The state made the system and the system made the state

Abstract

State systems vary in terms of the importance of declaratory versus constitutive forms of establishing statehood. We argue that interaction capacity explains this variance. Thinly developed systems with lower interaction capacity exhibit more declaratory properties of statehood, where states have less need to seek acceptance into an institutionalized “club” of states. Conversely, in systems with higher levels of interaction capacity, constitutive forms of recognition are more pronounced. Based on a combination of a qualitative analysis of historical state systems and a longitudinal network analysis of the contemporary international system, we demonstrate that increasing interaction capacity shifts the organizational principles of state systems from declaratory to constitutive practices. We theorize that this is how systems evolve. In thinner, low-capacity systems, states come first and make the system. But as the system thickens and constitutive practices are developed to manage crowding, it gradually determines who can be a state. Over the long run, the state made the system and the system made the state.

There is a prominent debate in international law and political theory about what makes a state (Agne et al., 2013; Crawford, 2006; Erman, 2013; Fabry, 2010; Grant, 1999). Do states exist prior to recognition and simply declare that they are sovereign? Or do states come into being by virtue of the collective judgment of the other states in the international system? These contrasting views are known as the declaratory and constitutive theories of statehood – what Grant refers to as the Great Debate (Grant, 1999). In a world where declaratory theory holds, states are empirical realities that are prior to the system. States come first; they make the system. But in a world where constitutive theory prevails, priority is given to the system of states upon whose collective judgment membership depends. In other words, the system makes the state. So, which comes first: the state or the system?

This is an important question because it specifies the basis on which states are created and how they construct and are constructed by the systems in which they interact. Although the Great Debate in international law is primarily a normative question about how states ought to be recognized, recent scholarship in International Relations (IR) has begun to examine this question on empirical grounds. The answer will shed light on how states are born, as well as the different relationships between states and systems across time and space. It will reveal variation in historical state systems. It will show how the international system has evolved and may continue to evolve.

We argue that the answer depends on the level of interaction capacity in a given state system. In low-capacity systems where the connectivity between states is sparse and where frontier regions and open land are common, the diplomatic relations between states lack constitutive processes to regulate membership. Here, where the diplomatic recognition of others means very little and may be irrelevant, statehood depends on the state in question; it is declaratory. In contrast, high-capacity systems take on constitutive properties as the increased interaction capacity creates club effects. Just as the benefits of recognition increase as the density of the system increases, system members will develop rules that control membership and limit crowding. We theorize that this is how systems evolve. In thinner, low-capacity systems, states come first and make the system. But as the system thickens, it develops constitutive practices for what counts as a state. Over the long run, the state made the system and the system made the state.

We advance and test this argument in the pages that follow. We begin with a coverage of the debate on declaratory versus constitutive theory. We argue that the relative priority of the state and the system depend on the level of interaction capacity. We next develop a yardstick for measuring the degree of diplomatic recognition practices ranging from highly declaratory to highly constitutive in nature. We then test our theory in two segments. First, we conduct a qualitative analysis of two 19th-century state systems in West Africa and East Asia. Second, we conduct a network analysis of the global system since 1945. We conclude with observations and directions for further research.

States and systems

What is the relationship between states and systems and is one prior to the other? Let us define two key terms. Following Butcher and Griffiths (2017: 330), we adopt a modified Tillian (1992: 1–2) definition of the state as “coercion-wielding organizations that . . . exercise clear priority in some respects . . . over all other organizations within” a territory and have control over their foreign relations. This definition takes Tilly’s broad conception of the state that includes national states, empires, and city-states, among others, and further requires that the polity in question has control over their diplomatic relations with other states. The depth of those diplomatic relations can vary from thin to thick – that is, from the occasional mission to dense networks of diplomatic rules, institutions, and bureaucracies. The key point is that the state has what we may call internal and external sovereignty (Krasner, 1999). ¹

Following Tilly (1992: 5) once more, we contend that “states form systems to the extent that they interact, and to the degree that their interaction significantly affects each party’s fate.” The basic idea here, and in the literature more broadly, is that state systems are composed of state units that interact and have effects on one another. ² This, of course, invites various questions regarding the intensity of the interaction, the scope, and the degree to which states are influenced by the system. As Buzan and Little (2000: 90–97) note, we can envisage a spectrum of system types ranging from those that are only minimally connected along sociocultural lines to those that include dense societies of states with thick institutional and economic linkages. Our approach is to utilize a general definition of the state system that permits an analysis of variation across systems.

The debate between declaratory and constitutive statehood provides a useful framework for thinking about the relationship between states and systems. Proponents of declaratory theory hold that “Recognition presupposes a state’s existence; it does not create it.” (James, 1986: 147) The recognition of other states merely confirms facts on the ground; states come first (Grant, 1999: 5). A common critique of declaratory theory is that it requires criteria for statehood. If recognition follows from the act of states declaring themselves “as a fact something which has hitherto been uncertain” (Dugard, 1987: 7), then there must be some criteria to sort proper states from other political forms. Perhaps the most noteworthy attempt to answer that question is the 1933 Montevideo Convention on the Rights and Duties of States. According to Article 1 of the Convention, “The State as a person of international law should possess the following qualifications: (a) a permanent population; (b) a defined territory; (c) government; and (d) capacity to enter into relations with other states.” (Crawford, 2006: 45–46) Furthermore, according to Article 3, “The political existence of the state is independent of recognition by the other states.” Overall, these criteria stress the functionality of states and assert that statehood should not depend on the recognition of others.

Proponents of constitutive theory assert that statehood only obtains through the recognition of other existing states. As Oppenheim (1955: 125) put it, “A State is, and becomes, an International Person through recognition only and exclusively.” Criteria such as territorial control, a viable population, and a stable government are of course elements of statehood. But the recognition of other states perfects and finalizes that statehood. According to Lauterpacht (1948: 38), the constitutive vision “deduces the legal existence of new States from the will of those already established.” By allocating that final decision to the system of states, the question of what counts for statehood is answered collectively. One common critique of constitutive theory is that by making recognition a “tool of statecraft,” it gives this process the tenor of realpolitik (Grant, 1999: 2–3).

If the international legal scholarship has approached this issue from a mostly normative perspective, the developing work in IR has examined it in empirical terms. Fabry (2010) argues that whereas the 19th-century European/Atlanticist state system was more declaratory in practice, it gradually developed into a more constitutive system during the 20th century. In her analysis of secession and international recognition, Coggins (2014) argues that regardless of how the system ought to be ordered, contemporary secessionist movements require the recognition of the international community. According to Ker-Lindsay (2012: 130), “UN membership has come to be regarded as nothing less than the ‘gold standard’ of international legitimacy.” Finally, Griffiths (2021) writes that the defining feature of sovereign statehood in the post-1945 period is obtaining a full seat in the United Nations General Assembly (UNGA). Doing so provides the state with an international legal identity that is useful for a range of economic and diplomatic purposes. But entrance into the UN is managed by existing states who behave like a “Sovereignty Club” that screens, processes, and controls admission. As the cases of Nagorno Karabakh, Northern Cyprus, and Somaliland show, failure to gain admission to the UN condemns states to a de facto unrecognized status on the margins of international society. ³ If UN membership is a necessary condition for sovereign recognition, and it is something that is voted on and determined by the collective judgment of other states, then it is a constitutive process.

In sum, a contribution of IR scholars to the states vs systems question is that the answer is variable. Sometimes declaratory practices drive outcomes, and sometimes the system has stronger constitutive properties. In the international law scholarship, this has primarily been a normative question, a question of oughtness, as opposed to the positive question focused on that which is. In this article, our aim is to advance and test a positively grounded theory about the relationship between states and systems and how it varies. What factors cause transitions between the two forms? To our knowledge, there is no theory explaining this relationship across historical systems.

A theory on interaction capacity and statehood

We argue that state systems with high levels of interaction capacity should correspond with constitutive patterns of membership and mutual recognition. In other words, the thicker the system, the more likely it is to develop complex and constitutive patterns of membership recognition. Conversely, state systems with low levels of interaction capacity should exhibit patterns of declaratory statehood.

Interaction capacity is the key independent variable. Following Buzan and Little (2000: 80), we define interaction capacity as “the amount of transportation, communication, and organizational capability” within a system. It measures the carrying capacity of a given system, the volume of exchanges and interactions between populations that a system can support. It is a function of many technological, social, demographic, economic, and environmental factors. For example, it might take an hour for a modern worker to travel 25 miles to their office during rush hour. It would take a truck loaded with several tons of goods about the same time. But throughout much of history, the same journey could take a day by foot, assuming good weather and roads, and although a camel or ox could marginally increase the carrying capacity, they could not increase the speed of travel.

A system where it takes less time to transport goods, people, and ideas has higher interaction capacity. Other factors will influence interaction capacity such as the number and density of people as they raise or lower costs of trade. More people mean more potential consumers, which lowers the costs of transporting goods from one place to another and increases the potential for interactions between those locations. Language barriers, trade routes, mountains, and rivers may all affect interaction capacity as well. It is thus the sum of exogenous factors – navigable waterways, river systems, traversable plains – and endogenous factors – roads, cities, ports. Systems with high interaction capacity are those where interactions are dense and frequent among a large number of actors. Systems with low interaction capacity are those where interactions are sparse, local, slow, and relatively infrequent.

Why should the level of interaction capacity shape the balance between the declaratory and constitutive properties of a system? Put differently, why should we not expect to find dense systems with declaratory properties or thin systems with constitutive properties? The answer is that, as Buzan and Little (2000: 85) argue, interaction capacity mediates the structural effects that the system has on the units: “when interaction is high (e.g. regular trade among the units) structural effects should be strong; when it is low (e.g. sporadic and small-scale trade) structural effects should be weak.” We argue that as interaction capacity increases, there is a point at which the system members begin to act as a club that determines who can be a member.

Consider the effects of interaction capacity with two stylized examples. Imagine a system with low interaction capacity that consists of two interacting states – if there was only one state or no interaction, then it would cease to be a system. The interaction between the two states is minimal. They are aware of one another, there exists a small amount of trade between the states by third parties, and there is a marginal exchange of ideas. There is no recognized linear border between the two states, but rather a vague frontier zone and wilderness into which the reach of the two states gradually ebbs. In a minimal system such as this, it would be absurd to argue that state A does not function as a state unless state B recognizes it if the vitality of A and the status of their ruler is completely independent of B. Here, statehood is truly declaratory. The question of what counts is decided by only the state itself.

Now imagine an extremely dense system with high levels of interaction capacity in which the states are integrated into complex networks of economic and cultural exchange, and where supranational organizations of states arise to address functional issues of the system. In this system, all land is claimed, apportioned to a specific state, and delineated by agreed-upon borders. The interconnectivity of the system is so high that system-level goods are generated in the realms of economics, security, and resource management. The member states and related supranational organizations administer and regulate these goods. Moreover, the density of the system creates crowding issues in the sense that goods are not unlimited, and any new state constitutes a subtraction in territory and population from at least one existing state. Here, the system takes on the features of a club where goods are excludable and can remain non-rivalrous as long as crowding is prevented. The members would now have an incentive to control who can became a state and effectively join the club. In a system such as this, statehood is truly constitutive. The question of what counts is decided by the members of the system.

Interaction capacity is useful for measuring and comparing between systems and within systems over time. That is, two or more systems can be compared by taking a static measure of their respective levels of interaction capacity. Or a single system can be examined by reviewing how the level of interaction capacity has changed temporally. In Gilpin’s (1981: 40) terminology, changes in interaction capacity lead to, as the name suggests, interaction change: “change . . . in the form of regular interaction or processes among the entities in an ongoing international system.”

Importantly, increased interaction capacity enables international change. ⁴ Increasing interaction capacity changes the system of states by increasing the structural effects of the system on the units (states). Put differently, the causal role of the system gains in relation to the units. The higher interconnectivity between states produces material benefits and system-level goods that the states have an incentive to manage. Increasing density gives rise to club effects in which membership is controlled to prevent crowding and over-consumption. ⁵ Club goods are excludable and non-rivalrous, but only to a point (Buchanan, 1965; Mueller, 1989: 131). With sufficient crowding, the goods become rivalrous, which is why the club members develop criteria for admittance. Consider the contemporary cases of Northern Cyprus and Somaliland, for whom access to the sovereignty club would grant an international legal personality, the sovereign right of territorial integrity, and access to formal financial aid. Since any additional state would imply a territorial reduction of at least one existing member – in this case, Cyprus and Somalia – as well as a potential drag on international resources, the sovereignty club has an incentive (as well as the ability) to limit membership (Griffiths, 2021).

The single most important factor separating a declaratory system from a constitutive system is the ability of the system members to exclude (i.e. deny entry). This is what makes it into a club. As Sandler and Tschirhart (1997: 338) argue, the distinguishing feature of a club good “is the presence of an exclusion mechanism that monitors utilization so that members can be charged tolls and non-members kept out.” Increasing interaction capacity increases the possibility of exclusion. It also creates incentives for limiting membership, though whether and how the system members respond can vary. ⁶

In sum, the relationship between states and systems is shaped by interaction capacity. The two examples above are static descriptions of that relationship at the extreme ends of the spectrum. However, we are also interested in movement along the spectrum. In low-capacity systems that likely characterized much of history, the priority of states is more visible. States make systems. But as systems thicken in connectivity, crowding becomes an issue, and the members will gradually behave like a club that controls membership. Declaratory statehood gives way to constitutive statehood, and systems makes states. Taking the long view – that is, in the original sense – states have to be prior to the systems they create, even if those systems may gradually come to take on causal attributes that supersede their creators. To paraphrase Tilly’s (1975: 42) observation on the endogenous relationship between war and state-making, the state made the system and the system made the state.

Our argument is partly related to a distinction that anthropologists make between primary and secondary states (Sandeford, 2018; Spencer, 2010). Primary (or pristine) states are thought to have arisen without contact with other states. They are first-generation states, isolates. Candidates for this small set of ancient states include Uruk, Harappa, Monte Alban, and early Egypt. Secondary states are the far larger set of second-generation states that arose in the context of, and in contact with, other pre-existing states. Scholars working in this area have used this distinction to map out divergent processes of state formation. In relation to our argument, primary states are clearly declaratory since there is no other state upon whose judgment their recognition depends. Initially, there is no system until other secondary states arise to create one. To be clear, the existence of secondary states does not imply constitutive practices because they can arise in conditions of low interaction capacity. They may be aware of, and influenced by, other neighboring states, but their statehood does not depend on them unless interaction capacity is high enough to create constitutive practices. However, any state that achieves statehood through constitutive practices is by definition a secondary state. Overall, there is a parallel relationship between the transition from primary to secondary states and declaratory to constitutive practices of statehood, but it is not perfectly synchronized.

Crucially, we are not arguing that the relationship between states and systems can be reduced to a simple “chicken or egg” question (Emirbayer, 1997; Jackson and Nexon, 1999). States shape systems and systems shape states in myriad ways. We argue that the weight of the system – that is, the degree to which it shapes states – increases as interaction capacity increases. Indeed, there is a point at which the system begins to determine, in a constitutive manner, what counts as a state.

Practices of recognition

To test our theory, we bring in practices of diplomatic recognition (Der Derian, 1987; Leira, 2021; Sending et al., 2015; Sharp, 2009; Wight, 1977). This is our dependent variable, and it is conceptually and analytically distinct from interaction capacity. Wight (1978: 113) maintained that diplomacy is a central feature of the international system. Although diplomatic practices are authored by states, they can have feedback effects on the states that created them (Jonsson and Hall, 2005: 13). They tell us what kind of system the states inhabit. We contend that some diplomatic practices are declaratory in form while others are constitutive. If those practices can be identified and delineated, then they can be compared with the level of interaction capacity in a given system.

In order to compare recognition practices across and within diverse states systems, we need a metric (or yardstick) of comparison. Following Viola (2020), we identify levels of diplomatic intensity and provide illustrations from classical Greece and the early modern period in Europe. Although the examples draw on the European experience, and thus exposes our theory to accusations of Eurocentrism, we argue (and show in the analysis) that other pre-colonial state systems used the same practices and forms of diplomacy, just with different names. We posit that these practices are common, even timeless solutions for diplomacy between states.

The simplest level of diplomacy is one that is composed, at most, of diplomatic agents who are directed by state leaders (the principal) to travel to and communicate with a different state leader. The purpose might be for trade, for information exchange, or perhaps for treaty making. Such agents might be more formally referred to as envoys, messengers, emissaries, etc. But importantly, their role is temporary and ends when the communication is complete. This level of diplomacy was common in the European Middle Ages, and much earlier in classical Greece. In fact, the word diplomacy is derived from the Greek word “diploma,” a folded and sealed document that envoys carried between different poleis (Knutsen, 2021: 391).

If a system achieved only this first level of diplomatic recognition, then it would lack any constitutive process of membership. Like the stylized example of the minimal system given earlier, it would be absurd to say that state A does not function as a state unless state B recognized it when the vitality of A is completely independent of B. We suspect that many historical state systems never rose higher than this level.

A second level arises with the role of the permanent resident ambassador. Here, the difference is that the agent/ambassador resides in the foreign state on an ongoing basis. Viola (2020: 99–102) maintains that the European concept of the permanent ambassador coalesced in the Italian city-states of the 15th century. The transition from temporary to permanent agents was natural as, for example, a Florentine merchant who resided for long periods in Milan might be continuously called on to represent their home state. The permanent position was beneficial and efficient because it provided a consistent line of communication. Importantly, these roles did not have to rise to the level of a professionalized core of ambassadorship. For example, the ancient Greeks had “proxeni,” men who lived in foreign cities in some other capacity (e.g. trade) but were called on to represent their polis in specific matters (Knutsen, 2021: 392).

The distinction between the first and second levels is meant to detect, in a broad sense, the degree (level) of diplomatic interaction between states. At the first level, state leaders have to send agents to the other state, a mission that might take considerable time. At the second level, leaders can call on agents that live and reside in the foreign state. This implies a higher level of connectivity and population exchange. Both levels capture the manner in which states may recognize others diplomatically. However, neither points to constitutive patterns of recognition. That is, a state could send and/or maintain resident ambassadors in another state and still not depend on the recognition of that state in any meaningful way. This does not imply a substantialist position in which the states are completely unaffected by their interactions (Emirbayer, 1997; Jackson and Nexon, 1999). It is just that the intensity and weight of those interactions do not rise to the level that they determine membership in the system.

A third level involves a series of diplomatic practices that are likely (but not necessarily) to occur together. These are the droit d’ambassade, accreditation, and the establishment of an interstate diplomatic corps. Droit d’ambassade translates roughly to “right of embassy.” It is the claim made by a ruler that they alone are permitted to receive an ambassador. In doing so, they exclude others within a given territory from doing the same, and thereby claim that they are the ruler and decision-maker. Essentially, it is using the recognition of an external state to legitimize authority domestically. For example, on ascending the throne in 1461, Louis XI claimed this right over rivals like the Duke of Burgundy (Viola, 2020: 108). The droit d’ambassade underscores the rising importance of the system in relation to the state units.

Accreditation is a related concept whereby ambassadors to a foreign court present themselves as the official representative of their home state and are acknowledged as such. They are accredited, or credentialed to put it differently. That credentialization amounts to recognition by the foreign state of an independent and sovereign status. There are many fascinating examples of this dynamic throughout diverse historical regions. For example, Wight (1977: 23) writes that during the Abbasid Caliphate, monarchs as far away as Delhi would request that the Caliph provide a diploma of investiture. Like the droit d’ambassade, the importance of accreditation highlights the increasing role that the system plays in determining membership.

Finally, the institutionalization of a diplomatic corps is the formalization of these processes, determining how they are conducted and who can conduct them. By attempting to rationalize interstate relations with a common vocabulary and set of practices, a diplomatic corps simultaneously sets boundaries for membership. Although the practices of the diplomatic corps are a common feature in contemporary international life, they did exist in earlier periods (e.g. the Concert of Europe).

Viola (2020: 106–107) sees the institutionalization of the droit d’ambassade, accreditation, and diplomatic corps as a threshold moment in the development of a state system. It heralds the closure of the system insofar as the member states begin to behave like a club that controls recognition and limits membership. Internationally, the legitimacy of each state depends on the accreditation (recognition) of others. Domestically, the right of embassy clarifies who rules and enables a ruler to marginalize the opposition. The diplomatic corps binds the system together through a common set of norms and social practices. Thus, the recognized states regulate club membership, giving the system a greater constitutive character than the earlier levels. The presence of these practices can vary in intensity. For example, the causal weight of these practices in 16th-century Europe or even during the Concert of Europe was less than with the contemporary UN system. Nevertheless, they capture key elements of what both Viola (2020) and Griffiths (2021) refer to as the club-like behavior of the community of states. Where functional statehood depends on processes of mutual recognition, the constituent members can decide who is eligible to join the club. The importance of that decision may vary from somewhat performative to essential. At the highest level, it can determine what states have access to the club benefits of statehood (Fazal and Griffiths, 2014).

The final level focuses on the development of membership organizations. These are international organizations that are restricted to recognized states. Their purpose is to manage interstate relations along functional lines such as trade and security. Although the contemporary international system is replete with examples like the United Nations (UN), the World Trade Organization, and the International Monetary Fund, there are earlier historical examples like such as the Zollverein and the Universal Postal Union. Importantly, these organizations represent a higher level and deeper institutionalization of diplomacy because they imply the third level. That is, they are restricted to recognized states that are accredited, maintain the right of diplomacy, and possess a diplomatic corps that can represent them at the organizational level. These organizations further the institutionalization of constitutive practices by creating an added layer of diplomatic interaction. They highlight the club-like effects of diplomatic practices; their purpose is to increase the efficiency of interstate relations, but they are restricted to the recognized members.

Let us make two additional points about this final level of membership organizations. First, we expect that membership in these organizations can gradually supplant the practice of bilateral interstate recognition. Imagine a system with 1,000 states. It would be a cumbersome and costly process for each state to seek the accreditation, claim the right of embassy, and send ambassadors to each of the other 999 states. A more efficient process would be to centralize these practices in key organizations that identify member states and provide the forums for their interaction. This is effectively what the UN and other prominent international organizations do. We hypothesize that increased interaction capacity should gradually shift the emphasis for individual states from bilateral recognition practices to participation in membership organizations, an action that implies recognition.

Second, it is important to remember that these organizations are not states; rather, they are comprised of states over whom they lack hierarchy. If the fictional system of 1,000 states became so thickly institutionalized that the membership organization combined to take on the qualities of a superstate, then we would no longer be talking about interstate diplomacy, we would be talking about the relations of administrative units within a state. In that sense, this level of recognition practices is perhaps the highest achievable level before system unification.

Figure 1 provides an illustration of our yardstick of recognition practices. It depicts the intensification in diplomatic practices as one moves from left to right through the levels of the diplomatic agent, the permanent ambassador, the three-part stage consisting of the droit d’ambassade, accreditation, and institutionalization of a diplomatic corps, and, finally, the development of membership organizations. The vertical double line labeled “Club Threshold” signifies the point at which club-like effects begin to produce constitutive properties, though this is a continuous and non-binary process. The lower dashed line represents the relative weight of declaratory and constitutive theories of statehood at each level. Note that the permanent Ambassador does not have to precede the subsequent levels. Although the second, third, and fourth levels imply the use of diplomatic agents, it is possible that a state could practice accreditation, possess a diplomatic corps, and/or claim the right to embassy without sending permanent ambassadors to other states. Such an outcome may be unlikely, but it is possible, and this is the reason for the two forked arrows beginning with diplomatic agents. The yardstick provides a tool for locating different systems in terms of their practices of recognition.

Figure 1.

Yardstick of recognition practices.

We argue that systems with higher levels of interaction capacity should be located further to the right, all else equal. In the pair of stylized cases discussed earlier we imagined a system with very low interaction capacity and one with very high interaction capacity. The low-capacity system never generated recognition practices beyond the first level and remained declaratory. Meanwhile, the high-capacity system developed club-like characteristics and took on constitutive practices. That pair could be located on the far left and far right of the continuum in Figure 1.

Importantly, Figure 1 should not be understood as a unidirectional process in which all systems gradually shift rightward along the spectrum. Systems can stall in the middle of the spectrum, unable to thicken and develop more constitutive properties. They can evolve beyond a system of constitutively determined states with membership organizations and unify into a single state that now controls the external relations of each of the former member states. But the possibility of leftward movement clearly exists as systems thin out and disintegrate, or as large imperial states dissolve into a system of states. A historical review of East Asia and Europe, among other regions, shows patterns of system centralization and decentralization (Buzan, 2023).

In sum, our theory yields testable implications. Our primary hypothesis is that interaction capacity should correlate positively with constitutive recognition practices. The higher the level of interaction capacity, the more likely we are to see diplomatic practices such as droit d’ambassade, accreditation, the institutionalization of a diplomatic corps, and the development of membership organizations. Moreover, the development of membership organizations can gradually supplant the need for bilateral recognition between states. Overall, we theorize that this is how systems evolve. They thicken as interaction capacity increases. In thinner, low-capacity systems, states come first and make the system. As the system thickens and its constitutive properties develop, it gradually makes states. Over the long run, the state made the system and the system made the state.

Analysis

Our analysis is divided into two parts. We first test our theory across two early-19th-century state systems. In each case, we determine the level of interaction capacity and compare it with the prevailing diplomatic practices in that system. In this paired-case, qualitative approach we are making static comparisons. We are also making the first such comparison, to our knowledge, of diplomatic practices across different historical state systems. It is in the second part of the analysis that we zoom in to examine one system over time, the post-1945 UN system, using a network analytic approach. Overall, this is a multimethod analysis of cases across space and over time.

Case studies of historical systems

This section offers a comparative analysis of two non-western state systems in the first half of the 19th century: West Africa and East Asia. For each system, we draw on the identified set of states according to the International System(s) Dataset (ISD) (Butcher and Griffiths, 2020), which utilizes a general definition of the state as “coercion-wielding organizations that . . . exercise clear priority in some respects . . . over all other organizations within” a territory and have control over their foreign relations (Butcher and Griffiths, 2017: 330; Tilly, 1992: 1–2). The analysis covers the period between 1816 and 1850. The year 1816 is the first year for which state data are available from the ISD. Although somewhat arbitrary, the year 1850 was chosen because we wanted to examine these systems on their own terms prior to their gradual integration with the expanding European-based system during the latter 1800s. Each case constituted a vibrant system of states that, with few exceptions, were extinguished with the coming colonial enclosure. This is the last period in human history during which different but concurrent systems existed; by 1900 most of the globe had been consolidated into one system.

Some readers may wonder whether these regions can be studied as discrete state systems given that they were not completely disconnected from the rest of the world during this time. After all, West African states were distantly connected to North African states across the Sahara, just as East Asia was connected to the states of Southeast Asia via maritime trade routes. If we allow for multiple intermediaries, both systems were ultimately connected to the states of South America. However, following Butcher and Griffiths, we contend the two regions represent dense clusters of states. ⁷ Surrounding areas were often sparsely populated on account of geographic features such as the Sahara, the dense rain forests of the tropics, or the mountains of Central Asia. There are geographic realities within each system that helps connect them such as the Sahel and the seaboard of East Asia. Moreover, there is precedence in the literature to think of these regions as zones of interaction. Although the European powers would gradually penetrate these regions later in the 19th century, their influence was weaker during the period under observation.

Our analysis in each case was structured by a comparison between the level of interaction capacity (our independent variable) and the level of diplomatic practices (our dependent variable) in the system (George and Bennett, 2005). Given that interaction capacity is a broad concept with multiple dimensions, we focused on two specific indicators: transportation capability and population density. Both factors increase the carrying capacity of a system. For the dependent variable, we looked for evidence of diplomatic practices. Did the relations between the states rise above the occasional emissary or permanent ambassador (levels 1 and 2)? Was their evidence of constitutive practices (levels 3 and 4)?

We begin with the West African state system. Bounded by the Sahara to the north and the dense rainforests of Central Africa to the south, the system encompassed 8.5 million square kilometers and 40 million people (Oliver and Atmore, 2001: 12). It ranged from states like Cayor on the far western coast of Africa, in what is now Senegal, to inland states like Wadai and Bagirmi, in the territory of modern-day Chad. The southernmost states on the Niger delta included Benin and Owo. The northernmost states like Massina stretched across the Sahel to the edge of the Sahara. In their dataset on independent states, Butcher and Griffiths (2020) identified 49 West African states in 1816.

Interaction capacity was low in West Africa. Our first indicator for assessing interaction capacity is transportation capability. Transport in the region was slow and expensive and communications by land were based “entirely on human and animal power.” (Hopkins, 2014: 72) Most states in the system were landlocked and lacked the ability to transport goods and people by ship. The Niger river did provide an important transportation artery, but it was prone to flooding and difficult to navigate (Oliver and Atmore, 2001: 26). Camels were used to transport bulky goods in the desert and the dry savannah to the north (Wilfahrt, 2021: 54). But further south, carriage was mostly on foot or using donkeys. For example, in the Oyo Empire, “the costs of transport were high and precluded the possibility of substantial trade over long distances of basic foodstuffs and other commodities.” (Law, 1977: 209).

In many regions, the use of animal power was further hampered by the prevalence of the tsetse fly, which carried a parasite that is harmful to horses and other domesticated animals (Alsan, 2015: 391). Goods south of the so-called Tsetse Belt had to be carried by human porters, and communications took place on foot. In practice, this meant that livestock could not be used for transport and that larger loads and information flows were limited by human foot-speeds (Hopkins, 2014: 72). As Alsan (2015: 387–388) states, “messages, carrying goods or military transport over land would have been hampered by the lack of large, domesticated animals.” In sum, transportation capability was low.

Population density, the second indicator of interaction capacity, was similarly low in West Africa. Herbst argues that the generally low population density across the region hindered state-making because it made it harder for rulers to project power over a thinly populated terrain. This meant that greater distances had to be traveled between locations where economic exchange could take place, raising the relative costs of transport. Indeed, state leaders would uproot and move to locations where denser populations had coalesced (Lovejoy, 2019: 149; MacDonald and Camara, 2012: 177). The result was a “vast pointillist landscape” of scattered independent states, where the authority of the ruler tended to dissipate as a function of the distance from the center (Herbst, 2000: 44).

There was, of course, variation across the region. The Hausa states in what is now northern Nigeria as well as the Yoruba and Igbo areas to the south did possess higher population densities. At the beginning of the 19th century, Kano, the so-called “Manchester of West Africa,” had a population of 30,000 (Hopkins, 2014: 19, 48). During the century, the population of Ibadan may have reached as high as 70,000. But these, along with a handful of similarly sized towns and cities, represented the densest areas. But these, as we will see, were far smaller than the cities of East Asia. By comparison, China has more than 10 times the population density by the end of the 19th century (Herbst, 2000: 16). In summary, interaction capacity in West Africa was relatively low.

We have argued that low-capacity systems are unlikely to yield practices of constitutive statehood. That is, diplomatic practices, to the extent that they exist, are unlikely to rise higher than the diplomatic agent, much less the permanent ambassador. West Africa is consistent with our expectations. There is historical evidence of diplomatic relations. For example, Sokoto and Asante sent trading delegations back and forth (Hopkins, 2014: 63). The Alafin of the Oyo Empire sent emissaries to negotiate with other states (Law, 1977: 96–112). Smaller states like Ilorin would send heralds to larger states for purpose of alliance making. But importantly, we found no evidence of permanent ambassadors, or the higher and constitutive practices of right of embassy, accreditation, and the widespread use of diplomatic corps. Interstate relations never rose above the level of the diplomatic agent or occasional envoy sent from one ruler to another. To a large extent, statehood was declaratory. That is, a ruler could establish a state in an empirical sense and the reactions and recognition of others may have mattered very little. States made the system.

The second case study centers on East Asia. It is an outlier because, unlike the West African system, the four constituent states are recognizable to a modern audience. They include (Butcher and Griffiths, 2020): (1) China (the Qing Dynasty); (2) Korea, sometimes referred to as Choson (Hwang, 2010: 5; Seth, 2016: 7); (3) Japan; (4) Vietnam, commonly called Annam during this period. All four were very large and populous states with centralized bureaucracies.

The East Asian system was one of comparatively high interaction capacity. First, transportation capability was robust. The states benefited from a communications network of roads, rivers, canals, as well as the natural seaway of East Asia (Kang, 2010: 63–65). China possessed excellent transport capability not just along its river basins, but also across them longitudinally via the imperial infrastructure and littoral trade. Similarly, Japan’s road system and shipping transport yielded “dense information networks.” (Lieberman, 2009: 462–469) Korea and Vietnam had multiple urban centers connected to the sea, and both states established linear borders with China that assisted with trade regulation. More generally, the Confucian system of laws and norms provided a kind of administrative/diplomatic lingua franca that further facilitated communication (Woodside, 1998). Transportation did slow down and was reduced to pack animals in the rural and interior portions of the states. Nevertheless, transportation capability was high for pre-industrial societies.

The second indicator of population density was very high. These were states with large and dense populations. China had a population of approximately 330 million in 1800, and would rise to 430 million by 1850. This is estimated to be more than a third of the global population. ⁸ As a point of contrast, consider that in 1800 the British Isles had a population of 16 million, France a population of 28 million, and Russia a population of 37 million (Kennedy, 1987: 99). Meanwhile, Japan had a population of 30 million in 1800 and a population of 33 million in 1850. ⁹ Korea’s population during the early 1800s was about 10 million. Finally, Vietnam’s population is estimated to be 7 million. By the 1800s, each of the states possessed dense clusters of cities with populations into the hundreds of thousands (Reba et al., 2016). Both Peking and Tokyo possessed had populations of roughly 1 million. Overall, these were states with high populations, typically quite dense.

Given the high level of interaction capacity, we might expect to find constitutive diplomatic practices. A fascinating aspect of the region is the constitutive practices of the so-called tribute system. These were key patterns to diplomatic life that involved the sending of envoys and the practice of investiture. Each peripheral state would send envoys to the Qing court at regular intervals (Ringmar, 2012). This reinforced the relationship and facilitated trade and cultural exchange (Kang, 2010: 56; Park, 2017: 54). It also assisted with investiture, which, as Kang (2010: 56) wrote, “involved the explicit acceptance of subordinate tributary status and was a diplomatic protocol by which a state recognized the legitimate sovereignty of another political unit and the status of the king in that tributary state as the legitimate ruler.” Lee (2017: 57–61) has argued that this process should be viewed as a two-level game. On one level, investiture symbolized recognition by the Qing court. But on a more local level, the right of embassy gave rulers legitimacy in the eyes of their domestic audience and enabled them to consolidate their power.

How do these practices map on to the yardstick in Figure 1? There is clear evidence of first-level diplomatic agents. This is a core aspect of the tribute system. However, we did not find evidence for second-level permanent ambassadors. That is, there was no record of Koreans living in the Qing capital who were periodically authorized to speak on behalf of the Korean state. Conversely, although there were plenty of Chinese traders living in foreign capitals, it is not clear that they were authorized to periodically speak on behalf of the emperor. But strikingly, the system did display third-level practices that resembled the droit d’ambassade and accreditation, practices that reveal club-like patterns. After all, the process of investiture is a form of credentialization. Moreover, as Lee argues, it performed a dual function in which state leaders could claim legitimacy at home by arguing that they alone were invested by the Qing court. Although these practices were more hierarchical arranged than in early modern Europe – that is, only the Qing court could invest – it amounts to a similar process of recognition. In all, the East Asian system displayed constitutive practices that corresponded with the higher level of interaction capacity.

This comparative analysis yields conclusions that are consistent with our theory. In the low-capacity system of West Africa, diplomatic practices did not exceed the first level. The system was declaratory; states made the system. The East Asian system had higher interaction capacity, and it displayed clear club-like patterns and constitutive practices of recognition. That is not to say that the system was as constitutive as the contemporary UN system, and that recognition for Japan in 1826 meant as much as it did for Japan in 2026. But investiture by the Qing Court played an important role.

Network analysis of diplomacy in the UN system

We now turn to a longitudinal network analysis of the international system since 1945 (the UN system). In relation to the 19th-century systems above, this would be a high-capacity system, one in which constitutive practices are regulated by international membership organizations like the UN. As Griffiths (2021) writes, the defining feature of sovereignty in the post-1945 period is obtaining a full seat in the United Nations General Assembly (UNGA), what he called the Sovereignty Club. Like the East Asian system, it displays clear constitutive practices. But not the same practices; in fact, the UN system appears to go further with the development of membership organizations. We theorize that as interaction capacity has increased throughout the UN system since 1945, we should expect to see a deeper institutionalization of diplomatic practices in membership organizations. Moreover, as we argued, that institutionalization can gradually supplant the need for bilateral relations between states.

A unique quality of the institutions of the modern international system is their universalism, even in spite of the recurrent systemic establishment of competing blocs of states. The central components of the international system which bestow legitimacy are universalist in nature. For example, while the UN is the most prominent example of a constitutive international membership organization, it exists in a broader universe of organizations that together compose the network of international organizations. Table 1 provides examples of the largest organizations in the system, according to the number of member states. Other organizations, such as the International Telegraph Union and the Universal Postal Union, play an important legitimizing role like the UN.

Table 1.

Largest international organizations (2014).

Organization	Members
United Nations	193
International Telegraph Union	192
Food and Agriculture Organization	191
Universal Postal Union	190
World Health Organization	190
UNESCO	189
Montreal Convention	187
International Monetary Fund	186
Interpol	185
International Labor Organization	183

We utilize the passage of time as a proxy indicator for an increase in interaction capacity. In the 20th century, technological developments facilitating interaction emerged and radically transformed the ability of states to engage with one another even at great geographic distance (Keohane and Nye, 2012, 194). The increase in potency of these technologies is roughly in line with the passage of time: interaction capacity has grown more effective year by year without any substantial decreases. This is in contrast to earlier historical periods, in which interaction capacity plateaus for long periods and even decreases during certain periods. For example, one major technological ecosystem with extensive impact on the interaction capacity of states is the development of digital information infrastructure. Like all technological systems, the digital information infrastructure developed unevenly since the mid-1990s, initially reaching only the rich countries of the developed world. Penetration of digital technologies throughout the international system, to an extent that could lead to a degree of systemic change, was a gradual process that took place over the course of decades (Herrera, 2003; Lemke and Habegger, 2018). These technological advancements in infrastructural capacity that continually expand across the international system directly contribute to the interaction capacity of the international system as a whole.

To see the effects of increasing interaction capacity on the international system, we compare the evolution of two networks of states. Specifically, we measure the average degree and density of international networks. Degree is the number of relationships a state maintains with other states. In the network of embassies, the degree statistic of a state is how many embassies they send and receive. For example, if a country sends 20 embassies (an out-degree of 20) and hosts 40 embassies (an in-degree of 40), the degree of the country would be 60. In the context of the network of international organizations, the degree is the number of organizations a state is a member of. If a state is a member of 60 international organizations, the state would have a degree of 60. The average degree statistic demonstrates the level of connectivity in the network: a low average degree implies low levels of connectivity among states. In contrast, density is the ratio of actual ties to potential ties. For example, in the case of the embassy network, if each state were to send an embassy to every other state, the density of the network would be 1. Conversely, if no states exchanged embassies, the density would be 0. A density of 0.25 in the network of embassies means that 25 percent of all potential embassies that could exist have been realized.

We examine how each network has evolved temporally to understand the effects of increasing interaction capacity. Deeper institutionalization of diplomacy through the system of membership organizations should be visible if state connections through international organizations are growing faster and increasing in density compared to direct connections through bilateral embassies.

For data on embassies, we relied on the Diplometrics project. Because of changes in the coding scheme in the mid-1980s, the original Correlates of War (COW) data is problematic for longitudinal analyses. The Diplometrics project remedies this deficiency by standardizing the coding schema, but with the limitation that data only stretches back to 1960. International organization membership data is drawn directly from the COW project (Pevehouse et al., 2020). Membership lists in international organizations are relatively straightforward to code in comparison to embassies.

Figure 2 demonstrates how the network of bilateral embassies has evolved since 1960. We can see that although the average number of embassies sent and received by states (the average degree) has increased dramatically from 30 to almost 90 embassies per state, the overall density of bilateral contacts among states has remained relatively constant. The density of the system has remained stable at between 0.2 and 0.24. This means that while the system has grown in absolute terms, the intensity of interaction among states has not increased. Since 1960, states have, on average, never established direct diplomatic contact with more than a quarter of other states, even though interaction capacity is increasing. This trend is visible in Figure 2(b); although there are ups and downs over the decades, particularly with the collapse of communism in the early 1990s, there has been no substantive change between 1960 and 2010.

Figure 2.

Embassy network. (a) Average degree. (b) Density.

In contrast to the network of embassies, the network of international membership organizations has grown dramatically in density since the 1990s. Figure 3 displays the same set of longitudinal statistics for the network of international organizations. The effects of increasing interaction capacity are visible in the skyrocketing connectivity between states through international organizations. From an approximately 10 memberships in 1950, states now maintain an average of over 60 memberships.

Figure 3.

International organization network. (a) Average degree. (b) Standardized valued density.

Notably, the average density of the embassy network remained relatively constant around 20–25 percent, despite the absolute growth in the number of sovereign states and average degree size of the embassy network since 1960. In contrast, the international organization network has increased in both degree and density, as seen in Figure 3. From 4 percent of potential memberships being realized in 1945, density has increased to nearly 30 percent of possible shared IGO memberships per state. The increase in density is particularly striking, since this measure accounts for the formation of new IGOs. We see a rapid increase in IGO membership density in the post-1990 period, correlating with the improvement in global digital infrastructure that contributed to increased interaction capacity in the international system.

Overall, the embassy network appears to be relatively ossified compared to the network of international organizations. Figure 4 displays a ratio of the previously analyzed statistics of the embassy and international organization networks. The trend lines indicate that the organization network is increasing and overtakes the embassy network in both size and density between the around 1990. States historically maintained more ties to other states through embassies than IGOs from 1960 to the 1990s. (Figure 4(a)). The density of the international organization network surpasses the density of the embassy network around the mid-1990s (Figure 4(b)). Given many memberships are impossible (e.g. Japan will never be a member of the Organization of American States) while any embassies are theoretically possible, a comparatively lower density in the international organization network is unsurprising. The fact that the ratio between the two surpassed 1 indicates the increasing connectivity of states through the international organization network in comparison to the embassy system.

Figure 4.

Comparing embassy and IGO networks. (a) Average degree. (b) Density.

In Figure 4, higher values of the ratios indicate better connectivity in the international organization network compared to the embassy network. For both statistics, the international organization network increases in relevance, though to differing levels. In terms of density, states are much better connected to other states through the IGO network. Where the embassy network has reached a potential density limit, the international organization network appears more responsive to increases in interaction capacity.

The international organization network is becoming more institutionalized, supplementing if not supplanting the embassy network. International organizations, because of their efficiency, represent a higher level of diplomatic evolution than the bilateral embassy network. In the short period of time we analyze, there is a gradual shift toward increasing connectivity through the network of international organizations, demonstrating how systems can move along the continuum between declaratory and constitutive as a function of increasing interaction capacity.

The qualitative difference in the importance of gaining membership in international organizations compared to establishing bilateral embassies is evident when we compare two episodes: three Arab states that entered the international system in 1946, and the case of South Sudan, which entered the international system in 2011. The three states that gained sovereignty in 1946 established extensive bilateral diplomatic missions with other states in multiple regions around the world, part of a conscious effort to diversify diplomatic ties in a context dominated by great powers. In contrast, when South Sudan gained its own independence, international organizations were the main avenue through which the country expressed its diplomatic presence.

In 1946, Syria, Lebanon, and Jordan emerged from mandate status as sovereign states. By 1950, within 4 years of gaining sovereignty, these countries maintained active consular representatives in countries across Europe, North America, the Middle East, and South America. All three countries sent missions to great power diplomatic hubs, such as the United States and the United Kingdom, a common practice among all states (Alger and Brams, 1967: 653). In addition, these three also established missions in other regions, particularly in Latin America. All three states established direct diplomatic ties with Argentina, which at the time was at loggerheads with both the United States and especially the Soviet bloc regarding membership in the UN. In the context of the UN membership debates underway in San Francisco, Argentina utilized diplomatic solidarity with the new Arab states to undercut Soviet opposition to Argentinian membership. For the Arab states, establishing a substantial diplomatic presence in Argentina was an effective way to consolidate their new status as independent states (Klich, 1995).

Bilateral diplomatic missions have always served a legitimizing purpose as an expression of diplomatic recognition. To send an embassy, or to withhold one, makes a normative statement from the sending state about the status of the receiving state. In the nascent years of the UN, when membership debates were active and contentious, establishing diplomatic missions could circumvent limitations on membership in universal international organizations. While diplomats from both Latin America and the Middle East met on the sidelines of early UN conferences, working bilaterally outside the confines of the organization allowed Argentina to pursue systemic integration without exclusively relying on the organs of the UN.

In addition to establishing diplomatic missions around the world, the three Arab states were founding members of the League of Arab States but otherwise exclusively joined universal international organizations like the UN. While the three states maintained an average of 13 embassies within 4 years of independence, they only joined eight international organizations. This pattern of establishing more embassies than joining international organizations would later reverse, as is evident in the case of South Sudan.

South Sudan was established as a seceding state from Sudan following a referendum in 2011. Within 3 years of independence, South Sudan established 15 embassies and joined 25 international organizations. Like the Middle Eastern states that gained independence in 1946, South Sudan established diplomatic missions in global hubs such as the United States, United Kingdom, and Belgium, arguing in the case of the United States that establishing an embassy in Washington was necessary because of its global importance (Kinne, 2014: 257). In addition to countries in the global north, South Sudan opened embassies in several East African countries, such as Ethiopia, Uganda, and Kenya. Outside the African continent and Europe, South Sudan did not set up more than one mission in any other region.

However, South Sudan integrated rapidly into regional and international organizations. Beyond the universal international organizations like the UN, South Sudan acceded to a variety of regional organizations including the African Union (AU) and the East African Community (EAC). Joining the EAC was itself part of a coherent effort to culturally and diplomatically integrate into a new geopolitical neighborhood (Mondesire, 2023). Unlike the Arab League, with its limited level of economic integration among member states, the EAC has a relatively high level of cultural and economic integration, including a common market.

Between 1946 and 2011, interaction capacity increased dramatically. The cases of Arab countries in the late 1940s and early 1950s demonstrate that establishing bilateral relationships across broad divides in the international system could undercut exclusion in international organizations. By the 21st century, however, as the case of South Sudan illustrates, international organizations have become effective gatekeepers of membership in the international system. Establishing legitimacy in the international system is defined by integrating into international institutions. As interaction capacity grows, international organizations, with their constitutive features, become more important to diplomatic activity than their bilateral counterparts.

In sum, the contemporary UN system is not only highly constitutive, but it has also become more so since its creation. On Figure 1, it would be located on the far right where membership organizations are prominent. Importantly, interaction capacity has increased during the UN era. We contend that this has deepened constitutive practices, gradually shifting them from bilateral relations to diplomatic interaction in international membership organizations.

Conclusion

Do states exist prior to systems, or do the system members determine who can be a state? We argued that practices of statehood exist on a spectrum ranging from declaratory to constitutive. The main factor shaping the nature of a state system in this regard is the level of interaction capacity. In systems with comparatively low interaction capacity, states need not seek the acceptance of a club of states. Comparatively, in systems with higher levels of interaction capacity, more constitutive practices dominate and institutions of recognition play a more significant role.

This theory was supported in the empirical analysis. In the cases of two 19th-century state systems, there is clear evidence that the system with higher levels of interaction capacity exhibited more constitutive practices. West Africa was a low-capacity system exhibiting declaratory practices of statehood. East Asia, the highest capacity system, displayed clear constitutive practices related to the tribute system. Finally, the effects of increasing interaction capacity within a single case are evident in the contemporary post-1945 UN system. As interaction capacity increased over time, constitutive international organizations became thicker and more connected. In contrast, lower-level diplomatic processes, namely the exchange of bilateral embassies, plateaued in development. This gradual shift within a case exemplifies how the process plays out on a micro-level, with more institutionalized forms of diplomatic activity becoming increasingly relevant for states as interaction capacity grows.

The question of which comes first – states or systems – can be answered by including the factor of interaction capacity. Over the long run, the state made the system and the system made the state. We contend that this is the basic pattern by which systems evolve. Further research can test this theory across different historical systems. It could also identify the potential for slip back. As the fall of Rome, the disintegration of the Mughal Empire, and the collapse of the Chinese dynasties showed, this is not a one-way historical process. It would be useful to know under what conditions system practices become less constitutive and more declaratory. The Chinese dynastic cycles suggest that constitutive practices can be sticky, providing an enduring diplomatic grammar. Looking at the past may give us a sense for what specific constitutive elements of the UN system would endure in the event of its collapse.