Aryabhata Said the Earth Spins in 499 CE. His Day Was 0.01 Seconds Off. India Spent 900 Years Un-Saying It

Kusumapura, Kali year 3600: a young man signs his book

The year is 3600 of the Kali age, which the Aryabhatiya counts from a sunrise at Lanka in 3102 BCE (other schools used the preceding midnight). In our reckoning it is 499 CE. The place is Kusumapura, "the city of flowers", which the commentator Bhaskara I identifies as Pataliputra in Magadha — the city under modern Patna, where the Mauryan emperors had held court seven hundred years before.

A man there is finishing a short book. He gives neither birthplace nor teacher, only his age. In the third chapter, verse 10, he writes that when "sixty times sixty years and three quarter-yugas" of the current age had passed, "twenty-three years of my life had passed". The editors of the 1976 critical edition work the date: the 3,600 Kali years ended at mean noon on Sunday, 21 March 499 CE, as the mean Sun entered Aries. He was twenty-three. So he was born in 476.

His name is Aryabhata. His book is the Aryabhatiya, "the composition of Aryabhata": in the words of its first complete English translator, "the earliest preserved Indian mathematical and astronomical text bearing the name of an individual author". Four sections, 121 verses. Pi as 3.1416, a sine table, equations in whole numbers, eclipse rules — and, packed into a verse written in a number-code of his own invention, the claim this Investigation is about.

The first section lists how many times each body in the sky goes round in a yuga, a great age of 4,320,000 years. Sun: 4,320,000. Moon: 57,753,336. Then a line no astronomer before him is known to have written: ku — the Earth — 1,582,237,500, prak — eastward.

The Earth is the thing you stand on. Why would it have a number of revolutions, like a planet? And why eastward?

He answers in the last chapter. "As a man in a boat going forward sees a stationary object moving backward," he writes, "just so at Lanka a man sees the stationary asterisms moving backward (westward) in a straight line."

That is the whole case. A boat. A riverbank that seems to slide past. A sky that only appears to turn. The rest of this Investigation is what happened to those sentences over the next fourteen hundred years.

At a glance

- 499 CE, age 23. Aryabhata dates his own book: "When sixty times sixty years and three quarter-yugas had elapsed (of the current yuga), twenty-three years had then passed since my birth". (Aryabhatiya III.10; Shukla–Sarma 1976, p. xix; Clark 1930, p. 54) - 1,582,237,500 eastward rotations of the Earth in a yuga of 4,320,000 years. (Aryabhatiya I.3 — Clark's "stanza 1"; Clark 1930, p. 9; Shukla–Sarma 1976, p. 6) - 23 h 56 min 4.1 s. The Earth-rotation implied by that number. The modern conventional sidereal day is 86,164.0905 s = 23 h 56 min 4.0905 s. Difference: about one-hundredth of a second. (our arithmetic below; IERS Earth-rotation constants; Dutta, Resonance 2006, part 2, pp. 65–66) - 628 CE, Brahmagupta's objection: "if it rotates (at the same place), why do tall lofty objects not fall down?" (Brahmasphutasiddhanta XI.17, as translated in Shukla–Sarma 1976, p. 15; Sanskrit in Clark 1930, p. 14) - The Earth stays in the middle. "The sphere of the Earth, being quite round, situated in the center of space, in the middle of the circle of asterisms …" He is not a heliocentrist. (Aryabhatiya IV.6; Clark 1930, p. 64) ## First, the words a beginner needs

Siddhanta. A Sanskrit treatise of mathematical astronomy: how to compute where the Sun, Moon and planets will be, and when eclipses fall. The Aryabhatiya is the oldest that survives with its author's name on it.

Yuga. A great age of 4,320,000 years: a very long clock-face on which every body in the sky completes a whole number of circuits, so no fractions are needed. "The Moon goes round 57,753,336 times in a yuga" says "27.32167 days per circuit" without decimals, which did not yet exist.

Revolution (bhagana). One full circuit. For the Moon, one trip round the sky. For the Earth, in Aryabhata's usage, one full turn on its own axis.

Asterisms (bha, naksatra). The fixed stars — "fixed" because, unlike the planets, they keep their pattern. Aryabhata's claim is that they do not move at all; we do.

Apparent motion. Motion you see that is not there, caused by your own movement: trees "moving" past a train window.

Sidereal day. The time for the Earth to turn once relative to the fixed stars: 23 h 56 min 4.09 s. It is shorter than the 24-hour civil day (sunrise to sunrise) because, while turning, the Earth also moves a little along its orbit, so it must turn slightly more than once to bring the Sun back overhead. Over a year the extra bits add up to one extra turn: 366.25 rotations in a year of 365.25 days.

Lanka. Not Sri Lanka. In Indian astronomy, the point on the equator due south of Ujjain — a reference point in the Indian Ocean, chosen because an observer there sees the stars rise straight up and set straight down.

Pravaha, "the provector wind". The great wind that, in the standard Indian model, carries the whole sky round the Earth once a day. Aryabhata mentions it too — which is a problem.

Commentary. A line-by-line explanation of an older text by a later scholar. The Aryabhatiya has at least a dozen in Sanskrit alone, from Bhaskara I in 629 to Ghatigopa in the nineteenth century. That is where the argument happened.

The claim we are testing

"Aryabhata discovered that the Earth rotates on its axis, a thousand years before Copernicus." Layer by layer:

1. The text states, in two chapters and in a number, that the Earth rotates eastward once a day, and explains the sky's motion as apparent. True, from the surviving text, from Brahmagupta's hostile quotation in 628, and from the 1976 critical edition.
2. The day-length implied by his number matches the modern sidereal day to a hundredth of a second. True, by arithmetic anyone can repeat — with the catch that his year, from the same numbers, is 3 minutes 20 seconds too long.
3. He was a heliocentrist, "a thousand years before Copernicus". Myth. He kept the Earth at the centre, in writing.
4. India accepted the discovery. Myth. Rejected by Varahamihira, attacked by Brahmagupta, dropped by his admirer Bhaskara I, explained away for nine hundred years. One commentator calls it "the false view".
5. The verse is unambiguous. Mostly. The next verse mentions the sky driven west by a wind, and the 1930 translator admits "it is difficult to see how the two stanzas can be brought into agreement".

The number first: what 1,582,237,500 actually says

Aryabhata hides his numbers in a code: each consonant is a digit, each vowel a power of ten, so a word spells an integer. The Earth's count is the five-syllable nisibunlkhr: 1,582,237,500. Clark's translation of the verse:

"In a yuga the revolutions of the Sun are 4,320,000, of the Moon 57,753,336, of the Earth eastward 1,582,237,500, of Saturn 146,564, of Jupiter 364,224, of Mars 2,296,824, of Mercury and Venus the same as those of the Sun." — Clark 1930, p. 9

Now the arithmetic a sixteen-year-old can do.

In a yuga the Earth turns 1,582,237,500 times relative to the stars. But sunrises are fewer than star-turns: every year the Earth goes once round the Sun, and that costs exactly one sunrise, because the Sun drifts a little eastward against the stars each day and the drifts add up to one full turn a year. So:

Civil days in a yuga = rotations − years = 1,582,237,500 − 4,320,000 = 1,577,917,500.

Aryabhata's verse III.5 defines both: "The conjunctions of the Sun and Earth are (civil) days. The rotations of the Earth are sidereal days", and the 1976 edition's note tabulates the two totals from his own rules — civil days 1,57,79,17,500, sidereal days 1,58,22,37,500 (Shukla–Sarma 1976, III.5 and note).

If 1,577,917,500 civil days of 24 hours contain 1,582,237,500 rotations, one rotation lasts

(1,577,917,500 ÷ 1,582,237,500) × 24 hours = 23 hours 56 minutes 4.10 seconds.

The modern conventional sidereal day, from the table of constants kept by the International Earth Rotation and Reference Systems Service (IERS), is 86,164.0905 seconds: 23 h 56 min 4.0905 s. Aryabhata's figure is 86,164.1012 seconds. The gap is 0.011 seconds a day — four seconds a year. The 1976 editors: "The accuracy of Aryabhata I's value is remarkable."

The same numbers give his year: 1,577,917,500 ÷ 4,320,000 = 365.25868 days, exactly 365 days 6 hours 12 minutes 30 seconds. The modern sidereal year is 365.25636 days. He is 3 minutes 20 seconds too long — worse than his day. We come back to this.

Verse I.6 says the same thing as a rate: "The Earth rotates through (an angle of) one minute of arc in one respiration" — a four-second breath; 21,600 minutes of arc × 4 s = one day (Shukla–Sarma, p. 13).

The boat: Gola 9, and the verse right after it

The fourth section is Gola, "The Sphere". It puts the Earth as a ball in the middle of a larger ball of stars, then explains what an observer sees. Verse 9:

"As a man in a boat going forward sees a stationary object moving backward just so at Lanka a man sees the stationary asterisms moving backward (westward) in a straight line." — Clark 1930, p. 64

The 1976 edition: "Just as a man in a boat moving forward sees the stationary objects (on either side of the river) as moving backward, just so are the stationary stars seen by people at Lanka (on the equator), as moving exactly towards the west" (Shukla–Sarma, p. 119).

The Sanskrit is anulomagatir nausthah pasyaty acalam vilomagam yadvat / acalani bhani tadvat samapascimagani lankayam. The key word is acala, "not moving", used twice — once for the riverbank, once for the stars. The stars do not move. What moves is the boat, and the boat is the Earth. Why Lanka? At the equator the stars rise straight up and set straight down, so the "backward motion" is a clean line, sama-pascima, "exactly westward" — the one place where the simile is geometrically perfect.

Then comes verse 10. Clark:

"The cause of their rising and setting is due to the fact that the circle of the asterisms, together with the planets, driven by the provector wind, constantly moves straight westward at Lanka." — Clark 1930, p. 66

Read cold, that says the opposite of verse 9: the sky is driven west by a wind. Clark does not hide it: "If the readings of our text are correct it is difficult to see how the two stanzas can be brought into agreement. The ninth stanza states unequivocally that the asterisms are stationary and implies the rotation of the Earth" (p. 67). He rejects the idea that verse 10 was inserted later; Brahmagupta already knew these sections as "The Hundred and Eight Stanzas", and 108 is what we have.

The 1976 editors resolve it in translation: "(It so appears as if) the entire structure of the asterisms together with the planets were moving exactly towards the west of Lanka, being constantly driven by the provector wind" (p. 119) — verse 10 describes the appearance verse 9 has just explained. Clark allows the move but calls it "a doubtful expedient", with no word in the verse to signal it.

Our view: the consistent reading is the one that fits the rest of the book — the rotation number in I.3, the rate in I.6, the sidereal day defined in III.5 as "the rotations of the Earth". Three plain statements against one ambiguous verse. But a reader who calls verse 10 a loose end has Clark on their side.

What Aryabhata did not say

It is tempting to make him Copernicus. He is not. Three verses before the boat:

"The sphere of the Earth, being quite round, situated in the center of space, in the middle of the circle of asterisms, surrounded by the orbits of the planets, consists of water, earth, fire, and air." — Aryabhatiya IV.6; Clark 1930, p. 64

In III.15 he calls the Earth khamadhyastha, "standing in the middle of space". Clark, answering a scholar who wondered whether Aryabhata meant the Earth's motion round the Sun, is blunt: "The latter is out of the question" (p. 66, n. 3). The Sun, Moon and planets go round the Earth; the stars sit still; the Earth turns on the spot. That is his universe. He gives no cause for the turning, no force, no mechanics. When Brahmagupta asked, there was nothing in the text to reply with.

628 CE: Brahmagupta asks about the tall things

Brahmagupta wrote the Brahmasphutasiddhanta in 628 CE, "just 129 years after the Aryabhatiya", as Clark counts it. Its eleventh chapter, the Tantrapariksa, "Examination of Treatises", is in Clark's words "chiefly devoted to criticism of Aryabhata", whom he names about sixty times.

Verse XI.17 is the one that matters. Clark prints the Sanskrit (p. 14): pranenaiti kalam bhur yadi tarhi kuto vrajet kam adhvanam / avarttanam urvyas cen na patanti samucchrayah kasmat. The 1976 INSA edition translates it:

"If the Earth moves (revolves) through one minute of arc in one respiration, from where does it start its motion and where does it go? And, if it rotates (at the same place), why do tall lofty objects not fall down?" — Brahmasphutasiddhanta XI.17, in Shukla–Sarma 1976, p. 15

Two objections in one verse. First: if the Earth "moves", where is it going? Second, the one everyone has thought of since: if the ground is spinning, why does a tower not topple?

Dutta's 2006 Resonance series paraphrases the argument in modern units, following S. R. Das (1926): a heavy body dropped from a mountain-top takes, say, thirty seconds to fall; the Earth, turning once in 24 hours, moves eight miles in those seconds; why does the body land at the mountain's foot? It is a good question. Copernicus could not fully answer it either. The answer needed Galileo's idea that a falling body keeps the sideways motion of the ground it left — a thousand years away.

Brahmagupta's verse is also, ironically, the best evidence that Aryabhata's text said what it says.

The syllable that was changed

The last word of Aryabhata's rotation rate (I.6) is, in Brahmagupta's quotation, bhuh — "the Earth". In the text Paramesvara used, printed by Kern, it reads kham — "the sky"; the 1976 editors report the commentators' reading as bham — "the asterisms". Either way, one syllable. Change it and the verse says the sky turns one minute of arc per breath — the orthodox view, with nothing to object to.

Clark: "The reading kham of our text must be incorrect. It is a reading adopted by Paramesvara who was determined to prove that Aryabhata did not teach the rotation of the Earth … The true reading is bhuh, as is proved conclusively by the quotation of Brahmagupta (XI,17)" (p. 14).

The 1976 edition agrees and adds two witnesses: "The correct reading is bhuh, which has been mentioned by Brahmagupta (A.D. 628), Prthudaka (A.D. 860) and Udayadivakara (A.D. 1073)" (p. 15). The same happened in III.5, where the commentators read "revolutions of the stars" and two of them record the alternative "revolutions of the Earth" — "evidently, the correct reading".

So: Aryabhata writes bhuh. Brahmagupta, attacking him, quotes bhuh. Later copyists who revere Aryabhata but cannot stomach a spinning Earth transmit kham. The hostile witness preserves the truth; the friendly ones edit it.

Nine hundred years of ingenuity

Some of the finest minds in Indian astronomy, working to make two plain sentences mean their opposite:

Varahamihira (died 587) rejected it; the 1976 editors say the theory "received a severe blow at the hands of Varahamihira (d. A.D. 587) and Brahmagupta (A.D. 628)" (p. 8). We did not open his verse and say so below.

Bhaskara I (629), the earliest commentator and a devoted follower, simply left it out. Dutta: "he omitted Aryabhata's phrases (in Gitika 3 and 6) pertaining to Earth's rotations". Loyalty by silence.

Lalla (eighth century), more than once called Aryabhata's disciple, attacked it with birds and arrows, in his Sisyadhivrddhida (20.42) as quoted by Dutta: "how do birds flying in the sky return to their nests … why do not arrows, shot towards the sky, fall westward?"

Prthudaka (c. 860) is the surprising exception: a commentator on Brahmagupta, the chief critic, who sides with Aryabhata. Dutta translates his verse: "The sphere of the stars is stationary. It is only the Earth that is regularly rotating once in a day, and thereby causing the daily rising and setting of the stars and the planets." He thought Aryabhata's own followers "were afraid of the public opinion" (Shukla–Sarma, p. 120).

Somesvara did the twisting in the open, and the 1976 edition prints him (pp. 119–120). First he explains the boat correctly: "it is the Earth that moves towards the east; the stars are fixed." Then: "This is the false view. For, if the Earth had a motion, the world would have been inundated by the oceans, the tops of the trees and castles would have disappeared … and the birds etc. flying in the sky would never have returned to their nests … Hence this stanza must be interpreted in another way." The other way: "the asterisms driven by the provector wind, due to their own motion, see the objects at Lanka as moving in the opposite direction, i.e., they see the stationary Earth lying below as if it were rotating." The stars look down and see the Earth seem to turn. The man in the boat has become a star.

Paramesvara (born c. 1380, writing 1408–1443) of Kerala, whose commentary Kern printed in 1874, used the same device and named it: mithyajnana, "false knowledge". The boat describes a state of mind, not of the world. Clark: Paramesvara "tries to prove that here and in IV,9 (which he quotes) Aryabhata does not really mean to say that the Earth rotates. His effort to bring Aryabhata into agreement with the views of most other Indian astronomers seems to be misguided ingenuity" (p. 9). On the boat he "is forced to take the words anuloma and viloma, which regularly mean 'ahead,' 'eastward,' and 'backward,' 'westward,' in exactly the opposite senses" (p. 65).

Kern had seen it in 1874. His preface calls the commentary "generally speaking, a useful commentary", then adds: "some of his explanations, e.g. at st. 9 of the 4th section, can not be admitted to be correct" (Kern 1874, Preface p. vi). Stanza 9 of the fourth section is the boat.

How it reached Europe: Arabic, Malayalam, Leiden, Chicago

The book travelled. Around 800 CE it went into Arabic: the Biographical Encyclopedia of Astronomers states that the "Aryabhatiya was translated into Arabic around 800 as the Zij al-Arjabhar". The 1976 editors name the translator as Abul Hasan al-Ahwazi and record a small comedy: "The Arabians misunderstood the exact significance of the title of the work and wrongly thought that it meant 'one thousandth part'" (p. lxii). Al-Biruni, in the eleventh century, repeatedly calls him "Aryabhata of Kusumapura".

In India the text survived in manuscripts, mostly southern, bundled with commentaries. Hendrik Kern at Leiden printed the first edition in 1874, and his preface says exactly what he had: "two manuscripts in Malayalim [sic] character, one of them belonging to Dr. A. C. Burnell, the other once in possession of Whish and now forming part of the library of the Royal Asiatic Society, London. The latter manuscript dates from A.D. 1820; the former from A.D. 1863" — Burnell's copied "fr. an Ola M.S. in the Chirakkil Raja library … at Calicut 1863". Both "show the same gaps" and "are derived from one source". He signed off "Leiden, July 10 1874".

"Of the rest of the work no translation has appeared," Clark wrote in 1930 from Harvard, and supplied one through the University of Chicago Press — "a preliminary study based on inadequate material", needing "a year's work in India with unpublished manuscript material and the help of competent pundits."

That work was done for the 1,500th anniversary: the Indian National Science Academy's three-volume critical edition of 1976, by K. S. Shukla with K. V. Sarma, which restores bhuh, prints Somesvara's "false view" in full, and puts Brahmagupta's objection into English. The year before, on 19 April 1975, India's first satellite was launched. ISRO named it Aryabhata.

Read the actual document

"In a yuga the revolutions of the Sun are 4,320,000, of the Moon 57,753,336, of the Earth eastward 1,582,237,500 …" — Aryabhatiya I.3, Clark 1930, p. 9

"As a man in a boat going forward sees a stationary object moving backward just so at Lanka a man sees the stationary asterisms moving backward (westward) in a straight line." — Aryabhatiya IV.9, Clark 1930, p. 64

"If the Earth moves (revolves) through one minute of arc in one respiration, from where does it start its motion and where does it go? And, if it rotates (at the same place), why do tall lofty objects not fall down?" — Brahmagupta, Brahmasphutasiddhanta XI.17, as translated in Shukla–Sarma 1976, p. 15

In plain words: the first gives the Earth a count and a direction, as if it were a planet; the second explains why the sky seems to move when it is we who move; the third is the earliest surviving reply, and it takes the claim literally. You do not ask why towers stand on a metaphor.

The strongest case for the other side

The other side is not "the Earth does not rotate". It is "you are making this cleaner than it is". Four parts, all real.

The text is thin and partly ambiguous. Four short statements in 121 verses, never argued for, with no cause and no reply to objections. Verse IV.10, on its own, says the sky is driven by a wind, and Clark called the rescue-reading "a doubtful expedient". A sceptic can fairly say Aryabhata held both pictures, or reported a rate without deciding what physically turned.

His year is wrong by more than his day is right. The same verse gives a year 3 minutes 20 seconds too long — a full day's error in about 430 years. The day-length is not a separate measurement; it is a by-product of the rotation count and the year count together, and a count ending in a round 500 may have been tuned to fit a year that was itself slightly off. The accuracy is real; what it proves about his method is less certain.

The objections were not stupid. Brahmagupta's stone and Lalla's birds are the questions Copernicus still could not fully answer in 1543. Without Galileo's principle, the common-sense case against a spinning Earth was the stronger one; the commentators who rejected it were weighing a bare assertion against everyday evidence and choosing the evidence. Dutta sees in the controversy "a critical scientific attitude and spirit of open-minded free enquiry".

It led nowhere. Rotation changes no calculation: mean motions, eclipses and planetary positions come out identically whether the sky turns or the Earth does — which is exactly why it could be dropped at no cost. A rotating Earth became useful only when joined to an orbiting Earth, and Aryabhata kept the Earth at the centre.

What we could not prove

• Varahamihira's own words against the rotation. The 1976 editors say the theory "received a severe blow at the hands of Varahamihira"; Clark cites Pancasiddhantika XIII.6–8. We did not open the Pancasiddhantika and quote nothing from it.
• Brahmagupta from an edition of Brahmagupta. We have XI.17 in Sanskrit from Clark (p. 14) and in English from Shukla–Sarma (p. 15), not from Sudhakara Dvivedi's 1902 edition. The modern-units version of the falling-stone argument is Dutta's paraphrase, following S. R. Das (1926), which we did not open.
• Somesvara's date. The 1976 edition (p. xxxvi) can only bound it: after Bhattotpala (968) and before Amaraja (c. 1200).
• Paramesvara's date. Clark says "sixteenth-century"; Kern could not date him; the 1976 edition infers a birth about 1380 and dates his Drgganita to 1431 and Goladipika to 1443, and cites him as "A.D. 1431". We follow 1976.
• The Arabic Zij al-Arjabhar. Not extant, as far as we can find. "Around 800" and the translator's name rest on the Biographical Encyclopedia of Astronomers (2007) and one paragraph of the 1976 introduction; we did not see Pingree's fuller account.
• Whether the rotation count was tuned to the year. Dutta suggests the year came first; the text gives no method. Unproven either way.
• A manuscript image of the Aryabhatiya. None on Wikimedia Commons with a usable licence. The folio shown is of Bhaskara II's Lilavati.

Why it still matters today

The usual lesson — "India knew first" — is the least interesting thing here. Heraclides of Pontus said it about 350 BCE and was forgotten too. What the documents show is more useful than a trophy: what happens to a correct idea that arrives without a mechanism.

Aryabhata gave a number right to a hundredth of a second and a picture — the boat — that was exactly right. He gave no reason the ground should move and no answer to the man who asked why towers stand. For nine hundred years the best people in the field, who revered him and copied his sine table, quietly deleted the one thing he said that was most true, because it could not be defended. The hostile critic preserved the word. That is not a story about ancient India. It is a story about how evidence and authority behave, anywhere.

It is also a story about how a text survives: two Malayalam-script copies, one lent from Madras and one lying in London, collated at Leiden, and an Indian edition on the 1,500th birthday restoring the syllable. The satellite carries the name. The syllable is the harder achievement.

Verdict

The Aryabhatiya says, in three places and in a number, that the Earth turns eastward once a day and the stars stand still — and the earliest witness, Brahmagupta in 628, quotes it saying so in order to attack it. That is as strong as sixth-century textual evidence gets. Clark 1930 and the 1976 INSA edition both read it that way, and the arithmetic gives 23 h 56 min 4.1 s against the modern 4.09 s. Those are facts.

Aryabhata did not say the Earth goes round the Sun; he said the opposite. He gave no physics and answered no objection. One of his verses reads, on its face, like the orthodox view. His year is three minutes wrong. Anyone who tells you he was Copernicus has not read IV.6.

Was reading the verse as "false knowledge" misguided ingenuity? Yes. Kern saw it in 1874, Clark named it in 1930, and the editors of 1976 proved it with the one syllable Brahmagupta had kept. But the men who explained it away were not villains. They had a bare assertion on one side and a thousand years of falling stones on the other, and the stones were, for their time, the better evidence. The verse waited. Verses can.