“Little research exists on this.” The sentence sits in tens of thousands of introductions, review reports and grant applications. It makes a claim about the state of knowledge. It rests on a list of hits somebody pulled before coffee.
What the database has not heard of does not enter the hit list. What is not in the hit list does not enter the literature review. What is missing there is called unstudied two pages later. A technical limit of reach becomes a scientific statement of fact, and nobody lied along the way.
What the registries are, and are not
Crossref is a registration service. It holds the metadata its members deposited for their DOIs, and it searches no full texts. What was never registered there, it does not know. 1
PubMed is a window, not a building: biomedical citations and abstracts, with the full texts held elsewhere. OpenAlex is broader and more open than either, drawing Crossref, PubMed, DataCite, arXiv and further tributaries into one index. That is precisely why it inherits their gaps. An aggregator joins rivers. It does not invent water. 2
None of these systems is the scholarly record. All three are read as one every day.
The language nobody counted
It can be counted. A survey across sixteen languages found that for a single year, 35.6 percent of scientific documents on biodiversity conservation were not published in English. 3
Five years later the same group screened 419,679 peer-reviewed papers in sixteen languages. It found 1,234 non-English studies carrying usable evidence on the effectiveness of conservation interventions. The English-language count was 4,412. More than a fifth of the evidence sat in languages the standard apparatus never queries. Those who passed it by called the remainder the state of the field. 4
The filter nobody notices
Queries to the registries like to carry a small field: type:article. It keeps out paratext, publisher forewords, erratum notices, book reviews. It also keeps out dissertations. Habilitations. Monographs. Research reports. Conference papers.
A field is then not thinly researched. It is thinly queried. Admit journal articles only, and you will not find a research gap. You will build one.
Publication type is context, not a title of nobility. An article can be methodologically thin, a dissertation can hold four documented years, a report can hold the only primary data in existence. Quality is judged on method, data and inspectability, not on a metadata field reading article.
Closed is not empty
Part of the world's research sits in systems that require an account, a licence or a contract: eLIBRARY.RU, CNKI, Wanfang, alongside national catalogues and repositories for theses. These holdings are not openly queryable, and nobody is going to scrape them quietly. That would not be scholarly ambition. It would be trespass with a press release.
A catalogue entry proves that a work exists. It does not prove what is in it. Turn that into “no research” and you have claimed more than the sources allow. Turn it into “proven effect” and you have done the same thing facing the other way.
Five states that must never collapse into one: read and checked; bibliographically established but sealed; plausibly attested through citations and catalogues; registry outage; and the genuine null result. A timeout looks on screen exactly like scientific emptiness. It is not.
What we are building from it
NotaVera is getting a world map. For a given subject it will show where the research happens, since when, in which languages, in which publication types, which part surfaces in the dominant registries and which part demonstrably exists without being accessible. What becomes visible is the difference between a standard search and an honest one.
It hands out no grades to countries. No bonus for being politically misread. No penalty for being well represented in a Western registry. The map shows differences. The verdict stays with the sources.
A sealed archive is not an empty one. A visible title is not proof. The difference is the entire profession.
- Crossref. “Metadata retrieval.” Accessed August 6, 2026. https://www.crossref.org/documentation/retrieve-metadata/
- Priem, J., Piwowar, H., & Orr, R. (2022). OpenAlex: A fully-open index of scholarly works, authors, venues, institutions, and concepts. arXiv:2205.01833. https://arxiv.org/abs/2205.01833. On PubMed: National Library of Medicine, “PubMed User Guide,” accessed August 6, 2026, https://pubmed.ncbi.nlm.nih.gov/help/
- Amano, T., González-Varo, J. P., & Sutherland, W. J. (2016). Languages Are Still a Major Barrier to Global Science. PLOS Biology, 14(12), e2000933. https://doi.org/10.1371/journal.pbio.2000933
- Amano, T., Berdejo-Espinola, V., Christie, A. P., et al. (2021). Tapping into non-English-language science for the conservation of global biodiversity. PLOS Biology, 19(10), e3001296. https://doi.org/10.1371/journal.pbio.3001296