The instrument · shows its working
The Ledger.
Every source this platform leans on, ranked by what you can check — peer review, journal standing, how the literature received it, and whether the funding is disclosed — never by the author's name or nation. Pick a question. See who says what, and how far each one can be trusted to hold.
Each source carries badges for the things that actually predict reliability: peer-reviewed vs preprint; the journal's standing; whether the citing literature contested it; and whether its funding is disclosed. The colour of the left edge is the platform's overall confidence. A famous name with a weak record ranks below an unknown one with a strong record. That is the whole point.
Choose a question
Why rank sources at all?
the argument behind the instrument
On contested history, the loudest claim usually has the least behind it, and the quietest, best-evidenced finding gets buried. The nationalisms with ministries behind them flood the space with confident narratives; an honest reader drowns. The Ledger is the antidote: it refuses to let confidence, fame, or volume stand in for reliability. It asks only the questions you could check yourself if you had the time — is it peer-reviewed, where did it appear, did other scholars contest it, who paid — and it shows the answers side by side. Judged by name or nation, sources become a shouting match. Judged by record, they sort themselves.
A source is never ranked up or down because of who the author is, where they were born, or which nation their finding flatters or offends. It is ranked on the record: review status, venue, reception, funding transparency. This is the same discipline the platform demands of every claim, turned back onto the scholars themselves. It is the one defence against motivated research that motivated research cannot fake — because faking it would mean disclosing the very things it needs to hide.
What a strong negative looks like.
and how this platform grades one
Most claims on this platform are about absence: a word that does not appear, a signal not detected, a survey not carried out. Absence is easy to assert and hard to evidence, so it needs a grading standard. The clearest available example comes from outside South Asia, and it is worth stating because it sets the bar.
In 2021 a large international team sequenced Bronze Age individuals from two basins in Xinjiang. In the southern basin they found no detectable Western Steppe ancestry — overturning three long-standing migration hypotheses at once. In the northern basin, a few hundred kilometres away, the same team using the same methods found Afanasievo-related steppe ancestry clearly.
A null result from a study that finds nothing anywhere is weak. A null result from a study that finds the signal next door is evidence. When this platform reports that a term is absent from a corpus, or a signal absent from a population, the question to ask is whether the same method detected the thing elsewhere. Where it did, the absence carries weight. Where it did not, the absence is only a detection limit — and is labelled as one.
The same study also separates two things this platform keeps distinct: that population was genetically isolated and culturally cosmopolitan at once, adopting dairying, foreign grains and irrigation from neighbours without absorbing their ancestry. Cultural adoption is not population replacement, and the two leave different traces.
Zhang et al. (2021), Nature 599: 256–261, 10.1038/s41586-021-04052-7. Max Planck (Jena and Leipzig), Jilin University, the Chinese Academy of Sciences, the Xinjiang Institute of Cultural Relics and Archaeology, Seoul National University and Harvard. Peer reviewers named in the publication; no competing interests declared. Not a Reich-lab output — an independent evidence stream from the ancient-DNA concentration noted above.
Every source, graded.
the register
The rules above are only worth stating if they are applied and the results published. Every source this platform relies on is graded on whether the scholar was audited, the venue peer-reviewed, and the evidence stream independent — with the largest group being those that are peer-reviewed but not yet checked at the author level, and that group's size published rather than hidden. The ranked source register →
Venue is not verification.
the scholar-level rule
A journal name is the easiest check and the weakest one. Nature on the masthead tells you a paper cleared peer review at a good venue. It does not tell you who wrote it, who funded it, what else they have argued, what criticism they carry — or, most importantly, whether the three papers you are treating as independent confirmation are actually the same laboratory publishing three times. This platform now requires the harder check on every load-bearing source: author, institution, laboratory, funding source, prior work, documented criticism, and shared authorship or shared samples across apparently separate studies.
This rule was not invented in the abstract. Applied to this platform's own genetics sources, it found a real defect in how they were being presented.
The David Reich Laboratory (Harvard Medical School / Broad Institute) had, as of 2020, produced the majority of all published ancient-DNA data in the world. Lazaridis, Patterson, Narasimhan, Moorjani and Chintalapati are all of that lab or its lineage. So when a claim about steppe ancestry is supported by "Narasimhan 2019, Lazaridis 2017 and Chintalapati 2022," that is not three independent confirmations. It is one evidence stream, with overlapping authors and overlapping samples, cited three times. This platform previously presented them that way. It no longer does.
This is not an accusation of bad work. The Reich lab's output is rigorous, peer-reviewed, and has genuinely transformed the field. The problem is structural: when one lab supplies most of the evidence, the field loses the capacity to check itself, and a reader cannot tell corroboration from repetition.
Two further flags, recorded because they bear on this platform's own argument: the lab has faced documented criticism — following a 2019 New York Times Magazine account — over monopolisation of ancient-DNA resources and under-representation of non-Western collaborators in publications. That is the same extractive dynamic this platform documents elsewhere, operating on the very evidence base the platform relies on. And its funding is politically exposed: the programme ran on an NIH grant renewed three times, and the 2025 cancellation of federal grants to Harvard put its continuation in doubt.
1. Independence is now a badge. Sources below carry INDEPENDENT or SHARED LAB so a reader can see at a glance whether two citations are really two sources.
2. Independent checks are named. Where the dominant lab makes a claim, this platform looks for a genuinely separate team. On the steppe question that is Damgaard et al. 2018 (Copenhagen), which partly contrasts with Reich-lab framing — finding no steppe ancestry in Bronze Age Anatolia and "limited direct genetic impact in Asia." A contrasting independent result is worth more than three concordant ones from the same source.
3. Genetics is not language. Steppe ancestry is not automatically Indo-European speech. Language shift can occur through elite dominance, trade or cultural transmission without genetic replacement — a caveat other scholars press, and one this platform states wherever ancestry and language are discussed together.