Falling through the gap

George C. Scott as General Buck Turgidson, famously opposed to gaps.

There is a hole. Right at the top of our science. In the introductions to our peer-reviewed papers, where we should explain the need for the new research results we are about to present, there is more and more often … an absence. An emptiness. A lacuna, even. Or, more conventionally: a gap.

I refer to a distressing trend — or at least, a thing I am noticing more and more in the papers I edit and review and read — in how my fellow researchers describe what we do. In introductions and abstracts from the pages of Science and Nature to graduate theses, authors explain that their motivation for doing scientific research is to fill “knowledge gaps.”

The Oxford English Dictionary record for the phrase “knowledge gap” has examples back to 1947 — but in the sense of “a disparity in levels of (esp. technological) knowledge”, not in the sense of “a thing no one (yet) knows”. A brief dig through Web of Science quickly reveals the “disparity” usage in social science contexts. However, restricting the search evolution and ecology and adjacent fields reveals something suspiciously like the trend I think I’ve noticed. With each passing year, more and more abstracts contain the phrase “knowledge gap.”

Knowledge gaps are proliferating, it appears. (Data from Web of Science)

Let’s look at how the phrase is deployed in the examples I linked above from the two most prestigious journals on the planet. (I trust any embarrassment this may cause the authors is offset by the fact that they attracted my cranky attention by landing papers in Science and Nature.)

Burin et al. (2021, Science) demonstrate that species of fruit-eating birds in lineages with longer evolutionary histories tend to be more centrally connected in the networks of seed-dispersers and the plants they help. Their “gap” appears in a general-audience summary of the work at the top of the online version of the article — it’s not clear who wrote it, because it talks about the authors in the third person, but it doesn’t have a byline.

The structures of many such networks and the ecological forces that shape them are well known, but their deeper evolutionary history has received little attention. Burin et al. address this knowledge gap in a study of frugivorious bird species in documented seed-dispersal networks around the world …

The “gap” here is that Burin et al. examined “deeper evolutionary history” in ecological networks, which “has received little attention.” It’s a claim of novelty, but a tentative one. It’s not that no one has paired phylogenetics with ecological networks, it’s just that it hasn’t been done much. The claim is, thus, defensible against someone debunking its priority by finding an older paper that uses phylogenies and network data, but it doesn’t really say what’s novel about the work. So, okay, this hasn’t been done much before … but why didn’t the prior examples come up with this result?

On to Cunha et al. (2022, Nature), who present experimental evidence that primary productivity in much of the Amazon rainforest forest is limited by the availability of phosphorus. They find a “gap” at the end of their first paragraph:

Thus, directly determining the extent to which soil fertility controls tropical forest growth and the elements that are most important, remains a key knowledge gap, and addressing this is critical for understanding forest growth dynamics and predicting responses to CO2 fertilization.

Up to this point Cunha et al. have explained that indirect evidence points to phosphorus as a critical limiting nutrient for Amazonian forest productivity, but these observations are confounded by covariance of multiple factors that could drive productivity — the sort of issue best resolved with a manipulative experiment. The “gap” is the point at which interpretation of correlations in nature becomes uninformative, and the claim to novelty seems straightforward: though the next paragraph lists nutrient-manipulation experiments in other tropical forests, it also notes there is no precedent for a “large-scale nitrogen, phosphorus, and base cation experiment … in Amazonia.” To call something a “knowledge gap” in this sense is to violate the storytelling rule of “show, don’t tell” as it applies to a research paper — you shouldn’t simply assert that there’s a need for your work, but actually explain what the need is and why the need exists.

Why do these cases, and the others I’ve seen in my reading, irritate me? Partly it’s because the “knowledge gap” framing hedges and diminishes the value of the authors’ work. Scientists who fill a “knowledge gap” are not claiming that their work advances human understanding beyond what we knew before, even if incrementally. Instead, they’re describing their work as slotting into a space surrounded by existing knowledge. It’s not extending the road of science towards the horizon — it’s patching a pothole.

There is safety in pothole-patching. As in the Burin et al. non-technical summary, it can make a claim of novelty so narrow and slight that it’s effectively unassailable. Yet it also weakens the justification for the project: “I haven’t see a lot of other people doing this” isn’t the worst reason to try something in science, but it’s hardly the best. That’s a pity, because the actual Introduction of the paper lays out the reason to link phylogenies with present-day ecology in a couple of efficient paragraphs: because we want to know whether species interactions over as short a period as a single growing season are shaped by the multi-million-year history of the species interacting.

In Cunha et al., we see another possible explanation: it’s just a way people are talking about scientific questions now, much as some of us try to frame work as testing specifically stated hypotheses, or as answering “research questions”. This, too, strikes me as a pity. I can imagine, I suppose, a lot of well-meaning mentors teaching students to identify a “knowledge gap” that their research will fill — maybe that framing eases the stress of coming up with a novel project in a field you’re still learning. But carrying it over into the explanation for your work that you present to the world is another thing altogether.

I was brought up in science to try to identify the new thing I would learn with each project I took on. Even if it’s as small and specific as figuring out whether the community of weeds in city park lawns varies across greater Los Angeles, or trying to see how far a particular pollinating moth carries pollen — both questions students in my lab have pursued — if the project succeeds, it’s added something new to the world’s knowledge. It feels ignoble, for lack of a better word, to describe the work of creating knowledge as filling in a hole, like spackle.

So this is my complaint about describing research in terms of a “knowledge gap”: at its least offensive, it’s an empty cliché, telling me nothing that won’t be explained in detail in a good paper’s abstract or introduction. And in the worst-case usage, to describe a project’s goal as filling a “knowledge gap” is to undercut the value of the very work you’re trying to justify. I am, perhaps, fighting a rising tide of “gaps” in the world’s knowledge, but I’m going to make the plea all the same. Scientists, for the love of science, forget about the gap.

About Jeremy Yoder

Jeremy B. Yoder is an Associate Professor of Biology at California State University Northridge, studying the evolution and coevolution of interacting species, especially mutualists. He is a collaborator with the Joshua Tree Genome Project and the Queer in STEM study of LGBTQ experiences in scientific careers. He has written for the website of Scientific American, the LA Review of Books, the Chronicle of Higher Education, The Awl, and Slate.
This entry was posted in community, journal club, modest proposals, Science Communication, science publishing and tagged , , , . Bookmark the permalink.