Seeing borders
Most of us picture borders by picturing maps. In our mind’s eye they are the lines that carve the world up into individual political states. By extension we can easily picture the borderlands that surround them - spaces running alongside borders like the margins that run along the sides of a piece of paper. This imagery gives us the most straightforward and most practical definition of borderlands – spaces close to international borders. That simplicity is part of the term’s power, and just thinking about it prompts a shift in perspective on territory and socio-economic development. Suddenly we have geographical spaces that cross border lines – a whole new unit of analysis outside of the state system:
Many of us also have an intuitive understanding of why living in a borderland is unlike living anywhere else. Being close to a border brings closer access to everything that is different on the other side. Cross-border asymmetries – in living conditions, job markets, the price of food and rent, natural resources, schools, churches, banks, bars, and sports stadiums – they all create opportunities, especially where residents can enjoy the ‘’best of both worlds’’.
But borderlands are also by definition far from the centre of the state. That physical distance tends to translate into political, cultural, and economic distance from central government, and from powerful interests in the capital city. Very often it leaves people living in border regions marginalised, lacking a voice in their own politics. The problem extends from states to development agencies, as seen in recent studies that show aid expenditure drop off significantly with distance from state capitals.
Practical Problems:
Throughout Africa, developmental initiatives tend to be ineffective in borderlands, especially where states lack the practical means or the political will to step out from the centre and better serve the periphery. Even where these obstacles are successfully overcome, most developmental project work – from digging boreholes to building roads to training healthcare workers – extends right up to the border… then stops the moment a foot is set on the other side. The moment the line is crossed, it becomes someone else’s country office, someone else’s funding stream, someone else’s political buy-in, operations team, currency, visa, and vehicle registration – in short, it becomes someone else’s problem.
Project work on one side of a border affects the other, and right now it does so in ways that we struggle to predict, understand, or plan for. This approach will never work well. Africa’s borderlands are integrated social and economic systems. There tends to be a lot more in common across lines at the edge of the map than there is between borderland regions and their respective capitals. Pretending that states are atomised, insulated units, creates an enormous blind spot around their outer edges – a space that too often fills up with violence and poverty.
Zooming in:
More technical work calls for a more technical definition. To say that that borderlands are ‘’spaces close to borders’’ may work at the level of political advocacy or fundraising, but for researchers and data scientists it is impossibly ambiguous. Put your back to a border and walk away from it --- at what point do you exit the borderland? It is a very simple question and one that is very difficult to answer.
Defining borderlands based on kilometre distance from borders quickly breaks down. Walk 50 east out of DRC into Rwanda and you are practically at Kigali in the heart of the state. But walk 50 miles south out of Ethiopia into Kenya and you have barely scratched the surface of Marsibit County, let alone Kenya as a whole. In one case you passed all the way through, in the other you barely got through the door. For more absurdities, think of walking 50 miles north of Senegal into The Gambia, or 50 miles east of South Africa into Lesotho.
An alternative to raw distance is to take the borderland as being made up of a state’s border-touching districts. This approach has some merit, since municipalities often contain at least some reflection of demographics – population sizes, densities, and ethnicities. They are social and political units and not purely geographical, which resonates with the ideas of the borderland discussed at the start. The problem this time is that the districts of neighbouring states tend to have very different sizes and populations. If you take them together on either side you often get mismatches – think enormous swathes of territory in Algeria bordering much more fine-grained administrative units in Morocco. Even when administrative units across a border have similar proportions, they are often drawn in a way that makes labelling one place a borderland and another not, practically impossible to defend. For example:
The upshot seems to be that while some places are unambiguously borderlands, very few places are unambiguously not borderlands. Taking a 30km buffer zone around a border or taking a list of counties that make physical contact with borders – these approaches will give you a borderland. They just won’t give you all of a borderland, and occasionally they might give too much. Either way it raises difficult questions about why some areas are included and others not.
Capital cities, or mega-cities, might be an obvious answer for what a borderland is not. But then what about Kinshasa, or Brazzaville? These cities sit right on top of borders, but their internal dynamics are defined to a much greater extent by the presence of central government – a completely different social and political force.
The fact that we often do not consider large urban spaces such as Lagos, or capital cities such as Lomé, to be borderlands suggests that dynamics may be the key --- that borderlands are regions in which the border has a significant influence over everyday life. But this also creates a serious practical problem in the form of a circular argument: Borderland spaces are those that exhibit borderland dynamics. And borderland dynamics? They are what goes on in borderland spaces.
In a sense we have clarity, but only because nothing whatsoever has been defined.
Turning to academia does not offer a great deal of encouragement. In the classic text ‘’Understanding Life in the Borderlands”, William Zartman writes: “The epistemological result of these sets of components of borderland life and analysis is a Rubic cube of interactions, complex but limited, spinning its own planes and traveling through time […] This fluid structure only sets up the possibility of identifying some salient interactions that stand out in determining the dynamics of borderland life.”.
The Rubic cube (my emphasis) is just a glimpse into what lies at the bottom of the definitional rabbit hole. It opens the door for an interpretation of borderlands that loses touch entirely with international borders. The result has been a kind of academic mission-creep that sees borderlands springing up between institutions, between people, and between ideas.
This is a bus stop for many of us, even those who initially found the concept very appealing. It may also be one of the main reasons why the borderlands movement has begun to fade, never quite reaching its potential as a fully-fledged, interdisciplinary area of study, nor as a practical entry point for governments and development practitioners. Lost in space (and time).
What next?
First, we need to be comfortable dealing with grey areas. Given the subject matter, this should not be at all surprising. Spaces “in-between” quite literally come with the territory.
That said, practical steps can be taken when it comes to examining, comparing, and ultimately better understanding Africa’s border areas. This starts by making use of more than one definition, then choosing appropriately between them depending on what the idea of the borderland is being used for. Some examples:
1. A 30km ‘’walking’’ borderland – a practical unit of analysis for looking at such things as informal cross-border trade and walking access to services. It is based on the simplified logic that a 60km round trip is the maximum that could be reasonably be walked in a 12-hour timespan. Naturally there will be exceptions, but the 30km figure would cover the vast majority of walking trade. It is one example of how a distance-based borderland can still offer a valuable lens, in spite the problems outlined above.
2. A municipal borderland, made up of the combined territories of local government administrative zones on both sides. This would most likely be what OCHA refer to as tier 2, often (but not always) labelling the “district” level. Most African borders make contact with five or so units of this size. The municipal borderland reveals a space that carries with it political and demographic characteristics that provide useful perspectives on borderland administration and governance.
3. Catchment borderlands: probably the best illustration would be water catchment zones surrounding borders – areas fed by the same sources. But the same logic could apply to services, drawing out borderlands around the catchment zone of things such as schools, hospitals, or airports. This approach can help map healthcare or education access in a cross-border fashion, something rarely attempted as things stand.
4. Limited borderlands: Put bluntly: mega-cities or governing capitals can simply be taken out of datasets to give a more representative view of populations who are influenced principally by borders, not by urban or other dynamics.
5. Proportional borderlands: Not much more than an experimental idea, but one that may provide practical workarounds in some circumstances. Take the landmass of a state, for example DR. Congo at 2,345,409km2. Model that as a circle and it’s easy enough to calculate the diameter. Take 10 percent of that diameter and you get a borderland zone of 273.1km. Do the same with neighbouring Rwanda (26,000 km2) and you get 28.9km. The flaws with this approach are obvious: states are not circles and borderlands are not uniform. But under certain circumstances this may offer a crude entry point for estimating such things as continent-wide borderland populations, especially where neighbouring states are vastly different in landmass and distance-based approaches break down.
Going back to Kenya and Somalia, you can layer this all together or isolate different borderlands however you want. Here is the ugly result of an all-in-one:
If that looks like a mess, it’s because it is a mess. But this is also intellectually what we have been trying to do for decades. The point is that we shouldn’t. There is no one-size-fits-all interpretation of the borderland.
Instead, we should pick, depending on what we’re talking about, the conception of the borderland best suited to it --- deliberately catering our delineation of the borderland to the subject matter at hand. With modern GIS tools, we can take data from any one of these layers, compare it to any other layer, or any other borderland, or any national average. We can know the hospitals per capita for a cross-border space between Togo and Ghana and see how it compares to other cross-border regions in East and Southern Africa. Or how much water access there are in the border zone compared to the national averages on either side. We can identify the cross-border areas of greatest need and develop risk maps that take on board the connected dynamics on both sides of the border. And therein is some real value.
Hugh Lamarque