Colonial epistemology: whose knowledge counts as knowledge in the global research system

By The Arcaneum

Category: Academic Commentary

Tags: colonial epistemology, global research system, open access, article processing charges, parachute science, knowledge production, epistemic justice, global south research, academic publishing, web of science

In January 1666, the Philosophical Transactions of the Royal Society of London published a short document titled "Directions for Sea-men, Bound for Far Voyages." It was drafted by the mathematician Lawrence Rooke and edited by the Society's secretary Henry Oldenburg, and it asked naval captains heading to the East and West Indies to record specific observations: tidal patterns, depth soundings, and the behaviour of compass needles at various latitudes. The Royal Society wanted the natural world catalogued, and it had a practical problem: most of its fellows would never see the coastlines it wished to describe. So it recruited the ships. Knowledge would flow from the periphery to the centre; the Society would analyse and classify what the sailors gathered. The people who actually lived on those coastlines, who had accumulated generations of precise knowledge about exactly the phenomena Rooke's questionnaire asked about, were not consulted. They were subjects of inquiry, not parties to it.

This is not a discovery. Historians of science have traced this arrangement carefully for half a century. But I keep returning to it, because the arrangement has not ended. It has been institutionalised.

The contemporary form runs like this. Two commercial databases, Web of Science, owned by Clarivate Analytics, and Scopus, owned by Elsevier, function in most of the world as the operative definition of what research counts. Academic hiring decisions, promotion criteria, grant allocations, and university rankings across country after country depend on whether a scholar's work appears in journals these two databases have chosen to index. Their indexing decisions are, in this sense, epistemological decisions, though they are almost never described that way. What they include exists. What they do not include does not, at least not in any way that registers in the systems that determine careers and distribute funding.

In 2020, a research team led by scholars at the University of Ottawa examined 25,671 journals that had published more than 5.8 million research articles over the previous decade, representing 136 countries, with nearly 80 per cent of those journals based in the Global South and publishing across 60 languages. Web of Science indexed 1.2 per cent of them. Scopus indexed 5.7 per cent.

The remaining journals are not invisible because they are bad.

Many have not been evaluated at all.

This is worth pausing on. When scholars discuss colonial epistemology, what the Argentine semiotician Walter Mignolo has called the "colonial matrix of power" as it operates in knowledge production, the claim is not merely that colonial-era science was racist, or that individual researchers harbour biases, or even that certain fields have long ignored certain populations. The claim is structural. The infrastructure of modern knowledge, which gets indexed, what gets cited, what gets funded, what gets recognised as the kind of thing that counts as research in the first place, was built to reflect a particular geography of intellectual authority. Europe and North America are at the centre; everywhere else is at the margin, supplying data.

The same 2020 analysis found that despite Scopus indexing publications in 40 languages, English accounts for 93 per cent of its content. What this means, in practice, is that a researcher whose first language is Indonesian, or Arabic, or Amharic is writing into a system that will not count what they write as part of the authoritative record, regardless of quality. The language of science is not neutral. It is the language the indexers have decided to read.

The evidence for this structural bias does not stop at database statistics. In 2019, a systematic review published in BMJ Global Health examined authorship on infectious disease research conducted in Africa between 1980 and 2016. Less than half of the publications had an African researcher in either the first or last position, the two positions that, in academic convention, signify primary intellectual contribution. The research was done in Africa, the diseases studied were African diseases, and the people writing the papers were mostly elsewhere.

This practice has several names. Parachute science. Helicopter research. Safari study. A 2003 analysis by the Hungarian Academy of Sciences found that 70 percent of articles about least-developed countries in a random publication sample did not include a local research co-author. Not insufficient. Not underrepresented. Absent.

And this is not only an ethical complaint, though it is that too.

When researchers arrive in a place, collect data, and leave, they determine the questions. That determination is not neutral. A researcher from a London laboratory studying malaria in sub-Saharan Africa arrives with assumptions about what needs explaining, which variables matter, and what a satisfying answer would look like. A researcher who has spent a working life inside the problem brings different assumptions, not better by default, but often more accurate precisely because of what they know without being entirely able to articulate how they know it. Ignoring that knowledge is not scientific caution. It is a loss.

The open access movement was supposed to correct some of this. The idea was clean: remove the subscription paywalls that kept research locked behind institutional licences, and knowledge becomes available globally. Something like equity might follow. What happened instead is worth describing carefully.

Publishers moved from charging readers to charging authors. To publish open access in a reputable journal, the kind indexed in Web of Science or Scopus, a researcher typically pays what is called an article processing charge. A 2024 preprint examining health and surgical journals found the median charge sitting at $3,700. Research published the same year found that in 63 countries, this median fee exceeded 95 per cent of gross national income per capita. A comparison drawing on Brazilian and Canadian data found that publishing open access in a top critical-care journal costs the Brazilian researcher the equivalent of between five and seven times the monthly minimum wage; the same fee for a Canadian researcher amounts to roughly one and a half times that figure.

The paradox is exact. The reform designed to widen access to research made it harder for researchers in the poorest countries to participate in producing it.

Here is where the argument becomes more difficult to make, and where I want to be careful about what I actually think. The standard critique of colonial epistemology tends, at this point, to shift into a call for inclusion: more diverse editorial boards, more Global South authors in print, more equitable funding arrangements. These are not wrong. But they leave the deeper question sitting there, unasked, whether the structure itself, the peer-reviewed journal and the citation-count and the impact factor and the database index, is the right vehicle for the full range of human knowledge, or whether it is a vehicle built for particular kinds of knowledge that is simply not suited to carry others.

Boaventura de Sousa Santos, the Portuguese philosopher and sociologist, uses a term that carries more than its literal weight: epistemicide. He means the systematic erasure, through colonial and post-colonial processes, of entire ways of knowing that did not fit the European framework for valid inquiry. When colonial administrators dismissed indigenous agricultural practice as superstition, they were not only making a cultural judgment. They were, and this is the part that interests me, often wrong about the practical question. The agriculture was frequently worked. The dismissal did not make it less functional. It made it harder to protect, to transmit, and to build upon.

Research on neglected tropical diseases offers a concrete case for how this persists. According to a 2024 analysis of tropical disease research funding, organisations based in countries where these diseases are mostly non-existent, the United Kingdom, the United States, and continental Europe, received 75 per cent of direct research funding for these conditions between 2007 and 2022. The research was conducted largely on populations living with the diseases, using local logistical support and local bodies as the site of inquiry. The funding, the publications, the institutional capacity that those publications generate, all of it remained largely at the centre. The arrangement of 1666 is not a historical curiosity. It is a funding model.

None of this means that science produced in wealthy countries is wrong, or that the researchers involved are acting in bad faith. Most are not. The question colonial epistemology asks is not about individual intention. It is about what gets to count, who gets to say so, and what is lost when those decisions follow old grooves worn by older arrangements.

That is a much harder problem to solve with better indexing criteria.