The Myth and the Truth
A Man in a Boat
You will have heard that Aryabhata, the great astronomer of fifth-century India, invented zero. His book has no zero in it, and the first rules for doing sums with zero were written by his fiercest critic. What the book does have is a young man of twenty-three saying the Earth turns, a verse about a boat, and a word that later hands quietly changed to make the Earth stand still again.
The myth
Now, you will have heard this one, and if you went to school in India you may have been asked it in an exam. Who invented zero? Aryabhata. A careers website in India sets it as a quiz question, with "Arybhatta" among the options. A mathematics-education blog at the Open University, posted in 2022, says that zero's promotion from placeholder "to a digit of its own" was "established in India by a man named Aryabhata in the 5th Century". A learning site called Cheenta answers a reader who asks, in all good faith, "How did Aryabhatta invent zero? How did he get this idea? Why did he give zero an oval shape?"
It is a proud story and not a silly one. India's first satellite, launched on 19 April 1975, was named after him, and the Indian Space Research Organisation still says so on its website. The story reminds the world that the numbers it writes every day came from India. That part is true, and nobody here will argue with it.
So let us treat it kindly and treat it as a case. We have a famous man, a famous number and a short book. The question is a simple one. Is zero in the book?
The cast
Aryabhata, who tells us his own age and very little else. Here is the telling detail: he wrote his whole treatise in about 121 verses, with a number code that seems to have been his own.
Brahmagupta, a younger astronomer who finished his great book in 628. He mentions Aryabhata again and again, and rarely kindly.
Bhaskara I, who wrote a commentary on Aryabhata's book in 629 and was loyal to it.
Lalla, a later follower of the same school, who disagreed with the master on one large point.
Paramesvara, a Kerala astronomer of the fourteenth and fifteenth centuries, who watched eclipses from 1393 to 1432 and was a fine scientist in his own right. Remember that; he will need it.
And Walter Eugene Clark, a Sanskrit scholar who put the book into English for the University of Chicago Press in 1930, and who called his own translation preliminary. He is our most careful witness, because he says plainly when he does not know.
What the young man says about himself
The record is thin, so let us take only what he gives us.
In the chapter on time, Clark's English has him say: "When three yugapadas and sixty times sixty years had elapsed (from the beginning of the yuga) then twenty-three years of my life had passed." If you count the ages as later astronomers did, that is the year 499, and he was born in 476. Clark is careful to put an "if" in front of that sum, and so shall we.
He opens his mathematics chapter by saying he sets forth "the science which is honored at Kusumapura", a city near modern Patna. Clark adds that Kusumapura "may or may not have been the place of his birth". You may also have read that he ran the great university at Nalanda. No source of his time says so, and the encyclopedias now call it speculation.
That is nearly all. A young man, a city, a school of thought and a book. Everything else is in the verses.
Is zero in the book?
Here is the awkward part.
Aryabhata did not write numbers with digits at all. In the opening chapter he sets out a system in which the consonants of the Sanskrit alphabet stand for numbers and the vowels shift them up by powers of ten. It is clever, and it let him pack enormous figures, such as the number of times the Moon goes round in an age of the world, into a line of verse. In the mathematics chapter he lists the names of numbers from one to a thousand million and says, in Clark's English, that they "are from place to place each ten times the preceding".
That is place value, plainly stated. Now look for the zero. There is no symbol for it. The best candidate is a single word in the code, kha, which later became one of the commonest Sanskrit words for zero. Some nineteenth-century scholars read it that way here. Clark did not. He took it to mean "place", and wrote that "it is still disputed whether a symbol for zero was actually in use in Aryabhata's time". He offered a guess: sums may have been done on a board ruled into columns, with an empty column standing for nothing.
There is another view, and you should hear it. Georges Ifrah, in his history of numbers, argued that knowledge of zero was implicit in Aryabhata's system, because the code would have been impossible to invent without it. That is a reasonable argument, and a good many people find it convincing. Notice what kind of argument it is, though. It says he must have known of zero. It does not show him writing one down.
So who did write the rules down? For that we must call our second witness. The earliest surviving rules for doing arithmetic with zero, treating it as a number in its own right, are in Brahmagupta's Brahmasphutasiddhanta of 628. He was not perfect either. As J. J. O'Connor and E. F. Robertson note in their MacTutor biography, he said that zero divided by zero is zero, which is wrong. But he wrote the rules, and he signed them.
And Brahmagupta, it turns out, had a good deal to say about Aryabhata.
The critic's verdict
Clark prints two verses from Brahmagupta's eleventh chapter, which is largely a review of earlier astronomers. Rendered roughly from Clark's Sanskrit (the translation is mine), they say that because Aryabhata does not know even one of mathematics, time-reckoning or the sphere, his faults have not been listed one by one; there are too many to count, so the wise reader should supply the rest. Clark calls this criticism "in the severest possible way".
What was he so cross about? Clark had a theory, and it is a good clue. Brahmagupta's "bitterest and most frequent criticisms", he wrote, "are directed against points in which Aryabhata was an innovator and differed from smrti or tradition".
Here is an example. Aryabhata's sphere chapter says, in Clark's English, that the Earth's shadow is made "of darkness", and that "the Moon obscures the Sun and the great shadow of the Earth obscures the Moon". That is a plain physical account of eclipses. The older account, as MacTutor's biographers put it, blamed a demon called Rahu, who swallows the Sun and Moon. Brahmagupta took Aryabhata to task on this subject, and yet, Clark notes, Brahmagupta himself, elsewhere in his book, puts eclipses down to Rahu. Make of that what you will. The usual reading is that he was keeping faith with scripture. I have not read Brahmagupta's chapter for myself, so I leave that reading where I found it.
But then comes the oddest clue in the case. Brahmagupta wrote a second, shorter handbook in 665, the Khandakhadyaka, at the age of sixty-seven. Clark prints its opening line, and in my rough translation it promises results "equal to those of the teacher Aryabhata". Clark found this "curious", and so do I. Thirty-seven years after saying the man knew nothing of mathematics, the old critic built his practical handbook to give the same answers. Some scholars have wondered whether a different Aryabhata is meant. Nobody can say for sure, and the trail goes cold there.
A man in a boat
Now for the verse I promised you. It is in the sphere chapter, and here it is in Clark's English.
"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."
Lanka here is not Sri Lanka. It is a point on the Earth's equator that Indian astronomers used as their reference. And the asterisms are the stars. So the stars stand still, and the man sees them sliding west because he is moving east. The Earth turns. Clark found the same idea in the opening chapter, where Aryabhata gives a number for the "revolutions of the Earth", and Clark took them to mean its turning on its axis.
This was not welcome. "Indian astronomy, in general, maintains that the Earth is stationary," wrote Clark, and later writers "attack him bitterly on this point". Even most of his own followers, "notably Lalla", went back to the old view. Lalla's objections, Clark records, sit in a chapter whose title means "false knowledge".
And then something happened to the text itself. In the opening chapter, Brahmagupta quotes Aryabhata using the word bhuh, "the Earth", for the thing that turns. In the text that came down with Paramesvara's commentary, the word had become bham, "the stars". Clark says the reading "has been changed by Paramesvara or by some preceding commentator" in order "to eliminate Aryabhata's theory of the rotation of the Earth". One syllable, and the Earth stood still again. Notice that the evidence for the original reading comes from Brahmagupta, the very man who could not bear Aryabhata. A hostile witness, in this case, is the one who kept the word safe.
Paramesvara went further. In his commentary he argued that the boat verse means only that the Earth seems to move, because the stars really do. To get there he had to read the words for "eastward" and "westward" the opposite way round. Clark called this "misguided ingenuity". Remember that this was a serious astronomer, the man who watched nearly forty years of eclipses. He was not a vandal. His effort, in Clark's words, was "to bring Aryabhata into agreement with the views of most other Indian astronomers", which is to say, to save him from what looked like a mistake.
Here is the honest loose end. The very next verse, as we now have it, describes the stars and planets being driven westward by a wind round a still Earth. Bhattotpala, another commentator, quoted it to refute the boat verse, which, as Clark remarks, meant quoting Aryabhata against himself. Was that verse changed too? Is it the popular view, set down for contrast? Clark wrote: "I see at present no satisfactory solution of the problem." Neither do I.
And lest you think our man a saint, Clark also records that two of Aryabhata's rules, for the volume of a pyramid and of a sphere, are simply wrong. His value for the ratio of a circle's circumference to its diameter is very good, 62,832 to 20,000, or 3.1416, and his verse says the result is "approximately" the circumference. He was a brilliant man who made mistakes, and he knew the difference between exact and close.
Why the truth is the better story
So here is the reveal. Zero is not in the book. The book has place value, stated clearly, and a code that may well have needed zero behind it, and the question of a written symbol was still open when Clark translated it. The earliest surviving rules for calculating with zero belong to Brahmagupta, and like most great ideas zero grew over generations, with no single inventor. Even the oldest written zeros are argued over: in 2017 Oxford's Bodleian Library announced that a page of the Bakhshali manuscript carried the oldest known written zero, dated by radiocarbon to around 300; a group of historians of Indian mathematics replied in History of Science in South Asia that the treatise was younger than that, and its zero more important.
Take away the invention of zero, and what do you have left? A young man of twenty-three who said that the sky stands still and the Earth turns. He said it with a picture anyone who has crossed a river can check. He explained eclipses as shadows, not a demon. Most of the astronomers who came after him rejected the turning Earth, including his own followers, and somebody, perhaps one of his own commentators, changed a word to make him agree with everybody else.
That is a far better story than an invention. The myth gives Aryabhata a gift everybody already accepts. The truth gives him an idea almost nobody would accept, and shows us how near it came to being edited out. We know what he really wrote about the Earth for a slightly comic reason. His fiercest critic quoted him correctly.
His own last verses, in Clark's English, say that "the precious sunken jewel of true knowledge has been rescued by me, by means of the boat of my own knowledge". He picked the boat twice. For a man whose best idea was that we are all moving and simply cannot feel it, that seems about right.
On sources. Aryabhata's own words are quoted from Walter Eugene Clark, The Aryabhatiya of Aryabhata: An Ancient Indian Work on Mathematics and Astronomy (University of Chicago Press, 1930), read in the Internet Archive scan and its machine-read text and checked against the Wikisource transcription: the age stanza (Kalakriya 10), the Kusumapura stanza (Ganitapada 1), the place-value stanza (Ganitapada 2), the circumference rule (Ganitapada 10), the eclipse stanza (Gola 37), the boat stanza (Gola 9), the stanza on the westward wind (Gola 10) and the closing "boat" stanza (Gola 49). All translations of Aryabhata are Clark's; the 1976 Indian National Science Academy edition by K. S. Shukla and K. V. Sarma was not used, because the only full copies found online were uploads of unclear rights. Clark also supplies: the uncertainty over the 499 date; "may or may not have been the place of his birth"; the reading of kha and the "still disputed" sentence on a zero symbol; the ruled-board guess; the "revolutions of the Earth" and the bhuh/bham change, with "Paramesvara or by some preceding commentator"; "misguided ingenuity" and "to bring Aryabhata into agreement" (his note on the opening chapter); Paramesvara's reversed reading of the boat stanza; Bhattotpala quoting Aryabhata against himself; "I see at present no satisfactory solution"; the remarks on Brahmagupta's "bitterest" criticisms, the "severest possible way", Lalla's chapter on false knowledge, Brahmagupta's eclipse criticism and his own attribution of eclipses to Rahu, and the errors on the pyramid and sphere. The two Brahmagupta verses (Brahmasphutasiddhanta XI, 43–44) and the Khandakhadyaka opening line are my rough translations of the Sanskrit as Clark prints it; I have not read either book in a scholarly translation, and an Indology mailing-list discussion (January 2021) reports that no full English translation of Brahmagupta's chapter XI exists. The suggestion that a different Aryabhata may be meant in the Khandakhadyaka comes from that same discussion and is a hypothesis only. "The usual reading" of Brahmagupta's Rahu passages as deference to scripture is my summary of Clark's remark about tradition, not a checked finding. Brahmagupta's date of 628, the Khandakhadyaka in 665 at sixty-seven, his zero rules and the zero-divided-by-zero error are from J. J. O'Connor and E. F. Robertson's MacTutor biographies of Brahmagupta and of Aryabhata (University of St Andrews), which also supply Rahu and the remark that later commentators changed the text; Ifrah's argument is as quoted there and in Wikipedia's "Aryabhata" article. That Aryabhata used no zero symbol, that the Nalanda link is speculation, and Bhaskara I's commentary of 629 are from that Wikipedia article, citing Sarma (2001), Ifrah (1998) and Kumar Jha (2023). Paramesvara's dates and his eclipse observations of 1393–1432 are from the Dictionary of Scientific Biography entry as summarised on Encyclopedia.com and from Wikipedia; his dates vary between sources, and Clark does not say for certain that he made the bham change. The satellite's name and launch date are from the Indian Space Research Organisation's website. The quiz is from careers360.com, where the answer was not visible; the Open University quotation is from Roshani Senior's post "The men who invented zero" on the OU MathEd blog (25 August 2022); Cheenta's question is from an undated page on cheenta.com, which itself says Aryabhata did not use the zero symbol. The Bakhshali dispute is from Medievalists.net (2 November 2017), reporting the Bodleian's announcement and the reply by Dominik Wujastyk, Takao Hayashi, Agathe Keller, Clemency Montelle and Kim Plofker; the reply was not read directly. The approximate verse count of the Aryabhatiya (121) is the conventional figure. Not confirmed: whether Aryabhata used or knew a written zero; where he was born; whether the westward-wind stanza was altered; which Aryabhata the Khandakhadyaka means; and who first changed bhuh to bham.