Introduction
Shortly after George W. Bush issued his ultimatum against Saddam Hussein’s Iraqi regime
to disarm and step down or face war with the United States, North Korea’s Kim Jong Il
slipped into a rare level of seclusion. Daily reports of his activities disappeared from North
Korea’s official media. In a culture where the primary focus of the society is on the Dear
Leader, such an absence of information is highly unusual. It is a short stretch to imagine
that leaders in Iran and Syria were paying close attention as well. Indeed, in the immediate
postwar period the Bush Administration seems to be counting on the assumption that US
actions in Iraq serve as a signal of resolve (Shanker, 2002).
Similar examples permeate world politics. The force displayed by the United States in its
air campaign against Serbia caught the attention of friends and foes alike. From France to
China, governments were forced to appreciate the consequences of divergent defense budgets.
Nor should we assume that only violent activities catch the attention of uninvolved parties.
Treaties and trade are likely examples of dyadic interstate behavior that is observed by other
states in the international system.
The claim that governments observe the behavior of their peers is easy enough to make.
But what remains puzzling is what these governments do with this information. How do
state leaders interpret the actions of their friends and foes? Does the war between the United
States and Iraq influence the North Korean government’s belief that war between the U.S.
and North Korea is coming? More generally, do nations alter their behavior with others
based on what these other nations are doing elsewhere? This broader puzzle is one of both
substance and research design. The dyadic level of analysis (pairs of states) has dominated
the study of international conflict for the last two decades (Bremer, 1992). It is a natural fit
with the parallel emphasis on strategic interaction, but it does a poor job of incorporating
information that lies beyond the dyad itself (Bueno de Mesquita, 1981; Fearon, 1994). While
we have recently made theoretical and empirical progress regarding the dependence of dyadic
observations across time (Raknerud and Hegre, 1997; Beck, Katz and Tucker, 1998; Crescenzi
and Enterline, 2001), we still have very little understanding of how dyads are influenced by
other states and other dyads. Scholars have recently renewed their econometric focus on this
spatial interdependence (Signorino, 1999; Heagerty, Ward and Gleditsch, 2002; Ward and
Gleditsch, 2002), but more theoretical progress has been absent.
There are two specific pieces of this puzzle to sort out. First, how does a state (more
specifically, its policy makers) process the information contained in international relations
where the state is not directly involved? Second, once this information is processed, what
1
Convention