Umbrella-Word Inflation in Research-Project Documents: A Computational Text Analysis of diversity, sustainable, multi- and collaboration
This computational text analysis of research-project documents from major funding agencies reveals that the rapid inflation of vague "umbrella words" like "diversity" and "sustainable" is driven by funder mandates and political shifts rather than scientific necessity, leading to reduced lexical specificity and a semantic drift from concrete objects to abstract goals.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the world of scientific research, getting a project off the ground usually starts with a document. Whether it is a grant application for a new medical study or a proposal for a global climate initiative, these texts are written to persuade funding agencies to invest money. Over the last few decades, a specific set of words has become ubiquitous in these documents. Terms like diversity, sustainable, collaboration, and innovation have appeared with increasing frequency, acting as a kind of shared vocabulary that signals a project is modern, ethical, and aligned with current values. Researchers and policymakers often assume these words reflect a genuine shift in how science is done, suggesting that projects are becoming more inclusive, more cooperative, and more focused on real-world impact. However, there is a lingering question about whether these words are simply describing reality or if they have become empty slogans used to win funding without necessarily changing the work underneath.
A researcher at Keio University set out to measure this phenomenon with the precision of a census taker. Instead of reading a few hundred documents by hand, the study analyzed approximately 1.84 million research project documents and journal articles from the United States and Europe, spanning from the 1980s to 2025. The goal was to track how often these "umbrella words"—broad, positive terms that can apply to almost anything—appeared in the text, and to see if their rise came with a cost. The investigation looked for signs that these words were losing their specific meaning, becoming mere decorations, or that their usage was driven more by the changing moods of funders and politicians than by the actual science being proposed.
The results revealed a dramatic surge in the use of these words. In grant applications from the US National Science Foundation, the share of documents mentioning diversity jumped from just 3.5% in the late 1980s to over 26% in the early 2020s. The word sustainable saw an even more explosive rise, appearing in less than one-tenth of one percent of documents in the late 80s and climbing to over 10% by the mid-2020s. This growth was far faster than the increase in standard, neutral words like experiment or method. The study found that these umbrella words had spread across almost every field of science, moving from their original homes in specific disciplines to become common in biology, engineering, and social sciences alike. This widespread adoption is exactly what one would expect if the words were acting as useful bridges between different communities, helping diverse groups find common ground.
However, the study also uncovered a more troubling side to this inflation. In the US National Science Foundation applications, the word diversity underwent a significant shift in meaning. Decades ago, when researchers used the word, they were almost always talking about biological diversity, such as the variety of species or genetic traits. By the 2020s, the word had largely lost that biological anchor. Instead, it was increasingly paired with verbs like increase, promote, or broaden, turning the concept into a promise to achieve a goal rather than a description of a specific object of study. The researchers confirmed this shift using computer models that analyzed how words relate to one another, showing that the word had drifted away from its scientific roots and toward institutional promises. In contrast, in published journal articles, the word diversity remained firmly tied to biology, suggesting that the change was specific to the language of funding applications.
The investigation also looked at whether these words reflected the actual structure of the research projects. The study found that claims of collaboration and interdisciplinarity were often disconnected from reality. In European Union projects, for instance, a significant number of applications from single researchers who had no partners still used words like collaboration and partnership. Conversely, many large groups of researchers working together did not use these words at all. While the use of such vocabulary did rise slightly with the size of the team, the link was loose. The words appeared to be a standard part of the application template rather than a precise description of who was working together. Furthermore, the study found that documents packed with these umbrella words were slightly less specific in the rest of their text, suggesting that a heavy reliance on broad, positive terms might crowd out concrete details.
Perhaps the most striking evidence of the fragility of this vocabulary came from a sudden political shift in 2025. Following executive orders in the United States regarding diversity and inclusion, the use of related words in National Science Foundation abstracts plummeted. Within a single year, the frequency of words like diversity, equity, and inclusion dropped by more than half, while the actual topics of the research awards remained largely unchanged. This rapid reversal demonstrated that the vocabulary was highly sensitive to political pressure and could be removed almost as quickly as it had been added. The study also noted that since 2023, documents containing words typical of artificial intelligence writing had begun to carry even more of these umbrella terms, hinting that automated tools might be accelerating the trend.
Ultimately, the research suggests that while these words have become a dominant feature of the research landscape, they do not necessarily signal a deeper change in how science is conducted. The growth of terms like diversity and sustainability is real, but it is often a reflection of what funders want to hear rather than a direct measure of what researchers are doing. The study concludes that the clearest cost of this inflation is the dependence on outside actors; because the vocabulary is so tied to the priorities of funding agencies and the political climate, it can be withdrawn just as easily as it is adopted. For the scientific community, this means that the presence of these words in a proposal is no longer a reliable signal of quality or genuine commitment, and that the true value of a project lies in its specific practices and outcomes, not in the buzzwords used to describe it.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.