James Madison contributed much to political science and to the art and
science of governing democratic regimes. A major contribution was his
recognition of the multiple threats facing the newly established
Confederation (see, e.g., Dougherty 2001). Madison
was willing to invest heavily in converting these threats into new
opportunities. In his “Preface to Debates in the Convention” ([1984] 1840, 19), Madison reflected on the causes
for alarm:
The failure however of the varied propositions in the Legislature,
for enlarging the powers of Congress, the continued failure of the
efforts of Congs. to obtain from them the means of providing for the
debts of the Revolution; and of countervailing the commercial laws
of G. B., a source of much irritation & agst. which the separate
efforts of the States were found worse than abortive; these
Considerations with the lights thrown on the whole subject, by the
free & full discussion it had undergone led to a general
acquiescence in the Resoln. passed, on the 21. of Jany. 1786, which
proposed & invited a meeting of Deputies from all the
States.
Madison devoted substantial effort to doing a deep analysis of prior efforts
to design robust democratic systems as well as actively participating in the
Convention (while keeping extensive—almost verbatim—notes). Upon the
completion of the Convention, he contributed key essays to The
Federalist Papers as the major effort to articulate the logic of
the new Constitution for those who would be deciding on its fate and for all
of us who have studied these essays as a core foundation of democratic
theory.
As a deep student of political life, Madison was among the most articulate
theorists to warn of the mortal diseases that threaten all popular
government. Factions are the most threatening of these diseases. In
Federalist 10, Madison defined a faction as “a number of
citizens, whether amounting to a majority or minority of the whole, who are
united and actuated by some common impulse of passion, or interest, adverse
to the rights of other citizens, or to the permanent and aggregate interests
in the community” (p. 78). As a man of conviction, Madison fought hard for
the constitutional provisions that he thought would alleviate the threats
facing the nation. He then gave unstinting support to the new Constitution
and articulated its principles in the Federalist Papers along with
Alexander Hamilton and John Jay (with whom he had previously disagreed on
key issues). The final product differed substantially from the Virginia plan
presented by Madison at the beginning of the Convention. Nevertheless, he
refused to organize a faction against the new Constitution (even though he
expressed pessimism about it in private correspondence [see Robertson 2005]). Madison joined in a spirited
defense of the effort he saw as a necessary step to convert the threats to
the long-term stability of the new nation into opportunities to move ahead
with a more successful mode of operation.
Madison also recognized that the “latent causes of faction are sown in the
nature of man” (p. 79). It would be “worse than the disease” to try to
destroy the liberty that is essential to the expression of diverse views.
Rather than destroy differences of views, the challenge for Madison was how
to design institutions so humans could learn from debate while not allowing
any one group to impose its will on others who oppose it.
Madison understood that the world he knew was undergoing major changes in its
underlying structure. He saw that it was necessary to learn from past
history, to engage in vigorous deliberation about how to avoid the threats
confronting a young nation, and to avoid participating in a faction against
the new Constitution even though he opposed some of the sections that were
included in it. Political scientists are also in an era that presents many
threats and challenges to our discipline. We shall not overcome these
threats unless we confront them, engage in serious deliberation about how to
convert them to opportunities, and change some of the ways of interacting
with one another and with our subject matters.
Those engaged in the Perestroika movement have generated some healthy debates
about the current state of political science (see Monroe
2005) as did those who were earlier involved in the Caucus for a
New Political Science. Many thoughtful scholars of our times have also
warned of threats to the future development of social sciences more broadly
that are of considerable relevance to political science. If we confront the
threats we face, we may be able to strengthen our overall abilities as a
discipline to understand and contribute to the increasingly volatile world
in which we live. This is our opportunity—let us not pass it by.
Many current challenges to the discipline of political science, if ignored,
could weaken our capabilities to understand and explain political phenomena
in a cumulative scientific manner and to offer useful policy advice for the
constitution of democratic regimes at multiple levels in an ever-changing
world. Let me introduce four of these disciplinary threats here and then
spend some time on each.
- The excessive energy devoted to factional fights. Given the
vast domain of our subject matter, we do need to specialize and focus on
particular questions or fields of analysis. On the other hand, we need
to avoid spending so much energy on trying to show that one approach is
right and the others are wrong (or even evil). Name-calling does not
create an environment where genuine learning from deliberation can
occur.
- The counterproductive search for universal laws of human
behavior. Some of those who believe that they have found
general laws are willing to write off all other approaches to our
discipline as non-science (or, even at times, non-sense). There are
those who would argue that political science is a quixotic
misnomer and that our task is to help understand specific contexts and
help reform them (Flyvbjerg 2001). Some of the
rest of us see the enterprise as identifying common building blocks for
doing theoretical analysis at multiple levels and scales of analysis.
The basic difference between the social world and the biophysical world
is that the biophysical world exists whether or not humans reflect on
it, but the social world is constituted by human thought, language, and
action. Given the importance of language, a more serious threat to the
future of our discipline than the lack of universal laws is our lack of
common definitions for key terms we use including power, norms, and
institutions. Can we ever escape from the “Tower of Babble” that we have
created?
- The validity (or, for some, the morality) of using rational choice
models of individual choice. There are many rigid defenders and
attackers of the narrow neoclassical model of rational choice. Many of
the attacks focus on a caricature of what was initially introduced into
our discipline from economics a half-century ago. Currently, a vigorous
effort is underway to broaden this theoretical tool so that it is a
useful foundation for analyzing behavior outside of the highly
competitive, repetitive, and static situations in which it was developed
and will continue to be used. It is as much a threat to further
understanding of human behavior to hold onto a narrow model that has
been shown repeatedly to fail in prediction and explaining behavior
outside of highly competitive settings as it is to reject all work that
is broadly characterized as “rat choice,” whether or not it is conducted
by the very scholars who are broadening the tool in light of empirical
research.
- The tendency to focus on one level of government and to ignore
other disciplines. We tend to narrowly focus on what happens in
national government and not how government decisions affect and are
affected by the biophysical environment as well as the citizens living
in a society and the other scales and means of governance. Political
science has, in the past, borrowed and gained much from history,
economics, and literary studies. We have much more still to learn from
ecology, geography, cultural history, and recent developments in
evolutionary theory by anthropologists, economists, and psychologists as
well as members of our own discipline. We also need to reach out to
participate in, as well as study, the ever-changing political world
around us.
Now let me discuss these challenges.
Factionalism versus Recognizing the Strength of Multiple
Methods
Many members of our discipline study different topics using a wide
diversity of methods. Instead of recognizing the strength that this
could give us, many departments and the discipline at large have devoted
substantial energy to factional (in the sense that Madison meant)
fights. The recent Perestroika challenge has generated many positive
reactions to the flurry of open debate with participants on all sides
communicating mostly in an open problem-solving manner (see Sigelman 2005; Alker
2005; and responses in Monroe
2005). Some earlier exchanges did not have such a friendly,
out-reaching orientation (Niou and Ordeshook
1999). Hopefully, in addition to the debate, we will continue
to develop further efforts to build on the diverse research
methodologies used by political scientists to study diverse
questions.
What is important along the way, however, is that we do not slip from “my
research methods is the best and only” to “anything goes.” Every
research method can be undertaken with rigor or sloppily. Let us not
confuse a sloppy case study, survey, or game-theoretical analysis with
what we can learn from all of these approaches when undertaken with care
and conclusions are modestly framed. Campbell, Stanley, and Gage (1966) have provided us all with a thoughtful
recognition that every research method faces threats to internal or
external validity (or both). King, Keohane, and Verba (1994) stress that research design issues are
essential and have strong similarities regardless of the particular
method we use. Michael Coppedge (1999) has
illustrated how case studies offer an opportunity to examine the
internal logic posited by theorists by tracing the causal processes that
occur in one or several well-chosen case studies (see also Campbell 1975). The method of “analytical
narratives” has demonstrated the feasibility of bridging and blending
formal theory with in-depth case studies (Bates et
al. 1998). In light of her own extensive research, Barbara
Geddes (2004) shows how study site selection,
as well as the type of methods used, strongly affect results. We need to
stop name-calling or arguing that there are only a few acceptable
methods for our discipline and get to work on serious studies that blend
the methods we use (see, e.g., the important reconciliation of two
traditions by Dryzek and List 2003).
Perhaps I can illustrate how one can build a research program that builds
on the strength of multiple methods. Amy Poteete, Marco Janssen, and I
are in the early stages of a book with the working title of Multiple
Methods for Analyzing Collective-Action Theory.1
Our central question
is: What have we learned about the theory of collective action in regard
to the governance of common-pool resources (CPRs) since the challenge
issued by Garrett Hardin (
1968)? Hardin
posited that all users of a commons were helplessly trapped in a causal
process leading to overuse, if not destruction, of the very resources on
which they were dependent.
Hardin's theory posited that it was impossible for resource
users themselves to overcome the collective-action problem they faced.
Only two solutions existed—the first, impose government ownership and
the second, impose private ownership. Case studies are an excellent
first step in examining the validity of that theory. When a theory
asserts that X is impossible, finding a single case where X occurs
negates the theory. Many more than one case study challenged this
“impossibility theorem.” A large number of case studies were written by
historians, anthropologists, ecologists, as well as the full array of
social scientists, describing how multiple groups of users overcame the
tragedy in relationship to pastures (Hardin's hypothetical case),
forests, irrigation systems, and inshore fisheries (see NRC 1986; Workshop CPR Library at www.indiana.edu/∼workshop/wsl/index.html).
Further, many scholars also found that government-imposed ownership as
well as imposed private solutions had not uniformly worked as optimally
as Hardin had posited.
Case studies did not establish that users always crafted rules
to overcome commons dilemmas or that government-imposed ownership or
privatization always failed. Rather, case studies strongly
challenged Hardin's strong derivation of impossibility (that had been
backed up by formal models of CPR settings consistent with Hardin) and
the necessity of government ownership or privatize ownership. Initially,
it was case studies that provided the essential challenge to a theory
that appealed widely to policy analysts, ecologists, and government
officials around the world.2
See the
special issue of Science (December 2003), which focuses
on the 35th anniversary of the publication of
Hardin's first article in Science.
Once case studies had refuted Hardin's impossibility theorem, however, an
even larger task arose: how to construct an explanation that would be
useful in distinguishing between the cases in which common-pool
resources survived and the cases where they had been degraded or
destroyed. Given the large number of case studies written by researchers
who knew the local history and culture, ecology, and the actions of key
actors, two forms of meta-analysis appeared to be the next reasonable
step.
In my Governing the Commons (1990), I
picked a set of cases that met several criteria (relatively small and
contained, local resources with a very long history) so that I could
examine the factors that appeared to account for the robustness of some
systems and the failure of others. The number of specific variables that
I encountered in examining these cases was overwhelming (and was
concurrently dealt with by the second form of meta-analysis described
below). Thus, I tried to account for the robustness of long-surviving
systems by a set of eight “design principles” that were consistent with
the robust systems. I also developed a flowchart causal model
identifying broad variables related to the resource, the resource users,
and their autonomy to make and enforce their own rules to begin the
theoretical effort to explain why some systems did govern themselves
while others overharvested, as Hardin had predicted. Many scholars have
now undertaken formal analyses of common-pool resource games to
understand how findings from the cases could be analyzed using game
theory (see Faysse 2005 for an excellent
review).
At the same time, a group of colleagues (Arun Agrawal, Bill Blomquist,
Edella Schlager, and S. Y. Tang) and I read hundreds of case studies
written by other scholars. After extended discussions and many attempts,
we created a coding form that Schlager used to code 30 inshore fishery
cases and Tang used to code 39 irrigation systems. That analysis is
brought together with that of Agrawal, who drew on his own fieldwork in
six forest communities in India, and Bill Blomquist, who compared three
groundwater basins in Southern California (see E. Ostrom, Gardner, and Walker 1994). We found that all of the
self-organized common-pool resource systems shared four characteristics:
(1) presence of at least one well-defined boundary rule, (2) presence of
various choice rules related to allocating resource units, (3) active
monitoring and sanctioning by local users, and (4) no use of the
grim-trigger mechanism posited by game theorists to be a solution to
repeated dilemma situations (Blomquist et al.
1994, 301). We also found that resource systems characterized
by relatively stationary resource units and storage capacities were more
likely to self-organize (Blomquist et al.
1994, 308–309; see also Schlager, Blomquist,
and Tang 1994).
But a further theoretical task remained: whether explanations constructed
ex post facto to account for past cases possessed any
predictive capability that would stand up to new situations. Our group
was fortunately joined by Roy Gardner, a game theorist, and Jimmy
Walker, an experimentalist. This enabled us to go back and forth from
reading individual case studies, doing the meta-analysis as well as
original fieldwork, to develop game-theoretical models of what we
thought we were observing, and to create simplified common-pool resource
experiments in the lab. In the first set of lab experiments, we were
able to show that Hardin's theory roughly predicts outcomes when
participants are truly trapped by an imposed requirement of no
communication.3
On average, subjects
approached the predicted Nash equilibrium after only a few
rounds of the experiment when they were not allowed to
communicate.
Simply allowing subjects to engage in
face-to-face communication (called cheap talk in game theory), which had
been predicted to have
no difference in any social dilemma
situation, enables all groups to improve outcomes. Some groups come very
close to optimal solutions after engaging in face-to-face
communication.
Achieving the control that is feasible only in well-crafted experiments,
where one actually knows what the optimal outcome is (which one rarely
knows in the field), gained for us a method strong in regard to internal
validity to complement the external validity achieved by our other
methods. Laboratory experiments also allowed us to examine a frequent
phenomenon that I had observed in the field (and read about in the case
studies). Resource users, who had crafted their own rules (a second
level collective-action problem), were frequently ready to enforce those
rules at a cost to themselves (a third level collective-action problem)!
We could test that field observation carefully in the lab. We found the
same willingness to pay a personal cost to impose a sanction on those
who deviated from agreements (see E. Ostrom, Walker,
and Gardner 1992). To explain the diverse findings from our
research, we posited that users developed heuristics in the sense of
bounded rational participants rather than continuing to use a thin model
of rationality.
We have also developed two large-N field studies. The first is of 231
irrigation systems in Nepal characterized by very substantial variation
in ecology and the cultural history of local peoples (see Shivakoti and Ostrom 2001; Joshi et al. 2000; Lam 1998; Regmi forthcoming). The second is an over-time
study of more than 200 forests located in 12 countries to examine the
impact of diverse forest institutions on forest users and official
behavior as well as on measured attributes of forest conditions (e.g.,
extent, basal area, biodiversity, etc.) (see Gibson,
McKean, and Ostrom 2000; Gibson, Williams,
and Ostrom 2005; Tucker and Ostrom
2005). It is our deep belief that the reliability, as well as
the use, of this data in developing countries is greatly enhanced if it
is carefully collected by scholars in their own countries, who are then
linked together in a network and use a common set of research
instruments. In the last several years, we have been able to use
agent-based models to pursue some of our key theoretical work about
settings that are too complex to use formal analytical tools (Janssen
and Ostrom forthcoming A; forthcoming B).
I hope this long story illustrates the feasibility, better yet, the
overall strength, of using multiple methods to attack two tough
analytical puzzles. First, we need to understand the conditions that
facilitate resource users to overcome collective-action problems.
Equally important is understanding why efforts fail. This has been a
story of hard work and scholars from diverse backgrounds willing to work
together. We have not found diverse methods to be threatening. We have,
I think, used the opportunity to build a stronger research program
because of our multiple methods approach. We have found in our research
many heterogeneous groups who have devised rules that enable them to
cope successfully with the problems they face (see Varughese and Ostrom 2001). Perhaps we political scientists
can learn from resource users around the world how to overcome our own
heterogeneities so as to achieve a better understanding of the world
around us. As the chair of the Political Economy Section, I was
delighted that our spring-summer newsletter had two excellent papers.
One stressed the importance of case studies for unraveling complex
theoretical puzzles (Alston 2005). The second
stressed the importance of rigorous quantitative analysis when one is
working on a puzzle where sufficient evidence is available (Granato 2005). It is great to see both
approaches made so ably and side-by-side.
Universal Laws or Basic Elements of Analysis (Building
Blocks)
A major dividing line among members of our discipline when I first
started to teach was whether our “charge” as members of a scientific
discipline was to produce universal laws of political process and
outcomes that hold anywhere and at any time (similar to the laws of
physics). Those schooled in the tradition of Carl Hempel (1965) and Ernst Nagel (1961) sought only nomological, universal findings as the
object of study and strongly criticized most of the earlier work of
political scientists as not fitting that criteria. This view of the
enterprise underlay the takeoff of the “behavioral tradition” in
political science (Dahl 1961). This strong
commitment to using quantitative analysis to test propositions was
criticized by Gabriel Almond and Stephen Genco (1977) as a mistaken effort to apply hard scientific
strategies so that basic research in the social sciences did not differ
too widely from that in the natural sciences.
Some social scientists still dream of developing universal laws. Some
call their results general laws, as for example, “the general law of
supply and demand.” The predictions made by these “laws,” however, hold
only when a large number of conditions are present. Efficient market
clearance, for example, occurs only when there are clearly developed
legal systems defining property rights, regularized markets where
exchange can occur, extensive information systems about the attributes
of relevant goods, well-functioning courts established to clarify
contractual obligations, an honest set of police to enforce these laws,
goods that have few externalities and are undifferentiated on all
relevant respects, and citizen demands that are focused entirely on the
quantity and price of these goods—to name some of the most relevant
conditions necessary for the “law of supply and demand” to operate as
predicted.
Those of us schooled in the tradition of John R. Commons (1968), John Searle (1969; 2002), and Vincent Ostrom
(1980; 1997)
recognize that the social world differs significantly from the
biophysical world. As Searle (2002, 17)
captures the difference: “Absolutely essential, then, to understanding
the nature of the natural sciences is the distinction between those
features of reality that are intrinsic and those that are
observer-relative. Gravitational attraction is intrinsic. Being a five
dollar bill is observer-relative.”
In an earlier discussion of observer-relative facts, Searle (1995, 1–2) reflects that:
there are portions of the real world, objective facts in the
world, that are only facts by human agreement. In a sense there
are things that exist only because we believe them to exist. I
am thinking of things like money, property, governments and
marriage. Yet many facts regarding these things are ‘objective’
facts in the sense that they are not a matter for your or my
preferences, evaluations, or moral attitudes. I am thinking of
such facts as that I am a citizen of the United States, that the
piece of paper in my pocket is a five dollar bill, that my
younger sister got married on December 14, that I own a piece of
property in Berkeley, and that the New York Giants won the 1991
superbowl. These contrast with such facts as that Mount Everest
has snow and ice near the summit or that hydrogen atoms have one
electron, which are facts totally independent of any human
opinions.
The biophysical world, or brute reality, exists whether or not humans
think about it or relate to it. The social world simply does not exist
without humans using language to constitute it and relate to each other
and to relevant physical objects. The buildings or monuments of an
ancient civilization can remain as mysterious physical objects for
centuries. It is relatively easy, given advanced scientific measurement
techniques, to determine the physical composition and age of these
physical remains. To decipher the meaning of these physical objects,
however, takes archaeologists extensive work trying to unravel the
purpose of structural remains and to decipher linguistic clues.4
And, in line with my general argument
that multiple methods are essential to gain full understanding,
modeling alternative explanations is also very helpful (see
Janssen and Scheffer
2004).
Without knowledge of a relevant symbolic
system, however, the best one can do is to speculate on the hypothetical
use of a physical object (e.g., this must have been used as a tool or
for shelter).
Some social scientists are greatly impatient with the work of others who
identify a large number of conditions surrounding their explanation for
some process or event or point to some small “unimportant” variable as
causing a particular outcome. In the main, I do not view our primary
enterprise as one of explaining individual events. Rather, given the
number of conditions that affect most political choices, our normal task
is to understand structures and resulting processes and how specific
combinations of elements affect the likely outcomes of these
processes.5
There is no single cause
for human behavior. To live, we need oxygen, water, and
nutrition, and all are key parts of the explanation of life.
Life itself operates at multiple levels. Genes underlie
phenotype. Selection processes operating primarily at an
individual level, but also in a multilevel process, select for
some genes and against others. Complementing the biological are
many developmental and social factors (see Alford and Hibbing 2004). At any one point in time,
one or two variables may become the limiting condition in the
sense that John R. Commons discussed it. At that point, that
particular variable is crucial, but it is still not the
only cause of the complex patterns
observed.
There are times, however, when the outcome of a
particular process is unexpected, important, or has serious consequences
that we must dig deeply into the full set of conditions that led to the
particular outcome. When this occurs, our task is similar to that of
engineers who are asked to determine the cause of a failure of some
complex equipment such as an airplane, ferry boat, or rocket. No one
criticizes them for identifying some very small part of a large and
complex structure as having caused the unexpected.
6 While writing this paper during the summer of 2005 at
our log cabin in the Canadian wilds of the Manitoulin Island, we
listened to the CBC news of a major ferry boat accident in
Vancouver Harbor that was eventually attributed, after a
week-long investigation, to a small part in a clutch that had
broken.
Similarly, when we explain why a street demonstration erupts into
violence or eventually into a rebellion, it may turn a small factor that
triggered a frequency-dependent process to flip from peaceful to violent
means (see Kuran 1989; Schelling 1978). Ecologists have long studied complex
systems and examined the types of disturbances that may cause a flip of
a system from one fundamental pattern to another (Berkes, Colding, and Folke 2003; Gunderson and Holling 2001). We are faced with the necessity
of understanding how the configuration of multiple variables affects
outcomes and to recognize that the variables that caused a flip in one
dynamic process may not change the outcome in other broadly similar
processes characterized by slightly different combinations of
variables.
Then, how do we understand institutional structures and resulting
processes? Drawing on an immense literature on institutional analysis
from multiple approaches including historical, sociological, and
rational-choice institutionalism,7
we can now begin to identify some of
the basic elements needed to conduct theoretical and empirical analysis.
By “element,” I mean a part used in the analysis of the processes and
outcomes obtained in
any decision structure including: a
legislative committee or a full legislature, court proceedings, clean-up
activities undertaken by local communities, market exchanges, or
principal-agent relationships within a private firm or government
agency. This is a bold assertion. I do mean
any decision
structure that humans have self-consciously designed or evolved over
time.
One of the challenges that we face is that the number of elements
involved in even the simplest structure is quite large. Further, most
elements can themselves be broken down into sub-parts and then further
broken down into sub-, sub-parts of nearly decomposable systems. Simon
(2000, 753) describes the characteristics
of nearly decomposable systems: “They are arranged in levels, the
elements at each level being subdivisions of the elements at the level
above” (see also Simon 1995). Koestler (1973), in discussing biological systems, uses
the term “holon” to refer to these part-whole elements.
Instead of resisting this effort to identify elements, their sub-parts,
and how they go together as being too complex, we should recognize that
the physical sciences started with an identification of four basic
elements: fire, earth, water, and air. Modern chemistry has come some
distance in positing somewhat over 100 basic elements. In this overview,
I cannot list all of the basic elements of political structures.
Instead, I will provide a brief overview of the Institutional Analysis
and Development (IAD) framework that scores of colleagues at the
Workshop in Political Theory and Policy Analysis have developed over the
years (Kiser and Ostrom 1982; E. Ostrom, Gardner, and Walker 1994; V. and E. Ostrom 1997). It is a framework for
analyzing nearly decomposable systems ranging from families, to
neighborhoods, to large communities, to governmental units at a local,
regional, national, or international level (see E. Ostrom 2005; Gibson et al.
2005).
The focal level of analysis that we start with in using the IAD framework
is that of an action arena. An action arena is composed of
participants in positions choosing among
actions at particular stages of a decision process in light
of their control over a choice node, the information
they have, the outcomes that are likely, and the benefits
and costs they perceive for these outcomes. These seven basic
elements (italicized above) can be used to build an immense array of
superficially different structures. For complex structures, it is, of
course, necessary to break down the structure further into components
and sub-components rather than building the entire structure from basic
elements.
Obviously, one of the key working parts of this focal level of analysis
is the participant. Since the IAD is a framework for doing theory rather
than a theory itself, scholars may use and compare any of a diversity of
theories of human behavior (including critical theory [Schiemann 2000], evolutionary psychology [Cosmides and Tooby 1994], heuristic choice
[Gigerenzer and Todd 1999], or bounded
rationality [Jones 2001]) to animate the
structure being studied. It is also possible to compare outcomes
predicted by diverse models of the same broad theory rather than being
committed to using one and only one model of a theory (see, e.g., Cox and Sadiraj 2004; Ahn,
Ostrom, and Walker 2003).
Underlying the focal arena are variables that are treated as temporarily
exogenous while focusing on a particular focal level of analysis: the
underlying biophysical world, the rules used to structure the arena, and
the broader community in which the arena is located. For political
scientists, the rules creating the arena (or game form) under study are
themselves of substantial interest.
After many years of studying the rules in use in a wide diversity of
arenas, we have developed a classification system that focuses on the
rules that directly affect each of the seven working parts of all
arenas, enabling one to begin to capture the broadest grouping of rules
needed for consistent study of the impact of diverse rules on behavior
and outcomes. This leads to a classification of seven broad types of
rules: position, boundary, choice, aggregation, information, payoff, and
scope. Crawford and Ostrom (2005, 190)
explicate the relationship between the broad classification of rules and
the earlier-mentioned working parts of all arenas.
Position rules create positions (e.g., member of a legislature or
a committee, voter, etc.). Boundary rules affect how individuals
are assigned to or leave positions and how one situation is
linked to other situations. Choice rules affect the assignment
of particular action sets to positions. Aggregation rules affect
the level of control that individual participants exercise at a
linkage within or across situations. Information rules affect
the level of information available in a situation about actions
and the link between actions and outcome linkages. Payoff rules
affect the benefits and costs assigned to outcomes given the
actions chosen. Scope rules affect which outcomes must, must
not, or may be affected within a domain.8
This classification of rules is broadly similar
to the five “dimensions” of rules that Koremenos,
Lipson, and Snidal (2001, 763)
propose as basic tools used in the design of
institutions.
The specification of seven broad types of rules does not imply that there
is a one-to-one relationship between any particular action and a
specific rule. Anthony Giddens (1979) long ago
pointed to the fallacy of statements like “the rule governing the
Queen's move in chess” since what moves a Queen can take in chess depend
on multiple rules that constitute the game of chess and where in the
stages of a game a participant is contemplating diverse moves.
Each of the seven broad types of rules can be unpacked into
sub-categories and then further unpacked into sub-, sub-categories just
as the seven working parts of a focal arena can be unpacked multiple
times to address relevant research questions that require deeper
categories than the higher levels. To add further to the challenge of
analyzing complex, nested arenas, underlying any particular set of rules
are deeper policy arenas where changes in the focal rules may occur, and
even deeper constitutional arenas where changes in the rules of the
policy arenas are themselves debated and crafted.
Many important analyses in political science have focused on procedural
rules related to agenda setting or final approval in electoral settings.
Since many other aspects of these electoral settings are held constant,
the sometime relatively large and sometimes surprising differences in
outcomes result from small changes in what might be considered “dull”
procedural rules. Alt and Shepsle (1998)
compare three very simple voting arenas where an odd number of voters
are posited to have single-peaked preferences over potential outcomes as
well as complete and common knowledge and no time limits. In the first
two cases, the final aggregation rule is simple majority rule. The first
case is the well-known, median-voter model of Duncan Black (1948) who posited rules related to agenda
setting that allow anyone to make a proposal pitted against the status
quo, any proposal winning becomes the new status quo, and the decision
is reached when no further proposals are made. The final outcome is
posited to be the one preferred by the median voter.
In the second model, Romer and Rosenthal (described in Rosenthal 1990) change only two rules—creating a new
position and giving the holder of that position the sole authority to
propose changes to the status quo. All of the other assumptions used in
the original Black model are held constant. With this change, they find
several logically possible outcomes instead of only one. Changing only
these two rules and keeping all the other rules and attributes
of an arena the same, however, makes it very rare for the final outcome
to be that most preferred by the median voter. In the third model,
Shepsle and Weingast (1987) drop the
agenda-setter rules but add a new position with the authority to veto
any decision after it is reached by a majority rule. This change
generates a result that the final outcome is the one preferred by the
veto player.9
See also Tsebelis (2002) and Herzberg and Ostrom (1991) for further explorations of the
role of veto players in a wide diversity of arenas.
What is useful about theoretical exercises such as these is that one can
demonstrate the changes in posited outcomes that are likely to result
from changes in particular choice rules added to a simple majority
voting rule. David Mayhew (2003) provides an
excellent illustration of the empirical importance of these rules
(regarding who picks the agenda or must agree to the finality of a
majority vote) in regard to the U.S. Senate. There, the aggregation rule
is a simple majority, but it is compounded by a cloture rule requiring
agreement by any 60 senators to end debate and accept the last version
of a bill that had been agreed upon by a majority of the senators.10
Ken Shepsle (2003, 28) recently reflected on the beauty and fun
of scholars using multiple methods who respect each other's work
in acknowledging “the unparalleled cross-fertilization among
soakers and pokers, history jocks, quants, and rat choicers. The
congressional scholarship field has always been a methodological
rainbow.” It is my hope that we can all search out ways to
produce “rainbows” rather than “thunderstorms.”
By being consistent in building a theoretical model that holds the
assumptions made about participants constant and changing a limited set
of rules at a time or vice versa, one can examine the likely
consequences for situations that closely approximate the ones in the
theoretical models. When one asserts that the outcome of using majority
rule is X or Y, one learns from these examples (and many more in the
literature) that one needs to specify much more about both the other
rules-in-use and about the knowledge, information, preferences, and
other important attributes of participants.
Beyond this very brief illustration, the entire framework is obviously
too complex to present fully on this occasion. Those of us who have been
working with it over a long period of time have more and more confidence
in its strength as a core organizing theoretical framework given the
wide diversity of research that has successfully been conducted using
this framework (for relevant cites, see E. Ostrom
2005, chap. 1). Recognizing the complexity of the building
blocks of political structures can help us understand why prediction and
explanation are so challenging in our discipline. Our task is to
identify the working parts, the grammar, and the alphabet underlying the
patterns of human social behavior as well as the underlying structure of
rules, of community, and of the biophysical environment and how these
combine to generate incentives, actions, and outcomes. To do this well
over time, we will all need to pay closer attention to how we define and
use key terms. For concepts such as strategies, norms, and rules, we
frequently talk as if on a Tower of Babble where few use the same
definitions for the terms we all use (see Crawford
and Ostrom 2005).
Rational Choice—An Evil Approach or a Theory Undergoing
Challenge and Development?
One of the basic disagreements in political science is over the model of
the individual that we use or whether we even need to think about
individual-level choice behavior (Green and Shapiro
1994). It is obviously impossible for micro-theorists to
avoid working with a theory of how individuals make decisions within
diverse situations. A self-conscious awareness of the theory that one
uses to generate explanations or predictions is far better than keeping
assumptions about the processes of individual choice hidden from one's
own and others' views. Rational choice theory, as developed in
neoclassical economics and game theory, is a powerful and useful tool of
analysis.11
It is important to
recognize that one can use game theory without working at the
individual level as well as without the need to assume a single
preference function. Game theory has been extraordinarily useful
in helping political scientists understand two-level games and
some of the limitations of international negotiations (McGinnis and Williams 2001; Putnam 1988) as well as how conflicts
can tragically move from one bad equilibrium to another (McGinnis 2005).
Many of us
appreciate the clarity and precision that one can bring to the analysis
of complex, competitive situations using this tool. Modeling processes
is one important method for checking out that one's intuitive vision of
why outcomes are generated is logically sound. Like all tools, however,
one needs to learn for what problems it is useful and where it leads to
erroneous predictions.
The battle within political science has, however, not really focused on
when and where this theoretical tool helps to better understand
behavior. Rather, far too much energy has been devoted to criticizing or
defending the use of the neoclassical version of rational choice as the
only way to explain individual choice. If it were our only
way to view political behavior, we could not explain why some people
vote, engage in voluntary charities, or invest time and effort to create
rules and enforce them related to the sustained use of common-pool
resources.
Mancur Olson (1965) did political science a
gigantic service by challenging the naïve theory of latent groups or
quasi groups held by theorists such as David Truman (1951) and Ralf Dahrendorf (1959).
Assuming that individuals will routinely devote time and energy to the
pursuit of group benefits whenever common interests are perceived as
important is a comforting but not very useful foundation for the study
of democratic systems. Using this assumption, how institutions may or
may not enhance the level of assurance and feedback that individuals
obtain about the likely and continued contribution of others is not even
considered relevant. And yet, the ways that individuals gain a
reputation as being trustworthy have repeatedly been shown to be
essential in many settings.12
On the other hand, the prediction of zero cooperation in Olson's theory
of collective inaction has been shown to be only one outcome
out of many in indefinitely repeated settings (Kreps
et al. 1982), and is not empirically supported in many
empirical settings (Dawes et al. 1986; Frohlich, Oppenheimer, and Young 1971; E. Ostrom
1990; 1998; and
many others). This unlikely theoretical prediction, nevertheless, has
stimulated an immense body of empirical research and thus has been a
great contribution to the development of all of the social sciences. We
can appreciate the stubbornness of scholars who do not give up a theory
or method easily just because it does not do everything.
Considerable differences exist among scholars who rely on a
self-conscious theory of individual choice when explaining or predicting
behavior. Many political scientists lump all scholars who use a thin
model of human rationality together with those who are developing
second- and third-generation behavioral theories (Camerer 2003; Cox 2004; Frohlich, Oppenheimer, and Kurki 2004; Orbell et al. 2004; E. Ostrom
1998; Scharpf 1997) as well as
important dynamic theories of belief formation and investment (North 2005). One does not just toss out a theory
that has proven valuable in many settings because it does not work well
in others. Many efforts to broaden the theory are well underway, and it
will continue to be usefully employed to address many interesting
questions in competitive situations. Some of the factionalism does stem
from the arrogance of those who consider the continued use of a narrow
model of human rationality the essential qualification for doing good
social science.13
The extreme devotion to
exclusive use of the thin model of rational choice continues in
spite of strong advice to the contrary by at least four Nobel
Prize winners in Economics: Reinhard Selten, Douglass C. North,
Herbert Simon, and Vernon Smith (see Selten
1990; Gigerenzer and Selten
2001; North 2005; Simon
1999; 2000; Smith
2000).
Context does matter! Alchian (1950) long ago
argued that the neoclassical model of rational choice is an effective
tool for predicting behavior in an open, highly competitive market
setting. Satz and Ferejohn (1994) have also
shown its usefulness in other types of highly competitive settings. We
should not toss away useful tools in the social sciences. We need any
and all tools that are helpful for explaining and predicting behavior in
some settings. Further, we need to develop even further tools to analyze
dynamic processes in a changing world (North
2005). We must avoid, however, relying on the strategy
heavily used by all scientists that might actually approach being a real
“law”—“The Law of the Hammer.” Scientists do seem to follow a pattern
that when we have found a useful tool, we determinedly use it in any and
all environments. To provide sound explanations, we need to pay
substantial attention to the structure of a situation and the underlying
cultural, institutional, and biophysical context. Then, in light of
understanding this context, we can choose the model of individual
behavior most relevant to predicting outcomes in this type of situation
(see Pasotti and Rothstein 2002).14
When we are interested in studying how
individuals interact in settings that are relatively new to all
or most of them, we must rely heavily on any of several learning
models of human behavior (see Busemeyer and
Myung 1992), but when we are studying patterns of
interaction among participants who have a long history of
interacting in a particular setting, how they learn about the
values of the relevant variables is no longer as important to
analysis.
The Silo Organization of Our Ivory Towers in an Ever-More
Connected World
The increasing specialization that is occurring in most universities is
leading more and more researchers to focus almost entirely on their own
specialty and to know little about related work even in their own
discipline, let alone to work with colleagues in other disciplines.15
Richerson and Boyd (2005, 245) contrast how biology is taught to
freshman students—covering an immense set of subdisciplines—as
contrasted to the social sciences that narrowly focus on a
subdiscipline even in the freshman year.
Further, many
scholars do not get out into the communities surrounding their
universities or participate in ongoing collective action in other
domains to study real problems on the ground. Instead, they rely on
census data or other data that are made available to them. There is much
to study in the field! Madison's presumption that creating a federal
system would stimulate considerable experimentation at state and local
levels has certainly been substantiated (
McGinnis
1999b). Victor, House, and Joy (
2005) point to the success of “A Madisonian Approach to
Climate Policy” in which innovative policies are being tried out within
nations, regions, and even by individual firms.
As a scholar who co-directs two interdisciplinary centers—the Workshop in
Political Theory and Policy Analysis (the Workshop) and the Center for
the Study of Institutions, Population, and Environmental Change
(CIPEC)—I am frequently asked how we overcome the extremely high walls
that have been built around contiguous disciplines to study complex
phenomena that requires a multidisciplinary approach. My answer focuses
on the questions we address.
A core interest of colleagues associated with the Workshop is the
question of how individuals can build and sustain workable democratic
systems and sustainable resources at local, regional, national, and
international levels in all parts of the world.16
Colleagues in political science have
frequently chided us for the many studies we have conducted on “dull,
unimportant local problems.” If one confines political science to the
study of national elections, national legislative behavior, and the
politics of the presidency, we are missing a great deal at both a local
and international level.
17 It is now
common for textbooks in political science to focus almost
entirely on national and international affairs. In Ginsberg,
Lowi, and Weir's (2003) introductory
textbook, We the People, not even a chapter is devoted
to state or local governance. A real challenge for the
continuance of an informed citizenry in the U.S. is the
disappearance of state and local governance from all but a few
upper-division courses taught in most political science
departments.
We must draw on all of the social sciences
as well as ecology, biology, history, cultural studies, area studies,
and law to slowly build a better, multidisciplinary base for our work on
all levels of political organization.
18 For an overview of the research program of the Workshop, see
the special issue of the Journal of Economic Behavior and
Organization (vol. 57, no. 2, June 2005) edited by
Peter J. Boettke on Polycentric Political Economy: Essays in
Honor of Elinor and Vincent Ostrom, and McGinnis (1999A; 1999B;
2000).
A core question for colleagues associated with CIPEC is what are the
major causes of environmental change at local, regional, national, and
international levels? One must draw on the environmental sciences and on
methods like Geographic Information Systems, the interpretation of
remotely sensed data, as well as gathering field data about ecological
processes as well as human processes affected by and affecting the
ecological world (Moran and Ostrom 2005).
Scholars involved in genuinely interdisciplinary research have to learn
something about the other disciplines necessary to address a common
puzzle while not pretending to become a specialist in any but their own
foci.19
In other words, to work on a
land-use problem involving the use of GIS and remotely sensed
data, all members of a team must learn the elementary “basics”
of this technology. Without experts, however, on the team, bad
errors can be made in the execution of key steps in analysis and
the interpretation of findings. This mixture of modestly
educated and real experts is essential to the success of
interdisciplinary teams.
Unfortunately, promotion, tenure, and hiring committees are frequently
composed of colleagues specializing in one discipline who evaluate the
work of others by the number of publications in their own discipline and
totally discount publications in interdisciplinary journals. The
fractionalization within our own discipline makes it even more difficult
for young political scientists who participate in interdisciplinary
teams to be considered for tenured positions at major universities.
Fortunately, some innovative major universities are developing rigorous
programs that examine major policy questions from an interdisciplinary
perspective—a growing need in the contemporary world.20
See especially the new school of Human Evolution
and Social Change at Arizona State University; the Donald Bren
School of Environmental Science and Management at the University
of California, Santa Barbara; the George Perkins Marsh Institute
at Clark University in Worchester, Massachusetts; and the School
of Social Ecology at the University of California,
Irvine.
Facilitating research teams composed of faculty from diverse disciplines,
as well as students, is one way of breaking down unnecessary barriers
within universities as well as the barriers between universities and the
public. We learn a lot from studying how people solve or fail to solve
core problems that are repeatedly found in many diverse settings.
Studying how citizens from very diverse backgrounds coped with the
creation of multiple institutions at many scales to solve (in some
cases, and not in others) complex groundwater basin problems gave both
Bill Blomquist and I a substantial appreciation for the entrepreneurship
citizens can show when they are in an open society with authority to
create private and public instrumentalities (see Blomquist 1992; E. Ostrom 1990).
And, along the way, we had to learn a little about geology and
engineering!
One gains substantial insight into the importance of obtaining accurate
information about the structure of the problems being faced and being
able to challenge the validity of that information. In the cases we
studied, the groundwater users threatened by overuse were able to use
equity jurisprudence and ask a court to assign a fact-finding
watermaster to make a report as a disinterested party on the conditions
of the groundwater basins. Many ways exist for problem-solving
individuals to approach this challenge—equity jurisprudence is only one
way.
The key variables, however, are (1) an open system, (2) the availability
of impartial information organized at the right scale for the problem
being addressed, (3) the capacity to challenge information as well as
the implications that stem from diverse interpretations of the problem
and potential solutions, and (4) the creation of fora at appropriate
scales where potential solutions can be posed, argued about, and
eventually adopted with mechanisms that assure conformance with
agreements and sanctioning of non-conformers. These elements are
important for any system facing threats if the threats are to
be turned into opportunities for future improvements and developments.
Many of the new committees established to discuss governance within the
APSA, our journals, our undergraduate and graduate-level curricula, and
our hiring and tenure practices may enable us to grow stronger because
we have recognized some threats, obtained more accurate information
about them, and debated openly about how to interpret this information
and about which solutions may be feasible.21
The new section organization of the APSA is also a very good
structure to enable folks from diverse traditions interested in
some intersecting foci to meet, prepare newsletters, and discuss
methods, so long as it too does not encourage
factionalism.
If we continue along these lines of development, the opportunities for
the future are substantial. If we regress into ever greater
factionalism, we risk becoming irrelevant in a world that has become
ever more complex, multi-leveled, and in need of skilled political
scientists working with scholars and practitioners from other
disciplines to put forward their best ideas derived from multiple
approaches for citizens to consider in democratic fora. Let us meet this
challenge, even welcome it as an opportunity.
***
Let us learn from Madison about the dangers of faction for any group that
has multiple views of the best way forward. We are much stronger if we
learn to respect the multiple approaches proposed, both for how we do
our science as well as how diverse policies can lead to better outcomes
when learning and adaptation are enhanced. As a discipline, we can move
forward when we embrace learning from multiple approaches. As citizens
of the world, we must learn there are no blueprints that fit all the
puzzling policy questions we face. We improve our chances for survival
if we encourage a Madisonian approach to local, national, regional, and
global problems.
Biography
Elinor Ostrom is Arthur F. Bentley Professor of Political Science and
co-director of the Workshop in Political Theory and Policy Analysis, and
the Center for the Study of Institutions, Population, and Environmental
Change, Indiana University, Bloomington.
%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
