premises about egoism and the non-excludability of the good in question.
With the second strategy, new testable propositions are derived from an initial set of
premises that include the rational choice mechanism and other bridging assumptions. For
example, by assuming that individuals make rational choices and that the premises of the theory
of collective action are true, analysts have logically arrived at the prediction that broad-based
cartels will not exist unless these organizations offer selective incentives to their members. This
is true because non-participators can free-ride and still enjoy the benefits of the cartel. Unlike
21
the first strategy, the connection to empirical research here does not involve testing premises, but
rather offering new testable propositions derived from an initial set of (untested) premises.
Finally, with the third strategy, the analyst works backward from existing correlations to
the rational choice mechanism and bridging assumptions in an effort to explain why these
correlations exist in the first place. For example, a number of group characteristics have long
been known to increase the probability of providing a public good. Included among these are
small size, previous organizational history, and an unequal distribution of resources. Again,
Olsons theory of collection action can explain why these associations exist in terms of a single
theoretical model (see Geddes 1995, pp. 95-96). For example, inequality is positively associated
with organization because powerful actors may have an incentive to provide the good by
themselves, even if most or all other actors free ride. Likewise, small groups can more readily
organize because each individuals contribution in such a group can make a difference to the
production of the good and because free riders can be more easily monitored.
In short, then, it is useful to speak of three kinds of rational choice theorists: outcome
explainers, proposition derivers, and theoretical integrators.
15
The outcome explainers are those
analysts who use bridging assumptions to derive the conclusion that the rational choice
mechanism produced a particular outcome of interest (e.g., Bates et al. 1998). These theorists
make empirical contributions insofar as the premises of their models can be substantively
justified. Second, the proposition derivers are those theorists who do not seek to explain any
specific outcome, but rather who seek to use the assumptions of rational choice theory to arrive
at new propositions (e.g., Kiser and Hechter 1991 on state autonomy). These theorists supply
empirical researchers with testable hypotheses that embody underlying assumptions about
rational choice. Lastly, the theoretical integrators are analysts who attempt to make sense of
22
pre-existing correlational findings by interpreting them in light of overarching assumptions about
rational choice theory (e.g., Geddes 1998 on findings about democracy). The contribution of
these researchers is to integrate existing knowledge into a single theoretical model.
Despite the proliferation of literature on the methodology of rational choice theory, these
three strategies have not been previously identified. As a result, rational choice theorists have
lacked a solid foundation for discussing the relationship between their theoretical models and
empirical analysis. In distinguishing between outcome explainers, proposition derivers, and
theoretical integrators, I hope to provide a stimulus for more empirical work from this
perspective.

Other Causal Mechanisms?
The quest for ultimate causes that generate empirical associations and outcomes animates
much scientific research. Following the realist literature, I have used the expression causal
mechanism to designate such ultimate causes. In the social sciences, however, there are very
few examples of causal mechanisms beyond the functionalist notion that system needs
ultimately drive important outcomes and the rational choice notion that the utilitarian cost-
benefit assessments of individuals are the final movers of social happenings. Given this limited
range of causal mechanisms, social theorists might consider building new theories of such
mechanisms.
16

Formulating theories of causal mechanisms requires abstract thinking and imagination, but
the process need not be purely deductive. Rather, it is possible to work inductively from existing
studies, asking what implicit mechanism informs a given work. In many cases, one will find that
social science arguments are simply not grounded in any clear mechanism. Yet, for other
23
studies, especially studies that group together to form a coherent body of literature, one may be
able to identify an underlying mechanism.
The field of comparative-historical sociology is a particularly fruitful area for formulating
ideas about new causal mechanisms. Although work in this field has been characterized as
antitheoretical (e.g., Kiser and Hechter 1991), its proponents maintain that they draw
extensively on general theoretical principles (e.g., Quadagno and Knapp 1992). If this is true, it
should be possible to identify the causal mechanism that guides research in this field. In what
follows, I attempt to do this for those comparative-historical works that emphasize power as an
unobserved but ultimate causal force.

POWER THEORY:
BASIC COMPONENTS AND APPLICATIONS

Much comparative-historical research assumes that the abstract power of actors ultimately
generates outcomes in the social world. Taking this assumption as a starting point, a new power
theory of society can be introduced to complement functionalist and rational choice theories.
Power theory makes an analogy between force and movement in the physical world and events
and processes in the social world. In this sense, it builds on both materialistic approaches that
assume causal effects in the social world may exist independent of consciousness (see Mann
1979; Porpora 1993) and conflict theories that view struggles between competing actors as major
determinants of social happenings (e.g., Dahrendorf 1958; Lenski 1966; Collins 1975).

Building Blocks of Power Theory
Like rational choice theory, power theory assumes that actors (not necessarily defined as
individuals) have certain essential properties that ultimately generate social outcomes. Two
24
characteristics in particular define actors: power and direction. Taken in isolation, both of these
characteristics are abstract and incomplete. Yet, by assuming their existence, one can build a
theory of society in which the basic causal imagery parallels the study of motion in physics,
including the idea of vectors moving through space.
Power itself can be usefully disaggregated into two sub-dimensions: mass and force.
17

Mass refers to the total latent capability of an actor to produce some end, acting as the equivalent
of weight in the physical sciences. Typically, the mass of an actor is defined in materialistic
terms and relative to other actors. For example, with respect to the end of winning an interstate
war, the United States has considerable mass, because it possesses the latent capacity to defeat
nearly any other individual country. Force refers to the extent to which the total latent capacity
of an actor is actually directed toward an outcome, acting like speed or acceleration in
physics. Actors with extensive mass have the potential to push aside less massive actors, but
they may lack sufficient force to do so. For example, Vietnam may be able to prevail over the
United States in a war if the latter country does not move with adequate force. In this theory, the
product of the mass and the force of an actor define its power; that is, mass x force = power.
The second component of actors is their direction, or the particular outcome toward which
they move. Actors may be directed at outcomes in a conscious or unconscious fashion. For
example, the Swedish government may consciously direct power toward cleaning up the
environment. By contrast, the citizens of the Sweden may unconsciously use power to damage
the environment, simply by virtue of driving automobiles. In power theory, it is highly possible
that final outcomes will not reflect the conscious preferences of any actor, given that outcomes
are often the result of actors moving in very different directions. Hence, the theory (like rational
25
choice theory) provides a basis for understanding the unintended nature of many social
occurrences.
To be linked to empirical analysis, the basic components of power theory must be specified
through bridging assumptions. This process is equivalent to adding empirical content to the
empty cores of functionalist theory and rational choice theory. Among the key assumptions that
often must be made in power theory are the definition of all relevant actors, the empirical
specification of the mass and force of these actors, and the identification of the outcomes at
which these actors are directed. Although formulating such assumptions may seem a large task,
it is no more demanding than the process of specifying bridging assumptions in functionalist and
rational choice theories.
Power theory leads to a particular conception of causation in the social world, one that
contrasts with both the natural selection imagery of functionalism and the utilitarian imagery of
rational choice theory. With power theory, social outcomes are the product of objects moving
with mass and force that may collide with one another; the vision here is one of billiard balls in
motion. Given this conception, vector analysis from physics (Durrant 1996) can help analysts
formally diagram their arguments. A vector refers to both the direction of an actor, which is
diagramed using an arrow, and the power (mass x force) of an actor, which is specified by the
length of the arrow. If the power and direction of all actors are known, standard rules of vector
addition can be used to calculate final outcomes. For example, if space is conceptualized as a
one-dimensional line containing two or more vectors, the lengths of each vector can be summed
together to determine a final location. If space is conceived as having two dimensions (an x and
y axis), the addition rules of vector mathematics can also be easily applied to determine final
locations.
26

Application to a Balance of Power Argument
If only implicitly, power theory has already underpinned empirical research in the field of
comparative-historical sociology. To illustrate this point, I consider in this section
Rueschemeyer, Stephens, and Stephens Capitalist Development and Democracy (1992),
arguably the most important recent work in the field. I show how the authors implicitly use a
power mechanism to carry out three basic tasks: explaining particular outcomes, deducing new
propositions, and subsuming existing correlational knowledge under a single theoretical account.
---------------------------
Figure 1 about here
---------------------------
Rueschemeyer, Stephens, and Stephens summarize the central argument of their study as
follows: power relations . . . determine whether democracy can emerge, stabilize, and then
maintain itself (p. 5). To develop this argument, the authors offer a balance of power model in
which the occurrence and stabilization of democracy depends on the relative power and direction
of different actors. Through bridging assumptions, the authors identify their main actors as
different social classes: workers, peasants, middle classes, the bourgeoisie, and landed elites (see
Figure 1). The authors also make assumptions about the direction of these actors (pro-
democracy or anti-democracy) based on whether they are likely to gain or lose from the
establishment of democracy. They assume that workers will be a consistently pro-democratic
force, while landlords will generally be an anti-democratic force. The other social classes
peasants, middle classes, the bourgeoisie could move in either direction depending on
historical contingencies. Hence, in the model, democracy hinges significantly on the relative
power of workers and landlords, and on the direction of the other class actors.

27
----------------------------
Figure 2 about here
----------------------------
Much of Capitalist Development and Democracy uses this model to explain particular
outcomes, especially the onset of democratic or authoritarian regimes in specific countries. In
offering these explanations, the authors implicitly introduce assumptions about the power and
direction of the various class actors. They then draw on historical narrative to justify these
assumptions. While the authors do not develop their model by assigning numerical values to
different actors, their qualitative assessments can be reconstructed in this fashion. For example,
Figure 2 offers a formal summary of the authors argument about Sweden and Italy (pp. 92-4,
103-6). The basic outcome to be explained is the development of democracy in Sweden versus
the onset of authoritarianism in Italy during the interwar period. The key difference between the
countries involves landed elites, the bourgeoisie, and middle classes. In both cases, landed elites
exerted full force in a negative direction, but their mass differed considerably: the Swedish
landed classes were tiny and had almost no mass (thus a power score of 1), while the Italian
landed elites were well developed and had considerable mass (thus a score of 5). As for middle
classes and the bourgeoisie, the difference concerned not their mass, but rather their direction;
that is, middle classes and the bourgeoisie were an anti-democratic force in Italy (-3 each) and a
neutral force in Sweden (0 each). All of these assumptions about the power and direction of
actors are backed up by considerable qualitative evidence.
By offering historical narratives that implicitly describe the relative mass, force, and
direction of class actors, Rueschemeyer, Stephens, and Stephens explain nearly forty cases of
democracy and authoritarianism. In this sense, even though the authors do not formally specify
28
the power of their actors with numerical values, their book illustrates how power theory can be
used to explain particular outcomes.
The authors also derive a number of testable propositions from the abstract model
presented in Figure 1 above. For example, they hypothesize that the size of the working class
will be positively correlated with democracy, while the size of landlords will be negatively
correlated with democracy. This hypothesis is built around the assumption that size is an
empirical indicator of the relative mass of workers and landlords. Likewise, because working-
class power will usually not surpass that of all other class groups combined, the authors
hypothesize that the working class will not be able to create democracy by itself, but rather will
require help from other classes. This reasoning leads to the bold prediction that pro-democratic
class coalitions that include the working class should be an almost necessary condition for full
democracy.
Several other variables should be correlated with democracy because they affect the power
of the working class or landed elites, or because they move other class actors in a particular
direction. For example, the presence of a powerful church that is strongly linked to landed elites
will be negatively correlated with democracy because it tends to weaken working class demands
for democracy and move other class actors in a negative direction. On the other hand, a state
apparatus that is autonomous from landed elites should be positively associated with democracy
because it tends to increase the force of working classes by enabling them to realize and act on
their interests.
In a similar vein, several variables should not be correlated with democracy because they
do not consistently alter power relations in one direction or another. For example, in contrast to
the expectations of much writing on democracy, the authors do not believe increases in the size
29
or strength of the middle classes will be positively associated with democracy. This belief is
grounded in the theoretical assumption that middle classes will not consistently gain or lose from
democracy. Likewise, the authors assume that war, political dependence, and economic
dependence will not consistently move social classes in one direction or the other, and thus will
not be correlated with democratization.
18

Finally, the authors use the power mechanism to integrate existing correlational knowledge.
Most importantly, they use their model to explain theoretically the longstanding statistical
correlation between economic development and democracy. They do so by suggesting that this
correlation exists because economic development alters the power of different class actors. In
particular, economic development generally increases the power of the working class and
weakens the power of landed elites, thereby activating the underlying process that generates and
sustains democracy. Although the association between development and democracy is but a
single correlation, it is probably the most important finding to emerge from the quantitative
literature on democracy, a finding that previously had been without a well-developed theoretical
explanation.
This discussion suggests that comparative-historical works are neither inductive nor
antitheoretical. Rather, these analyses implicitly follow an abstract theoretical program that is
not unlike rational choice theory in basic form. The problem is that comparative-historical
analysts, for all their many substantive contributions, have not said enough about how they use
abstract principles to guide their research. As a result, they have made it more difficult than
necessary to understand and evaluate the implicit power theory driving many of their arguments.


30

Future Applications of Power Theory
Power theory opens up exciting new possibilities for comparative-historical sociology and
other fields in which power can be meaningfully treated as an ultimate causal force. For one
thing, scholars can attempt to elucidate the general theoretical principles underlying classic
works in this area, such as Moores Social Origins of Dictatorship and Democracy (1966) or
Skocpols States and Social Revolutions (1979). Studies such as these have been criticized
precisely for their silence about ultimate theoretical principles (Kiser and Hechter 1991). By
locating these theoretical principles, scholars will not only better clarify the logic of the
arguments offered in these analyses, but they will help identify new models that could be used to
study other substantive topics. In this sense, a key agenda of power theorists should be to
accumulate a set of models equivalent to those that already exist in rational choice theory. The
advantage of working backward from existing studies is that the potential for empirical
application of the model is already addressed; that is, the analyst knows that the model can
actually be used in empirical investigation.
Reconstructing the implicit models used to generate empirical findings for entire bodies of
literature focused on particular outcomes is also a fruitful area of research. To take one possible
example, consider the literature on the origins of social revolutions (see Foran 1993; Goldstone
1980; Skocpol 1994 for reviews). Scholars working in this field have identified several
conditions that make social revolution more likely, such as the presence of a personalistic
regime, divisions among elites, and sustained fiscal crisis. Likewise, research has pointed to
conditions that make successful social revolution nearly impossible, including the presence of an
industrial economy or a regime type such as democracy, populism, or bureaucratic-
authoritarianism. To explain these correlations, one may hypothesize that social revolution is
31
ultimately driven by the power and direction of three basic actors: the state, dominant classes,
and subordinate classes. Countries with personalistic regimes are vulnerable to revolution
precisely because that kind of regime is associated with a weak state, the actor most committed
to preventing a revolution. Furthermore, personalistic regimes may disillusion dominant classes
and lead them to defend the existing order with only marginal force or perhaps even move in the
direction of promoting social revolution. By contrast, countries with industrialized economies or
military regimes are nearly invulnerable to revolution because of the overwhelming mass of the
state in comparison to subordinate classes who might lead revolutionary movements. Divisions
among elites are correlated with social revolutions precisely because these divisions diminish the
mass and force with which dominant classes might move to prevent revolution; and fiscal crises
have the obvious effect of weakening the mass of the state, thereby also encouraging revolution.
In short, power theory can be used to uncover the general theoretical principles that guide
empirical explanation in this field.
Power theory could make contributions to other literatures as well. For example, it might
help put Marxist analysis and feminist sociology on a stronger theoretical foundation. Although
Marxist scholars use both functionalist and rational choice mechanisms in their work (e.g.,
Cohen 1982; Elster 1982; Roemer 1982), they often do so reluctantly because these mechanisms
do not seem to fully capture the explanatory thrust of much of Marxs original writings. By
turning to power theory, analysts could revive the Marxian concern with actor capability and
force as the ultimate cause of societal development, an orientation that seems to provide a much
better fit with important parts of Marxs original program than either functionalism or rational
choice theory. Likewise, feminist work in sociology routinely conceptualizes social outcomes as
the product of aggregate group actors defined significantly by their relative power. If feminist
32
scholars explicitly draw on power theory, they might better connect their research to the social
science concern with assessing empirically testable hypotheses, and thereby more actively
influence the mainstream of sociology.
Literatures that seek to identify the sources of actors capabilities also lend themselves to
empirical work using power theory. For example, scholars working in the area of resource
mobilization theory have long hypothesized about the effects of organizational, coercive, and
other resources on the ability of social movements to move with force and mass toward certain
outcomes (see McAdam, McCarthy, and Zald 1996 for a review). These ideas have been
usefully supplemented by scholars who emphasize the role of non-material resources such as
cognitive orientations and cultural framings in the generation of social movement power. The
effort of scholars in the field of network analysis to uncover the relational sources of behavior is
also consistent with power theory (see Emirbayer and Goodwin 1994 for a review). In this case,
the power of an actor is conceived not as a categorical attribute, but rather as a product of the
actors social ties and position within a social network. These and other literatures can help
scholars develop better empirical measures for studying the force and mass of real actors ranging
from countries to movements to individuals.
Other extensions of power theory entail new modeling strategies. In some cases, for
example, it may be useful to model actors in a two-dimensional space that contains both an x and
y axis. With such a model, the outcomes of interest would be specified along two separate
dimensions, and actors would move according to the settings on a compass (e.g., 45 degrees west
of north). For example, scholars who work on the politics of social provision are often interested
in explaining both the public funding of welfare and the public delivery of welfare (e.g., Huber
and Stephens 2000). Yet, actors may not move in a consistent direction and with the same level
33
of power toward each of these outcomes. As a result, to model the mechanism generating
welfare outcomes, one must sketch a diagram with two axes and trace how the vectors for all
actors radiate from the center of the space along both dimensions according to their relative
power and direction. The final outcome will correspond to the point in space representing the
linear combination of all vectors.
19

It bears emphasis that scholars interested in using power theory in these or other ways can
learn from the experiences of functionalists and rational choice theorists. Most basically, they
must recognize that power theory does not contain any fully specified proposition, and it need
not be held up as a theory of everything (Wallerstein 2001). At the core of power theory is a
simple causal mechanism, one that asserts that the power and direction of actors ultimately
generate social outcomes. This causal mechanism is empirically empty and, by itself, embodies
no testable propositions. As a result, power theory is useful only insofar as it can be combined
with bridging assumptions to help inform empirical research. Rather than asserting the universal
application of power theory, therefore, analysts would be better off trying to identify a specific
class of social outcomes appropriate for the theory.

CONCLUSION


Major social theorists argue that the study of causal mechanisms can substantially improve
empirical work. Unfortunately, scholars have not clearly defined the concept of causal
mechanism in way that unambiguously differentiates the study of such mechanisms from
standard correlational research. In this paper, I have argued that causal mechanisms are
unobserved entities or processes that act as ultimate causal forces in the social world. By
themselves, causal mechanisms are inherently underspecified; to be of use, they must be given
34
empirical content though bridging assumptions. In conjunction with such bridging assumptions,
causal mechanisms can help scholars explain particular outcomes of interest, derive testable
hypotheses, and integrate existing correlational findings.
The mechanisms associated with three theoretical traditions were explored: functionalist
theory, rational choice theory, and power theory. Each of these mechanisms identifies a
particular entity and causal process that is understood to generate occurrences in the social
world. Functionalist theory sees underspecified social systems as the crucial entity of the social
world, and it holds that the needs of these systems ultimately produce major social outcomes. By
contrast, rational choice theory suggests that individuals are the essential entity of the social
world, and that the rational utilitarian behavior of these individuals is the social worlds ultimate
causal force. Finally, power theory argues that social actors are the most appropriate unit of
analysis, and that the power capabilities of these actors ultimately produce social outcomes.
I believe there is ample room for the use of multiple causal mechanisms in the social
sciences; one need not necessarily see functionalist theory, rational choice theory, and power
theory as direct competitors. Rather, each mechanism appears to be especially relevant for
certain substantive questions. With regard to functionalist theory, ideas about natural selection
are especially helpful in explaining social structures that are found widely across societies and
that have persisted over long periods of time. For example, natural selection at the level of social
systems may explain the nearly universal existence of nation-states with an economic division of
labor, government, private property, and organized religion. By contrast, rational choice theory
seems most appropriate for explaining outcomes in institutional settings where the rules of the
game are fixed and transparent. Within these settings such as legislatures, formal economic
markets, and bureaucracies one can more readily model the ways in which individuals make
35
choices based on careful utilitarian considerations. For its part, power theory is well-suited to
explain large-scale political and social outcomes in which a range of macro actors are relevant to
the analysis. For instance, the notion that outcomes depend on the relative power of aggregate
actors seems especially appropriate for topics such as national economic development,
revolution, and the formation of different kinds of states and regimes.
This discussion implies that scholars should first select a substantive research area and then
choose the causal mechanism that is most appropriate for that area. Yet, in practice, the causal
mechanism that scholars favor will likely lead them to study certain substantive questions and
not others. In other words, attraction to a causal mechanism often dictates the kind of
substantive research that scholars carry out rather than the reverse. One important reason why is
that causal mechanisms carry substantial normative entailments. Causal mechanisms are not
only tools for formulating and explaining correlations; they also serve as the basis for broad
worldviews, for overarching ideological and philosophical orientations. For example, rational
choice theory and power theory suggest quite different visions of the fundamental nature of
society, including how one might conceptualize a good society. These normative
underpinnings are precisely why scholars sometimes show a religious-like devotion to a
particular mechanism and seek to reveal its universal application. They are also a major reason
why scholars committed to a given mechanism may unfairly dismiss or otherwise not take
seriously empirical work that is grounded in an alternative mechanism.
Scholars would do well to separate the normative-philosophical implications of causal
mechanisms from their potential applications in empirical research. In other words, different
debates should be held over mechanisms as philosophical worldviews and mechanisms as tools
for stimulating empirical research. By doing so, scholars can state more precisely what they do
36
and do not endorse about a given mechanism. For example, it is possible that a particular scholar
will reject the normative implications of a mechanism but still find the mechanism useful for
addressing certain substantive topics.
If social scientists can successfully bracket the normative entailments of causal
mechanisms from their empirical entailments, they will be better positioned to follow the
example of the natural sciences, where theorists propose mechanisms without raising weighty
normative issues. Yet, even in the natural sciences, theorists debate exactly how many basic
forces exist in the world (Jasso 1988, p. 5). Although scholars conventionally asserted that
there are four basic forces in the physical world, some now believe that there may be five or even
six forces, and others insist that these various forces emerge from a single basic force.
20
In
sociological theory, it is worthwhile to explore whether functionalist, rational choice, and power
mechanisms could be reduced to a single ultimate mechanism (Harr 1970, chaps. 10-11). It is
also worthwhile for theorists to try to invent new mechanisms. Yet, for the time being, the most
important activity is that scholars link causal mechanisms with correlational analysis.
Sociological theory ultimately needs to be made empirically relevant.
37

References
Abell, Peter. 1994. Sociological Theory: What Has Gone Wrong and How To Put It Right, A
View from Britain. Pp. 105-120 in Formal Theory in Sociology: Opportunity or Pitfall?,
edited by Jerald Hage. Albany: State University of New York Press.
Aberle, D. F., et al. 1950. The Functional Requisites of a Society. Ethics LX (January): 100-
111.
Alexander, Jeffery C. 1998. Neofunctionalism and After. Malden, Mass.: Blackwell.
Almond, Gabriel. 1991. Rational Choice Theory and the Social Sciences. Pp. 32-52 in The
Economic Approach to Politics, edited by Kristen R. Monroe. New York: Harper Collins.
Althusser, Louis. 1969. For Marx. Translated by Ben Brewster. New York: Pantheon.
Barber, Bernard. 1956. Structural-Functional Analysis: Some Problems and
Misunderstandings. American Sociological Review 21:2 (April): 129-35.
Bates, Robert H., Avner Greif, Margaret Levi, Jean-Laurent Rosenthal, and Barry R. Weingast.
1998. Analytic Narratives. Princeton: Princeton University Press.
Bennett, Andrew, and Alexander L. George. 1997. Process Tracing and Case Study Research.
Paper presented at the MacArthur Foundation Workshop on Case Study Methods, Harvard
University, October 17-19.
Bhaskar, Roy. 1975. A Realist Theory of Science. London: Leeds Books.
Bhaskar, Roy. 1979. The Possibility of Naturalism: A Philosophical Critique of the
Contemporary Human Sciences. Atlantic Highlands, N.J.: Humanities Press.
Blalock, Hubert M. 1961. Causal Inferences in Nonexperimental Research. Chapel Hill:
University of North Carolina Press.
Boudon, Raymond. 1998. Social Mechanisms without Black Boxes. Pp. 172-203 in Social
Mechanisms: An Analytical Approach to Social Theory, edited by Peter Hedstrm and
Richard Swedberg. New York: Cambridge University Press.
Boyd, Richard N. 1973. Realism, Underdetermination, and a Causal Theory of Evidence.
Nous 7:1 (March): 1-12.
Buckley, Walter. 1957. Structural-Functional Analysis in Modern Sociology. Pp. 236-259 in
Modern Sociological Theory: In Continuity and Change, edited by Howard Becker and
Alvin Boskoff. Holt, Rinehart and Winston.
Cohen, Bernard P. 1989. Developing Sociological Knowledge: Theory and Method. Second
Edition. Chicago: Nelson-Hall.
Cohen, G. A. 1982. Reply to Elster on Marxism, Functionalism, and Game Theory. Theory
and Society 11: 4 (July): 483-95.
Collins, Randall. 1975. Conflict Sociology: Toward an Explanatory Science. New York:
Academic Press.
Cowen, Tyler. 1998. Do Economists Use Social Mechanisms to Explain? Pp. 125-146 in
Social Mechanisms: An Analytical Approach to Social Theory, edited by Peter Hedstrm
and Richard Swedberg. New York: Cambridge University Press.
Cullen, John B., and Shelley M. Novick. 1979. The Davis-Moore Theory of Stratification: A
Further Examination and Extension. American Journal of Sociology 84: 6 (May): 1424-
1437.
Dahrendorf, Ralf. 1958. Class and Class Conflict in Industrial Society. Stanford: Stanford
University Press.
38
Davis, Kingsley. 1948. Human Society. New York: Macmillan.
Davis, Kingsley. 1959. The Myth of Functional Analysis as a Special Method in Sociology
and Anthropology. American Sociological Review 24:6 (December): 757-772.
Davis, Kingsley, and Wilbert E. Moore. 1945. Some Principles of Stratification. American
Sociological Review 10: 242-49.
Denzin, N. K. 1990. Reading Rational Choice Theory. Rationality and Society 172-92.
Dessler, David. 1991. Beyond Correlations: Toward a Causal Theory of War. International
Studies Quarterly 35:337-355.
Dore, Ronald Philip. 1961. Function and Cause. American Sociological Review 26:6
(December): 843-53.
Durrant, A. V. 1996. Vectors in Physics and Engineering. London: Chapman and Hall.
Elster, Jon. 1982. Marxism, Functionalism, and Game Theory: The Case for Methodological
Individualism. Theory and Society 11:4 (July): 453-82.
Elster, Jon. 1989. Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University
Press.
Elster, Jon. 1998. A Plea for Mechanisms. Pp. 45-73 in Social Mechanisms: An Analytical
Approach to Social Theory, edited by Peter Hedstrm and Richard Swedberg. New York:
Cambridge University Press.
Emirbayer, Mustafa, and Jeff Goodwin. 1994. Network Analysis, Culture, and the Problem of
Agency. American Journal of Sociology 99:6 (May): 1411-54.
Fallding, Harold. 1963. Functional Analysis in Sociology. American Sociological Review
28:1 (February): 5-13.
Foran, John. 1993. Theories of Revolution Revisited: Toward a Fourth Generation?
Sociological Theory 11(1): 1-20.
Friedman, Debra, and Michael Hechter. 1988. The Contribution of Rational Choice Theory to
Macrosociological Research. Sociological Theory 6 (Fall): 201-18.
Friedman, Milton. 1953. Essays in Positive Economics. Chicago: University of Chicago Press.
Gambetta, Gudmund. 1998. Concatenations of Mechanisms. Pp. 102-124 in Social
Mechanisms: An Analytical Approach to Social Theory, edited by Peter Hedstrm and
Richard Swedberg. New York: Cambridge University Press.
Geddes, Barbara. 1991. Paradigms and Sand Castles in Comparative Politics of Developing
Areas. Pp. 45-75 in Political Science: Looking to the Future, volume two, edited by
William Crotty. Evanston, Ill.: Northwestern University Press.
Geddes, Barbara. 1995. Uses and Limitations of Rational Choice. Pp. 81-108 in Latin
America in Comparative Perspective: New Approaches to Methods and Analysis, edited by
Peter H. Smith. Boulder: Westview Press.
Geddes, Barbara. 1998. What Do We Know about Democratization after Twenty Years?.
Paper presented at the Annual Meeting of the American Political Science Association,
Boston.
Glennan, Stuart S. 1996. Mechanisms and the Nature of Causation. Erkenntnis 44: 49-71.
Goldstone, Jack A. 1980. Theories of Revolution: The Third Generation. World Politics 32
(April): 425-53.
Goldthorpe, John H. 2000. Causation, Statistics and Sociology. Pp. 137-60 in On Sociology:
Numbers, Narratives, and the Integration of Research and Theory, by John H. Goldthorpe.
Oxford: Oxford University Press.
39
Gould, Roger V., ed. 2000. The Rational Choice Controversy in Historical Sociology. Chicago:
University of Chicago Press.
Green, Donald P., and Ian Shapiro. 1994. Pathologies of Rational Choice Theory: A Critique of
Applications in Political Science. New Haven: Yale University Press.
Greene, Brian R. 1999. The Elegant Universe: Superstrings, Hidden Dimensions, and the Quest
for the Ultimate Theory. New York: W.W. Norton.
Hacking, Ian. 1983. Representing and Intervening. Cambridge: Cambridge University Press.
Hage, Jerald. 1994. Constructing Bridges between Sociological Paradigms and Levels: Trying
to Make Sociological Theory More Complex, Less Fragmented, and Less Politicized. Pp.
152-168 in Formal Theory in Sociology: Opportunity or Pitfall?, edited by Jerald Hage.
Albany: State University of New York Press.
Harr, Rom. 1970. The Principles of Scientific Thinking. Chicago: University of Chicago Press.
Harr, Rom. 1972. The Philosophies of Science. London: Oxford University Press.
Hedstrm, Peter and Richard Swedberg. 1998. Social Mechanisms: An Introductory Essay.
Pp. 1-31 in Social Mechanisms: An Analytical Approach to Social Theory, edited by Peter
Hedstrm and Richard Swedberg. New York: Cambridge University Press.
Hempel, Carl G. 1959. The Logic of Functional Analysis. Pp. 271-310 in Symposium on
Sociological Theory, edited by Llewellyn Gross. New York: Harper and Row.
Hempel, Carl G. 1965. Aspects of Scientific Explanation and Other Essays. New York: Free
Press.
Hernes, Gudmund. 1998. Real Virtuality. Pp. 74-103 in Social Mechanisms: An Analytical
Approach to Social Theory, edited by Peter Hedstrm and Richard Swedberg. New York:
Cambridge University Press.
Horowitz, Irving Louis. 1962. Consensus, Conflict, and Cooperation: A Sociological
Inventory. Social Forces 41 (December): 177-88.
Huaco, George A. 1986. Ideology and General Theory: The Case of Sociological
Functionalism. Comparative Studies in Society and History 28:1 (January): 34-54.
Huber, Evelyne, and John D. Stephens. 2000. Partisan Governance, Womens Employment,
and the Social Democratic Service State. American Sociological Review 65: 3 (June):
323-42.
Hume, David. 1748/1988. An Enquiry Concerning Human Understanding. Buffalo:
Prometheus Books.
Isaac, Jeffrey C. 1987. Power and Marxist Theory: A Realist View. Ithaca: Cornell University
Press.
Jasso, Guillermina. 1988. Principles of Theoretical Analysis. Sociological Theory 6 (Spring):
1-20.
Johnson, Chalmers. 1964. Revolution and the Social System. Stanford: The Hoover Institution,
Stanford University.
Kanazawa, Satoshi. 1998. In Defense of Unrealistic Assumptions. Sociological Theory 16:
193-204.
Keat, Russell, and John Urry. 1982. Social Theory as Science. London: Routledge.
Kelle, Udo, and Christian Ldemann. 1998. Pp. 112-25 in Rational Choice Theory and Large-
Scale Data Analysis, edited by Hans-Peter Blossfeld and Gerald Prein. Boulder: Westview
Press.
King, Gary, Robert O. Keohane, and Sidney Verba. 1994. Designing Social Inquiry: Scientific
Inference in Qualitative Research. Princeton: Princeton University Press.
40
Kiser, Edgar and Michael Hechter. 1991. The Role of General Theory in Comparative-
Historical Sociology. American Journal of Sociology 97:1 (July): 1-30.
Kiser, Edgar, and Michael Hechter. 1998. The Debate on Historical Sociology: Rational
Choice Theory and Its Critics. American Journal of Sociology 104:3 (November): 785-
816.
Koslowski, Barbara. 1996. Theory and Evidence: The Development of Scientific Reasoning.
Cambridge, Mass.: MIT Press.
Lenski, Gerhard E. 1966. Power and Privilege. New York: McGraw Hill.
Levi, Margaret. 1999. Producing an Analytic Narrative. Pp. 152-171 in Critical
Comparisons in Politics and Culture, edited by John R. Bowen and Roger Petersen.
Cambridge: Cambridge University Press.
Levy, Marion J., Jr. 1952. The Structure of Society. Princeton: Princeton University Press.
Lieberson, Stanley. 1985. Making It Count: The Improvement of Social Research and Theory.
Berkeley: University of California Press.
Lindert, Peter H., and Jeffrey G. Williamson. 1985. Growth, Equality, and History.
Explorations in Economic History 22: 341-77.
Little, Daniel. 1991. Varieties of Social Explanation: An Introduction to the Philosophy of
Social Science. Boulder: Westview Press.
Lenski, Gerhard E. 1966. Power and Privilege: A Theory of Social Stratification. New York:
McGraw-Hill.
Lukes, Steven. 1974. Power: A Radical View. London: MacMillan.
Mahoney, James. 2000. Path Dependence in Historical Sociology. Theory and Society 29:4
(August): 507-48.
Mann, Michael. 1979. Idealism and Materialism in Sociological Theory. Pp. 97-199 in
Critical Sociology: European Perspectives, edited by J. Freiberg. New York: Irvington
Publishers.
Mann, Michael. 1986. The Sources of Social Power: A History of Power from the Beginning to
A.D. 1760. Cambridge: Cambridge University Press.
McAdam, Doug, John D. McCarthy, and Mayer N. Zald, eds. 1996. Comparative Perspectives
on Social Movements: Political Opportunities, Mobilizing Structures, and Cultural
Framings. Cambridge: Cambridge University Press.
McMullin, Ernan. 1984a. Two Ideals of Explanation in Natural Science. Midwest Studies in
Philosophy 9: 205-220.
McMullin, Ernan. 1984b. A Case for Scientific Realism. Pp. 8-40 in Scientific Realism,
edited by Jarrett Leplin. Berkeley: University of California Press.
Merton, Robert. 1949. Social Theory and Social Structure. New York: Free Press.
Miller, Richard W. 1987. Fact and Method: Explanation, Confirmation, and Reality in the
Natural and Social Sciences. Princeton: Princeton University Press.
Moore, Barrington Jr. 1966. Social Origins of Dictatorship and Democracy: Lord and Peasant
in the Making of the Modern World. Boston: Beacon Press.
Moore, Wilbert E. 1978. Functionalism. Pp. 321-61 in A History of Sociological Analysis,
edited by Tom Bottomore and Robert Nisbet. New York: Basic Books.
Morton, Rebecca B. 1999. Methods and Models: A Guide to the Empirical Analysis of Formal
Models in Political Science. Cambridge: Cambridge University Press.
Munck, Gerardo L. 2001. Game Theory and Comparative Politics: New Perspectives and Old
Concerns. World Politics 53:2 (January): 173-204.
41
Nagel, Ernest. 1957. Logic without Metaphysics. Glencoe: Free Press.
Nagel, Ernest. 1961. The Structure of Science: Problems in the Logic of Scientific Explanation.
New York: Harcourt, Brace and World.
Nielsen, Francois, and Arthur S. Alderson. 1995. Income Inequality, Development, and
Dualism: Results from an Unbalanced Cross-National Panel. American Sociological
Review 60:674-701.
Olson, Mancur. 1965. The Logic of Collective Action. Cambridge: Harvard University Press.
Parson, Talcott. 1937. The Structure of Social Action. Glencoe: Free Press.
Parsons, Talcott. 1966. Societies: Evolutionary and Comparative Perspectives. Englewood
Cliffs: Prentice-Hall.
Porpora, Douglas. 1993. Cultural Rules and Material Relations. Sociological Theory 11:2
(July): 212-29.
Quadagno, Jill, and Stan J. Knapp. 1992. Have Historical Sociologists Forsaken Theory?:
Thoughts on the History/Theory Relationship. Sociological Methods and Research 20:4
(May): 481-507.
Roemer, John E. 1982. Methodological Individualism and Deductive Marxism. Theory and
Society 11: 4 (July): 513-26.
Rueschemeyer, Dietrich. 2001. Can One or a Few Cases Yield Theoretical Gains? Paper
presented at Conference on Comparative-Historical Analysis: Achievements and Agendas,
Harvard University.
Rueschemeyer, Dietrich, Evelyne Huber Stephens, and John D. Stephens. 1992. Capitalist
Development and Democracy. Chicago: University of Chicago.
Salmon, Wesley. 1984. Scientific Explanation and the Causal Structure of the World.
Princeton: Princeton University Press.
Schelling, Thomas C. 1998. Social Mechanisms and Social Dynamics. Pp. 32-44 in Social
Mechanisms: An Analytical Approach to Social Theory, edited by Peter Hedstrm and
Richard Swedberg. New York: Cambridge University Press.
Skocpol, Theda. 1979. States and Social Revolutions: A Comparative Analysis of France,
Russia, and China. Cambridge: Cambridge University Press.
Skocpol, Theda. 1994. Reflections on Recent Scholarship about Social Revolutions and How
to Study Them. Pp. 301-44 in Social Revolutions in the Modern World, by Theda
Skocpol. Cambridge: Cambridge University Press.
Somers, Margaret R. 1998. Were No Angels: Realism, Rational Choice, and Relationality in
Social Science. American Journal of Sociology 104:722-784.
Srensen, Aage B. 1998. Theoretical Mechanisms and the Empirical Study of Social
Processes. Pp. 238-266 in Social Mechanisms: An Analytical Approach to Social Theory,
edited by Peter Hedstrm and Richard Swedberg. New York: Cambridge University Press.
Steinmetz, George. 1998. Critical Realism and Historical Sociology. Comparative Studies in
Society and History 40:170-186.
Stinchcombe, Arthur L. 1968. Constructing Social Theories. Chicago: University of Chicago
Press.
Stinchcombe, Arthur L. 1993. The Conditions of Fruitfulness of Theorizing about Mechanisms
in Social Science. Pp. 23-41 in Social Theory and Social Policy: Essays in Honor of
James S. Coleman, edited by Aage B. Srensen and Seymour Spilerman. Westport, Conn.:
Praeger.
42
Stinchcombe, Arthur L. 1998. Monopolistic Competition as a Mechanism: Corporations,
Universities, and Nation-States in Competitive Fields. Pp. 267-305 in Social
Mechanisms: An Analytical Approach to Social Theory, edited by Peter Hedstrm and
Richard Swedberg. New York: Cambridge University Press.
Tilly, Charles. 1992. Coercion, Capital, and European States, AD 990-1992. Malden, Mass.:
Blackwell.
Tilly, Charles. Forthcoming. Historical Analysis of Political Processes. In Handbook of
Sociological Theory, edited by Jonathan H. Turner. New York: Plenum.
Tsebelis, George. 1990. Nested Games: Rational Choice in Comparative Politics. Berkeley:
University of California Press.
Tumin, Melvin M. 1953. Some Principles of Stratification: A Critical Analysis. American
Sociological Review 18:4 (August): 387-94.
Turner, Jonathan H., and Alexandra Maryanski. 1979. Functionalism. Menlo Park, Calif.:
Benjamin/Cummings Publishing Company.
Wallerstein, Immanuel. 1974a. The Modern World System I: Capitalist Agriculture and the
Origins of the European World-Economy in the Sixteenth Century. New York: Academic
Press.
Wallerstein, Michael. 2001. Does Comparative Politics Need a TOE (Theory of Everything)?
APSA-CP 12: 1-2, 31.
Wendt, Alexander. 1999. Social Theory of International Politics. Cambridge: Cambridge
University Press.
43
Table 1. Glossary of Definitions of Mechanism

I. Mechanism as a cause of an outcome

Boudon (1998, p. 172): A SM [social mechanism] is, in other words, the well-articulated set of causes responsible for a given
social phenomenon. With the exception of typical simple ones, SMs tend to be idiosyncratic and singular.

Cowen (1998, p. 125): I interpret social mechanisms (defined in greater detail below) as rational-choice accounts of how a
specified combination of preferences and constraints can give rise to more complex social outcomes.

Elster (1989, p. 3): nots and bolts, cogs and wheels that can be used to explain quite complex social phenomena.

Tilly (forthcoming): Mechanisms are events that alter relations among some specified set of elements.



II. Mechanism as an intervening process, event, or variable

Bennett and George (1997, p. 1): the processes and intervening variables through which causal or explanatory variables produce
causal effects.

Goldthorpe (2000, p. 149): some process existing in time and space, even if not perhaps directly observable, that actually
generates the causal effect of X on Y and, in so doing, produces the statistical relationship that is empirically in evidence.

Hedstrm and Swedberg (1998, p. 11): Mechanism-based explanations usually invoke some form of causal agent that is
assumed to have generated the relationship between the entities observed.

Hedstrm and Swedberg (1998, p. 13): Mechanisms . . . are analytical constructs that provide hypothetical links between
observable events.

Keat and Urry (1982, p. 30): causal explanations require the discovery both of regular relations between phenomena, and of
some kind of mechanism that links them. . . . In describing these mechanisms and structures we will often, in effect, be
characterizing the nature, essense, or inner constitution of various types of entity.

King, Keohane, and Verba (1994, p. 85): Some scholars argue the central idea of causality is that of a set of causal
mechanisms posited to exist between cause and effect. This view makes intuitive sense: any coherent account of causality needs
to specify how the effects are exerted.

Kiser and Hechter (1991, p. 5): A complete explanation also must specify a mechanism that describes the process by which one
variable influences the other, in other words, how it is that X produces Y.

Koslowski (1996, p. 6): A causal mechanism is the process by which a cause brings about an effect. A mechanism is a theory
or an explanation, and what it explains is how one event causes another.

Little (1991, p. 15): A causal mechanism, then, is a series of events governed by lawlike regularities that lead from the
explanans to the explanandum.

Mahoney (2000, p. 531): Causal mechanisms are the intervening processes through which one variable exerts a causal effect on
another variable.

Somers (1998, p. 726, citing Coleman [1986, p. 1328]): meaningful connection between events as the basic tool of description
and analysis.

Srensen (1998, p. 240): My definition of mechanism is simple: It is an account of how change in some variable is brought
about a conceptualization of what goes into a process.


44
III. Mechanism as an underspecified causal process

Elster (1998, 45, his emphasis): Roughly speaking, mechanisms are frequently occurring and easily recognizable causal
patterns that are triggered under generally unknown conditions or with indeterminate consequences.

Gambetta (1998, p. 102): I take mechanisms to be hypothetical causal models that make sense of individual behavior. They
have the form, Given certain conditions K, an agent will do x because of M with probability p. M refers either to forms of
reasoning governing decision making (of which rational choice models are a subset) or to subintentional processes that affect
action both directly (as impulsiveness) or by shaping preferences or beliefs.

Schelling (1998, pp. 32-33): I propose . . . that a social mechanism is a plausible hypothesis, or set of plausible hypotheses, that
could be the explanation of some social phenomenon, the explanation being in terms of interactions between individuals and
other individuals, or between individuals and some social aggregate.

Stinchcombe (1998, p. 267): I have defined mechanisms before as bits of sometimes true theory or model that represent a
causal process, that have some actual or possible empirical support separate from the larger theory in which it is a mechanism,
and that generate increased precision, power, or elegance in the large-scale theories.

Rueschemeyer (2001, p. 31): Incomplete theoretical propositions . . . a causal hypotheses, but one whose conditions are
insufficiently specified.



IV. Mechanism as an unobserved entity that necessarily generates outcome

Bhaskar (1979, p. 15): the construction of an explanation for . . . some identified phenomenon will involve the building of a
model, utilizing such cognitive materials and operating under the control of something like a logic of analogy and metaphor, of a
mechanism, which if it were to exist and act in the postulated way would account for the phenomenon in question.

Harr (1970, pp. 101, 102, 104): the structures, states and inner constitutions from which the phenomena of nature flow . . . . the
permanent or enduring conditions under which a certain kind of phenomenon will occur. The inner constitutions, structures,
powers, encompassing systems, and so on, of which natural generative mechanisms are constituted, and of which the connection
between cause and effect usually consists.

Steinmetz (1998, pp. 177-178): Within critical realism, a law is not a constant conjunction of events but the characteristic
pattern of activity, or tendency, of a mechanism. More specifically, real structures possess causal powers which, when triggered
or realized, act with natural necessity and universality as generative mechanisms. [Accompanying footnote: Generative
mechanisms are tendencies rather than powers because they are not just potentialities but potentialities that may be exercised
without being manifested.]

V. Other definitions

Hernes (1998, p. 74): A mechanism is a set of interacting parts an assembly of elements producing an effect not inherent in
any one of them. A mechanism is not so much about nuts and bolts as about cogs and wheels the wheelwork or agency by
which an effect is produced. But a mechanism or inner workings is an abstract, dynamic logic by which social scientists render
understandable the reality they depict.

Stinchcombe (1993, pp. 24-25): As I use the word, mechanism means (1) a piece of scientific reasoning which is independently
verifiable and independently gives rise to theoretical reasoning, which (2) gives knowledge about a component process (generally
one with units of analysis at a lower level) of another theory (ordinarily a theory with units at a different higher level), thereby
(3) increasing the suppleness, precision, complexity, elegance, or believability of the higher-level theory without excessive
multiplication of entities in it, (4) without doing too much violence (in the necessary simplification at the lower level to the
make the higher-level theory go) to what we know as the facts at the lower level.
45
Table 2
Bridging Assumptions, Mechanisms, and Testable Propositions



A. Using a Mechanism to Explain an Outcome


Assumptions formulated after Premise 1
conclusion is known; directly Premise 2
tested as hypotheses. Premise n
-----------------------
Conclusion that logically Mechanism Outcome
follows from premises;
not directly tested.




B. Using a Mechanism to Deduce New Propositions


Mechanism
Assumptions assumed to be Premise 1
true; not tested. Premise 2
Premise n
------------------------
New testable propositions Proposition 1
that follow logically from Proposition 2
premises. Proposition n




C. Using a Mechanism to Integrate Existing Correlational Findings


Mechanism
Assumptions assumed to be Premise 1
true; not tested. Premise 2
Premise n
------------------------
Existing correlations that Proposition 1
follow logically from Proposition 2
premises; already tested. Proposition n
46
Table 3. Modified Version of the Davis-Moore Argument


Basic Functionalist Mechanism: The social structures of a system exist because they are (or
were) functional for that system.

Social Structure to be Explained: Status and income stratification.

Conclusion to be Deduced from Premises: Differences in status and income are functional for
society.


Proof:

Premise 1: Individuals desire status and/or income.

Premise 2: Ceteris paribus, individuals will not occupy social
positions that require training.

Premise 3: Access to status and/or income can compensate for
training.

Premise 4: Society needs individuals to occupy social positions
that require different levels of training.

Conclusion: Differences in status and/or income are functional for
society.


47


Figure 1. Rueschemeyer, Stephens, and Stephens Balance of Class Power Model



Landed Elites

Bourgeoisie

Middle Classes

Peasants

Working Class



Authoritarianism Democracy


48
Figure 2. Model Specification for Sweden and Italy




Landed Elites


SWEDEN

Bourgeoisie
Solution:
-1v + 0v + 0v + 3v + 5v = 7v
(Democracy)
Middle Classes




Peasants




Working Class





Authoritarianism Democracy






Landed Elites


ITALY

Bourgeoisie
Solution:
-5v + -3v + -3v +3v +5v = -3v
(Authoritarianism)
Middle Classes




Peasants




Working Class




Authoritarianism Democracy
-1v
0v
0v
3v
5v
-5v
-3v
-3v
3v
5v
49


1
This mechanism constitutes the core assumption of rational choice theory as employed in the social sciences. See,
for example, Friedman and Hechter (1988, p. 202); Geddes (1995, pp. 83-85).
2
The best representative of functionalism is Davis (1948) and perhaps Levy (1952), not Merton or Parsons. Much
of the work of Parsons is not functionalist, including his The Structure of Social Action (1937) and his writings on
unilinear modernization (e.g., Parsons 1966). Likewise, as Huaco (1986, p. 38) suggests, Mertons (1949)
introduction of the idea of dysfunction and other modifications are not logically consistent with functionalism.
3
The division between empirical research and sociological theory is not new. For example, more than fifty years
ago, a concern with overcoming this division led Merton (1949) to advocate theories of the middle range.
4
This is true regardless of whether qualitative analysis (e.g., narrative) is used to analyze these intervening
processes.
5
There is some debate over the rules for introducing bridging assumptions that are not tested (see Morton 1999, p.
23, 37-8). For example, scholars disagree whether bridging assumptions must be realistic, either in the sense of
true or the in the sense of complete (see Friedman 1953 for a defense of untrue assumptions; see Kanazawa
1998 for a defense of incomplete assumptions).
6
Some theorists believe that this third strategy is less valuable than the second strategy because the analyst evaluates
the mechanism in light of empirical relationships that are known in advance (see Abell 1994; Cohen 1989; Hage
1994).
7
Causal mechanisms cannot be proven wrong. Their success depends above all on their capacity to inspire
hypotheses that are empirically confirmed (see Boyd 1973, p. 2; Jasso 1989, p. 4).
8
Confusion over whether functionalism is simply a tool to generate hypotheses or itself an explanation of outcomes
is found in several overviews of functionalist theory (e.g., Buckley 1957, pp. 242-43; Nagel 1957; see also the
discussions in Dore 1961; Fallding 1963; Hempel 1965).
9
There are other criticisms of functionalism, including especially its difficulty analyzing change and its bias toward
the status quo (see Barbar 1956; Hauco 1986; Horowitz 1962; Turner and Maryanski 1979). However, I am inclined
to believe that functionalism as a causal mechanism could be used to analyze change. For example, one could
formulate a model that begins with the following conclusion: Change X occurred because it was functional for
society. The challenge would then entail logically deriving this conclusion from a set of premises. As for the
50

conservative bias, it is worth emphasizing the close relationship between many Marxist programs of societal
transformation and functionalism (e.g., Althusser 1969; Cohen 1982; Wallerstein 1974).
10
Biologists debate whether these mutations are truly random. But that debate need not detain us here.
11
Durkheim himself seemed to support this position, but many subsequent functionalists assumed that traits initially
come into existence because of the functional consequences they serve. For a good discussion of the problems that
arose from this position, see Turner and Maryanski (1979), pp. 17-18, 119.
12
This kind of reasoning figures prominently (if only implicitly) in certain well-known works of contemporary
sociology, including Charles Tillys Coercion, Capital, and European States, AD 990-1992 (1992). Tilly seeks to
explain why national states proliferated in Europe at the expense of other forms of territorial organization, such as
city-states and empires of domination. To do so, he develops a functionalist explanation that highlights the adaptive
advantage of national states in an anarchical and war-ridden environment in which political units rose or fell
according to their ability to gain and maintain territory. The ability of national states to generate both capital and
coercion enabled them to spread through means of warfare and territorial occupation at the expense of both city-
states (which accumulated and concentrated capital, but not coercion) and empires of domination (which
accumulated and concentrated coercion, but not capital). Over time, national states proliferated to the point that they
appeared to be a natural form of organization, relegating other forms of political organization to the dustbin of
history.
13
Although Davis and Moore never formally justified this premise, one can argue that societies that failed to
develop a division of labor in which some social positions required more extensive training would have been
selected against given the threat of starvation and the need to survive in warfare.
14
In contrast to Davis and Moores position, evidence suggests that far less selection pressure exists for
contemporary societies than for societies of the past; that is, contemporary societies are not in danger of extinction
through warfare or starvation in a manner analogous to ancient agricultural and hunter-gatherer societies. However,
for the sake of deriving hypotheses, I treat this assumption as true.
15
In practice, scholars may combine these types.
16
It is possible that only one ultimate mechanism drives all social events in the world, in which case it would be
inappropriate to call for the development of multiple theories of causal mechanisms. This idea is explored further in
the conclusion.
51

17
There have been many efforts to develop useful definitions of power; good overviews can be found in Lukes
(1974) and Isaac (1987). Here I offer an abstract definition that is intentionally underspecified.
18
In any single country, war or dependency could greatly affect the prospects for democracy by shifting the power
and direction of class actors. However, across a large number of countries, the model predicts that there should be
no consistent relationship between these variables and democracy.
19
Power theory can also be used to model three-dimensional outcomes, in which case the analyst conceptualizes a
space with x, y, and z axes. For the most complex outcomes with multiple dimensions, separate models may need to
be formulated.
20
The four basic forces are: gravitational force, electromagnetic force, strong nuclear force, and weak nuclear force.
In the past, scholars often believed electricity and magnetism were separate forces with distinct carriers, but they
now generally agree that the two are different aspects of the same (electromagnetic) force. The most promising
candidate for a unified theory that reduces all forces to a single force is superstring theory (e.g., Greene 1999).

También podría gustarte