From Baghdad to Cordoba — how al-jabr reshaped the language of number
Between the 8th and 13th centuries, scholars working in Arabic transformed mathematics from a collection of geometric tricks into a systematic science of the unknown. This exploration traces that transformation — and argues that it was the single greatest turning point in the history of mathematical thought.
The Persian scholar whose name became the word "algorithm," and whose book gave us the word "algebra."
Muḥammad ibn Mūsā al-Khwārizmī (c. 780 – c. 850 CE) was born in Khwārazm, in present-day Uzbekistan, and worked in Baghdad at the court of Caliph al-Maʾmūn during the early ninth century. He was an astronomer, geographer, and mathematician attached to the Bayt al-Ḥikma ("House of Wisdom"), where Greek, Indian, and Persian scientific texts were being systematically translated, criticised, and extended.[1]
Around 820 CE, al-Khwārizmī completed Kitāb al-mukhtaṣar fī ḥisāb al-jabr waʾl-muqābala — "The Compendious Book on Calculation by Restoration and Balancing." Its title gives us two technical terms that still define the discipline:
"That fondness for science... has encouraged me to compose a short work on calculating by al-jabr and al-muqābala, confining it to what is easiest and most useful in arithmetic." — al-Khwārizmī, preface to Kitāb al-jabr, c. 820 CE[2]
Before al-Khwārizmī, equations were solved case-by-case, usually through geometry (as in Euclid's Elements) or through verbal arithmetic puzzles (as in Diophantus's Arithmetica, c. 250 CE). There was no general procedure applicable to a whole class of problems. Al-Khwārizmī's innovation was to classify all linear and quadratic equations into six standard forms — for example, "squares equal to roots" (ax² = bx) and "squares and roots equal to numbers" (ax² + bx = c) — and to give a systematic recipe (an algorithm) for solving each one.[3]
Although al-Khwārizmī used no symbols — every step was written in full Arabic prose, and every proof was justified geometrically — the method he created was symbolic in spirit. He treated "the unknown" (shayʾ, "thing") as an object that could be manipulated according to fixed rules. This is the conceptual seed from which all later algebra grew.
A practical pressure drove this work: Islamic farāʾiḍ (inheritance) law, derived from the Qurʾān, required precise fractional divisions of estates among many heirs. Al-Khwārizmī devotes the longest section of his book to inheritance problems that simply could not be solved by Greek geometry alone — they demanded a calculus of unknowns.[4]
A base-10 positional system, transmitted from India through Baghdad to Europe — and the reason we can write 2026 in four symbols instead of MMXXVI.
Around 825 CE, al-Khwārizmī wrote a second treatise, Kitāb fī ḥisāb al-hindī ("Book on Calculation with Hindu Numerals"), describing a base-10 positional system imported from Indian sources — especially Brahmagupta's Brāhmasphuṭasiddhānta (628 CE), which had already formalised zero as a number with its own arithmetic rules.[5] The Arabic word for zero, ṣifr (صفر, "empty"), passed into Medieval Latin as zephirum, becoming both "zero" and "cipher" in English.
In this system, the value of a digit depends on its position. The same symbol "2" means twenty in 26, two hundred in 200, and two thousand in 2026. Zero is essential — it holds a place open when a power of ten contributes nothing to the total.
| Numeral | Thousands (10³) | Hundreds (10²) | Tens (10¹) | Units (10⁰) | Positional expansion |
|---|---|---|---|---|---|
| 1 | 0 | 0 | 0 | 1 | (1 × 10⁰) = 1 |
| 7 | 0 | 0 | 0 | 7 | (7 × 10⁰) = 7 |
| 19 | 0 | 0 | 1 | 9 | (1 × 10¹) + (9 × 10⁰) = 10 + 9 |
| 60 | 0 | 0 | 6 | 0 | (6 × 10¹) + (0 × 10⁰) = 60 + 0 |
| 144 | 0 | 1 | 4 | 4 | (1 × 10²) + (4 × 10¹) + (4 × 10⁰) |
| 1000 | 1 | 0 | 0 | 0 | (1 × 10³) + (0 × 10²) + (0 × 10¹) + (0 × 10⁰) |
| 2026 | 2 | 0 | 2 | 6 | (2 × 10³) + (0 × 10²) + (2 × 10¹) + (6 × 10⁰) |
Notice the role of zero in 60, 1000, and 2026: it is not "nothing" — it is a placeholder that keeps the other digits in their correct columns. Without it, "26" and "206" and "2006" would be indistinguishable.
| Feature | Roman numerals | Greek alphabetic | Hindu-Arabic (positional) |
|---|---|---|---|
| How "2026" looks | MMXXVI | ͵ΒΚϚ | 2026 |
| Zero as placeholder | Absent | Absent | Present (ṣifr) |
| Position carries value | No (additive: V + I = VI) | No (each letter = one fixed value) | Yes (the 2 in "20" means 2 × 10) |
| Easy long multiplication | Extremely difficult — requires counting board | Difficult — no carry mechanism | Routine — algorithmic columns |
| Representable size | Bounded by available letters | Bounded by alphabet | Unbounded — any integer with finitely many digits |
This is why the system spread. Through Latin translations of al-Khwārizmī's arithmetic — beginning with Algoritmi de numero Indorum in the twelfth century — and through Fibonacci's Liber Abaci (1202 CE), the Hindu-Arabic numerals reached European merchants, who quickly abandoned Roman numerals for bookkeeping.[6]
The very first worked example in Kitāb al-jabr — solved twice. Once in al-Khwārizmī's geometric prose, once in modern symbolic algebra.
Al-Khwārizmī chose this equation as his canonical illustration of "squares and roots equal to numbers" (ax² + bx = c). The methods below look superficially different, but they describe the same underlying mathematical operation: completing the square. What modern notation compresses into three lines, al-Khwārizmī had to write — and prove — geometrically.
Rhetorical & geometric, no symbols
"A square and ten roots are equal to thirty-nine dirhams." — Kitāb al-jabr, Ch. I
Al-Khwārizmī recognised only the positive root; in his geometric framework, a "negative length" was meaningless. The answer is therefore x = 3.
Symbolic, with negative roots
Method 1 — Completing the square
x² + 10x = 39 x² + 10x + 25 = 39 + 25 (x + 5)² = 64 x + 5 = ±8 x = 3 or x = −13Method 2 — Quadratic formula
x² + 10x − 39 = 0, a=1, b=10, c=−39 x = (−10 ± √(100 + 156)) / 2 x = (−10 ± √256) / 2 = (−10 ± 16) / 2 x = 3 or x = −13Modern notation recovers both roots automatically because the symbol "±" makes no geometric commitment. Al-Khwārizmī's diagram was the completing-the-square method — he simply did not yet have the symbols to write it in two lines.
The geometric figure above is the algebraic identity (x + 5)² = x² + 10x + 25. Al-Khwārizmī did not call it that — he had no parentheses, no variables, no "²" symbol — but the logical move is identical. We can therefore say something sharper than "al-Khwārizmī invented algebra": he discovered the procedure of completing the square, justified it geometrically, and generalised it into a recipe. The symbolic notation that lets us compress his diagram into "(x + 5)² = 64" arrived gradually — through Viète in the late 16th century and Descartes in 1637 — but the method was already complete in Baghdad more than 700 years earlier.
A library, a translation bureau, and an observatory — the institution that made the Golden Age possible.
Baghdad was founded in 762 CE by the Abbasid Caliph al-Manṣūr as the new capital of an empire that stretched from Spain to Central Asia. Within a generation it had become the largest city in the world outside China. Its circular plan placed the caliph's palace at the centre, but its intellectual heart was the Bayt al-Ḥikma (بيت الحكمة), the "House of Wisdom" — a library and research institution that flourished especially under Caliphs Hārūn al-Rashīd (r. 786–809) and his son al-Maʾmūn (r. 813–833).[7]
Al-Maʾmūn reportedly sent diplomatic missions to Constantinople specifically to acquire Greek manuscripts, and offered translators their weight in gold. The Graeco-Arabic translation movement that followed — spanning roughly 750 to 1000 CE — rendered Aristotle, Euclid, Ptolemy, Archimedes, Apollonius, Galen, and Diophantus into Arabic, often with critical commentaries that surpassed the originals.[8] Indian astronomical and arithmetical works (the Siddhāntas) and Persian texts on algebra and astronomy were translated in parallel. The House of Wisdom was thus not merely a passive library but the world's first sustained cross-cultural research programme.
Founder of algebra as a discipline. Wrote Kitāb al-jabr and the arithmetic treatise that introduced Hindu numerals to the Arabic-speaking world.
Three brothers — Muḥammad, Aḥmad, and al-Ḥasan — who funded translators, commissioned manuscripts, and wrote the Book on the Measurement of Plane and Spherical Figures.
Sabian translator-mathematician who produced the definitive Arabic Euclid and extended Greek number theory, including a famous theorem on amicable numbers.
Astronomer whose trigonometric tables — using sines rather than Ptolemy's chords — later guided Copernicus and Kepler.
In al-Fakhrī and al-Badīʿ, he freed algebra from geometric reasoning and developed polynomial arithmetic and an early form of mathematical induction.
Solved cubic equations geometrically by intersecting conic sections; reformed the solar calendar to an accuracy rivalling the modern Gregorian one.
The House of Wisdom's golden age ended with the Mongol sack of Baghdad in 1258 CE, when the libraries were destroyed and the surviving scholars dispersed. By then, however, the mathematical achievements of the tradition had already begun travelling westward through al-Andalus (Islamic Spain) into the Latin-reading universities of Europe — seeding the Scientific Revolution that would follow four centuries later.
The thesis of this exploration: that the invention of al-jabr — the systematic, symbolic-in-spirit manipulation of unknown quantities — is the single transformation from which all of modern mathematical thought descends.
Many candidates compete for the title of "greatest turning point": the Babylonian base-60 place-value system; Greek deductive proof in Euclid; the calculus of Newton and Leibniz; Cantor's set theory; Gödel's incompleteness. Each is profound. But the argument made here is that al-Khwārizmī's al-jabr is more fundamental than any of them, because each of them presupposes the move that al-jabr made first: treating an unknown quantity as an object that obeys rules of manipulation. Five premises support this thesis.
Diophantus, writing around 250 CE, solved roughly 150 individual equation puzzles — each with its own ingenious trick. Al-Khwārizmī instead classified all linear and quadratic equations into six standard forms and provided an algorithm for each. The difference between "I can solve this problem" and "I can solve every problem of this type" is the difference between a craft and a science.[3]
Before al-Khwārizmī, "the thing we don't know yet" was a gap waiting to be filled by geometric construction. Al-jabr reverses this: the unknown (shayʾ) is itself the subject of operations — you can add to it, subtract from it, square it, balance it across an equals sign. This is exactly the conceptual move that later permits the symbol "x" of Descartes, the variable of calculus, and ultimately the abstract objects of group theory and topology.
Al-Khwārizmī's six standard forms come with explicit step-by-step recipes — what we now call algorithms, a word descended from his own name. This is why a modern computer can be programmed to solve a quadratic: not because Babylonian scribes solved one in 1800 BCE, but because al-Khwārizmī articulated a procedure that could be written down, taught, and mechanically followed.[9]
Greek deductive geometry, Indian decimal arithmetic, and Persian astronomical practice had never been brought together. Al-Khwārizmī's work fuses them: Indian numerals to write the numbers, Greek geometry to justify the proofs, Persian practical problems to motivate the questions. Every later mathematical tradition — European, Ottoman, Mughal — inherits this synthesis rather than any single ancestor.
Calculus is the algebra of changing quantities; Newton's Principia (1687) is unthinkable without symbolic manipulation of unknowns. The same applies to coordinate geometry, linear algebra, statistics, and machine learning — all of them descend technically and conceptually from the move al-Khwārizmī made when he wrote that a square and ten roots equal thirty-nine, and proved how to find the root.
Could one not argue that the Hindu-Arabic numerals are the real turning point? They are certainly indispensable, but they are a notational invention: they let us write numbers efficiently. Al-jabr is a conceptual invention: it lets us reason about unknown numbers. Notation without reasoning is bookkeeping; reasoning without notation is still mathematics (al-Khwārizmī did it in prose). Of the two, the reasoning is primary.
Could one not argue that Greek deductive proof matters more? Proof is the criterion by which a mathematical claim is judged true. But proof needs propositions to prove. Al-jabr generates an unending supply of propositions about unknowns — and so gives proof something new and powerful to work on. The two are complementary; but after al-jabr, the scope of what could even be proven expanded enormously.
"Al-Khwārizmī's algebra is regarded as the foundation and cornerstone of the sciences. The subject deals with what is general in number — that is, with the abstract." — Roshdi Rashed, The Development of Arabic Mathematics, 1994[10]
Conclusion of the argument: the invention of al-jabr in early ninth-century Baghdad — systematic, general, algorithmic, cross-cultural, and conceptually freeing the unknown to become an object of operation — is the transformation from which the rest of modern mathematics flows. It is the moment mathematics ceased to be a collection of solutions and became a method.
Primary sources are listed first; secondary scholarship follows. Citations use a combination of MLA-style and IB-appropriate descriptive form.