a Pre-Westphalia (before 1648)

It makes sense to begin in the pre-Westphalian window, as the conceptual roots of borders form here. In medieval Europe, a ruler’s dominion was not evaluated based on the expanse of rule, but rather on the quantity and quality of material it encompassed. This included people, places, and property; the medieval model of authority is often referred to as population based (i.e., population-as-jurisdiction). Where authority claims were territorial, they tended to be concentrated narrowly around the capital, petering out into indistinction in the periphery – nebulous zones in which people often claimed fealty to more than one ruler. For the most part, exterior and interior boundaries were indistinguishable – external boundaries “were fundamentally similar in kind to feudal limits within the kingdom” (Sahlins, Reference Sahlins1989: 6).

The idea that authority should be linked to territory and delimited by borders – the first stage, conceptually at least, of territoriality (1a) – dates from the thirteenth century. By one account, the critical years were 1212–1221, during which period “the notion itself of boundaries was established” (Sassen, Reference Sassen2006: 44). This window also saw the first steps toward the development of ports of entry, such that the gateway to the territory could be a tool in extracting customs – an early version of the facet of control I call extraction (2b).

Shifts in the logic of boundaries were aided greatly by the expansion of European powers overseas, where colonial dominions had to be divided quickly and efficiently; an inchoate form of linearity (1c). For example, in the Treaty of Tordesillas (1494), Spain and Portugal divided their territories such that “[a] boundary or straight line [una rraya o linea derecha] be determined and drawn, from pole to pole” (cited in Elden, Reference Elden2013b: 242–243).

New ideas about territory and boundaries were aided hugely by maps, beginning in the fifteenth century. But while the map was essential to the adoption of thinking about authority in the language of space – territoriality (1a) – this is deceptive, because while many early maps show how states were distinguished by boundaries, these were in fact largely cartographic inventions, creating linear borders where they didn’t previously exist. As Jordan Branch writes: “Key characteristics of modern statehood – such as linear boundaries between homogeneous territories – appeared first in the representational space of maps and only subsequently in political practices on the ground” (Branch, Reference Branch2014: 77).

In their initial manifestation, borders were contested zones where states defended against each other (i.e., where one state “stood face to” another). But the boundary also acquired symbolic purchase at this time, including as places that should be loyal to a single metropole, i.e., exclusivity (1b). In 1564–1566, Charles IX of France did a two-year tour of the frontiers of France designed both to solidify their dominion and forge local identification with the center. It was a spectacular display of state power, with feasts and ceremonies. But the sheer fact that such a bond needed to be created suggests little existed earlier.

b Westphalia to Paris (1648–1815)

Scholars usually trace the formation of the modern state system to the seventeenth century in Europe, following the end of the Thirty Years War (1618–1648) and the Peace of Westphalia. There is evidence to suggest the weakness of this historical marker as the birth point of sovereignty (Osiander, Reference Osiander2001). Nevertheless, the seventeenth century saw a shift in the language of treaties from “a listing of places and non-territorial jurisdictions to a careful delineation of spaces separated by discrete boundaries” (Branch, Reference Branch2014: 6). But this took a considerable amount of time – in fact, medieval forms of rule lasted well into the nineteenth century (undercutting (1a)). Here again, while maps portrayed an aspirational uniformity (in which territory was demarcated and fixed, and authority was exclusive and nonoverlapping), this did not yet exist on the ground.

In seventeenth-century France, for example, leaders resisted borders that were either linear or fixed (1c) because they impeded expansion. Louis XIV (1643–1715), in particular, developed sophisticated legal tools to justify territorial growth:

Thinking of borders in this way was good for justifying expansion. But French authorities did make contemporaneous efforts at achieving territorial exclusivity (1b). In a memorandum in 1673, Marquis de Vauban, a key figure in Louis XIV’s statecraft, wrote: “The king ought to think a little of squaring his field. This confusion of friendly and enemy fortresses mixed together does not please me at all” (Sahlins, Reference Sahlins1989: 68). Vauban spent the next decades putting a plan into action: “[Vauban’s] ‘iron frontier’ consisting of two lines of fortified sites. The idea was to abandon the most advanced fortresses and towns, relinquishing more distant outposts in the interests of a more enclosed space … ‘The enemy,’ writes Vauban, ‘will almost never know what is going on behind our backs’” (Sahlins, Reference Sahlins1989: 68–69). This passage reveals how revolutionary the concept of nonoverlapping territories really was. Previously, interweaving territories – the enemy behind one’s back – were the norm. This falls short of linearity (1c) as these are areas remained zonal, but they were increasingly nonoverlapping (1b).

A considerable step in the development of the modern border took place at the Treaty of Utrecht (1713), with the division of British and French colonial holdings in North America. This was the first time that a boundary commission was appointed to determine where a border should actually lie. This was possible in the New World as this land (as far as the European powers were concerned) was uninhabited and thus easily divisible. At the level of principle, linearity (1c) was here favored; but because they couldn’t agree on what maps to use, it remained unattainable in practice.

The first real efforts at linear borders in Europe came later, in 1775, when the French Ministry of Foreign Affairs developed a strategy for “establishing and fixing the limits of the kingdom” (Sahlins, Reference Sahlins1989: 93) During the subsequent decade, France negotiated almost two dozen “treaties of limits” with their neighbors – designed to be “rational” (i.e., homogeneous, straightened out, and free of enclaves) (Goettlich, Reference Goettlich2019: 213). By one account, this expressed France’s need to “carefully close its territory, as a peasant would enclose its field” (Sahlins, Reference Sahlins1989: 95).

The view on the ground was of course more complicated, a point rendered clearly by Peter Sahlins in his seminal study of the Cerdanya, on the border between France and Spain. The claims of each country were officially determined at the Treaty of the Pyrenees (1659–1660), which used the language of linear borders, but in fact peoples, rather than territory, remained the source of political loyalty – a link to the medieval bonds whence the early state arose – and the border operated as a complex amalgam of lines and zones. The location of the border itself was not fixed until centuries later, at the Treaties of Bayonne (1866–1868): “The history of the boundary between 1659 and 1868 [can] hardly be summarized as the simple evolution from an empty zone to a precise line, but rather as the complex interplay of two notions of boundary – zonal and linear – and two ideas of sovereignty – jurisdictional and territorial” (Sahlins, Reference Sahlins1989: 6–7). Moreover, during this period, the border was jointly managed by the French and Spanish authorities, especially in their common pursuit of smugglers – another example of the border becoming a site of extraction (2b), albeit in the most overlapping of manners (contra 1b).

As the eighteenth century came to a close, the process of border linearization was clearly underway and the complex schemes of authority that characterized the Middle Ages were slowly replaced by mapped, defined, and increasingly (or at least relatively) homogeneous logics of rule. But it would only be after the French Revolution, and the subsequent defeat of Napoleon, that many of the remaining pieces would fall into place.

c Napoleon to the Present (post-1815)

The first mention of linear borders in an actual treaty in Europe was in 1815, at the Congress of Vienna. Article II of the General Treaty includes the phrase: “That part of the Duchy of Warsaw which His Majesty the King of Prussia shall possess in full sovereignty and property … shall be comprised within the following line…” (cited in Branch, Reference Branch2014: 5). This phrasing – whereby linear borders were established as the core principle defining political authority – was possible in large part because Napoleon had run roughshod over the continent, providing postwar elites a clean slate by which to carve up territory in ways previously reserved for the colonies.

And it is in this period, especially after the 1850s, that we begin to see real changes on the ground, owing to expanded state capacity, the rise of bureaucracy, police activities aimed at population control, and state use of statistics – an important forebear to contemporary data protocols – which enabled “the supervision of human activities” (Giddens, Reference Giddens1987). This expansive administrative state led to radical changes at the ports of entry, which became hubs of taxation and surveillance (2c), as it was only through “documents such as passports and identity cards, along with elaborate registration and information systems” that states could effectively distinguish “who is who” and “what is what” (Torpey, Reference Torpey2018: 37).

At this point the Westphalian border began to come into view, at least in Europe. But in other parts of the world, it was a norm for borders to be territorial but not linear well into the nineteenth century, largely because European powers were loath to make linear borders in colonized areas as this might put a hedge on their expansion – as evidenced perhaps most spectacularly in the so-called scramble for Africa, 1884–1885. It was only at the turn of the twentieth century that this began to change – mostly because of the mutual exhaustion of territory across the globe. On this point, Thomas Holdich, in charge of many of the borders of the British Empire, wrote in 1899, “Truly this period in our history has been well defined as the boundary-making era” (cited in Goettlich, Reference Goettlich2019: 211).

The twentieth century saw the rapid building up of borders as physical spaces, following upon and buffered bytreaties, such as the Treaty of Versailles (1919) and the UN Charter (1945). The purpose of these militarized borders was principally defensive (2a), especially in the Cold War years, 1950–1990. This changed in the 1990s with the rise of global migration, at which point borders generally stopped being places where states faced off against each other and instead became places where states controlled migrants (2c) – that is, the border changed from being a military space (concerned with armies) to a police space (concerned with migration) (Bigo, Reference Bigo, Bigo and Guild2005: 55–56). And so, while the nineteenth century saw states obsess over border delineation, consecrating the Westphalian border-as-authority, the twentieth century saw them build up borders as physical institutions – that is, borders-as-control.

d Borders as Control: The Case of the US Border with Mexico

To discuss border-as-control, I will use the example of the US–Mexico border as it represents one of the most built-up borders in the world. Given the struggles of the US government in attempting to establish control, it is illustrative of how rare the phenomenon really is.

The US–Mexico border came into existence primarily as a result of the Treaty of Guadalupe Hidalgo, 1846–1848. But the movement from delimitation to demarcation was a complex one. The boundary commission was not only faced with negotiating inhospitable environments and climates, but there was also political opposition to the idea that the border should be fixed in place at all – in the nineteenth century, many Americans felt it was their “Manifest Destiny” to expand across the continent (contra 1c).

Once the border was demarcated, with boundary stones set at intervals across the linear expanse, it nonetheless remained largely unpopulated for most of the nineteenth century. Technically there was a customs office set to manage cross-border trade (2b), but in practice this was all but impossible. This made for a climate, as the International Boundary Commission reported in 1896, of “open, yet lawful, evasions of customs duties” (cited in St. John, Reference St. John2011: 90). And anyway, the border was so vast it was nearly impossible to determine which side of the line people were really on (and thus which authority should be apprehending them).Footnote ⁵

It wasn’t until the 1910s, corresponding with the Mexican Revolution – which produced a huge volume of migrants fleeing violence – that any serious infrastructural development took place. In 1917, Congress passed a comprehensive Immigration Act which tightened restrictions against immigrants (2c), including by imposing literacy tests, taxes, and enforcing the prohibition of certain kinds of laborers. By the time of the Immigration Act of 1921, the border was no longer fully open. Most Mexicans still came and went without trouble, but were increasingly subject to invasive physical protocols. US health officials required Mexican border crossers to surrender their baggage and clothing for sterilization, and endure intrusive examinations and disinfection processes.

These changes at the ports of entry naturally pushed people to cross illegally in the desert. In response, following the Johnson-Reed Immigration Act of 1924, the Border Patrol was formed. Congress disbursed funds sufficient for 472 officers to patrol America’s borders with Mexico and Canada (2a) – still a pittance, but a massive increase given what had existed before. This new system plodded along until the 1930s when the Great Depression sparked fears that Mexican migrants would take jobs from American workers. At this point, border officers began enforcing immigration restrictions more tightly, beginning a process of trying to assert control over the border that continues to this day. The border could at this point be considered to be Westphalian, at least in its rough contours. These policies went alongside fencing, which grew incrementally through the decades.

But to what degree was control really attained? The move toward fences and walls took off with renewed vigor with the Immigration Reform and Control Act of 1986 and expanded in 1996, through the Illegal Immigration Reform and Immigrant Responsibility Act, and again after 9/11. But the effectiveness of walling is famously hard to measure, owing to the displacement effect – walling in one area simply pushes illicit trade and traffic elsewhere (McCarter, Reference McCarter2010). Moreover, walls fail to stop tunnels that go underneath them and unmanned aerial vehicles (or drones) that go over them.

All told, despite their political appeal, border security practitioners have broadly come to accept that the idea that walling would generate operational control was “unreasonable and unsustainable” (Basham, Reference Basham2013). As late as 2011, the Government Accountability Office reported that of the 873 miles along the southwest border, only about 129 miles (15 percent) were classified as “controlled”; the remaining 85 percent were classified as merely “managed.” If you include the northern border, only 69 miles, or 1.7 percent, of 4,000 total miles were under “operational control” (Kimery, Reference Kimery2011).

Rather than pursue unilateral attempts at border control, contemporary border security practitioners increasingly understand that we must be “widening” the border and engaging in cooperative agreements with Mexico and Canada. Thus, while the US border took on its Westphalian form at least in the 1930s, control in any meaningful sense has been evasive. Indeed, the more the line was bolstered by infrastructure, the more government officials realized that a wider, more zonal model was desirable – bringing us back to premodern versions of the border, whence we started.

a Looking Backwards

The linear border is typically contrasted with the zonal frontier; the former is for states what the latter was for empires and medieval proto-states. But from the study presented here, we can see that aspects of these older forms of rule lingered far longer than we usually assume and linearity was slow in developing. This suggests we ought to view frontiers and borders less in opposition than in dialogue – situating them on a spectrum of linearity – where the frontier represents a zonal form of political rule and the border a linear one.Footnote ⁶ Before developing this point, it is helpful to say a bit about this frontier/border dichotomy and its origin in imperial rule.

We usually think of empire as a kind of authority that is spatial but not sovereign, because sovereignty is territorially bounded while imperial authority isn’t. This latter authority – what in Rome was called imperium – was broad enough to encompass modes of rule over territories outside the metropole. “The imperium populi Romani was the power Romans exercised over other peoples, viewed in its widest sense … If the limits to the imperium of a Roman magistrate on the boundaries of Roman power were not strictly defined, this implies that the boundaries of the imperium Romanum itself were uncertain” (Lintott, Reference Lintott1981: 53–54; 64). By this logic, the center can expand or contract without territorial limitation. What is interesting for our purposes is that this is the same logic of rule that undergirded both the French and American processes of state-building, each of which resisted fixed borders well into the nineteenth century. Despite the language of sovereignty, both metropoles pursued polities that more closely resembled imperium. The authority they pursued was spatial, but not bounded. As such, it viewed jurisdiction as much in the medieval sense (qua population) as the modern sense (qua territory).

Just as we might think of borders and frontiers as positioned on a spectrum of linearity, we might think of authority similarly – that is, on a spectrum of boundedness – with dominion more or less fixed to the territories over which rulers preside. Allen Buchanan makes something akin to this point with regard to sovereignty, critiquing the idea that it is ever “an all-or-nothing affair,” but rather “a matter of degree” (Buchanan, Reference Buchanan, Buchanan and Moore2003: 236). Osiander (Reference Osiander2001: 277) too contends that rather than posit a transition from empires to sovereign states, we should instead view them in tandem – that is, that they “do not represent mutually exclusive paradigms. Instead, they are part of a spectrum.”

In sum: different forms of authority have persisted under the surface throughout the modern era, obscured by the way we have conceptualized the state/territory/border triad. As the history presented here has shown, in the process of moving from zonal to linear boundaries, states have operated in flexible, homeostatic ways toward their peripheries, using both imperial and sovereign modes of authority. Looking at the linear border as merely the recent stage in this evolution – in no way natural unto itself – furnishes us with a new frame by which contemporary changes at the border might be viewed.

b Looking Forward

In recent years, the border has become a hot topic in political and legal theory, with a special focus on how security protocols are being deterritorialized. My point is not to revisit these claims or to evaluate the moral harms they uncover but rather to contextualize them. The historical material presented here suggests that polities embrace more zone-like or line-like strategies of boundary maintenance at different times and according to different needs. States in this rubric are not positioned as part of a fixed grid, but rather an adaptive system of expansion and contraction. I call this way of thinking about borders the Accordion Model.

Thinking of borders in this way opens several doors for reconsideration. The first pertains to the scope of authority. While in the last centuries we have seen the border contract from a zone down to a line, the Accordion Model suggests we shouldn’t see this as a natural end-state of the state system, but rather a conditional resting point – with contemporary borders now reexpanding back into a zone (on the linearity spectrum) and authority taking a more imperial positioning (on the boundedness spectrum). Historically, linear borders made sense as a strategy of peace-making, responding to conditions of interstate war; just as deterritorialized borders make sense given the conditions of globalized mobility we see today.

The Accordion Model also captures the oscillation between the (medieval) population-jurisdiction model and the (modern) territory-jurisdiction model. For most of recent history, we have been moving away from population and toward territory as our defining idiom. But contemporary evidence suggests we are moving back. This is perhaps most clear with regards to digitalization – a centerpiece of current border security policies – which reinstitutes the population-jurisdiction model across a wide range of deterritorialized spaces. If you begin with the assumption that borders/states/territory are naturally linked, the world of data governance seems like a radical rupture. But if instead you view borders as oscillating between territory and population models, the new data protocols fit perfectly into this course of evolution, with states adapting to new circumstances by reclaiming old principles.

Indeed, conceiving of the border as akin to an accordion forces us to ask different questions: not whether new border protocols deviate from prior norms, but whether they revisit them; and if so, under what conditions these older territorial/deterritorial, linear/zonal markers of authority might thrive. These questions deserve consideration beyond what can be offered here. The hope is that by starting from the assumptions that borders are contingent – not natural – we can perhaps better understand the political and moral significance of these contemporary changes, opening avenues of inquiry which have been previously obscured.