The Contribution of Rational Choice Theory to Macrosociological Research
Author(s): Debra Friedman and Michael Hechter
Source: Sociological Theory , Autumn, 1988, Vol. 6, No. 2 (Autumn, 1988), pp. 201-218
Published by: American Sociological Association
Stable URL: https://www.jstor.org/stable/202116
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Wiley and American Sociological Association are collaborating with JSTOR to digitize, preserve
and extend access to Sociological Theory
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
THE CONTRIBUTION OF RATIONAL CHOICE THEORY
TO MACROSOCIOLOGICAL RESEARCH*
DEBRA FRIEDMAN AND MICHAEL HECHTER
Russell Sage Foundation and University of Arizona
Because it consists of an entire family of specific theories derived from the same first
principles, rational choice offers one approach to generate explanations that provide for
micro-macro links, and to attack a wide variety of empirical problems in macrosociology.
The aims of this paper are (I) to provide a bare skeleton of all rational choice arguments;
(2) to demonstrate their applicability to a range of macrosociological concerns by reviewing
a sample of both new and classic works; and (3) to discuss the weaknesses of current
rational choice theory and the possibilities for its future development.
Until recently, microsociologists and macro-
Our aims in this paper are threefold: (1) to
sociologists have fundamentally disagreed provide the bare skeleton of all rational choice
about the proper level of analysis for
arguments; (2) to demonstrate their applicabilsociological inquiry. As members of these ity to a range of macrosociological concerns;
two camps increasingly have come to appre- and (3) to discuss the weaknesses of rational
ciate the theoretical and empirical payoffs ofchoice and the possibilities for its future
joint venture, however, there is a new call fordevelopment.
micro-macro research (Collins 1981; Hechter
A case for rational choice theory could be
1983; Coleman 1986a; Alexander, Giesen,
made on epistemological grounds, but we
Munch and Smelser 1987). One of the
shall not make it; such philosophical discus(several) starting points from which sions
it ishave flowered at least since the time of
possible to generate explanations that provide
Thomas Hobbes. Nor will we argue for its
for micro-macro links is rational choice
merits in accounting for social-psychological
theory.1 Beyond its reliance on the concept
of group processes; in sociology this
and small
purposive action (the idea that all action
is the province of exchange theory
has been
intentional) and its commitment to methodo(Heath 1976; Emerson 1981; Cook 1987).
logical individualism (the doctrine that
allwe argue for the development of the
Nor will
social phenomena are, in principle, theory
onlymerely for theory's sake; the search for
explicable in terms of the action of
Nash equilibria and for saddle points we leave
individuals)2, the properties of rational choice
to others. Instead, we shall argue that rational
theory are poorly appreciated by many
choice offers an alternative theory to explain,
sociologists.3
and sometimes to predict, empirical observations that have traditionally been of concern
to macrosociologists. In future contests with
* A previous version of this paper was presented at the
other kinds of explanations of these same
panel on Micro-Macro Linkages at the Annual Meetings
phenomena, we will be content to let
of the American Sociological Association, Atlanta, 1988.
empirical analyses tip the balance in favor of
We are grateful to Mary Brinton, Carol Diem, Douglas
Heckathor, David Jacobs, Edgar Kiser and two
one theory or the other4.
anonymous reviewers for their comments on an earlier
draft.
' The rational choice approach to the micro-macro
problem is epistemologically positivistic. Non-positivist
attempts to address this problem can be found in Giddens
(1979), Habermas (1983), and in some of the essays in
Alexander, Giesen, Munch and Smelser (1987).
2 Some rational choice analyses do not use individuals
as the units of analysis, per se, but use states or firms as
(corporate) actors. In so doing, theoretically they treat the
state or firm as if it were an individual while at the same
I. A SKELETAL RATIONAL
CHOICE MODEL
What follows is a basic skeleton of all rationa
choice models, inclusive of the bones found
in nearly every rational choice argument, and
exclusive of the flesh of any specific rational
time appreciating that in actuality states and firms do not
act.
4 We recognize that paradigms often appear impervi-
3For a sample of critical reviews of rational choice
ous to contravening data. Nonetheless, judging a theory,
written by sociologists, see Oberschall and Leifer (1986),
or set of theories, against the available evidence remains
and Hirsch, Michaels, and Friedman (1987).
the scientific ideal to which we subscribe.
Sociological Theory, 1988, Vol. 6 (Fall:201-218)
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
201
SOCIOLOGICAL THEORY
202
choice model. For the moment, the work
(andof
the preferences implied b
must
those who would rearrange the skeletal
framebe specified in advance.
prespecification
of actors' en
will be set aside. A diagram of the
main
points of the argument that followschoice
is to be explanations are liable to
cal.
found in Figure 1.
Figure 1 is intended to be a heuristic Yet
device individual action is no
product
of intention. It is a
to accompany the discussion in the first
and
constraints
second sections of the paper. The solid
lines that derive from at least two
independent
sources. The first set of conrepresent the explanatory paths of
typical
rational choice theories. Some theories,
for is due to a scarcity of resources.
straints
possession of and access to
example, link purposive actors Differential
to social
outcomes through the mechanism ofresources
opportu-make some ends easy for individuals to
attain, some more difficult, and
nity costs, while others link purposive
actors
preclude the attainment of others altogether.
with social outcomes through the mechanism
On account of scarcity, then, the opportunity
of institutional constraints, or preferences.
costs-those
The paths are numbered for the purpose
of costs associated with foregoing
the next most
identification. The aggregation mechanism
is attractive course of action-will
enclosed in broken lines to indicate that it is
often implicit in these theories.
vary considerably for different actors. Hence,
actors will not always choose that course of
action which satisfies their most valued ends.
In seeking to reap maximum benefit, actors
Basic Assumptions
keep a wary eye on opportunity costs, for
Rational choice models always rely on
conceptions of actors as purposive and
intentional. These actors are conceived to
these lower the net benefit of any contemplated action.
Social institutions can be a second source
of constraints.6 The modal individual will
have given preferences, values or utilities
find his or her actions checked from birth to
(hereafter termed preferences in this essay).
death by familial and school rules; laws and
They act with the express purpose of attaining
ordinances; firm policies; churches, syna-
ends that are consistent with their hierarchy of
gogues and mosques; and hospitals and
preferences. In general, rational choice theory
funeral parlors. By restricting the feasible set
is mute about what these preferences might be
of courses of action available to individuals,
and where they come from.5 In any specific
enforceable rules of the game-including
rational choice theory, however, actors' ends
norms, laws, agendas, and voting rules-
systematically affect social outcomes. These
5The issue of preference-formation is discussed
constraints serve to provide sanctions of both
further below.
a positive and a negative kind that raise or
lower the net benefit of any contemplated
course of action.
Therefore, within rational choice models,
variations in outcomes can be due logically to
6 There is considerable confusion about the relation-
ship between opportunity costs and institutional constraints, especially around the issue of whether institutional constraints are reducible to opportunity costs. We
believe that it is useful to maintain an analytic distinction
between the two. Institutional constraints are invariant
AGGREGATION MECHANISM /'
7
SOCIAL OUTCOME
across the individuals subject to them, while opportunity
costs often vary considerably for different individuals
subject to those same institutional constraints. To cast a
ballot in a U.S. presidential election, everyone must
register to vote, arrive at a polling place during specified
hours, and punch a computer card. These are institutional
constraints, and they affect all eligible voters. Yet the
opportunity costs associated with casting a balloton resources
such as occupation,
access to
Figure 1. The Various Paths depending
to Social
Outcomes
in
Rational Choice Explanations
transportation, literacy, and so on-vary widely.
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
RATIONAL CHOICE THEORY
203
(Gottfredson & Hirschi 1986) because, in
variations in preferences, in opportunity
costs, and/or in institutional constraints.7
We they have less to lose if they are
general,
will discuss examples of research demonstratapprehended (that is, they have lower opportunity costs). Finally, it has often been
ing each of these three sources of variation.
observed that hard science is a young person's
For example, Thomas Schelling's (1978)
model of residential segregation is based on game. This is because the opportunity costs
the assumption that individuals have a weak associated with retooling each time a technopreference for living with their own kind (this logical advance occurs are often prohibitively
kind of explanation corresponds to path 4 in high.
The third major source of variation in
Figure 1). This low level preference, he
shows, rapidly results in a chain of actions outcomes is due to institutional constraints
that leads to extreme segregation-a social (path 6). It is by now well appreciated that
outcome that is unintended by any of the different voting rules-majority, plurality, or
individual actors who participate in the unanimity-can influence both the time it
system. Take away the weak preference takes to arrive at a decision as well as the
assumption and Schelling's predicted out- actual decision itself (Buchanan and Tullock
1962). As any member of an academic
come disappears. Another example of a
model in which variation in social outcomes
department knows, the order of items on an
is attributed to variations in preferences
agenda
is
may have implications for which
illustrated in Anthony Downs's (1957:118-25)
items hold the day (Riker 1986).
discussion of the causes of two-party as Opportunity costs and institutional constraints are more often the cause of variation
against multi-party systems. For Downs, the
distribution of single-peaked voter preferin rational choice explanations than are
ences across a left/right political dimension
preferences.
is
Less is known about prefera major determinant of the number of parties
ences-about their origin, persistence, or
malleability-than about either opportunity
in a given society.8
Opportunity costs also can be the source costs
of or institutional constraints. There are
variation in outcomes (path 5). These are not
several reasons that the role of preferences
tends to be slighted in rational choice models.
distributed equally, either over every category
of actor, or for the same sort of action.
First, opportunity costs and institutional
Consider the relationship between age andconstraints are more reliably measured than
opportunity costs. Older people (especiallyare internal states. Second, while price theory
older women) tend to attend church more
undergirds opportunity cost arguments, and a
frequently than younger people do: the multitude of empirical generalizations support
opportunity costs associated with foregoing institutional constraint arguments, no compapossible salvation rise as the prospect of death rable theoretical or empirical structure supnears (Azzi and Ehrenberg 1975). In general, ports arguments based on preferences. Until
these costs drop in the post-child-rearing, we have a robust theory of preference-
post-retirement phase of life. Young people formation, or a rich body of data, the
engage in crime more often than their elders persuasiveness of explanations based upon
preferences will hinge on reader's perceptions
of their intuitive appeal.
There are two other elements common to all
7 Even if preferences, opportunity costs and the effects
of social institutions all could be exactly specified, some rational choice models. The first is an
unexpected variation in outcomes might still be noted due
principally to the effects of uncertainty, either objective aggregation mechanism by which the separate
or subjective. One anonymous reviewer suggested that individual actions are combined to produce
strategic interaction also might be an independent source the social outcome. Often rational choice
of variation. For us, however, it is not an independent theories read as if there were no aggregation
source but an important subset of the category of social
problem.9 This is because in most of the
institutions. Strategic interaction describes the action of
participants in a game, and a game is a social institution.
8 The other main determinant is an institutional
constraint that Downs called the electoral structure
9 In a very limited set of cases (especially in
(basically, whether there are single-member districts
or
legislatures),
an institutional constraint carries with it an
proportional representation). It is significant that
subseaggregation
mechanism-thus any kind of voting rule
quent rational choice research on the determinants
of a collective decision. Many of the institutional
will yield
party systems has focused almost exclusively onconstraints
the role
that macrosociologists are concerned with do
not share this feature, however.
of institutional factors as against preferences.
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
204
SOCIOLOGICAL THEORY
original applications the market is taken
to be (Path 4), opportunity costs (Path
preferences
5), or institutional constraints (Path 6). The
the mechanism that aggregates preferences.
bulk
of our discussion therefore concerns
In non-market applications there is an
assumpissues.
tion, which is frequently hidden,these
that
all
Theover
studies that are described in this
actors have similar preference orderings
section constitute
the best evidence we can
a given set of choices. The usual defense
of
this assumption relies on the law muster
of large
to try to convince a skeptic of the
numbers (Stinchcombe 1968:67-8; Hechter
merit of this approach. (Those who do not
1987:31-3).
need convincing may wish to proceed directly
to Section
The alternative to making an assumption
ofIII.) The rational choice theories
homogeneous preference orderings
that is
areto
utilized in these studies compete with
specify an aggregation mechanismestablished
that regeneral theories (Marxist or structural theories, for instance) and with a wide
flects the preferences of the constituent
actors. It has been difficult to formulate such
variety of inductive theories in accounting for
a mechanism: when actors have heterogetheir subject. In attempting to secure a place
neous preference orderings, no determinatein macrosociology alongside these other
collective outcome may be able to be
theories, many of these studies analyze
predicted (Arrow 1951). The most extensive
classic empirical problems, the outcomes of
work on this problem has been done as part of
which are well-known. These are, then, post
the scholarship on voting. The work on thehoc explanations. Yet in the competitive
Condorcet paradox is relevant here (for an
arena of macrosociological theory it is
overview, see Mueller 1979: Ch. 3).
essential that a general theory be applicable to
The final element is information. Initially, the historical phenomena that are the staple of
this branch of sociology (economic develophad perfect or sufficient information necesment, voting patterns, protest movements,
stratification). When rational choice theories
sary for making purposive choices among
alternative courses of action. In much of the
have demonstrated their worth in elucidating
most recent work, however, the quantity and
these subjects, the case for rational choice as
a predictive theory then can be persuasively
quality of available information is taken to be
a variable, and a highly significant one at that
advanced. Some of this predictive work is
(Hirshleifer and Riley 1979).
already underway, and is also detailed in the
following pages.
rational choice models assumed that actors
II. EXEMPLARY AND NEW RATIONAL
CHOICE EXPLANATIONS OF
MACROSOCIOLOGICAL PROBLEMS'l
Examples of Research Investigating Path 4
(Hierarchy of Preferences -- Social
Outcome)
If this is the skeleton upon which all rational
There
are as many hierarchies of preferences
choice arguments are built, how can
such
as there are actors in the world. On the one
explanations be used to illuminate macrosohand, this allows for infinite gains from
ciological problems? We will illustrate their
exchange. On the other hand, when a narrow
applicability by discussing examples of re-
range of ends is considered, large numbers of
people may share similar preferences. Whether
plary studies and from new ones-illustrating
a given
social outcome results from exchange
the various paths represented in Figure
1.
or from commonality, the dynamic element in
Thus far most research has concentrated on
these kinds of analyses is the content of the
the connections between social outcomes and
preferences held by relevant actors.
search-drawn from both classical or exem-
a. The similarity of American political
party platforms. Why do the American
choice theory by sociologists and for sociologists.
Democratic
and Republican parties tend to
Because the use of rational choice theory in sociology
is
adopt virtually indistinguishable platforms
relatively new, however, work that serves as exemplars
10 This paper explicitly focuses on the uses of rational
in this tradition was generally not done by sociologists.
relative to the parties in Western European
So we draw on the work of economists and political
democracies, which seem to have much more
scientists to provide classic or exemplary models of
distinctive political programs?
scholarship, but only on the work of sociologistsideologically
to
One possible explanation is the lack of
illustrate the range of new applications to macrosocio-
logical phenomena.
sophistication of American voters relative to
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
205
RATIONAL CHOICE THEORY
parties strive to distinguish their party plattheir European counterparts. Another ascribes
the difference to the lower salience of class
forms and to maintain the purity of their
identification in the United States as againstpositions. 12
the European democracies (Campbell et al. b. The origin of collective action. In 1973
1966).
James S. Coleman published The Mathematics of Collective Action (Coleman 1973) in
The exemplary rational choice explanation,
which he proposed a market solution to the
as presented by Anthony Downs (1957:Ch.
problems of coordination and information
8), holds voter preferences to be responsible
for the type of party system (two- or
exchange that occur in every instance of
multi-party system), together with the type of collective action. In any given group, some
electoral rules (single-member district or individuals will prefer to pursue one collecproportional representation). In Downs' model, tive good while others will prefer to pursue
(1) voters support a given party because they another. Yet in most instances, the efforts or
expect that enactment of its policies will make resources of those who are less interested are
them better off; (2) legislators are principally as important to the attainment of the good as
interested in re-election; and (3) parties are are the efforts of those who are most
teams seeking to control the governing interested. Since all of the actors appreciate
apparatus by winning national elections. Each that successful collective action (that which
party wants to garner the maximum number they are interested in, as well as that which
of votes. Using Hotelling's (1929) general others are interested in) requires the contribu
theory of optimal location, Downs explains tions of a wide range of group members, the
that in a two-party system where all voters sometimes seek to exchange their control ove
vote, each party can attract the maximum events. Thus, there is a supply of control ove
number of votes only by appealing to the a given collective action held by those le
center of the ideological spectrum. Thus, interested in the outcome, and a demand fo
each party will intentionally alter its platform that control emanating from those with
to attract the median voter. " Not only will the significant interest in the outcome. Thi
parties come to resemble Tweedledum and market solution leads to predictions about th
Tweedledee, but their platforms will be as probability of a successful collective action a
ambiguous as they can manage.
well as to predictions about the expected
In contrast, multi-party systems create value of an event for involved actors. Unlike
conditions favorable to one party at each most treatments of collective action, Coleideological mode (and perhaps to balancing man's contribution recognizes that most
parties between modes). In such systems instances of collective action involve collective decision-making processes. (For other
papers written from these theoretical pre1 Downs presents arguments both for how electoral mises, see Marsden [1981] and Coleman
rules can be determined by voter preferences and for how [1986b].)
voter preferences can be determined by electoral rules.
What can be taken from these studies as
On the one hand, if the distribution of voter preferences
is normal (that is, unimodal and symmetric), it is likely
general lessons about the relationship between
single-member districts are preferred, and these, in turn,
noted that the rational choice perspective
that no large group will be ignored politically. Thus, preferences and outcomes? First, it can be
tend to produce two party systems. If, however, voter directly challenges the sociological tradition
preferences are polymodal, lawmakers may choose
on this subject. The effect of preferences on
multi-member districts-that is, proportional representation-in order to allow sizeable extremist groups to have social outcomes long has been a central
a voice. This tends to produce multi-party systems. On concern in sociology. The rise of the sample
the other hand, Downs also argues that the number of survey-surely a major achievement of modparties is likely to affect the distribution of voter ern sociology-was, after all, predicated on
preferences: in a two-party system, voter preferences are
likely to converge in the long run, whereas the opposite
effect may occur in a multiparty system with proportional
representation. The relationship between preferences and
number of parties also depends upon assumptions
the notion that given better information,
government officials-as well as the producers of other kinds of goods--would formulate
policies-or other products-that were tail-
regarding voter turnout. So long as all voters vote, the
ored to preferences, as reflected in public
parties will cater to the median voter regardless of the
preference distribution. If some voters abstain, however,
12 For more recent work in this tradition, see Enelow
then the median voter result is upset if the preferences are
asymmetric or multimodal (Mueller 1979:98-105).
and Hinich (1984).
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
206
SOCIOLOGICAL THEORY
opinion polls. In contrast, these rational
Rational choice theory offers quite a differchoice studies suggest that preferences
rarely,
ent
explanation for the failure of socialist revif ever, determine social outcomes
in any
olutions
to occur in advanced capitalist societies. A socialist revolution is in the collective
such simple, intuitive fashion.
interest of those workers who believe that their
Schelling's (1978) study of residential
segregation, for example, implies that
(1) the
welfare
will be improved by it. Yet, even though
preferences that are reflected in social
theyoutare potential beneficiaries, rational workcomes are likely to be more extreme
than
ers will
be unlikely to help bring about a sorevolution. Since each individual makes
those held by the constituent actors.cialist
Downs's
(1957) analysis implies that, at least
ina the
only
negligible difference to the likely suc-
cess of the revolution, the best individual stratAmerican political context, (2) a political
party that is self-consciously responsive
thethat combines non-participation with
egy to
is one
preferences of a constituency will produce
the (hoped-for)
a
consumption of the fruits of the
new order. If, however, most workers follow
vapid set of policies. And from Coleman's
(1973) work on collective action comes
the
this individualist
logic and actually do not parcounterintuitive finding that (3) ticipate
in prein the revolutionary movement, then it
is clear
that few such social transformations
existing groups, successful collective
action
will ever occur.
depends on mobilizing the disinterested,
rather than those whose interest is strongest
This explanation for the failure of revoluand least ambiguous.
tion in capitalist societies is based on a
general rational choice theory known as the
theory of public goods. As formulated by
Mancur
Examples of Research Investigating Path
5 Olson (1965), a public good is one
that,
(Opportunity Costs - Social Outcome) once produced, can be consumed by
anyone in the relevant public. The difficulty
withvary
public goods is that rational individuals
Like preferences, opportunity costs
will consume them without contributing to
across individuals. Nonetheless, individuals
sharing social characteristics-age, sex, mar- their provision. In other words, rational
ital status, income-often have roughly com- individuals will act as free riders. So if a
parable opportunity costs as well. What is given group is composed of a large number of
common to these explanations is the idea that free riders, then a less than desirable amount
social outcomes derive from the alternative
of the public good will be produced. This
courses of action available to similarly
kind of logic can account both for the rarity of
situated actors.
collective actions in the absence of organiza-
tions providing significant private benefits (or
a. The failure of class consciousness. One
what Olson calls selective incentives), and for
of the oldest predictions of social science
its
holds that people who occupy structurally likelihood in the presence of such
similar positions will act collectively to organizations.14
further their common interests. This expecta- b. Explaining variations in patterns of
collective protest. Working to propose a
tion underlies Thomas Hobbes' explanation
for the rise of the state, Adam Smith's
version of rational choice theory closer to the
richness of the empirical evidence on collecaccount of the existence of protectionist
tive action, Karl-Dieter Opp (1986a; 1986b)
tariffs, and Karl Marx's dictum that advanced
suggests that the selective incentives needed
capitalist societies would undergo socialist
revolutions. Yet there are many instances to
ininduce participation in protest-of both the
which this principle does not hold. Forlegal and illegal varieties-are of two types.
The first includes external selective inceninstance, what has kept Marx's prediction
tives, including normative expectations of
from being fulfilled? One possibility is that
the failure of socialist revolution is due to
'false consciousness.'13
others, and positive and negative sanctions.15
not place their scholarship in the rational choice
tradition).
13 It should be noted that (1) some recent Marxist
14 The best overall summary of the literature in
scholarship (Elster 1985; Przeworski 1985) uses rational
response to Olson is Hardin (1982); see also Oliver,
Marwell and Teixeira (1985).
choice arguments, as well; and (2) many other Marxist
15 There is some debate in rational choice circles about
scholars rely on structural constraints of all sorts to
explain the absence of revolution (even though they
thedo
consequences of countenancing non-material incen-
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
RATIONAL CHOICE THEORY
207
In addition, there are internal incentives,
expected on the basis of Olson's theory. In a
including norms of participation, norms about related article Oliver (1980) has argued that
violence, intrinsic value, entertainment value, there are positive and negative selective
and catharsis value. Together these go far in incentives and that these have different
explaining participation in anti-nuclear protest implications for producing collective action:
activities in the United States and in West
rewards are effective in motivating small
numbers of participants while punishments
Germany by elucidating the relationship
between community integration and participaare effective only when the situation requires
unanimous cooperation.
tion. Opp (1986a; 1986b) tests these hypothd. Political influence and tax rates. In a
eses derived from rational choice theory and
holds them to be superior to those derived
study of political influence and tax rates,
from other approaches-relative deprivation,
David Jacobs (1988) traces one implication of
resource-mobilization and demographic modthe rational choice theory of collective
els-in accounting for observed empirical
action-namely, that economic concentration
relationships.
should bring forth political influence. When
Also dedicated to explaining variousassets
in- are concentrated and large proportions
stances of collective action is an elaborate
of the productive resources in one economic
rational choice model of social conflict in
sector (such as manufacturing) are controlled
Anthony Oberschall's 1973 book, Social
by a relatively small number of firms, each
Conflicts and Social Movements. Expanding large firm will have a greater stake in the
on the work of Olson (1965), Oberschall
provision of a public policy that benefits all
the firms in that sector. For under these
(1973) focused on recruitment processes that
conditions, each firm will be better able to
facilitated movement mobilization and ap-
plied these hypotheses to a wide field of
monitor and sanction potential free riders.
collective actions, most notably to the CivilJacobs argues that tax rates offer an indicator
Rights movement.
of political influence. Since firms cannot
More recently, Oliver, Marwell and Teieasily pass the cost of increased taxes on to
xeira (1985) have investigated the effect consumers,
of
when they pay greater taxes it is
group heterogeneity on production functions
unlikely that this is due to their expanding
for collective action. They distinguish two
political influence. But if effective tax rates
types of production functions-accelerating
diminish in response to an increase in
and decelerating ones-and show that colleccorporate resources, this suggests that the
tive actions characterized by accelerating
corporations in question have succeeded in
production functions hardly ever succeed in
translating their resources into political influence. Jacobs finds that the concentration of
getting off the ground, while those characterized by decelerating production functions are
assets among the largest firms (regardless of
more likely to begin, but yield reducedtheir industry) is an important predictor of
marginal returns of subsequent contributions.
effective tax rates on manufacturing.
c. Rewarding and punishing participants ine. Political regionalism. Why do the
collective action. Pamela Oliver and Gerald
peasants in western France vote for right-
Marwell, together with their associates,wing
have
parties, whereas those in the Mediterra-
sought to expand on Olson's theory of
nean region vote for left-wing parties? In
collective action in order to bring it more into contrast to prevailing interpretations that
line with empirical evidence about the
emphasize such normative factors as the
probability of collective action. In a series oftransmission of traditional party loyalties,
experiments, Marwell and Ames (1980) havecollective memories, or religious values,
demonstrated that contributions to collective
William Brustein (1988a) links specific re-
action were far more prevalent than would gional
be
structures to the actors (here, the
peasants) who are subject to them. His theory
proposes that different modes of production
tives, particularly admittedly fuzzy ones like norms.
Some argue that adding non-material incentives brings
are chiefly responsible for political regional-
these theories dangerously close to the precipice of
ism in France (see also Brustein 1988b). The
tautology. Others-like Opp-argue that the most
mode of production shapes voting behavior
important criterion for a theory is whether it captures
social reality and if non-material incentives are a key to both directly and indirectly. Individuals'
participation in protest, then they must have a place in a perceptions of their interests are derived
theory of protest (Klosko, Muller and Opp 1987).
directly from their economic activity and
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
SOCIOLOGICAL THEORY
208
property rights, and are determined indirectly
by the dominant structure of the economic
activity and property rights they possess. The
social outcome of these processes is that
particular modes of production of western and
Mediterranean France have produced divergent regional constellations of economic
interests. Right-wing voting is fostered by a
district, and whether the vote is on a
procedural or a substantive issue) contribute
to differences in party solidarity within
political systems.
g. The corruption of officials. Why do the
legislative representatives of western European socialist parties so often compromise the
interests of their constituents by supporting
mode of production characterized by a
bourgeois initiatives (Michels [1911] 1962)?
subsistence-oriented economy, medium-to- Why are narcotics police routinely on the take?
large sized tenancy or sharecropping, popula- And why are prison guards so often corrupted
tion dispersion, and the presence of social by the prisoners they guard (Sykes 1958)? One
elites. In contrast, left-wing voting is fostered kind of explanation suggests that those who
by a mode of production characterized by a are so easily corrupted were morally defective
market-oriented economy, small-scale owner- in the first place. Once again, rational choice
cultivation or salaried agricultural labor, logic offers a different sort of explanation for
population agglomeration, and the absence of the corruptibility of officials.
Consider the case of the prison guards.
social elites.
f. The solidarity of political parties. What
Whereas the employers (the wardens) direct
them to maintain order within the walls, the
explains cross-national variations in the soliguards can do this in two quite different
darity of major political parties? Or variations
in solidarity among parties of the sameways. Guards can enforce prison regulations
to the hilt (as the wardens would like them to
nation? Or variations in party solidarity from
do), or they can make deals with their
one legislative session to another? Whereas
prisoners. Why would they choose the latter
many previous attempts to answer these
strategy over the former? The simple reason is
questions emphasized the ideology of party
members, in Michael Hechter's book, Princithat the net benefit of making deals is likely to
exceed that of enforcing the rules because it
ples of Group Solidarity (1987), these questions are taken to be an instance of the more
provides an additional source of income
without threatening the loss of guards' jobs.
general problem of group solidarity.
Roll-call voting provides a means to assess Why then are the guards not punished for this
variations in party solidarity. The roll-callegregious behavior? Only because wardens
vote compels the legislator to make decisions have no means of monitoring the guards'
which test his or her compliance with party corruption. This case is a specific instance of
obligations. Since it is voters rather than the a more general phenomenon known as the
party leaders who can sanction the represen- principal-agent problem (Jensen and Mecktative by preventing his or her re-election, ling 1976). In a principal-agent relationship
differences in party control capacity are apt toagents will act in the principals' interests only
when they are sufficiently rewarded or
to do so. Hence principals' ability
variation in solidarity is not due to constrained
the
to monitor
agents is a key determinant of
different control capacities of the parties,
but
be minimal. Thus most of the observed
agents' compliance with principals' demands.
rather to differences in members' dependence
on their party. The greater the value h.
ofThe autonomy of Absolutist kings.
Explanations
of the role of Absolutism in the
party-derived benefits that aid legislators
in
transition from feudalism to capitalism have
their quest for re-election and career advance-
typically
ment, the greater their dependence on
the concentrated either on internal
political
party. Conversely, the greater the value
of and cultural dynamics of these
benefits available to them from alternative
societies, or on exclusively economic aspects.
Seeking to provide a link between these two
sources, the lower their dependence on the
kinds of arguments, Edgar Kiser (1987)
party. There are many factors which would be
inquires into crown autonomy, state policies,
expected to influence legislators' dependence
and economic development in Western Euroon the party. Some of these, such as
pean Absolutisms. Using principal-agent thelegislative rules, lead to differences in party
solidarity between political systems (or in aory, he argues that variation in autonomy of
Absolutist rulers is explained by differences
single political system, over time), and others
in the relative resources of nobles and
(including incumbency, characteristics of the
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
RATIONAL CHOICE THEORY
209
monarchs and in the control capacities
ofgames circumscribing actors'
strategic
nobles. These variations lead, in turn, to
choices-like Prisoners' Dilemma games-to
different state policies, policies which have contractual forms, to stratificational barriers
important ramifications for economic developto mobility. What is common to all of these
ment. In England and France, both countriesarguments is that the choice set is limited by
with ample resources, high crown autonomy
factors that are beyond the control of the
generally hindered economic development. In
actors themselves.
the less advantaged countries of Sweden and
a. The superpower arms race. Since the
Spain, however, high crown autonomy was a
leaders of both the United States and the
necessary but insufficient condition for ecoU.S.S.R. are well aware that nuclear stockpilnomic development.
What can be concluded about the effect of
ing is potentially destructive of the entire
world, why do they continue to participate in
opportunity costs on social outcomes from
these various studies? The most important
a nuclear arms race? One popular answer is
insight is that (1) social structural commonalthat key decisions in both countries are made
ity alone is insufficient to produce collective
by military-industrial elites whose private
action. Another insight about collective action
interests differ from that of the public, here
that can be drawn from Olson's (1965) book,
including the world at large (see Pursell 1972;
as well as from other studies in the same
Sarkesian 1972; and Rosen 1973). Another
tradition (Opp 1986a, 1986b; Oliver, Marwell
kind of explanation is that both the American
and Texeira 1985) is that (2) in amorphous
and Soviet economies depend on military
groups (those without well-defined boundexpenditures
to avoid recessions or depresaries), successful collective action depends on
sions and that this functional necessity
mobilizing potential beneficiaries. The exper-
iments of Marwell and Ames (1980) and
inevitably leads to an arms race (Boulding
Oliver's (1980) research implies, further, that 1973; Udis 1973; Leontieff and Duchin
(3) rewards and punishments do not follow1983).
expected patterns in motivating behavior Quite a different kind of answer is
oriented toward the attainment of a collective
suggested by rational choice logic. The
good. Brustein's (1988a) study, which ex-relationship between superpowers may be
plains why the political behavior of peasantsanalyzed as a Prisoner's Dilemma game
is so heterogeneous, provides graphic evi-
(Luce and Raiffa 1957). In this kind of game
dence that (4) many structural categories (e.g.
the costs or benefits of each player's
the peasantry) are so broadly defined that they
strategy -armament or disarmament - depend
are inadequate to explain political outcomes.
The structural parameters that actually will on the opponent's strategy. The best joint
affect outcomes can be determined only by solution (meaning the one that is most
carefully considering the individual's oppor-beneficial to both players simultaneously) is
tunity costs. One lesson inspired by Hechtermutual disarmament. Yet the best strategy for
(1987) is that (5) dependence overcomes each individual player is to arm while hoping
ideology. From Jensen and Meckling (1976)that the other disarms. Since in rational
we can infer that (6) one cannot automaticallychoice theory self-interest is generally
rely on wages to produce compliance. Fi-presumed, each player will follow this
nally, Kiser's (1987) research reminds us thatstrategy, and both will end up arming (Brams
(7) if the economic well-being of a country is1985).
of concern, rulers with large amounts of It turns out that this solution will hold even
resources at their disposal cannot be left alone
to rule.
when both leaders prefer to disarm. The
reason is that neither can be sure about the
other's intentions and actions (witness the
emphasis on verification in the arms talks),
and lacking any assurance, each seeks to limit
Examples of Research Illustrating Path 6
(Institutional Constraints -- Social Outcome) losses by arming. The non-cooperative strategy wherein both countries arm-either to
A plethora of institutional constraints canmaximize their gains or limit their lossesalter social outcomes. These range from
produces an equilibrium wherein neither
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
210
SOCIOLOGICAL THEORY
player has an incentive to deviate unilaterally,
attainment. Guidance and sponsorship are
more crucial
therefore to individual sociofor each either would do worse, or would
do
no better. 16
economic success in Japan than in the United
b. The tenacity of gender stratification
States, but since educators and employers are
patterns. In contrast to traditionalmotivated
Marxianto sponsor individuals who are sure
and Weberian accounts of the persistence
of as committed employees in firms
to remain
stratification patterns, Mary Brinton
(1988)
across
time, and since women are less likely
seeks to explicate the mechanisms that
to meet such a qualification than men (due to
perpetuate gender stratification. She chooses family responsibilities), then this system
two industrial societies with different socialimposes greater disadvantages. Additionally,
institutional patterns, norms, and degrees when
of
parents perceive that employers engage
gender stratification-Japan and the United
in sex discrimination, they will be motivated
States-and argues that distinctive features to
of invest, psychologically and economically,
the school-labor market configuration in
in the education of sons rather than daughters.
Japan condemn Japanese women to more
In these ways employers and parents unwit-
severe gender stratification than their Ameritingly conspire to perpetuate gender stratificacan counterparts. Although education istion in education and employment.
essential for status attainment in each society,c. The production of insurance contracts.
both the type of education necessary for such
Last, Carol Heimer inquires as to the origins
of different kinds of social institutions-in
attainment and the conditions for entry into
these institutions differ in important ways.
this instance, insurance contracts. She argues
1985) that if the reactiveness of risk
Employment in high-status firms in Japan (Heimer
is
is a constraint on insurance markets, and if it
reserved largely for the graduates of universities who receive a general education and,
shifts according to different types of losses,
after hiring, are subsequently given considerthen the contractual arrangements associated
able on-the-job training. But admission with
to
these kinds of losses should vary
universities in Japan is determined irrevocasystematically on this dimension. By effecbly in late adolescence. Youths who fail tively
to
transforming reactive risks into fixed
ones, such contracts make insurance markets
earn a place in a Japanese university typically
possible. Heimer shows that this is indeed the
forego this kind of employment forever. In
the United States, however, much of the
case by comparing the contractual arrangements in fire and marine insurance, and in
technical training for specific jobs occurs in
the educational institutions, entry to whichfidelity and surety bonding.
can occur at many different points in a What, then, can be learned about the
relationship between institutional constraints
person's life.
In consequence, the Japanese institutionaland social outcomes from these studies? In
rational choice suggests that the
structure gives educators and employers general,
a
effects of institutions on outcomes can never
much greater role in determining who will be
be taken for granted; the mere existence of an
favored in human capital investment than in
the United States, where self-investment is, to
institution (a norm, for example) has no
some extent, an alternative route to status
necessary behavioral implications. Institutions affect outcomes not by virtue of custom
16 Another example of the application of the Prisoners'
or of hallowed tradition, but only to the
degree that they carry with them the capacity
Dilemma game logic to macrosociological problems is
to reward or punish the individuals who are
offered by Douglas Heckathorn (in press) who investigates the effects of collective punishments and rewards in subject to them. Hence the study of instituinducing compliance with group norms. Using this logic tions necessarily becomes an analysis of the
he shows that the effect of control depends upon the incentive structures that they entail.
attributes of the agent who controls collective incentives
In this sense, Prisoners' Dilemma games
(that is, the agent's monitoring ability, vulnerability to
(Luce and Raiffa 1957; Brams 1985) have
revolt, and the strength of the agent's sanctions), as well
as the attributes of the group (such as group size, degree been among the most intensively studied of
of intergroup control, proportion of the group composed any social institutions. Analyses of these
of potential violators, and the costliness of normative games have a host of implications, two of
control and revolt). Collective sanctions are demonstrated
which are that (1) collective outcomes that
to be a potentially unstable means of behavioral control
benefit the whole are difficult to achieve
because the same factors that motivate compliance may
also motivate rebellion.
through individual actions; and (2) prefer-
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
RATIONAL CHOICE THEORY
211
ences do not determine actions. By treating
tures are theoretically relevant, however, and
the gender stratification system as an the
instituinterested reader can turn to a variety of
sources
tion, Brinton (1988) shows that, despite
thefor further discussion.17
best of intentions, (3) individuals are responsible for perpetuating societal-level discrimiof Rational Choice Applications
natory practices. Finally, Heimer'sPayoffs
(1985)
study indicates that (4) certain kinds of
In the previous sections a few of the
collective goods (especially those involving
substantive
payoffs that rational choice applimoral hazard) can be provided only in the
cations already have brought to macrosociolpresence of institutions that enforce stringent
ogy
were highlighted. Some of the insights
controls.
drawn from the individual studies are coun-
terintuitive; some challenge other general
Research Exemplifying the Remainingtheories;
Paths and some recall established knowl-
edge. Some are merely provocative. Despite
Research focusing on the connection between
the substantive specificity and diversity of the
information and constraints (Path 3) generally
studiesinreported, then, they yield general
has considered the role of uncertainty
insights
complicating actors' perceptions of the
con- that can easily be transformed into
propositions
and into testable hypotheses with
straints under which they operate. There is a
new
substantive
content. In this way each
large corpus of research in economics on this
piece
of research contributes not only to
issue, starting with the classic work of
Frank
understanding
a given empirical problem, but
Knight (1921). In sociology, uncertainty
has
also
to
the
cumulative
body of sociological
been principally the province of those who are
Of course, the adequacy of any
interested in the environmental context knowledge.
within
explanation can only be determined after
weighing the findings of many different
In another kind of application, Friedman
which organizations make decisions.
empirical studies.
(1987) proposes that the rational choice
Yetand
there are also important methodologimodel is appropriate for explaining
payoffs. One of our purposes in highlightpredicting collective action only undercal
condiing these applications is to demonstrate that
tions of objective certainty-that is, when the
rational
future can be predicted with some degree
ofchoice offers a very general means of
attacking a wide variety of problems in
confidence. Under such conditions, individu-
macrosociology. Rational choice has this
als can make judgments about the relative
not because it is merely an overarching
benefits of joining a collective action ability
and can
metatheoretical
perspective, but precisely
act accordingly. Under conditions of
objec-
it consists of an entire family of more
tive uncertainty, however, actors are because
robbed
specific theories derived from the same first
of the implements necessary to make a
rational decision-information key principles
among that can be employed to explain a
variety of macrosociological phenomthem-and they are thereby motivated wide
to seek
ena.
the advice and counsel of their fellows.
The key to this methodological approac
Uncertainty thus provides a special kind of
to
impetus for collective behavior that can be locate any specific empirical phenome
as an instance of a more general problem
marshalled into collective action. In an
application to strikes, Friedman showsthen
that to apply the theory that is appropria
the analysis of the general problem to
strikes occur predominantly under conditions
particulars of the specific case. Thus
of objective uncertainty when that uncertainty
failure of socialist revolution in a partic
is combined with certain contractual loss.
This explanation offers an alternative to time and place is identified as a (m
macroeconomic, political, and earlier rationalgeneral) collective action problem and
choice accounts of strikes.
theory of pubic goods is applied to it. A
There are significant bodies of research
investigating paths 2 (Information-+ Opportu- 17 See Arrow (1951) for the classic work on the su
nity Costs) and 7 (Aggregation Mechanism-* of the relationship between aggregation and s
Social Outcome), but these address issues
rather far removed from the traditional
outcomes. For a sample of work inspired by Arrow
Barry and Hardin (1982). On information, see,
example, Stigler (1961); Simon (1957); and Arr
(1987).
concerns of macrosociologists. These litera-
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
212
SOCIOLOGICAL THEORY
same time, there are many other kinds of
empirical phenomena that can be considered
to be collective action problems and thereby
can be profitably analyzed in an analogous
fashion. Hence, the causes and frequency of
strikes, wars, rebellions, and nationalist
movements are all amenable to treatment as
(for a good discussion of the importance of
this goal, see Coleman 1986a). Other distinguishing virtues include the increased possibility for fruitful exchange across social (and
perhaps even biological) sciences; the large
body of formal modelling that undergirds
many of the theoretical insights; and rational
collective action problems.
choice's agnosticism with respect to types of
Similarly, Downs explained the ideological
data and techniques of data reduction.
homogeneity of American political parties asWe have argued that the adoption of
a specific instance of an optimal location
rational choice theory can improve the quality
problem and then applied Hotelling's (1929)
of empirical analysis in macrosociology. At
general theory of spatial competition tothe
it. present time it cannot be claimed,
Hotelling illustrated his theory by considering
however, that there is a coherent corpus of
research in this tradition. Nevertheless, after
how one could locate grocery stores optimally
in a given territory, but was imaginative
considering already completed works, we
believe that the outlines of such a research
enough to understand that the same principle
which could aid investors to select grocery
program are beginning to emerge. Yet there is
store sites could help account for the behavior
no question that these new studies will have
of voters. Downs picked up Hotelling's idea
to pay close attention to the limitations of
twenty-six years later and incorporated it into
current rational choice explanations. It is to
these that we now turn.
his general theory of democracy.
So too with the contemporary Soviet-
American arms race. This is easily classified
III. LIMITATIONS OF CURRENT
as a non-cooperative game for which the
RATIONAL CHOICE EXPLANATIONS
Prisoners' Dilemma is the appropriate model.
All sorts of bargaining problems are amenable
Despite their ability to explain a variety of
to this kind of analysis, as are some problems
empirical phenomena, rational choice models
of collective action (Hardin 1982). The
are by no means exempt from criticism.
guards' deviation from the wardens' direcCriticism comes from many different sources.
tives is treated as an instance of a principalThe advocates of interpretive analysis (phenoagent problem and agency theory (rudimenmenologists, hermeneuticists, and neotary though it may be) is applied to it. functionalists, among others) often argue for
Lawyer-client relations, labor-management
a social science that is antipositive; as noted
bargaining outcomes (Friedman 1986), king-earlier, we shall not deal with their work
ruling class disputes, and owner-management
here. Some Durkheimians, network analysts,
conflicts all provide further empirical grist
Marxists and other structuralists object on
for this particular theoretical mill.
principle to any approach that is based on
It should be emphasized that these separate
methodological individualism. A considerexplanations all share a common theoretical
ation of this objection is also beyond the
maternity. They employ the same firstscope of this paper. What we discuss here are
the criticisms levelled at rational choice either
principles-all are based on common metafrom those who work within the tradition, or
theoretical assumptions of methodological
individualism and purposive action. In addi-from those who explicitly advocate an
tion to their agreement about how individuals
alternative to it without, however, rejecting
make choices, each explanation also attributes
methodological individualism out of hand
the differences in outcomes to variations in
(see Hogarth and Reder [1987] for a state-ofpreferences, opportunity costs, and/or instituthe-art collection of papers critical to various
aspects of rational choice).
These claims for rational choice theory are, Most of the critics object to the highly
to some extent, claims for any general theory.
unrealistic behavioral assumptions of rational
tional constraints.
Why, then, prefer rational choice? Perhaps
choice models. Especially in the earliest
the most compelling reason is that it is models, most real-world complexities were
explicitly concerned with linking micro andsuppressed in order to derive determinate
solutions. That which could not be modelled
macro levels of analysis rather than asserting
the analytical supremacy of one or the other
was not considered. Thus, heterogeneity of
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
213
RATIONAL CHOICE THEORY
utility of
theory that underlie standard rational
preferences and interests among groups
choice behavioral assumptions (as conceived
actors, strategic interaction (such as sophistiof andallformalized by von Neumann and
cated voting), and the costs of information
Morgenstern 1947; see also Allais and Hagen
were assumed away.
1979; Kahneman
and Tversky 1979; Machina
It was no surprise that the inadequacy
of
1983; and Hogarth and Reder 1987). For
these simple models was revealed as empirieconomists
interested in accounting for marketcal tests proceeded. One famous example
was
that Americans vote in much greater numbers
oriented behavior, these findings are not so
than Downs could account for under
any for in the market setting wealth
troubling,
extension of his, or (for that matter) any
other
maximization
can be realistically substituted
maximization: any participant in a
plausible rational choice voting model.for
It utility
was
perfectly
terribly clear to analysts and voters alike
that competitive market whose behavior
hardly any given individual voters ever
is would
not motivated by wealth maximization will
feel themselves to be casting the deciding
not long survive. Yet for social scientists
ballot, which is the only readily allowable
interested in accounting for non-market-
motivation in rational choice models of
oriented behavior, these results were both
voting. Why then did they bear opportunity
distressing and intriguing.
costs to take an action that could not
They were distressing because the kinds of
predictions
materially affect the outcome? Another
fa- that could be derived from the
mous example was that there is simplistic
far more
assumptions of expected utility
collective action than Olson-or any
theoryother
all became suspect; they were intrigu-
rational choice theorist of collective action-
ing because some of these findings provided a
could easily explain. While the lack of
basis for explanations of empirical findings
selective incentives seemed to provide a that otherwise seemed to fly in the face of
powerful explanation for the failure of some of rational choice theory's key predic-
collective action to occur, collective action
sometimes did occur in the absence of
tions.
For example, the work of Tversky and
sufficient selective incentives. So strikes,
Kahneman (1987), which demonstrates the
riots, anti-apartheid demonstrations in South
inconstancy of preferences about risk (the
Africa and in New York City, civil rightssame people are risk-seeking when choosing
marches, and even contributions to publicamong losses and risk-averse when choosing
goods in laboratory settings, all took place withamong gains), may be used to account for
a frequency that defied explanation in stricthitherto mystifying behavior about gambling
rational choice terms.
and insurance (see also Brenner 1983: Ch. 1;
Resolving the growing disjuncture between
Einhorn and Hogarth 1985). Thus, what had
theory and data became a problem of
seemed to be clearly non-rational behavior
paramount importance. Three tacks to more could, in fact, be accounted for when
successfully account for real world phenom-simplistic assumptions about risk preferences
ena seemed promising. The first was to fiddlewere modified.18
with the behavioral assumptions of rational While improving the fit of the theory to the
choice theory-especially those of its mostdata, complicating the behavioral assumptestable form, expected utility theory (see
tions nonetheless begs a fundamental quesArrow 1987:204-206). The second was to
tion. Where do preferences come from? If
maintain the standard behavioral assumptions
people are said to act rationally to pursue their
of rational choice, but to explore in depth the
most preferred ends, what are these ends and
effects of incomplete information on subsewhy do individuals hold these ends dear as
quent action. The third was to elaborate the
contribution of structural constraints in deter-
mining social outcomes.
18 There is a lively debate between traditional and
behavioral economists about the boundaries of rational
choice theory (Hogarth and Reder 1987). Are these new
Fiddling with the Behavioral Assumptions of explanations consistent with the older versions of rational
behavior or do they represent the beginnings of a new
Rational Choice Theory
paradigm? From the perspective of most sociologists,
however, traditional and behavioral economists may be
A growing body of research has indicated that
considered to be in the rational choice camp since they
there are significant and systematic empirical both accept the two key assumptions of purposive action
deviations from the hypotheses of expected and methodological individualism.
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
214
SOCIOLOGICAL THEORY
against others? As long as the behavior in fairness, or justice, as against efficiency.
question is market-oriented behavior, this Such outcomes are not at all far-fetched from
question can be skirted simply by inserting rational choice premises, for no single
wealth-maximization for the "most preferred individual stands to benefit from efficiency as
ends" part of the model.19 With any kind of a state of affairs (because it is a public good),
non-market oriented behavior such a substitu-
whereas all individuals may well find that
tion can be made, but only in a post-hocfairness is in their own self-interest.
fashion, and only with considerable qualifica- Until the time that significant progress is
tion. For instance, in Friedman's (1987)
made toward an understanding of preferences,
model of collective action, it is assumed that the scope or power of rational choice analyses
people prefer certainty to uncertainty. She can
is clearly limited. Since preferences are given
provide a theoretical reason for this particular
rather than explained in rational choice
assumption (e.g. people cannot act rationally analyses, these analyses are far better suited
under conditions of uncertainty), and canfor social phenomena that are the outgrowth
assemble considerable supportive empirical
of individual preferences that are strong
evidence, yet it is quite conceivable that a(relative to competing), stable (over time),
different logic and a different body of and uniform (across actors).21 Hence, social
evidence might be mustered to support the
outcomes dependent on preferences that result
opposite claim.
from fundamental biological drives or from
That there is no theory of preference strongly sanctioned social conventions are
formation has implications not only for thegood candidates for rational choice analyses.
behavior of individuals and groups, but also Social outcomes dependent on unstable,
for social outcomes.20 The nature of anyweak, or variable preferences-consumer
social outcome clearly depends upon the set
behavior,22 for example-are by this token
of underlying preferences. For instance, if
poor subjects for rational choice analyses.
people are wealth-maximizers, they will
desire social outcomes that provide for the
greatest economic efficiency. Even so, there Elaborating the Role of Information
are a multitude of possible Pareto-efficient
Yet another way to increase the empirical fit
equilibria, depending upon the initial distribuof rational choice models is to appreciate that
tion of resource endowments across individu-
variable amounts and types of information
als. Other kinds of behavioral assumptionswill have systematic effects on outcomes.
lead to different social outcomes, however.
These outcomes might emphasize certainty,
These models typically require some assump-
21 We are grateful to Douglas Heckathorn (personal
19 Simon (1985) and Arrow (1987) object even to this communication) for his suggestions in this regard.
solution on the grounds that any kind of maximization is
22 One anonymous reviewer of this paper suggests that
impossible due to computational and informational consumer behavior is one of the greatest success stories
limits.
of macroeconomics, for as the price of a good goes up its
20 None of the social sciences has contributed much in consumption generally goes down. The ambitions of the
the way of a positive theory of preference formation. economic theory of consumer behavior go considerably
Pleas for greater attention to this issue have been beyond the making of price/quantity predictions for the
emanating recently in nearly all of the social science consumption of a single commodity, however, for in this
disciplines. A select list would include Scitovsky (1976), case the critical problem of the substitutability of goods
Sen (1977), Hirschman (1985) and Arrow (1987) in
simply does not arise. The theory of consumer behavior
economics; Douglas and Isherwood (1979) in anthropol-seeks to explain nothing less than how individuals
ogy; March (1978) and Wildavsky (1987) in politicalallocate their budget given the totality of commodities
and other goods available in their environment. Since the
science; Zajonc (1980), Simon (1985) and Ainslie (1986)
proponents of the ordinal theory of utility (most
in psychology; Elster (1983) and Nelson (1986) in
philosophy, and Emerson (1987), Friedman (1987) and mainstream economists) eschew the possibility of indeEtzioni (1986) in sociology. Most of these writers,
pendently measuring individual preferences-let alone
however, do not claim to be constructing an explanatory satisfaction-apart from actual behavior, they are comtheory of preference formation themselves, and the pelled to assume the constancy of preferences in order to
models of the few who do make such a claim are far from
derive any empirical implications at all from their
elaborated. The most ambitious attempts to address the models. Since this assumption is known to be highly
issue of preference formation theoretically have not come unrealistic, models of this kind can have little to say
from social scientists at all, but from evolutionary about the empirical determinants of consumer behavior.
biologists (Alexander 1987; Boyd and Richerson 1985). For a recent assessment of the poverty of current
This is not the place to discuss the merits of these works, economic approaches to consumer behavior, see Roth
however.
(1987).
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
215
RATIONAL CHOICE THEORY
tion of "sufficient information" in order for
size of one's generational birth cohort). All
individuals to choose successfully among structural constraints affect behavior by deteralternatives. This assumption is increasingly mining the objective probabilities that an
regarded as problematic, however. To illus- individual's most preferred end can be
trate, consider a situation in which two people realized. In this way they act as inducements
to the pursuit of one end as against another.
who are subject to the same structural
constraints behave differently. We can ac- Not a great deal of effort has gone into this
count for this by saying either that they have area of rational choice analysis (see, howdifferent preferences, or that they have ever, Easterlin 1981), but sociologists know a
different information about the consequences great deal about how it is that locations within
of their actions. To claim that behavior is
social structures limit individual choices. One
explicable by reference to different prefercannot marry an Eskimo if there are none
ences is tautological, and undermines
around
the (Blau 1987:79). Among those who
standard methodology of rational choice.
wish
The
to work full-time, everyone would
second explanation is more satisfying:
clearly
theprefer to work in a primary rather than
amount of information that agents have
can
a secondary
labor market. That not everyone
affect behavior independently of constraints
does is often explicable by race, class, and/or
or preferences. It may well be that informagender position. Much of the new work by
tion is the crucial intervening variable sociologists
in all
using rational choice theory to
understand macrosociological problems falls
rational choice explanations.
Indeed, information has become an increasinto this category (Brustein 1988a; Brinton
1988; and Kiser 1987).
ingly key variable in studies of contracting,
Another way to elaborate the role of
bargaining, and organization. Both principal
agent and transaction-cost models startstructural
with constraints in determining outcomes
the assumption that informational asymmeis to inquire as to the origins of the
tries are significant. What seems to be
institutions that produce such constraints.
missing in these literatures, however, is a This line of research is just beginning
theory of optimal information investment.23 (Coleman 1989; Hechter 1989; Kliemt 1989).
Such a theory would have to answer a host of
questions. How much information is it
rational to collect and use in making a
CONCLUSION
specific decision? How does the agent decide
when to stop searching? How do we under- Rational choice provides a number of discrete
stand the case in which two individuals have
benefits to researchers who work within its
precisely the same information but draw
confines. Once scholars can identify the
different conclusions from it? Finally, supappropriate problem to which their research
pose that the information that is requiredquestion
to
belongs, they are assured of finding
make a decision rationally is inherently
a determinate solution. Armed with a given
unavailable (that is, the situation is markedtheory,
by
the investigator knows which facts are
critical for the solution and which facts are
uncertainty). What then should we expect
rational actors to do?
Elaborating the Contribution of
largely irrelevant. Since this often leads to a
demand to collect new kinds of data, rational
choice has a built-in tendency to create new
research agendas out of old ones. Further,
researchers beginning from rational choice
premises never need to build their theory
In addition to modifying behavioral assumpanew, starting from first principles and
tions to increase their verisimilitude, it is also
defining and justifying the use of each
possible to introduce other structural con-
Structural Constraints
concept. Not only does this save a good deal
straints than purely institutional ones explic-
itly into these models (we refer here to
of intellectual energy, but it also enables
constraints such as those produced by macro- scholars with quite different substantive
interests and areas of expertise to communieconomic fluctuations, wars, and even the
cate with one another. All of these conditions
23 This must go beyond the limits of Stigler's (1961) combine to help sustain a research program in
search theory, however, for reasons discussed by Arrow which knowledge cumulates relatively effi(1987:207).
ciently and progress is swift.
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
216
SOCIOLOGICAL THEORY
Debates about the merits or demerits of
rational choice theories often take on an
Allais, M. and 0. Hagen. 1979. Expected Utility
Hypotheses and the Allais Paradox: Contemporary
Discussions of Decision under Uncertainty with Allais'
ideological cast, and sometimes rational
Rejoinder. Dordrecht: D. Reidel.
choice theories are rejected becauseArrow,
they
Kenneth J. 1951. Social Choice and Individual
seem to have unpalatable political implicaValues. New Haven: Yale University Press.
tions. In this regard we think thatArrow,
it isKenneth J. 1987. "Rationality of Self and Others
in an Economic System." Pp. 201-16 in Robin M.
important to note two characteristics
of
Hogarth and Melvin W. Reder, eds., Rational Choice:
The Contrast between Economics and Psychology.
theories, individuals are accorded signifiChicago: University of Chicago Press.
Azzi, Corry and Ronald Ehrenberg. 1975. "Household
cantly more respect than in most other
rational choice theories. First, in these
Allocation of Time and Church Attendance." Journal
sociological models. Rational choice theories
of collective action, for instance, take into
account contextual constraints on action but
also leave room for the informed choices of
actors. This is in contrast to those who would
of Political Economy, 83, 1:27-55.
Barry, Brian and Russell Hardin, Eds. 1982. Rational
Man and Irrational Society? Beverly Hills, CA.: Sage.
Blau, Peter. 1987. "Contrasting Theoretical Perspectives." Pp. 71-86 in J.C. Alexander, B. Giesen, R.
argue that participants are merely sheep Munch, & N. Smelser, eds., The Micro-Macro Link.
Berkeley: University of California Press.
heeding the call of the shepherd, or individuBoulding, Kenneth E., ed. 1973. Peace and the War
als who ignore reason when their sensibilities Industry. New Haven: Yale University Press.
are excited. Second, arguing that individuals
Boyd, Robert and Peter J. Richerson. 1985. Culture and
will often follow their self-interest in choos-
the Evolutionary Process. Chicago: University of
ing among actions is not the same thing asChicago Press.
saying that they ought to do so. While this
was the position of Bentham and the
Brams, Steven. 1985. Superpower Games. New Haven:
Yale University Press.
Brenner, Reuven. 1983. History: The Human Gamble.
utilitarians, positive and normative analyses Chicago: University of Chicago Press.
Brinton, Mary. 1988. "The Social-Institutional Bases of
have no necessary connection.
As rational choice explanations come to Gender Stratification: Japan as an Illustrative Case."
American Journal of Sociology 94:330-34.
include more realistic behavioral assump-Brustein, William. 1988a. Social Origins of Political
tions, more knowledge about the effects ofRegionalism: France 1849-1981. Berkeley: University
social structures, and greater appreciation of of California Press.
the role of information, they cannot fail toBrustein, William. 1988b. "The Political Geography of
Belgian Fascism: The Case of Rexism." American
provide more satisfactory empirical accounts.
Sociological Review 53(1):69-80.
Scholarship is required on all of these
Buchanan, James, and Gordon Tullock. 1962. The
separate fronts. Sociologists have a compara- Calculus of Consent. Ann Arbor: University of
tive advantage in contributing to at least one Michigan Press.
Angus, Philip E. Converse, Warren E. Miller,
aspect of this work, namely, to the study ofCampbell,
and Donald E. Stokes. 1966. Elections and the
the effects of social structural constraints on
Political Order. New York: John Wiley and Sons
social outcomes. Sociologists also know a Coleman, James S. 1973. The Mathematics of Collec
great deal about how organizations control Action. Chicago: Aldine.
information. At the same time, cognitiveColeman, James S. 1986a. "Social Theory, Social
Research, and a Theory of Action." American Journal
scientists enjoy a comparative advantage in of Sociology 86(5):984-1014.
the search for more realistic behavioral
Coleman, James S. 1986b. Individual Interests and
assumptions, and indeed much research
Collective
on
Rationality. New York: Cambridge University Press.
this subject is already underway. Together
Coleman, James S. 1989. "The Emergence of Norms."
these two strands of research may well
In M. Hechter, K.D. Opp, & R. Wippler, Social
contribute to the development of positive
Institutions: Their Emergence Maintenance and Ef-
theory in social science.
REFERENCES
fects. New York: Aldine de Gruyter
Collins, Randall. 1981. "On the Microfoundations of
Macrosociology." American Journal of Sociology, 86,
5:984-1014.
Cook, Karen,
Ainslie, George. 1986. "Beyond Microeconomics:
Con- ed. 1987. Social Exchange Theory.
Beverly Hills: Sage Publications.
flict Among Interests in a Multiple Self as a DetermiDouglas,
Mary and Baron Isherwood. 1979. The World
nant of Value." Pp. 133-76 in Jon Elster,
ed., The
of Goods.
Multiple Self. Cambridge: Cambridge University
Press. New York: W.W. Norton and Company.
Downs,
Anthony. 1957. An Economic Theory of
Alexander, Jeffrey C., Bernhard Giesen, Richard
Munch,
and Neil J. Smelser, eds. 1987. The Micro-Macro
Democracy. New York: Harper & Row.
Link. Berkeley: University of California Press.
Easterlin, Richard A. 1980. Birth and Fortune: The
Alexander, Richard D. 1987. The Biology of Moral Impact of Numbers on Personal Welfare. New York:
Systems. New York: Aldine De Gruyter.
Basic Books.
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
RATIONAL CHOICE THEORY
Einhorn, Hillel J. and Robin M. Hogarth. 1985.
"Ambiguity and Uncertainty in Probabilistic Inference." Psychological Review, 92, 4:433-61.
Elster, Jon. 1983. Sour Grapes: Studies in the Subversion
of Rationality. Cambridge: Cambridge University
217
Economic Discourse." Economics and Philosophy
1(1):7-22.
Hirshleifer, J. and John G. Riley. 1979. "The Analytics
of Uncertainty and Information-An Expository
Survey." Journal of Economic Literature, 27:13751421.
Press.
Elster, Jon. 1985. Making Sense of Marx. Cambridge:
Cambridge University Press.
Emerson, Richard M. 1981. "Social Exchange Theory."
Pp. 30-65 in Morris Rosenberg and Ralph H. Turner,
eds., Social Psychology: Sociological Perspectives.
New York: Basic Books.
Hogarth, Robin M. and Melvin W. Reder, eds. 1987.
Rational Choice: The Contrast between Economics
and Psychology. Chicago: University of Chicago
Press.
Hotelling, Harold. 1929. "Stability in Competition." The
Economic Journal 39, 1:41-57.
Emerson, Richard M. 1987. "Toward a Theory of Value
Jacobs, David. 1988. "Images of the State and Corporate
in Social Exchange." Pp. 11-46 in Karen S. Cook,
Economic Power: Aggregate Concentration and Busied., Social Exchange Theory. Newbury Park, CA:
ness Tax Rates." American Journal of Sociology
Sage.
93(4):852-81.
Enelow, James and Melvin Hinich. 1984. The Spatial
Jensen, Michael C. and William H. Meckling. 1976.
Theory of Voting. New York: Cambridge University "Theory of the Firm: Managerial Behavior, Agency
Press.
Costs and Ownership Structure." Journal of Financial
Etzioni, Amitai. 1986. "The Case for a Multiple-Utility Economics, 3, (4):305-360.
Conception." Economics and Philosophy 2:159-183.
Kahneman, Daniel and Amos Tversky. 1979. "Prospect
Friedman, Debra. 1986. "The Principal-Agent Problem
Theory: An Analysis of Decision Under Risk."
Econometrica, 21:263-91.
in Labor-Management Negotiations." Pp. 87-104 in
Edward J. Lawler, ed., Advances in Group Processes,Kiser, Edgar. 1987. Kings and Classes: Crown AutonVol. 3. Greenwich, CT.: JAI Press.
omy, State Policies and Economic Development in
Friedman, Debra. 1987. "Notes on 'Toward a Theory of Western European Absolutisms. PhD Dissertation,
Value in Social Exchange.' " Pp. 47-58 in Karen S.
Department of Sociology, University of Arizona.
Cook, ed., Social Exchange Theory. Newbury Park, Kliemt, Hartmut. 1989. "The Costs of Organizing Social
CA: Sage.
Cooperation: Some Remarks About the Game of
Friedman, Debra. 1987. "Uncertainty and Collective
Creating a Game." In M. Hechter, K.D. Opp, & R.
Action." Unpublished paper.
Wippler, eds., Social Institutions: Their Emergence,
Giddens, Anthony. 1979. Central Problems in Social
Maintenance and Effects. New York: Aldine de
Theory: Action, Structure and Contradiction in Social
Gruyter.
Analysis. Berkeley: University of California Press.
Klosko, George, Edward N. Muller, and Karl-Dieter
Gottfredson, Michael and Travis Hirschi. 1986. "The
Opp. 1987. "Rebellious Collective Action Revisited:
True Value of Lambda Would Appear to be Zero: AnAn Exchange." American Political Science Review 81
Essay on Career Criminals, Criminal Careers, Selec(2):557-64.
tive Incapacitation, Cohort Studies, and Related
Knight, Frank. 1921. Risk, Uncertainty, and Profit.
Topics." Criminology, 24, 2:213-234.
Boston: Houghton Mifflin.
Habermas, Jurgen. 1983. The Theory of Communicative
Leontieff, Wassily W. and Faye Duchin. 1983. Military
Action. Boston: Beacon Press.
Spending: Facts and Figures, Worldwide Implications,
Hardin, Russell. 1982. Collective Action. Baltimore:
and Future Outlook. New York: Oxford University
Johns Hopkins University Press (for Resources forPress.
the
Future).
Luce, R. Duncan and Howard Raiffa. 1957. Games and
Heath, Anthony. 1976. Rational Choice and Social
Decisions. New York: John Wiley.
Exchange. Cambridge: Cambridge University Press.
Machina, Mark J. 1983. "Generalized Expected Utility
Hechter, Michael. ed. 1983. The Microfoundations ofAnalysis and the Nature of Observed Violations of the
Macrosociology. Philadelphia: Temple University Press.
Independence Axiom." Pp. 263-293 in B.P. Stigum
Hechter, Michael. 1987. Principles of Group Solidarity.
and F. Wenstop, eds., Foundations of Utility and Risk
Berkeley and London: University of California Press.
Theory with Applications. Dordrecht: D. Reidel.
Hechter, Michael. 1989. "The Emergence of CooperaMarch, James G. 1978. "Bounded Rationality, Ambigutive Social Institutions." In M. Hechter, K.D. Opp, ity,
& and the Engineering of Choice." The Bell Journal
R. Wippler, eds., Social Institutions: Their Emerof Economics 9(2):587-608.
gence, Maintenance and Effects. New York: Aldine deMarsden, Peter. 1981. "Introducing Influence Processes
Gruyter.
into a System of Collective Decisions." American
Heckathorn, Douglas. In press. "Collective Sanctions
Journal of Sociology 86(6):1203-35.
and the Creation of Prisoner's Dilemma Norms."
Marwell, Gerald, and Ruth E. Ames. 1980. "Experi-
American Journal of Sociology.
ments on the Provision of Public Goods. II. Provision
Heimer, Carol A. 1985. Reactive Risk and Rational
Points, Stakes, Experience, and the Free-Rider ProbAction: Managing Moral Hazard in Insurance Conlem." American Journal of Sociology, 85, (4):926-937.
tracts. Berkeley: University of California Press.
Michels, Robert. 1962 (1911). Political Parties: A
Hirsch, Paul, Stuart Michaels and Ray Friedman. 1987. Sociological Study of the Oligarchical Tendencies of
" 'Dirty Hands' versus 'Clean Models': Is Sociology
Modern Democracy. New York: Free Press.
in Danger of Being Seduced by Economics?" Theory Mueller, Dennis C. 1979. Public Choice. Cambridge:
and Society 16:317-36.
Cambridge University Press.
Hirschman, Albert 0. 1985. "Against Parsimony: Three Nelson, Alan. 1986. "New Individualistic Foundations
Easy Ways of Complicating Some Categories of
for Economics." Nous 20:469-90.
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms
SOCIOLOGICAL THEORY
218
Oberschall, Anthony. 1973. Social ConflictSchelling,
and Social
Thomas C. 1978. Micromotives and Macrobe-
havior. New York: W.W. Norton.
Movements. Englewood Cliffs, N.J.: Prentice-Hall.
Scitovsky, Tibor. 1976. The Joyless Economy: An
Oberschall, Anthony and Eric Leifer. 1986. "Efficiency
Inquiry into Human Satisfaction and Consumer
and Social Institutions: Uses and Misuses of Economic
Reasoning in Sociology." Annual Review of Sociology Dissatisfaction. New York: Oxford University Press.
Sen, Amartya K. 1977. "Rational Fools: A Critique of the
12:233-53.
Behavioral Foundations of Economic Theory." PhilosOliver, Pamela. 1980. "Rewards and Punishments as
ophy and Public Affairs 6(4):316-34.
Selective Incentives for Collective Action: Theoretical
Simon, Herbert A. 1957. Models of Man. New York:
Investigations." American Journal of Sociology, 85,
Wiley.
(6):1356-1375.
Simon, Herbert A. 1985. "Human Nature in Politics: The
Oliver, Pamela, Gerald Marwell and Ruy Teixeira. 1985.Dialogue of Psychology with Political Science."
"A Theory of the Critical Mass. I. Interdependence, American Political Science Review, 79:293-304.
Group Heterogeneity and the Production of Collective
Stigler, George. 1961. "The Economics of Information."
Action. American Journal of Sociology, 91, 3:522-56.Journal of Political Economy, 69:213-25.
Olson, Mancur. 1965. The Logic of Collective Action.
Stinchcombe, Arthur. 1968. Constructing Social Theo-
ries. New York: Harcourt and Brace.
Cambridge: Harvard University Press.
Sykes, Gresham M. 1958. The Society of Captives.
Opp, Karl-Dieter. 1986a. "Soft Incentives and Collective
Action: Participation in the Anti-Nuclear Movement."Princeton: Princeton University Press.
Tversky, Amos and Daniel Kahneman. 1987. "Rational
British Journal of Political Science, 16:87-112.
Opp, Karl-Dieter. 1986b. The Rationality of PoliticalChoice and the Framing of Decisions." Pp. 67-94 in
Robin M. Hogarth and Melvin W. Reder, eds.,
Protest. Unpublished manuscript.
Rational Choice: The Contrast between Economics
Pursell, Carroll W., Jr., ed. 1972. The Militaryand Psychology. Chicago: University of Chicago
Industrial Complex. New York: Harper and Row.
Press.
Przeworski, Adam. 1985. Capitalism and Social DemocUdis, Bernard, ed. 1973. The Economic Consequences of
racy. Cambridge: Cambridge University Press.
Reduced Military Spending. Lexington, Mass.: D.C.
Riker, William H. 1986. The Art of Political Manipula-Heath.
tion. New Haven: Yale University Press.
von Neumann, John and Oskar Morganster. 1947.
Rosen, Steven, ed. 1973. Testing the Theories of theTheory of Games and Economic Behavior. Princeton:
Military-Industrial Complex. Lexington, Mass.: D.C. Princeton University Press.
Heath.
Wildavsky, Aaron. 1987. "Preferences by Constructing
Roth, Timothy P. 1987. The Present State of Consumer Institutions: A Cultural Theory of Preference FormaTheory. Lanham, Md.: University Press of America.
tion." American Political Science Review 81(1):3-21.
Sarkesian, Sam C., ed. 1972. The Military-Industrial Zajonc, R.B. 1980. "Feeling and Thinking: Preferences
Need No Inferences." American Psychologist
Complex: A Reassessment. Beverly Hills, Calif.: Sage
Publications.
35(2):151-75.
This content downloaded from
152.118.24.31 on Wed, 15 May 2024 08:20:18 +00:00
All use subject to https://about.jstor.org/terms