Much of international behavior is linked spatially and temporally. Yet, analyses of interstate interactions generally either assume independence among units or resort to technical solutions to dependence that “throw away” relevant information. We detail a more informative and satisfying approach to modeling spatial dependence from extra-dyadic linkages in alliance ties and geographical proximity as specific pathways of conflict contagion. Beyond deterrence, the purpose of alliances is to draw other parties into dyadic contests, but most existing research on conflict onset generally only considers alliance ties within an individual dyad or external intervention in the same dispute. We develop new measures on third- and fourth-party alliance ties, demonstrating direct and indirect spatial effects of alliances on conflict onset. Similarly, ongoing contests can spread geographically, but dyads in some locations are much more at risk for conflict onset than others. We provide a new theory of geographic “inbetweenness” in conflict and show that dyads involving specific locations and ties to the ongoing conflict are much more likely to see dispute onset, even accounting for other purely dyadic factors. Beyond the intrinsic interest in the impacts of extra-dyadic position and alliances on conflict, our spatial approach can be applied more broadly to other extra-dyadic ties.