Method page · attribution
Not to whoever thought of it. The rule is mechanical, it operates in every direction, and it is the single most common error in popular history of ideas — including on earlier drafts of this platform.
Earliest surviving attestation is treated as origin. A compiler becomes an author. A translator becomes a discoverer. A transmitting culture becomes a source.
It is not usually deliberate fraud. It is a preservation bias with a political afterlife: whoever's records survived gets the credit, and whoever's records were oral, undeciphered or destroyed gets nothing — after which the outcome is retold as a difference in achievement.
The three stages, with what each one drops.
Green = source · gold = carrier · red = who gets the credit
The zodiac
MUL.APIN is not by anyone. It is a scribal compilation, in Akkadian, using Sumerian logograms for the star names — MUL star, APIN plough. Four layers, and the credit lands on the fourth. Even the Greek astrological texts attributed themselves to Egyptians — Nechepso, Petosiris, Hermes Trismegistus. Only later readers insisted the material was Greek.
The numerals you are reading this in
The clearest case in existence, because everyone in the chain was honest. Al-Khwārizmī titled his book on the Indian method of calculation; Latin Europe called the result algorismus after him. Nobody claimed to have invented it. The name still records the carrier and not the source — and “Hindu-Arabic” is a modern correction that only got halfway.
The Fibonacci sequence
Piṅgala's Chandaḥśāstra gives the binary enumeration of metres and the Meru-prastāra. The recurrence itself — each term the sum of the two before it — is stated by Virahāṅka, roughly eight centuries later, then by Gopāla and Hemacandra. Credit drifted backwards to the more famous earlier name, which is the same mechanism that moved it forwards to Fibonacci.
And the first box is the real one. The sequence arises from counting arrangements of short and long syllables — which poets were doing before anyone wrote a sūtra about it. Virahāṅka codified a practice; he did not invent the pattern. That practitioner has no name and never will.
Infinite series
Series expansions for π, sine and cosine, two to three centuries before Europe. Whether the results were transmitted is argued; that they were obtained first is not.
Chess
Here the transmission is uncontested and the naming still records the middle of the chain, not the start.
| Called | Actually from |
|---|---|
| Arabic numerals | India, via Arabic |
| Damascus steel | wootz crucible steel from South India and Sri Lanka, traded through Damascus |
| The Pythagorean theorem | Babylonian tablets and the Indian Śulbasūtras both state it before Pythagoras |
| The Greek alphabet | Phoenician, letter names and order intact |
| Roman numerals | Etruscan |
| Turkey (the bird) | the Americas, via merchants trading through Turkey |
A name is a record of the last leg of the journey.
Which is what makes it an argument rather than a grievance
| Credited | Prior or joint |
|---|---|
| Harvey and Servetus, pulmonary circulation | Ibn al-Nafīs, Damascus, 13th c. |
| Newton and Descartes, optics | Ibn al-Haytham, 11th c., whose Book of Optics established the experimental method for the field |
| Watson and Crick | Rosalind Franklin, whose diffraction data was shown to them without her knowledge |
| Hewish, Nobel Prize 1974 | Jocelyn Bell Burnell, who discovered the pulsar |
| Russell and Eddington on stellar composition | Cecilia Payne, whose thesis was talked down before being confirmed |
Attribution failure is a general property of how credit propagates, not a special vice of Europeans. It follows power, proximity to the archive, and institutional standing. Colonialism did not invent it; it industrialised it and gave it a civilisational story.
1 · Preservation asymmetry. Clay survives; birchbark, palm leaf and papyrus do not. India's oldest manuscripts are the Gāndhārī scrolls because Gandhāra is dry, not because nothing was written earlier. The Indian textual record looks late for climatic reasons.
2 · Decipherment asymmetry. We can read Akkadian and Greek. We cannot read Indus or Linear A. Both silences are read as absence, and only one of them is ever described that way out loud.
3 · The oral tradition penalty. A culture that transmits by memory for a thousand years leaves no dated artefact, so its antiquity is “reconstructed” while a written one is “documented.” The Rigveda was preserved with tone and metre intact for millennia and still counts as weaker evidence than a tax receipt.
4 · The compiler is legible. A synthesiser writes one findable document. The tradition behind it is diffuse and anonymous. Citation goes to whatever can be cited.
“Earliest attested” is the only claim a documentary record supports. Any origin claim needs either a continuous decipherable record or an argument that survives the absence of one. If a claim cannot be stated in earliest-attested form, it should not be made.
A developmental sequence is different from a first attestation, and stronger. Babylonian tablets show the zodiac being built in stages — MUL.APIN's eighteen unequal lunar constellations pruned toward twelve equal signs across centuries. That is evidence of local construction, not merely of first writing-down. It is still not proof that nobody elsewhere had something comparable earlier.
And undeciphered is not negative. The Indus script and Linear A are unknown in both directions. Neither may be cited as evidence of absence.
This platform has broken all three of those rules and published the corrections. The list is on the index page. That is the only reason the rules are worth stating here.