Tilly Goes to Church: The Religious and Medieval Roots of European State Fragmentation

The starting point for many analyses of European state development is the historical fragmentation of territorial authority. The dominant bellicist explanation for state formation argues that this fragmentation was an unintended consequence of imperial collapse, and that warfare in the early modern era overcame fragmentation by winnowing out small polities and consolidating strong states. Using new data on papal conflict and religious institutions, I show instead that political fragmentation was the outcome of deliberate choices, that it is closely associated with papal conflict, and that political fragmentation persisted for longer than the bellicist explanations would predict. The medieval Catholic Church deliberately and effectively splintered political power in Europe by forming temporal alliances, funding proxy wars, launching crusades, and advancing ideology to ensure its autonomy and power. The roots of European state formation are thus more religious, older, and intentional than often assumed.


INTRODUCTION
T he medieval church was a powerful, if neglected, force in European state formation. Canonical accounts of state building focus on war and violent conflict between secular rulers in the early modern era (1500-1800 CE). Taking the territorial fragmentation of Europe as a starting point, this "bellicist" literature agrees that early modern warfare consolidated the state system into fewer, larger states. The need to fund these wars also necessitated the rise of state institutions. Rulers who succeeded in waging war and extracting taxes went on to consolidate their territorial gains and ensure the survival and sovereignty of their states. Charles Tilly's summary is as succinct as it is well known: "war made the state and the state made war" (Tilly 1975, 42).
Yet empirical incongruities challenge this account. First, the fragmentation of Europe was extraordinarily persistent. It lasted well through the mid-nineteenth century, contrary to the bellicist argument that early modern warfare winnowed out and consolidated states (Spruyt 2017, 87). Second, many of the institutions that the bellicists claim were the inadvertent result of early modern warfare, such as taxation, courts, central administrations, or parliaments, arose long before the pressures of war supposedly made them necessary (Blaydes and Paik 2016;Stasavage 2010;. Third, conflict has not uniformly led states to consolidate. The early onset of military competition translated into a primitive and patrimonial administrations in Europe, rather than more effective and formalized ones (Ertman 1997). Wars produced crises: ancien regime France was exhausted by its military ventures, as was eighteenth-century Poland, so that "war unmade these states" (Teschke 2017, 45). More broadly, war can hinder state formation, ending intensive economic growth, spreading disease, and depleting the labor supply (Fouquet and Broadberry 2015;Ober 2015;Saylor and Wheeler 2017;Voigtländer and Voth 2013). State formation and warfare did not go hand in hand in other regions (Centeno 2002;Doner, Ritchie, and Slater 2005;Herbst 2000;Mazzuca 2021;Thies 2005;Hui 2005).
To explain these anomalies, we need to look beyond early modern warfare. The foundational era of European state formation 1 goes back centuries earlier, to the medieval period (1000-1350 CE). The Roman Catholic Church (or "the Church") was the dominant political actor in the Middle Ages, and it was most powerful from 1100 to 1300. 2 As a result, the strongest rival for an ambitious medieval ruler was not another monarch, but the Church. The Church wielded its enormous wealth, human capital, and moral authority to ensure its own autonomy and preclude the rise of a rival dominant power. It sought to fragment those rulers it saw as a direct threat to its autonomy by using temporal alliances, wars by proxy, and ideology. It 1 State formation is the process by which rulers amass authority over territory and populations. The "state" is an anachronism in the Middle Ages, when "lordship" would have been more intelligible (Davies 2003). Nonetheless, we can still meaningfully discuss the stated goals of these rulers: (a) a more effective set of mechanisms through which they could exercise authority, such as the nomination of officials, a legal apparatus, and taxation and resource extraction (state building) and (b) the assertion of that authority over people and territory, free from internal rivals or external interference (sovereignty). 2 This was also the critical period for the rise of cities, the growth of commerce and trade networks (including the founding of the Hanseatic League and the expansion of trade to Asia), the development of law and reinterpretation of Roman law, the rise of universities, and new scientific and technological advances. challenged would-be hegemons and indirectly gave local notables and independent cities opportunities to grab authority.
As a result, the territorial fragmentation we observe is no accident: instead, it was deliberately instigated by the Church, in an attempt to protect its autonomy. Constant conflict with the papacy contributed to the fragmentation of authority in Europe. The Church's chief target and rival was the Holy Roman Empire. 3 There, constant conflict with the papacy meant that central state power could not consolidate: while German emperors were frequently abroad fighting, local princes and cities assumed control over their local territories. In other areas of Europe, such as England or France, rulers could consolidate central authority more easily, either because the papacy left them alone or aided them. The material resources and institutional innovations of the Church also provided models for secular taxation, legal frameworks, and administrative division of labor (Grzymala-Busse 2023a), but the Church could also hinder the adoption of these institutions by fragmenting the central authority of rulers.
In short, I argue that the Church fragmented the territorial authority of powers it saw as hostile (chiefly the Holy Roman Empire), and helped to consolidate and to strengthen states in other areas. By emphasizing the role of medieval religious authorities, this analysis also contributes to a more recent literature that has emphasized the deep (and secular) history of the European state: the influence of the Crusades (Blaydes and Paik 2016), the development of legal systems (Cantoni and Yuchtman 2014;Møller 2019), cities and communes (Abramson 2017;Møller 2017), representative assemblies (Boucoyannis 2021;Stasavage 2010), and urban self-government and interdependence (Bosker, Eltjo, and van Zanden 2013;Møller and Doucette 2022). This study extends the framework first developed by Hintze ([1931] 1975) on the role of the Church in transmitting Roman precedents of the rule of law and the formalization of assemblies, as well as subsequent work exploring how the Church promoted the rule of law and diffused selfgovernment (Bendix 1978;Doucette and Møller 2021;Ergang 1971;Fukuyama 2011;Kiernan 1965;Møller 2019;Møller and Doucette 2022;Poggi 1990).
Scholars have noted the importance of the conflict with the Church, but there has been far less focus on the Church's direct role in fragmenting territorial authority. Thus, Møller (2019) focuses on the "crisis of church and state" as critical to subsequent "rise of Europe" (by which he means the multistate system, medieval parliaments, and early bureaucratic institutions) but devotes less attention to the mechanisms by which the papacy fragmented territorial rule. Møller and Doucette (2022) focus on the spread of the Cluny reform program, as critical to fragmentation, urban selfgovernment, and the rise of national assemblies. My argument here is that papal effort was the antecedent, leading to fragmentation, communes, and institutional adaptation.
Below, I argue that the Church was a critical force in medieval state formation in Europe. I first review the dominant "bellicist" account of state formation. I then take Charles Tilly to church, and examine the medieval papacy as a powerful player and adversary. The church wielded its wealth, spiritual weapons, and military alliances to deliberately fragment territorial authority and ensure its own autonomy. I examine two alternative explanations: the rise of communes and primogeniture. I conclude that the Church contributed to the lasting fragmentation of authority-and the rise of a European state system characterized by multiple competing states.

THE BELLICIST ACCOUNTS
The august bellicist accounts share several attributes. First, these analyses start with the fragmentation of territory and authority in post-Carolingian Europe, and view the decline of fragmentation as evidence for state consolidation. The starting point for state formation in Europe is the territorial fragmentation after the collapse of the Carolingian dynasty in 888 (Ertman 2017, 63;Gorski and Sharma 2017, 99;Mitteraurer 2010;Teschke 2003, chap. 3;Wickham 2016). Europe was a raft of principalities, ill-defined kingdoms, and territories controlled by local warlords. No empire arose in Europe that could compare to the Roman one: it was simply too difficult to sustain (Scheidel 2019). The potential causes of this initial fragmentation vary. Scholars have pointed to the uneven emergence of urban life (Abramson 2017), the rise of local warlords and bands of knights (Bisson 1994), and the low levels of religious legitimation that made European rulers weak (Rubin 2017; see also Fischer 1992).
Second, this fragmentation declined thanks to constant interstate conflict, according to the bellicists. The relentless pressures of warfare eventually meant fewer and bigger states, a change from as many as 500 independent states in Europe in the year 1500 to 30 four centuries later (Bean 1973, 204;Tilly 1992, 45-6). Repeated invasions and conflicts winnowed out weaker states and led to vigorous new efforts to tax and extract resources. Favorable geographic location meant that some states, such as Switzerland or England, could forego a military buildup. Those without such advantages, such as Poland, eventually disappeared (Downing 1992).
More broadly, bellicists emphasize that secular conflict drove state formation. Kings, princes, and emperors fought to consolidate territory and control people and resources. Those who succeeded developed as states. War consolidated larger states and forced the building of state institutions. Following in the footsteps of Hintze, who argued that the threat of war led to the consolidation and centralization of European states, historians, sociologists, and political scientists such as Anderson ([1974] 2013), Bean (1973), Downing (1992), Mann (1986), McNeill (1982, Parker (1992), and Tilly (1992) emphasized the fierce pressures of military competition. 4 Third, the bellicist approach tends to view state institutions as the incidental "byproducts" of these preparations for war (Tilly 1992, 26, 75). With the Military Revolution of the sixteenth to seventeenth century, war became increasingly costly, and necessitated the formation of state administrations to extract resources (Downing 1992;Mann 1986, 486;Tilly 1992). 5 Taxes, tributes, and rents then allowed these states to wage war with greater force and success. 6 The collection of these taxes required surveillance, which in turn prompted the growth of state administrations (Tilly 1992) and the rise of national assemblies as sites of negotiation about taxes (North and Weingast 1989). In more fine-grained bellicist accounts, the timing and context of war shaped regime development: early military competition led to patrimonial administrations and relatively weak local governance facilitated absolutist regimes (Ertman 1997). Geopolitically exposed areas such as France or Russia required massive economic mobilization, and thus abolished medieval constitutionalism in favor of militarized absolutism (Downing 1992).
Finally, the prevailing view is that the peak of state building took place in the early modern era, from roughly 1500 to 1800, also characterized by the rise of sovereignty and institutions such as parliaments. Thus, "the state" was invented as a corporate entity only in early modern Europe. That is when war became both costly and intense, increasing the pressures to consolidate and to extract resources. In the conventional periodization, state formation dates to the early modern era, from the mid-sixteenth to late eighteenth centuries, taking off only after 1600 (Ertman 2017, 54;Tilly 1975, 170). Scholars of international relations often echo these claims, and argue that the modern state arose with the Treaties of Augsburg (1555) and Westphalia (1648), which established the principle of state sovereignty. 7 Thus, states grew and strengthened in Europe as a result of vicious early modern warfare, the competition for land and people that it entailed, and the mobilization of resources and people that war demanded.

THE MEDIEVAL CHURCH AS A FORCE IN STATE FORMATION
Shifting the focus to the medieval era reveals distinct aspects of state formation that the bellicist perspective may obscure: the fragmentation was persistent and not reduced by warfare, the relevant rivalry involved religious authorities, not just secular ones, and state institutions of taxation, parliaments, and justice arose long before the early modern era, often patterned on church models. As a result, many aspects of state formation date back to the medieval period, rather than the early modern era.
First, territorial fragmentation took off in the twelfth century, it was persistent, and it was unevenly distributed. Figure 1 shows the overwhelming and persistent fragmentation of the former Holy Roman Empire in comparison to the rest of Europe. The graph plots the number of states that existed over the twelfth to twentieth centuries in Europe. I take the 1900 borders of Europe, and calculate the number of states that existed within these borders over time. To construct this measure, I added historical European country borders to data on state size from Abramson (2017). 8 Contrary to the bellicists, this fragmentation did not end in the early modern era with the onset of intense warfare. The fragmentation of the Holy Roman Empire was exceptionally durable. 9 Tilly's observation that the number of states decreased from 500 in 1500 to 20 in 1900 may need a broader context: two enormous territories, Germany and Italy, remained fractured until their political unification in the late nineteenth century (Spruyt 2017, 87). This fragmentation persisted because cities and local princes gained power and precluded the assertion of central authority in these territories. As a result, there is no steady pattern of consolidation of ever-larger states through the warfare of the early modern era.
Second, religious rivalry, not interstate conflict, was critical in this earlier period of state formation. The Church, more than any secular monarch, was a powerful geopolitical force in the medieval Western Christendom, as we will see below. It assiduously sought to foment fragmentation. As a consequence, the 4 Thomas Ertman differentiates Hintze's earlier work, with its emphasis on the geopolitical context and war, from his later scholarship, which emphasized uneven state development: rulers in the core of the former Carolingian empire, such as France, built bureaucratic administrative institutions with which they could challenge local lords. The periphery developed strong local governments and lords that could either accompany a powerful monarch (as in England) or dominate weak ones (as in Poland, Hungary, or Bohemia) (Ertman 2017, 63-5). This balance of power and local assemblies leading to the emergence of representative institutions are important themes in Downing (1992), Ergang (1971), Ertman (1997), and Kiernan (1965). 5 Tilly and Mann both also examine the interplay of capital and war, acknowledging that bargaining processes with social classes, most notably capitalists, were critical in generating revenue and thus variation in state building. See also Anderson ([1974] 2013). 6 Historians working in this multidisciplinary tradition focused on the intentional development of specific institutions of the fiscal state, analyzing the early modern regimes of taxation, extraction, and warmaking, rather than viewing them as incidental (Bonney 1999;Brewer 1989;Glete 2002;Stone 1994). 7 See Held (1995), Morgenthau (1985), Philpott (2001), and Watson (1992). Others dispute the idea that Westphalia marked the rise of state sovereignty (see Krasner 1993;Osiander 2001;Teschke 2003). Yet, as de Carvalho, Leira, and Hobson (2011) note, the textbook consensus on the importance of 1648 remains. 8 Data from the Mosaic historical maps project: https://censusmosaic. demog.berkeley.edu/data/historical-gis-files (accessed August 2021). 9 The one dip is in the aftermath of the Thirty Years' War, when Sweden temporarily conquered large swathes of German territory. fundamental rivalry of the medieval era was the struggle between popes and rulers over authority and supremacy (and its converse, sovereignty). The conflicts between the Church and various monarchs in the early medieval era were recurrent and unrelenting. Church efforts thus helped to prevent the rise of a hegemon in Europe, and instead maintained a more polycentric equilibrium. In maintaining this balance of power, the Church targeted what it saw as hostile rulers, allowing others, such as the Capetians in France or the Normans in England, to consolidate their power (see also Downing 1989, 214). Both popes and monarchs were as ambitious as they were relatively weak: neither could fully enforce laws or agreements, nor claim full control of territory. 10 Moreover, spiritual and secular authorities were intermingled, as were morality and law. As a result, these conflicts were not interstate rivalries, but personalized struggles over authority.
Third, state institutions arose much earlier than required by early modern warfare. Figure 2 summarizes the pattern of institutional development across several European political entities. The bands summarize when several major state institutions arose: chanceries, cameras (accounting chambers), taxation, legal courts, and national councils and assemblies. Several patterns are evident. First, the Church was an institutional pioneer, developing these institutions as early as the eleventh century. Second, the Holy Roman Empire stands out as a late adopter. No central taxation, parliament, courts, or chancery emerged in the Empire until 1500 and they were weak and unstable once they arose. The Empire developed these institutions long after late-developing peripheral countries, such as Sweden and Poland, did. Italian institutions arose relatively early on-but these were not on the level of central government, but on the level of communes. Finally, where the pope needed rulers as allies, monarchs had more opportunity to develop state institutions in the late twelfth and thirteenth centuries. For example, successive popes relied on English kings to remain neutral in the papal conflicts on the continent, if not to aid the papacy outright. Accordingly, central state institutions in England, including a judiciary and court system, taxation, and local governance emerged in the twelfth century with little papal interference. Monarchs favored by the papacy could more easily adopt church templates for courts, chanceries, and tax collection and we see steady institutional gains in England, Spain, and France.
The church itself was the source of multiple institutional templates: legal advances, administrative precedents, and conceptual innovations. The extensive history of institutional borrowing from the Church is explored elsewhere, 11 but to give a few examples: canon scholars preserved and reinterpreted Roman law, and the idea of a state based on law. The papacy showed rulers how to collect taxes, answer the flood of petitions, and keep records and accounts. Concepts such as proctorial representation, supermajority rules, and binding consent all followed from church conciliar practice and theory. In short, the papacy and the Church provided valuable institutional prototypes. The church also provided the conduits: these templates were transmitted through church documents and legal innovations, by clerics serving in the courts and chanceries, and by bishops, who regularly sat in the royal FIGURE 1. The Fragmentation of Europe, 1100-1900 10 As Gorski and Sharma (2017) and Sharma (2015) point out, rulers struggled over authority and control of people more than over territory per se. 11 In a series of works, Møller (2015; emphasizes that modern legal institutions and parliamentary practices are rooted in the Middle Ages. Grzymala-Busse (2023) traces at the emulation and adoption of state institutions, such as courts, parliaments, and administrations. councils and national assemblies, served as judges and chancellors, and who had both the training and the access to effectively transmit these templates. 12 Only some rulers could adopt and develop these institutional models: in areas targeted by the papacy, such as the Holy Roman Empire, these central state institutions did not develop.
This precocious institutional development also suggests that "contractarian" accounts of institutional formation also may need revision. In these accounts, institutions arise as a result of bargaining among elites in the early modern era. Specifically, these accounts argue that fiscal and representative institutions arose as commitment devices: when nobles could withdraw resources, monarchs and spending were constrained through institutions. Thus, early modern parliaments gained powers of consent and imposed constraints on the rulers in exchange for taxation and revenue (Barzel and Kiser 1997;Blaydes and Chaney 2013;Hoffman and Rosenthal 1997;Levi 1988;North and Weingast 1989). Yet many of these institutions predate the early modern era: for example, medieval parliaments were powerful sites of consent, legitimation, and judgment (Boucoyannis 2021 andStasavage 2016). 13

The Church Gains Autonomy
The Church gained autonomy and power in the late eleventh to early twelfth century. Until then, under the system of proprietary churches, lords and kings built churches, named clergy, and profited from church lands and revenues on their territory. The system made churches a lucrative source of income, and provided rulers with both revenue and military support (Joachimsen 1978, 13). These relations were especially prevalent in the Holy Roman Empire. The Emperor also claimed the right to name the Pope: Henry III (r. 1045-56) had essentially appointed four loyalists as popes. After his death and the ensuing succession struggles, a power vacuum opened up at Rome, which the papal reformists used to ensure in 1059 that cardinals, not emperors, would now elect the pope in the newly founded College of Cardinals.
Starting in the 1050s, the papacy asserted its power within the Church (Morris 1989, 33). The arrival of Pope Gregory VII (r. 1073-85) heralded a new era of reform, Church autonomy, and papal authority. Gregory quickly launched an ambitious reform program that freed the Church from secular interference and instilled greater discipline among the clergy. Papal power grew immensely during the twelfth and thirteenth centuries. As the papacy consolidated power within the Church, popes gained new confidence in demanding autonomy and even obedience from secular rulers.
A major episode was the Investiture Conflict (1075-1122), a conflict over papal and imperial authority that was nominally a dispute over the naming of bishops. (The "investitures" in question were the rights to name bishops.) 14 Bishops were important agents for both rulers and popes. They served as administrators for secular monarchs and the spiritual deputies of the pope. They held high spiritual and secular office, kept order and defended territory, collected taxes, issued local judgments and petitions, mediated disputes, and served as papal emissaries (Angelov and Herrin 2012, 170;Møller and Doucette 2022;Robinson 1990, 423). The bishops' loyalty was thus of paramount importance to both monarchs and popes and naming bishops was an exercise in ensuring both fealty and effective administration.
FIGURE 2. The Emergence of Central State Institutions in Europe 12 These templates did not comprise a wholesale importation, as earlier work might have suggested (Strayer 1970;Ullmann [1955] 1965). As Hintze notes, "procedures and ideas of these chancelleries passed from country to country, from court to court, and that in this way a certain uniformity of thinking about politics and administration was established which gave way only much later to the advancing differentiation of national characters" (Hintze [1931. 13 That said, the rise of administrative institutions is not equivalent to the rise of the state (Sharma 2017, 217) and state capacity is distinct from the territorial state (Spruyt 2017, 84).
14 Popes endowed bishops with their spiritual powers, whereas secular rulers endowed the clergy, bishops, and abbots as vassals to monarchs. As part of the ceremony, the king would present a bishop with the symbols of religious office: the staff and the ring, and with rights and privileges (regalia). The clergy would then swear fealty to the ruler who named them.
For the papacy, the Conflict served to assert papal power and to liberate the Church from secular interference (Schatz 1996, 81). Pope Gregory VII prohibited investiture by lay rulers as part of his reforms. Meanwhile, for the new German King Henry IV, controlling the bishoprics was critical to consolidating authority in Germany. The stakes were fundamental: "much of the emperor's power depended on his investiture right, since it linked high church officials to the crown as a counterweight against German territorial nobles" (Clark 1986, 668). When Henry IV began to name bishops, the Pope excommunicated him in 1075 and called on his lords to abandon Henry. Five decades of multiple excommunications, conflict, and negotiations over investiture and the delineation of temporal and ecclesiastical authority ensued.
The Concordat of Worms formally settled the controversy in 1122. The fundamental consequence of Worms was that it differentiated church from state, and helped the church to gain relative autonomy from temporal rulers. The Church selected the bishops and invested them with spiritual authority: the monarch could confer secular (but not spiritual) privileges (Robinson 1990, 437). The pope did not enforce the investiture agreements equally: outside of Germany, Italy, Burgundy, and France, Pope Gregory VII did not enforce lay investiture, since "he had no wish to alienate powerful rulers of the periphery" (Cowdrey 1998, 550). 15 Neither the papacy nor secular rulers could claim a decisive victory, but the Church now gained formal autonomy.
Having liberated the Church, the papacy now assumed a new "power of intervention and direction in both spiritual and secular affairs" (Southern 1970, 34). A spectacular example is Pope Innocent III (r. 1198-1216). A proponent of papal supremacy and an ambitious leader, Innocent III threw himself into temporal politics, crowning and deposing kings, and settling disputes. His successors asserted even more authority: Innocent IV (r. 1243-54) argued that popes were above human law, and Boniface VIII (r. 1295-1303) proclaimed papal authority that extended over all beings. Popes used this power to assert the autonomy of the Church and prevent the rise of another imperial threat. 16 Papal power eventually peaked in the fourteenth century, starting with the conflict between Philip IV the Fair (r. 1285-1314) and Pope Boniface VIII (r. 1295-1303) and the Great Schism that followed (1378-1417). 17 It then waned both because the popes demanded ever more resources and jurisdiction from various monarchs, and because in the meantime, state capacity had increased thanks to the adoption of institutional models from the Church. For over two centuries, however, from roughly 1100 to 1300, the Church exercised unparalleled power in European politics.

Sources of Church Power
The medieval church was so powerful because it had considerable resources. First, there was its wealth. The medieval Church was the single biggest landowner in Europe (Spruyt 1994, 44). A large portion of central Italy was a papal domain. By the time of the Reformation, over half the land in Germany was held by the Church and by the ecclesiastical princes (Goody 1983, 131). Immediately before Henry VIII dissolved the monasteries in 1536-41, the English church held 25% of English land, whereas the crown had only 6%. These enormous land holdings were the result of earlier accumulation, in the seventh to tenth centuries, with voluntary offerings, property transfers, and bequests. Tithing entitled the Church to collect a 10% tax on all income, generating huge revenues (Morris 1989, 388). Given this wealth, "one can hardly overestimate the importance of the Church as an economic entity in preindustrial Europe" (Cipolla 1993, 45).
A second resource for the Church was its human capital: literate clerks, legal expertise, extensive documentation and archives, and administrative experience. Taxation required both authority and administrative capacity and the Church developed financial and accounting offices, as well as a network of enforcement officers. Popes sent legates across Christendom to monitor religious and fiscal discipline (Riley-Smith 2005, 175). Clergy served at royal courts, writing letters and writs, and keeping accounts. They enforced local contracts, collected taxes, and recorded births, deaths, and wills in cathedral records. Cathedral chapters founded schools to study canon law, and kept records and relied on written documents, lowering transaction costs and allowing information to spread (Blum and Dudley 2001). Monasteries were a source of theological advances, a literate culture, and reformist zeal (Doucette and Møller 2021). The church further chartered universities and promoted the study of law starting in the late eleventh century.
Bishops were especially important, as the Investiture Conflict suggests. They held both high secular and ecclesiastical office, "providing both sacral authority and literate clerics for his chancellery, backing his judicial authority with legitimacy and efficiency" 15 Bueno de Mesquita (2000;2022) argues that Worms and lay investiture drove a permanent wedge between kings and popes, by incentivizing the former to promote economic development and the latter to hamper growth. Yet the struggle between popes and kings continued and lay investiture itself applied inconsistently and unevenly. For example, some kings, such as the French Louis VII in 1149 or Philip in 1203, voluntarily withdrew their rights to name bishops (Baldwin 2004, 518, 524). 16 That said, the Church was not a monolithic institution. Clergy initially opposed the elimination of clerical marriage and concubinage introduced by Gregory VII in the late eleventh century (though they came around by the II Lateran Council of 1139). Several schisms racked the Church, as competing popes emerged in 1130 and 1159 and famously during the Great Schism of 1378-1417. Bishops often "interpreted" papal requests or declarations in their favor (Dorin 2021). 17 The eventual consequence of this conflict was the exile of the papacy to Avignon, an area firmly under Philip's control, where it remained from 1309 to 1376. The Great Schism then ensued from 1378 to 1417, 40 years that saw competing papacies, highly politicized claims, and the eventual resolution of the crisis by secular rulers, rather than councils or clergy (Kaminsky 2000, 680). (Mann 1986, 382-3). As late as the fourteenth century, English bishops exercised discretionary justice in parliaments, councils, and chanceries (Dodd 2014, 216). There was no "clear area of separate governmental responsibilities that could be termed secular" (Morris 1989, 18). The Church provided legal arbitration and judgments for both clerics and lay people (Gilchrist 1969, 9) In short, the "great achievement of medieval civilization of the eleventh to thirteenth centuries would not have been possible but for the learning, example, and progressive character of the clergy and monks of the time" (Gilchrist 1969, 69).
Finally, the Church's power derived from its moral authority. The Church was deeply present in everyday life as both religious and secular authority (Mann 1986, 380). It "governed birth, marriage, and death, sex, and eating, made the rules for law and medicine, gave philosophy and scholarship their subject matter. Membership in the Church was mandatory: expulsion was tantamount to a social death" (Tuchman 1978, 32). Above all, the Church offered salvation: the promise of an eternal life and divine mercy that no secular ruler could possibly match. Conversely, the Church could also exclude the faithful from this ultimate benefit, through excommunication and interdicts.
In short, its wealth, administrative capacity, and its spiritual authority made the medieval church uniquely powerful. These advantages also meant that the Church would be an institutional pioneer-and that secular rulers would adopt ecclesiastical models. The Church became an ambitious and influential political actor and it fought to retain its autonomy, preclude the rise of rival superpowers, and consolidate its own administration over souls and territory. It formed alliances with secular rulers, financed military campaigns, and deployed spiritual weapons. The result was the peculiar pattern of the fragmentation of territorial authority in some areas of Europe, and the early consolidation of central states in others.

THE IMPACT OF PAPAL RIVALRY: THE FRAGMENTATION OF EUROPE
The broad scholarly consensus is that the fragmentation of territorial authority is the foundation for subsequent political and economic modernity in Europe, a point of departure for state formation and economic development. 18 Yet this fragmentation was no accident and it was deliberately sustained. As the papacy sought first to free the Church from secular influence, and then to prevent imperial resurgence, a special target of the popes was the Holy Roman Empire, and its ruling dynasty, the Hohenstaufens, who greatly expanded its territory from 1138 to 1254. These emperors repeatedly sought to rebuild the Carolingian empire by controlling both northern Italy and Sicily. Had they succeeded in this pincer movement, they would leave the papal states surrounded by a powerful rival, and the Church again under imperial control. The papacy sought to ensure that Germany would remain fragmented and Italy under papal control.
The Church marshaled its material, human, and spiritual resources to gain an advantage against its rivals. It deliberately played rulers against each other, and used both spiritual weapons and wars by proxy to ensure that no powerful rival could arise that might threaten its political or territorial interests. Popes tried to take successive states out of Emperor's sphere and into their own. These efforts drained the resources and attention of imperial rulers, hindering their ability to consolidate central power. These papal campaigns were so successful that Innocent IV (r. 1243-54) destroyed the imperial authority in Italy that had been already atomized in Germany, and the Hohenstaufen dynasty collapsed in 1268. Both Italy and Germany remained politically fragmented until the nineteenth century (Oakley 2012;Ozment 1980, 144;Stollberg-Rilinger 2018).
This fragmentation persisted because newly ascendant cities and princes precluded the consolidation of central authority long after the Church's power declined. Distracted by their attempts to annex territory in Italy, successive emperors failed to consolidate their central power. Instead, cities and princes gained regional authority relative to the emperor. First, the power vacuum after Henry III's death in 1056 meant 20 years without a ruler in Italy, giving the initial impetus to the "communal revolution" and the rise of increasingly autonomous cities in Italy, such as Pisa, Milan, or Lucca (Hyde 1973, 49). These cities then began to control neighboring territory and in effect became regional powers. Second, local lords grew in power, gaining greater control over serfs and taxation at the expense of the emperor (Clark 1986, 668;Mitteraurer 2010, chap. 5). Bishops and abbots used the political and financial authority bestowed upon them by church to strengthen their lordship rights (Stollberg-Rilinger 2018, 22). While the German emperors focused on the conflict with Rome, they had neither the time nor the resources to stem this leakage of authority. Communes and powerful regional princes who "grew in strength as a result of the conflict between kings and popes...could defeat any imperial plans to centralize administration or tax collection" (Hay 1995, 317). The emperor lost power to princes, towns, and bishops, who had a vested interest in maintaining fragmentation and preventing the imposition of central state authority or institutions.
In contrast, where popes sought the cooperation of rulers, monarchs could consolidate territorial authority more easily. The papacy supported the Spanish unification and Reconquista. It sought English neutrality in its conflicts on the continent, and largely left English politics alone. Medieval England was able to centralize the state and develop its own endogenous institutions, such as common law. France until the late thirteenth century cooperated with, and benefited from, the papacy. In France, popes supported the unification and consolidation efforts of Louis VI (1108-37), Louis VII, (1137-80), and Philip Augustus (1179-1223) (Baldwin 2004, 510). Since French monarchs tended to be allied with the Pope, the nobles could not challenge the monarch as successfully in the twelfth and thirteenth centuries and a more centralized French administrative state could develop. The French monarchs proved recalcitrant in the late thirteenth to early fourteenth century, when the papacy had grown weaker (and had moved to Avignon, from where it supported the French monarchy). In areas where the rulers posed little threat to the papacy, such as the Spanish territories, France until the end of the thirteenth century, Scandinavia, or East Central Europe, the Church was more influential in transmitting institutional templates-but also had less interest in fragmentation.
The general pattern is of fragmentation and hindering of central state development in Germany, Burgundy, and Italy, endogenous state development in England, and lower fragmentation and higher transmission of church templates elsewhere. Fischer (1992) argues that the Church failed to maintain European unity (Fischer 1992, 438). Rubin (2017) agrees that Europe was fragmented because relatively feeble religious "propagating agents" failed to legitimate secular rulers and strengthen them. Yet it is not that the Church failed to legitimate monarchs. Instead, the Church deliberately sought to balance them against each other and precluded any from gaining too much authority. Its main target and enemy was the Holy Roman Empire, which kept Italy and Germany fragmented, even as it allowed other states, such as England or Spain, to consolidate.

Weapons of the Meek? Papal Strategies of Confrontation
To protect the Church's autonomy from secular powers and ensure its status within Europe, popes relied on both spiritual weapons and armed conflict to target individual rulers and to undermine their authority over territory and people. They excommunicated and deposed monarchs, princes, and nobles, cutting off hostile or disloyal rulers from the community of the faithful and releasing them from loyalty to the monarch (Helmholz 1994;. They placed entire communities under interdict, cutting them off from sacraments in the hopes of fomenting disloyalty to the kings and monarchs that provoked papal displeasure. Yet these were surprisingly weak weapons. They did not consistently weaken rulers or remove them from office: excommunications destabilized monarchical rule and made it less predictable-but threatened only newer, more vulnerable rulers, and only weakly (Grzymala-Busse 2023, chap. 2). This is partly because to impose excommunications successfully, popes still had to rely on secular support compliance: the "consent and cooperation of [other] secular rulers" (Southern 1970, 20).
And, these supporters could be fickle: for example, when German emperor Henry IV was excommunicated for the first time in 1076, his nobles began to abandon him and his enemies elected an anti-king. He had to beg forgiveness of the pope. However, when he was excommunicated a second time in 1080, his princes stayed loyal-and he began to again encroach on papal territories in earnest. As Møller (2019) points out, excommunication is "only effective if it creates opposition against the targeted ruler" (217).
As a result, popes used temporal weapons to protect their secular interests. They allied with secular rulers, using favors, financial subsidies, exemptions, and dispensations to cement these coalitions. They funded wars by proxy, armies, and joint ventures, including the Crusades, to attack their enemies. They launched ideological salvos against imperial hierarchy and hereditary monarchy. In short, the papacy used a full range of privileges and prerogatives to punish imperial holdouts and reward papal loyalists (Whalen 2019, 180). These strategies destabilized some rulers, maintained the fragmentation of Europe, and affected the development of central state institutions.

Papal Conflict: Alliances and Wars by Proxy
To protect their interests, popes entered into strategic coalitions with secular rulers. These alliances shifted as the balance of power changed. The papacy was opportunistic: the Lombard League of city-states was traditionally an enemy, but Pope Alexander III allied with the League to prevent Frederick I from taking over Italy, and again to battle Frederick II under Pope Gregory IX. Urban II (r. 1088-99) and Pascal II (r. 1099-1118) found it expedient to favor the Normans: when Roger I conquered Sicily, Urban recognized him and named him the papal deputy on the island (Fried 2015, 195). Despite its overall hostility to the Hohenstaufens, the papacy could even ally with the German emperors: after the Concordat of Worms, hostilities ceased and the pope broke their costly alliance with the Normans and instead allied with the Empire (Brooke 1938, 264). Similarly, popes sought German help subsequently to contain Angevin ambitions in Sicily.
Popes offered protection and legitimation to allied rulers. Monarchs entered into a "feudal" relationship 19 by surrendering land to pope, and receiving it back as a fief. The benefit was that a vassal kingdom could not be legitimately offered to another ruler and any injury to king or country was an injury to the Church (Ullmann [1955(Ullmann [ ] 1965. In return, kings then had to pay annual tribute or perform military duties. At different points in time, Scandinavia, Hungary, Poland, Bohemia, and the British Isles were under papal lordship (Mundy 2000, 200). Popes also recognized the territorial grabs of their allies, but not the conquests of their foes (Hoffman 2015, 132).
Popes invested heavily in their defense, and subsidized their allies. Pope Gregory IX (r. 1227-41) excommunicated Frederick II on multiple occasions, but for good measure, both he and Innocent IV (r. 1243-51) raised armies to fight the emperor and to prevent the union of Sicily with the Empire that would have consolidated imperial power on the Italian peninsula. Innocent IV sold the rights to invade Sicily to the English King Henry III for the ungodly sum of 90,000 pounds. When Henry III failed to come up with the money, the papacy negated the sale and let Charles of Anjou take Sicily instead. The papacy then supported Charles in his fight against the Empire and the Hohenstaufen line ended in defeat in 1268. The result was continued violent hostilities in Italy and the eventual retreat of imperial ambition beyond the Alps, leaving behind a fractured Italy and an atomized Holy Roman Empire (Watts 2009, 65).
The Crusades were also an exercise in political power -and fragmenting of authority. Blaydes and Paik (2016) argue that these expeditions contributed to state formation by promoting the emergence of taxation, sales of feudal land to finance the expeditions, the reintegration of Europe into global trade networks, and the elimination of rivals to ruling monarchs (Blaydes and Paik 2016). Yet these joint ventures between popes and monarchs also had explicitly temporal and local aims. Popes summoned the Crusades to defeat their rivals, rather than to defend the faith. 20 The Baltic Crusades, designed to convert Northern Europe to Christianity and to gain the pope political influence, began in 1147 and lasted through the sixteenth century. The papacy subsequently blessed the Stedinger crusade against peasants who refused to pay the tithe (1233-34), and the political crusade against the Colonna, their enemies in Rome, in 1298. In 1241, shortly before his death, Gregory IX (r. 1227-41) commissioned a crusade against German Emperor Frederick II. Pope Innocent IV (r. 1243-54) also launched a crusade against Frederick II and then against Frederick's son, Conrad IV, excommunicating and deposing both. He summoned Germany, Lombardy, and Sicily, and offered indulgences to the crusaders as if they had been going on a crusade to the Holy Land, "creating an equivalency between the two theatres of holy war" (Whalen 2019, 186). Subsequently, Pope Alexander VI and then Urban IV called for a crusade against Manfred, the last Hohenstaufen king of Sicily and son of Frederick II, who invaded papal territories in 1258. This crusade was launched in 1255 and lasted until 1266, with enormous privileges granted to Charles of Anjou, an ally of the pope (Jedin 1993, 166). It became increasingly obvious that the aim of these military ventures, especially after the thirteenth century, became "less religious than hierarchical; it implied the domination of Church over State, and of clergy over laity, the demonstration of the civil power's derivation from ecclesiastical" (Smith 1964, 54).
Finally, popes used their authority to launch ideological salvos. First, they used law as a political weapon. Using new interpretations of Roman law, Pope Gregory VII fought with legal arguments, using the papal archives to buttress his arguments. When the law faculty at Bologna was founded at 1088, both popes and emperors then invested heavily in furthering legal expertise (Zacour 1976, 224). Second, popes articulated early concepts of state sovereignty, 21 or a ruler's right alone to control his territory and defend it from external demands, even if no ruler at the time could actually exercise this sovereignty. Drawing on precedents from Roman law, papal decrees and canonical reinterpretations underlined the principle that a ruler need not recognize any superior, including emperors. Pope Innocent III declared in his 1202 decretal Per Venerabile that "every king [is] an emperor in his kingdom." This doctrine was a move against imperial ambition, effectively replacing imperial hierarchy and deference with the equal standing of states. By the thirteenth century, drawing on Roman precedents and canon law, jurists recognized the sovereignty of the French, English, and Spanish kingdoms and the practical sovereignty of many city states (Canning 1983, 4;Rigaudiere 1995, 21). These concepts, thus, arose long before their sixteenth-century articulation by early modern theorists such as Jean Bodin or their supposed statement at Westphalia in 1648.
All of these weapons were deployed to defend a fundamental papal aim: hindering the rise of a hegemony that could once again subordinate the Church. The outcome of these balancing tactics was the fragmentation of territorial authority in Europe.

EMPIRICAL TESTS
To test the proposition that papal conflict undermined and fragmented territorial authority in Europe, I collected data on European state boundaries over time and the conflicts that took place over the years 1000-1800 within the boundaries of Western Christendom. 22 I also include existing data on cities (Bairoch 1988), sites of conflict (Dincecco and Onorato 2016), primogeniture (Kokkonen and Sundell 2014), and parliaments (Van Zanden, Buringh, and Bosker 2012). The data consist of over 105,000 grid cell-year 20 Popes also launched crusades against domestic religious dissent. Thus, Innocent III (1198-1216) launched Crusades against Muslim Spain and the Albigensian Crusade against the Cathars in 1209. 21 Concepts of medieval political authority, statehood, and sovereignty (autonomy from outside interference and control over internal affairs) are debated extensively (Costa Lopez 2020;Friedrichs 2001;Hall and Kratochwil 1993;Little and Buzan 2002;Ruggie 1983). 22 The Great Schism of 1054 split the Church into a Western and Eastern rite. The Church in the Eastern rite never gained its autonomy, did not function independently of the state, and instead continued to be controlled by the Byzantine emperor. The territories of modern Bulgaria, Greece, Romania, Russia, and Serbia all followed the Eastern rite, and are not part of the analysis.

observations (see the Supplementary Material for a full description of the dataset and sources).
The measure of papal conflict, the critical explanatory variable, consists of the number of clearly identifiable wars by proxy funded or directed by popes against papal enemies, papal depositions of secular rulers, political crusades, and attacks either led or financed by the popes directly, over a rolling 5-year period. Each of these is a distinct category: wars by proxy involve rulers and clergy acting on behalf of the papacy, with tacit papal approval, but no special privileges conferred by the pope. Political crusades target specific rulers, but like other crusades, still needed to be authorized publicly by the pope, involved preaching to arms, and conferred privileges on the combatants (see Riley-Smith 2005). Papal depositions involved the pope, usually acting in alliance with domestic barons or other secular rivals for power, to remove a ruling king of prince. Finally, there are a few cases of popes themselves leading battles. 23 Conflicts involving the Papal States, the territories directly governed by the papacy, are also coded as papal conflicts: for example, Pope Julius II (r. 1503-13) led armies to successfully defend the papal territories.
I count and add the number of such incidents involving the papacy per grid cell, without weighing the different conflict types: there is little theoretical reason to think that a war by proxy would be costlier or more effective than a political crusade, for example. An incident could also involve more than one type of conflict: as a result, if there is a war by proxy and a deposition effort against a ruler invading the Papal States, that incident would be given a value of 3. The main sources are the New Cambridge Medieval History, Brecke (2012), and Dupuy and Dupuy (1993).
The measures of secular conflict come from Dincecco and Onorato (2016), who code whether armed conflict occurred in a given site in the preceding 150 years among secular parties. In pre-modern Europe, religious and secular authority could overlap or blur (see the ecclesiastical princes in the Holy Roman Empire, or the consistories in early modern Protestant Europe). Therefore, by "secular conflict," I mean only the actors involved: monarchs and princes, acting in their secular capacity as rulers against other secular forces. 24 Figure 3 plots the incidence of papal conflict with temporal rulers and secular conflict among temporal rulers. The rise in papal conflict in the twelfth and thirteenth centuries coincides with the takeoff of fragmentation of territorial authority, as the papacy reached the acme of its ambitions. Papal conflict drops in the fourteenth century: during the Avignon papacy and then the Great Schism that lasted until 1417, the weakened papacy did not launch conflicts at anywhere near the same rates. The conflicts spike again in the sixteenth century, thanks to the proxy wars of the Reformation. Secular conflict takes off later,  (Jordan 2001, 87). When Innocent II took the field himself against Roger of Sicily in 1139, he, too, was taken prisoner and had to accede to Roger's demand for control of Sicily and Apulia (Brooke 1938, 276). History repeated in 1156 (Robinson 1990, 367). 24 Papal conflict and secular conflict, as measured here, do not overlap significantly: the pairwise correlation is 0.025, at p ¼ 0:05. consistent with the bellicist account of more intense early modern warfare.
The main dependent variable is fragmentation, or the number of political authorities within a given territory. I proxy territorial fragmentation with the number of state boundaries in a given area. I divide Western Christian Europe into 100 Â 100 km grid cells, and calculate the number of states within each grid cell. The number of state borders that cross the grid cell is the measure of fragmentation. Grid cells that fall entirely in one state have a value of 0 (there are no borders within the cell). 25 The results are robust to other specifications of fragmentation, such as the number of states within a given radius of a centroid of a state. 26 The literature identifies the presence of parliaments and economic development as two important forces in medieval state development. First, parliaments may help to consolidate authority: parliaments constrain monarchs and lead to power sharing, thus stabilizing rule (Blaydes and Chaney 2013;Ertman 1997;Stasavage 2010). I use the indicators of parliamentary presence from Van Zanden, Buringh, and Bosker (2012), which measure the number of years per century that parliaments met in a given state. Second, urbanization serves as a standard proxy used in the literature for economic development. Cities themselves may also contribute to fragmentation, since the rise of cities drove the survival of smaller units in Europe (Abramson 2017), and dense urban populations drove both economic growth and institutional development (Acharya and Lee 2019). I therefore use the number of cities with population over 5,000, using data from Bosker, Eltjo, and van Zanden (2013). I include whether or not a given cell belonged to the Holy Roman Empire, to ensure that the results are not simply driven by papal hostility to the emperor. Moreover, the shape of the Holy Roman Empire changed over time: Burgundy became part after 1032, the empire withdrew from Northern Italy after 1250, and the Empire itself was dissolved in 1806. I also include a coding for Protestant in the early modern period, in case the results are driven by the struggles of the Reformation. 27 The measure for Protestant consists of the number of cities coded as Protestant by Cantoni (2012) within the grid cell.
In all analyses, I use ordinary least squares regressions with two-way grid cell and year fixed effects, which approximate a difference-in-difference design with observational data and control for time-and space-invariant factors such as climate, elevation, agricultural suitability, and proximity to ports and coasts. Robust standard errors are clustered around grid cells. To capture the theorized difference between the medieval and early modern periods in papal strength, I split the sample into medieval (1000-1350) and early modern (1450-1750) periods to reflect the distinct periods of papal power. This periodization is substantiated by structural break tests (see Section 3 of the Supplementary Material). I report the results for both the entire time period, and the two specific eras.
If the Church deepened the fragmentation of territorial authority in Europe, we would expect conflict with the popes to be strongly and positively associated with the fragmentation of authority. As Figure 4 reports, that is indeed the case for the entire period examined, even when we include several potential other factors that might have led to fragmentation. The horizontal lines represent 95% confidence intervals for a given variable. Table 1 reports the results from Figure 4 (model 1), and from the medieval and early modern periods. Over the entire time period examined, from 1000 to 1800, papal conflict is consistently and positively associated with the fragmentation of territorial authority (model 1). In the medieval period, papal conflict is strongly, positively, and consistently associated with fragmentation. These results are robust to the inclusion of both parliaments and urbanization. Parliaments are negatively associated with fragmentation and cities are positively associated, in keeping with the existing literature. Parliaments indicate some measure of state institutional consolidation. Medieval cities, on the other hand, arose where central powers were weak. The coefficient on medieval papal conflict is even higher when these two variables are included, and retains its significance (compare models 2 and 3). 28 Status as part of the Holy Roman Empire is not independently associated with fragmentation in the Middle Ages. These results are robust both to other measures of fragmentation, and to placebo tests (see Tables A.2 In the early modern period, papal conflict initially appears to be positively and even more strongly associated with fragmentation (models 4 and 5). However, once we include status as Protestant for a given cell of territory, this relationship disappears (model 6). The enormous upheaval of Protestant Reformation, the papal efforts to fight it, and the wars that followed meant that there was a great deal of papal conflict, and nearly all of it was tied to the religious status of a territory. Contrary to bellicist predictions, secular conflict appears to be positively associated with territorial 25 The data come from MPIDR (Max Planck Institute for Demographic Research) and CGG (Chair for Geodesy and Geoinformatics, University of Rostock) 2013: MPIDR Population History GIS Collection, Mosaic Census Collection. 26 See Tables A.2 and A.3 in the Supplementary Material for two specifications using the number of state borders within the 100-km and 250-km radii of state centroids as a measure of fragmentation. 27 The spread of Protestantism itself in the early modern era may have been endogenous to medieval fragmentation: it took off in the fragmented territories of the Holy Roman Empire, the Low Countries, and East Central Europe. Various princes protected the nascent religious reforms within their territories, whereas others maintained a commitment to Catholicism. 28 Oster sensitivity tests report the delta, or the estimate of proportional bias due to unobservables. Oster (2019) proposes 1 as a conventional threshold. Negative signs suggest that the effect of the unobservables would have to run in the opposite reaction for the beta to be 0. fragmentation during this period. These patterns are not simply an artefact of the Holy Roman Empire, where the Reformation took off and where several massive wars took place. 29 That said, membership in the Holy Roman Empire, unsurprisingly, is also associated with increased fragmentation.
Contrary to the bellicists, then, temporal warfare does not seem to consolidate states. This may be because many states either consolidated before the early modern period (England, Spain, and Portugal) or long after (Germany, Italy, and Poland). Further, devastating early modern wars, such as the Thirty Years' War (1618-1848), were fought mostly on the already fragmented territory of the Holy Roman Empire, and did little to consolidate it. Instead, religious conflict instigated by the popes is closely  Note: *p < 0:05, **p < 0:01, ***p < 0:001. 29 The collinearity between Protestant and Papal Conflict disappears when I reran the same models on state-level data. All the other variables retained their previous relationship (see Tables A.2 and A.3 in the Supplementary Material). This is because the data on Protestantism are measured on the city level. The state-level dataset aggregates these into state-level averages, but the raster dataset does not. associated with the fragmentation of territorial authority in Europe.

Alternative Religious Explanations
Two other potential sources of fragmentation (and consolidation) are both related to the Church itself, and could serve as religious confounders. First, inheritance regimes may explain fragmentation. Specifically, primogeniture (the inheritance of all land and office by the oldest son) acts to prevent territorial fragmentation, whereas partible inheritance divides the land under a ruler's control among several heirs. Primogeniture and female inheritance can be traced back to earlier changes in the Church family and marriage laws designed to funnel wealth into the Church (Goody 1983; see also Møller 2019, 217). The Church introduced changes in marriage law, beginning in earnest in the tenth century, including monogamy, stricter definitions of legitimate children, constraints on marriage among distant relatives, and prohibitions on adoption and divorce (Acharya and Lee 2019;Goody 1983;Gorski and Sharma 2017, 204). Many of these allowed the Church to benefit financially, since they limited the number of potential heirs-and increased the probability the assets would revert to the Church. These changes in family law, in turn, prompted the rise of primogeniture (Kokkonen and Sundell 2014;Møller 2019).
Primogeniture emerged around 1000, and was adopted by the Capetians in twelfth-century France and within many families within the Holy Roman Empire by the thirteenth to fourteenth century (Goody 1983, 118;Wilson 2016, 425). Along with other changes in family law, primogeniture stabilized monarchical rule in Western Christendom and beyond (Acharya and Lee 2019;Brundage 2009;Goody 1983;Sharma 2015). Kokkonen and Sundell (2014) find that primogeniture extended and stabilized ruler tenure. Since primogeniture is also the basis for dynastic unions of territory, we might expect it to lower fragmentation (Sharma 2015, 169;Teschke 2003, 225). In contrast, other scholars emphasize that such unions, and therefore primogeniture, did not consolidate territorial gains (Bonner 2003;Fichtner 1976;Joseph 2015;Sharp 2001;Wilson 2016, 436). 30 Second, the rise of communes from the late eleventh century to the twelfth century would prevent rulers from establishing and centralizing territorial authority (Abramson 2017;Rokkan 1975;Tilly 1992;Tilly and Blockmans 1994;Wickham 2015;. In the "city belt" that stretched from central Italy to North Germany, self-governing towns arose where the central government was too weak to assert control (Møller and Doucette 2022;see also Stasavage 2010;2020, chap. 5). They arose where rulers were preoccupied with papal conflict: communes filled in the vacuum left behind by the withdrawal of imperial authority in Italy. They are thus an indirect result of papal conflict with the emperors (Hyde 1973;Wickham 2015). Further, as Doucette and Møller show, the Church itself diffused norms of local selfgovernance through bishops and monastic reform (Møller and Doucette 2022). Thus, cities expanded greatly in the thirteenth and fourteenth centuries, after the fragmentation of the eleventh to twelfth century. 31 I therefore include both communes (data from Van Zanden, Buringh, and Bosker 2012) and primogeniture (data from Kokkonen and Sundell 2014) in the analyses. Table 2 reports the results from Figure 5 (model 1), and from the medieval and early modern periods. Papal conflict has a very strong and positive association with fragmentation over the entire time period examined, from 1000 to 1800 (model 1). Secular conflict is not associated with territorial consolidation or fragmentation. The coefficients on parliaments are negative and on cities positive, as expected. Primogeniture and communes also have a positive relationship with fragmentation. As in Table 1 and throughout, I use two-way cell and year fixed effects and robust standard errors clustered around the cells.
In the medieval period, papal conflict is strongly associated with territorial fragmentation, even as parliaments, urbanization, primogeniture, and communes continue to play a role (model 2). Primogeniture does not lower fragmentation, consistent with the argument that medieval dynastic unions did not permanently unite territory, and did not lead to the expansion of authority (Joseph 2015). Communes also have a strong and positive relationship, consistent with their arising in territory where central authority was weak.
Subsequently, in the early modern period, we initially see papal conflict has a strong and positive relationship with fragmentation (model 3), which drops out once we include a cell's status as Protestant (model 4). Papal conflict, then, disappears as a factor in early modern fragmentation. Neither primogeniture nor communes are associated with early modern fragmentation: the single most important factor becomes membership in the Holy Roman Empire, which in the early modern era is powerfully associated with fragmentation.
In short, medieval conflict with the papacy is strongly associated with territorial fragmentation, independently of parliaments, urbanization, primogeniture, and communes. In the early modern period, papal conflict is so strongly associated with the wars of the Reformation that its independent association with fragmentation drops out. Secular conflict is not associated with the consolidation of territorial authority. 30 Despite prominent examples such as the union of Aragon and Castile through the marriage of Ferdinand and Isabella, or the uxorious expansion of early modern Austria, the main reasons for contracting such marriages were alliances rather than uniting territory. Territorial acquisition through dynastic unions only emerged after the fifteenth century, and could not be responsible for medieval territorial fragmentation or consolidation. Even in Habsburg Austria, the rapid accumulation of dynastic possessions in 1477-1536 did not expand the empire. Dynastic gains in the other most likely case, England, were only temporary. 31 For example, Germany went from 250 cities in 1200 to over a thousand a century later (Jedin 1993, 322ff).

Reverse Causation?
Conflict with the popes is thus closely associated with medieval fragmentation, consistent with the proposition that medieval popes actively fragmented territorial authority. An alternative interpretation may be that popes were more likely to enter into conflict with rulers in already fragmented areas, opportunistically targeting rulers who were weak, rather than going after powerful rivals.
Yet the historical evidence suggests that this reversed causation is implausible. First, there are the papacy's stated intentions. Medieval popes deliberately and consistently targeted what they considered to be their biggest threat: the Holy Roman Empire. Among their first targets in the twelfth century were the Hohenstaufen emperors (1079-1254), whom they denounced as a "brood of vipers" for their designs on Italian territories (Abulafia 1999, 506;Toch 1999). Second, there is sequencing: the Empire fragmented after papal efforts  Note: *p < 0:05, **p < 0:01, ***p < 0:001.
began in earnest in the late twelfth century, but not before (see Figure 1 and Doucette and Møller 2021). Popes did not target already fragmented territories. Third, there are the geographic patterns of papal conflict: in the peripheral areas, popes tended either to stay out of domestic politics (as in England), or offered assistance to their allies. Since smaller and weaker rulers did not pose a threat to church autonomy, they were not the main target of papal efforts. Indeed, the papacy served a very different role in the periphery of Europe. For example, Polish kings allied with the papacy and joined in the fight against the empire, and the papacy actively helped state builders to consolidate rule (Ertman 1997, 279-80;Wyrozumski 2004, 281). Instead of meddling, the papacy even allowed these rulers to name their bishops and exert secular control over the Church, in the name of preserving their alliances (Kotecki 2018, 308;Obertyński and Kumor 1974, 60, 360).

Empirical Implications: The Holy Roman Empire
Since the Holy Roman Empire posed the greatest threat to the papacy and the autonomy of the Church, it became a principal target of papal efforts. As a result, we should see greater fragmentation in the areas of the Holy Roman Empire as a result of papal conflict. To test this empirical implication, I interact papal conflict with a given grid cell belonging to the Holy Roman Empire. Figure 5 shows that consistent with the explanation advanced here, papal conflict has a strong and positive marginal association with fragmentation in the Holy Roman Empire, while it has a much lower (if still positive) association with fragmentation in non-imperial lands. This is consistent with the proposition that papal conflict fragmented authority everywhere: but where the papacy targeted specific rulers (here, the Emperors), the result was particularly intense fragmentation.

Empirical Implications: The Rise of Communes
This literature further suggests that papal conflict led to both the fragmentation of territorial authority and the rise of communes. As Max Weber and Otto Hintze have already noted, towns seized autonomy and control during the conflict between popes and emperors in the late eleventh century, and in the subsequent imperial power vacuum (Ertman 2017, 61;Ringer 2004, 206;Rokkan 1999;Weber 1958). Cities could carve out a sphere of independent activity, of relative autonomy and burgher rights. Local self-governance started to take off after 1000, and then grew further in the shadow of imperial-papal conflict: self-governing towns arose where the central government was too weak to assert control (Doucette and Møller 2021;Møller and Doucette 2022, 55ff;Wickham 2015, 9;2016, 148ff;Watts 2009, 99). To curry local support, both popes and emperors granted new charters with substantial privileges to lay officials and new political rights, which "only strengthened the sense of agency that the urban population felt," giving new agency, urgency, and responsibilities to communal self-government (Witt 2012, 206).
For their part, bishops often cooperated with the creation of communes, and medieval communal institutions then mirrored religious ones (Coleman 1999, 394-5;Schwartzberg 2014, 51; see especially Møller and Doucette 2022). But bishops could also have highly antagonistic relationship with urban self-government, one that could even erupt in violence and murder: the Roman commune arose in the 1140s to rebel against papal authority, also eventually resorting to violence, while the good burghers of Laon in France murdered their overweening bishop in 1112. Communes often served as "an instrument of liberation from the captivity of worldly and abusive bishops" (Malegam 2013, 231). Figure 6 shows that in the entire sample, papal conflict, secular conflict, and fragmentation all show a positive relationship to the rise of communes. Parliaments and the presence of cities are also positively and strongly associated, in keeping with the existing literature. Table 3 shows the results from Figure 6 (model 1), and from the medieval and early modern periods. Once we disaggregate the data into medieval and early modern eras, two distinct patterns emerge. In the medieval era, papal conflict is consistently and positively associated with communes. Fragmentation and the presence of bishops are also positively and independently associated with communes. Secular conflict is not. This is consistent with the argument that the power vacuum associated with papal conflict and territorial fragmentation allowed medieval communes to flourish. In the early modern period, in contrast, these variables no longer predict the rise of communes. This is because communes themselves changed: Italian communes became regional powers, expanding their authority over surrounding territory and ceding governance to prominent families and cartels. Elsewhere, communes fell under the renewed control of local nobles and princes. Borders stabilized and power vacuums disappeared, so that conflict and fragmentation were no longer a factor. Many local parliaments simply disappeared.
These empirical regularities are consistent with the core argument of this paper: that medieval papal conflict with emperors, monarchs, and princes fragmented territorial authority, allowing other forms of autonomous governance to arise and escape imperial control in the Middle Ages.

CONCLUSION: MISSA FINITA EST?
Bellicist accounts argue that the medieval fragmentation of Europe gave way to state consolidation through warfare. In these accounts, secular conflict winnowed out small states and institutions arose in response to the pressures of war. Early modern warfare ended the territorial fragmentation of Europe and incidentally established state institutions.
This article argues instead that the roots of the European state reach back further, to the medieval era, and to a different set of actors. Fragmentation was deliberate, persistent, and popes helped to sustain it. States did not consolidate uniformly: the Church contributed to the fragmentation of territorial authority of powers it saw as hostile (chiefly the Holy Roman Empire), and helped to consolidate the development of the central state in other areas. War was neither necessary nor sufficient to build states, in both the medieval and early modern eras: some highly fragmented states survived until the nineteenth century, whereas others consolidated without the pressures of war. Indeed, papal conflict increased fragmentation in the medieval period, and secular conflict in the early modern. By the same token, state institutions arose centuries before early modern warfare would have necessitated them; in those lands medieval rulers could consolidate power and adopt ecclesiastical innovations.
By shifting attention to the greatest geopolitical rival of the medieval era, the Catholic Church, we gain a new perspective on state formation in Europe. State formation began earlier, and this medieval state development differed from early modern in the key protagonists, motivations, and mechanisms. Medieval state formation was shaped by religious authority. Popes and Note: *p < 0:05, **p < 0:01, ***p < 0:001. bishops, rather than kings alone, were critical. That said, the motivation for the popes was not religious ideology or doctrine, but the institutional autonomy of the Church. The medieval papacy was consumed with preventing the resurrection of a rival superpower, the German Empire, and used everything from legal arguments, to proxy wars and crusades to achieve its aims. The mechanisms of this earlier episode of state formation thus consisted of both rivalry and emulation, of both temporal alliances and ideological weapons, rather than interstate conflict that somehow necessitated state institutions. This argument builds on both an older tradition that notes how the Church helped to diffuse the rule of law and administrative norms throughout Europe, and more recent work that emphasizes the importance of the Church to medieval state building and its legacies (Hintze [1931Bendix 1978;Doucette and Møller 2021;Ergang 1971;Fukuyama 2011;Grzymala-Busse 2020;Kiernan 1965;Møller and Doucette 2022;Poggi 1990). The contribution of this analysis is to compare the impact of the Church to other sources of fragmentation, specify the secular and temporal tactics used by the Church, and draw our attention to the direct mechanisms that helped to fragment territorial authority (papal conflict) and the indirect ones (empowering nobles and communes).
The irony is that with these successes, the Church ordained its own fall from grace. The very political fragmentation that it fomented meant that subsequently, when the Protestant Reformation took off, individual princes and lords could protect the new rival religion from a reassertion of Catholic monopoly. It is no accident that the Reformation took off in fragmented Germany, or that Frederick III, the Elector of Saxony, could successfully protect Luther against the vengeance of both the Pope and the German Emperor. In battling monarchs with both laws and arms, the Church led these rulers to sharpen their own legal arguments and buttress their own administrative and legal infrastructure. Secular nobles replaced bishops and clerics in the administration. Within kingdoms, rulers increasingly decided who would serve and who could govern the Church.
In winning battles, the Church lost the war. Yet this eventual supremacy of the state would not be possible without the medieval church, the clashing ambitions of medieval popes and rulers, and the early state formation they engendered.

SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/S0003055423000278.

DATA AVAILABILITY STATEMENT
Research documentation and data that support the findings of this study are openly available in the American Political Science Review Dataverse at https://doi.org/10.7910/DVN/DWQLIB.