Why Arabic Numerals Began in India and Reached Europe Through Arabic Scholarship

The numerals 0–9 are often called Arabic numerals in English because Europeans learned important forms and methods through scholarship written in Arabic. The decimal place-value system behind them developed in India. Neither name tells the whole history: Indian mathematicians developed ways to write and calculate with place value and zero; scholars across the Arabic-speaking world explained and adapted those methods; and Europeans encountered them through manuscripts, teaching and trade over several centuries. The University of St Andrews history of Arabic numerals documents this route while warning that older finger reckoning and sexagesimal computation continued alongside Indian-style arithmetic.

The word Arabic therefore describes a major transmission route, not a claim that one Arab scholar invented the ten digits. The word Indian identifies a crucial earlier development, not a claim that glyphs moved west unchanged. To understand the history, follow four separate questions: What does position do in a number? How was an empty place marked? Which texts taught the method? When did particular communities begin using it?

The opening image is an original AI-assisted reconstruction for educational context. It does not depict a documented meeting or reproduce a specific surviving manuscript.

What makes 0–9 a place-value system?

In 507, the 5 means five hundreds, the 0 records that no tens are present, and the 7 means seven units. In 57, the same 5 means five tens. A written digit has a value partly because of its position. The decimal system groups places in powers of ten: units, tens, hundreds and so on. This is different from merely having ten attractive symbols or using a base-ten word for a large quantity. The method lets a reader distinguish 507 from 57 without remembering a new sign for every multiple of ten.

The zero sign has a special practical role in that example. Leaving a blank between 5 and 7 can work in a tightly controlled setting, but blanks are easily lost when a calculation is copied. A visible mark keeps the empty place stable. Rules for carrying and borrowing then allow repeatable written operations on the columns. Try adding 507 and 26: the units make 13, so write 3 and carry 1 ten; the empty tens place in 507 still has to be counted, giving 3 tens after the carry; the five hundreds remain. The result is 533. The example is modern pedagogy, not a claim that an ancient manuscript used these exact glyphs or layout.

Digits, notation and arithmetic are related but not identical. A manuscript may display a row of nine symbols without proving that every reader used a zero sign, positional calculation or the same multiplication rules. Conversely, a culture may calculate positionally with tokens or a dust board while written signs differ from today's printed figures. The St Andrews account of zero separates the placeholder problem from later efforts to treat zero as a number on which one can calculate.

What evidence connects the system to India?

Indian mathematical traditions developed decimal place-value techniques and rules involving zero. Brahmagupta's Brāhmasphuṭasiddhānta, dated to 628, states rules for calculations with zero and negative quantities. The St Andrews biography of Brahmagupta makes an important distinction: some of those rules were productive, while his treatment of division by zero does not match modern arithmetic. Crediting his work does not require pretending every statement was later accepted unchanged.

The Bakhshali manuscript, preserved at Oxford, contains a dot used in numerical notation. In 2017 the Bodleian Libraries announced radiocarbon results from different birch-bark leaves, including material assigned a surprisingly early range. The announcement presented this as evidence for an early zero symbol. Yet the age of a piece of bark is not automatically the date when its surviving writing was put on it, and separately dated leaves need not all belong to one moment of copying. A published response by Kim Plofker and four other historians disputes the announcement's inference that the earliest tested leaf dates the whole written text. The authors also argue that some zero marks in the manuscript do more than hold a place. A responsible history reports both the laboratory finding and the historical disagreement; it does not promote one contested manuscript to an uncontested single birthday for zero.

St Andrews notes that the Gwalior inscription of 876 provides a firmly dated Indian example with zero signs in numbers associated with a garden. This does not mean zero was invented in 876. It means the inscription is an especially clear dated witness. A surviving artifact fixes what can be observed at that place and time; it rarely identifies an unknown first inventor or settles all earlier manuscript dates. India is central to the system's development without reducing a long, varied mathematical history to one person and one day.

What did al-Khwarizmi actually transmit?

Muhammad ibn Musa al-Khwarizmi worked in the ninth-century Abbasid scholarly world. A work associated with him explained reckoning with Indian numbers in Arabic. Its Arabic original has not survived; later Latin material associated with his name is an important witness, but it is not a photograph of an intact original. The St Andrews source study explains that the Latin text was substantially changed during transmission and that even the Arabic title is uncertain. It therefore supports a claim about a powerful line of teaching, while limiting claims about every exact sentence al-Khwarizmi wrote.

His book on algebra is a separate work from the arithmetic tradition of Indian numerals. The St Andrews biography of al-Khwarizmi discusses the two works separately and notes how little secure biographical detail survives. Our article on al-Khwarizmi's algebra, algorithms and surviving works examines those distinctions further. Calling him the inventor of the decimal digits because his name survives in algorithm confuses an author's influence with the origin of the notation he described.

The Latinized form of his name helped produce the medieval term algorism for calculation by the new figures and ultimately the English word algorithm. That etymology does not mean that every step-by-step procedure was invented in Baghdad. It identifies a historical route by which a person's name became attached to a method of reckoning. Likewise, the term algebra comes from a word in the title of his algebra treatise, not from the shape of the nine digits. These distinctions make the vocabulary more informative than a story of a lone mathematical genius.

Did the Arabic-speaking world use one set of figures?

No. The St Andrews historical survey describes several kinds of arithmetic in use in Arabic-speaking societies, including finger reckoning for practical work, alphabetic or sexagesimal notation in some contexts, and calculation with Indian-derived decimal figures. The presence of a treatise does not prove universal use by merchants or astronomers. The story is one of coexisting methods, changing uses and different regional forms.

The numeral shapes that developed in western parts of the Arabic-speaking world had a particularly important relation to the forms later familiar in western Europe. Eastern Arabic forms remained distinct and continue to be used. A much later British Museum record for a sixteenth-century Ottoman compass and sundial transcribes visible Eastern Arabic digits on its hour scale. This artifact cannot explain the ninth-century origin of the system. It does show why Arabic numerals should not make readers imagine that every Arabic-language document used only today's western printed 0–9.

Different written forms may express the same positional rule. Compare English 507 with a number written using another digit set: the outward signs can change while hundreds, tens and units still organize the quantity. Written Arabic can also carry letters, words and number signs with different functions. Our guide to Ajami manuscripts and African languages written in Arabic script gives a parallel warning: script, language and numerical method are not interchangeable labels. A digit's cultural journey cannot be read from its modern printed silhouette alone.

How did manuscripts and working surfaces affect calculation?

Early calculations could be made on a dust board, where figures were written, moved and erased as intermediate steps changed. The St Andrews account notes the practical value and limitation of this surface: it supports rearranging a calculation, but an erased stage does not leave the same record as a manuscript page. By the middle of the tenth century, al-Uqlidisi's surviving discussion of Indian reckoning shows a scholar explaining practical methods in writing. The text is evidence for continued development after al-Khwarizmi, not proof that the whole region switched to one method on a single date.

Written media matter because copying can preserve examples, terminology and errors. A manuscript may be produced long after the work it transmits was first composed. A Latin translation can preserve the outline of an Arabic teaching tradition even when the Arabic original is lost and the translation has changed it. A later copy of an algebra treatise can illustrate a scholarly lineage without being the author's autograph. Our guide to sizing and burnishing Islamic manuscript paper explains one part of the material history of prepared writing surfaces; its techniques should not be projected onto every dust-board calculation or every surviving arithmetic manuscript.

Mathematics also served several needs beyond a merchant's account book. Astronomers used numerical tables and procedures, though not necessarily the same written figure forms in every table. Our profile of al-Battani's observations and computational astronomy is context for that broader scientific culture, not a claim that one of his observations introduced 0–9 to Europe. Keeping the use of arithmetic separate from the provenance of a particular digit shape prevents a false chain of invention.

Were the numerals known in Europe before Fibonacci?

Yes. The St Andrews survey identifies the Codex Vigilanus, copied in northern Spain in 976, as an early surviving European document displaying Indian-derived numerals. It is an attestation in a manuscript, not evidence that every European trader was already keeping accounts with decimal notation. Medieval Europe contained varied schools, languages, recording practices and tools. A figure's appearance in one book and its routine adoption across workshops are different events.

Latin works connected with Arabic scholarship offered another route. Translators, teachers and readers worked with material associated with al-Khwarizmi; Mediterranean travel and commerce exposed people to practices used farther south and east. The number system did not pass through one sealed handoff from India to Baghdad to Pisa. The same historical survey explicitly warns against a simple “one invention, one transmitter, one conversion” story: sources show individual texts and communities at different dates. Where a source only proves that figures were visible in a manuscript, it should not be stretched into a claim about universal day-to-day use.

This slow spread helps explain the name. Europeans encountered useful arithmetic through Arabic texts and neighboring regions, so Arabic became a convenient label for a system whose earlier development was Indian. Some modern historians prefer Hindu-Arabic numerals to name both major phases. Neither label should be interpreted as a nationality test for every glyph, author or user. The actual history includes scholars of different places and faiths working in multiple languages.

What was Fibonacci's role in 1202?

Leonardo of Pisa, known as Fibonacci, presented Indian-style figures and methods in his Liber abaci of 1202. The St Andrews biography of Fibonacci also records a revised edition in 1228. He used practical problems as well as mathematical demonstrations to make the methods useful to readers concerned with trade and calculation. He was a consequential advocate and teacher, not the first person in Europe to see the numerals and not their inventor in India.

The famous rabbit problem is one part of his book, but it should not replace the book's role in the history of arithmetic. A merchant learning exchange, profit or measurement methods needed reliable operations on numbers; a manuscript's ability to teach those operations helps explain its influence. Even so, a publication date is not the date of universal adoption. The St Andrews biography notes that Fibonacci's direct influence lay especially in practical numerical methods, while other parts of his work had a narrower immediate readership.

The strongest answer to “Who brought Arabic numerals to Europe?” is therefore plural. Manuscript evidence precedes Fibonacci, Arabic and Latin teaching traditions crossed languages, and Fibonacci gave a prominent thirteenth-century presentation. Each stage has different evidence: an inscription or folio establishes a visible sign, a treatise explains a method, and records of wider use show adoption. Compressing all three into a single named hero makes the story shorter but less accurate.

How can a reader check a claim about the first zero or numeral?

Ask first what is being dated. Radiocarbon analysis can date the plant material of a manuscript leaf; a colophon may date a copy; an inscription can date a use of symbols in a particular place. None automatically dates the invention of the underlying idea. Next ask what the mark does: does it hold an empty column, operate as a number in arithmetic, or merely resemble a modern zero? Finally ask how far the evidence travels: a Tunisian, Iraqi, Spanish or Italian witness belongs to a specific location and object, not to every user of a language.

The contrast between Oxford's 2017 announcement and the published historical response is a useful example of method. Both concern the same Bakhshali leaves; they differ over what the tests prove about the whole written text. Reading both is more informative than repeating an absolute “oldest zero” headline. The contrast between the 976 Codex Vigilanus and Fibonacci's 1202 Liber abaci is similar: one is early visible evidence, the other a major teaching text. A chronology becomes clearer when each source is allowed to answer the question it actually documents.

The ordinary phrase Arabic numerals remains understandable, provided we use it precisely. The decimal positional system and zero have deep Indian roots; scholarship in Arabic developed and transmitted the methods; and European readers gradually adapted them. Today's printed figures are the result of that long history, not an unchanged set created by a single mathematician.

Sources

• University of St Andrews, “The Arabic numeral system” — historical survey for competing calculation methods, al-Khwarizmi's lost original and changed Latin witness, al-Uqlidisi, the 976 manuscript and Fibonacci's 1202 text.
• University of St Andrews, “Zero” — distinguishes a placeholder from arithmetic about zero and documents the dated Gwalior inscription without making it an invention date.
• University of St Andrews, Brahmagupta biography — identifies the 628 work and its rules for zero, including the limits of its account of division.
• Bodleian Libraries, 2017 Bakhshali announcement — reports radiocarbon testing and the institution's interpretation of the early bark dates.
• Plofker, Keller, Hayashi, Montelle and Wujastyk, History of Science in South Asia — published research response disputing an early date for the whole Bakhshali text and analyzing the zero marks.
• British Museum, Ottoman compass and sundial — later physical example whose transcribed Eastern Arabic digits show that numeral forms differ; it is not evidence for ninth-century invention.
• University of St Andrews, Fibonacci biography — documents the 1202 and 1228 Liber abaci editions and describes the scope of their practical influence.
• University of St Andrews, al-Khwarizmi biography — separates the algebra book from the arithmetic tradition, identifies the lost Arabic work and cautions against speculative biographical claims.