Brahmagupta's bhāvanā — the composition lemma chakravala stood on

# Brahmagupta's bhāvanā — the composition lemma chakravala stood on

Take any two integer solutions of an equation of the shape $N x^2 + k = y^2$ — call them $(x_1, y_1)$ with additive $k_1$ and $(x_2, y_2)$ with additive $k_2$. Brahmagupta wrote down, in 628 CE, a single rule for combining them into a third solution:

$$(x_3, y_3) = (x_1 y_2 + x_2 y_1, \; y_1 y_2 + N x_1 x_2)$$

This new pair satisfies $N x_3^2 + (k_1 k_2) = y_3^2$.

That's it. Two old solutions in. One new solution out. The additive multiplies. The operation is repeatable: feed the output back in with itself or another seed, get another solution, again and again.

Brahmagupta called the operation **bhāvanā** — "composition," "making right by combination." It is the engine of every later Indian solution to Pell-like equations. The chakravala algorithm Bhāskara II used to solve $61 x^2 + 1 = y^2$ in 1150 CE — yielding the famous $x = 226{,}153{,}980$ — is a controlled iteration on top of bhāvanā. Without bhāvanā there is no chakravala. Without chakravala there is no European-named "Pell's equation," because Lagrange's 1768 solution recovers the same composition structure that Brahmagupta wrote down 1,140 years earlier.

The verse, with Colebrooke's commentary

Colebrooke's 1817 translation renders Brahmagupta's bhāvanā at §77 of chapter 18:

> The "greatest" and "least" roots are to be reciprocally multiplied > crosswise; and the sum of the products to be taken for a least > root. The product of the two "least" roots being multiplied by the > given coefficient, and the product of the "greatest" roots being > added thereto, the sum is the corresponding greatest root; and the > product of the additives will be the [new] additive. > > — *Brāhmasphuṭasiddhānta* XVIII §76-77, trans. H. T. Colebrooke > (1817)

Translated into modern notation: "greatest root" = $y$, "least root" = $x$, "additive" = $k$, "coefficient" = $N$. Then "reciprocally multiplied crosswise" is the cross-product $x_1 y_2 + x_2 y_1$ (which becomes the new $x_3$). And the new $y_3$ comes from $y_1 y_2$ plus $N$ times $x_1 x_2$.

The next verse (§78) gives the *minus* variant:

$$(x_3, y_3) = (x_1 y_2 - x_2 y_1, \; y_1 y_2 - N x_1 x_2)$$

— useful when you want *smaller* roots from larger seeds rather than larger from smaller. Colebrooke's footnote labels these *samāsa-bhāvanā* (sum composition, §77) and *antara-bhāvanā* (difference composition, §78), and notes when each is the right move.

How chakravala is built on top

Bhāskara II's chakravala iteration ([covered in its own claim](/c/e9b99c06-981f-5f4d-8628-0357df5417f0)) starts from any trivial Pell-like seed and uses bhāvanā repeatedly to drive the additive $k$ down to 1 in a controlled way.

Concretely, for $N = 61$ (Bhāskara's worked example, replayed by Fermat as a challenge problem 507 years later):

1. **Seed.** $(x_0, y_0) = (1, 8)$, with $k = 3$, since $1 \cdot 61 + 3 = 64 = 8^2$. 2. **Apply bhāvanā** to compose $(1, 8, k=3)$ with the trivial $(1, m, m^2 - 61)$ for various $m$, generating new triples. 3. **Choose $m$ at each step** so the new additive divides cleanly (this is the chakravala-specific rule layered on top of bhāvanā). 4. **Iterate**, watching the additive march toward 1. 5. **At additive 1**, you have a Pell solution.

After the chakravala iteration converges on $N = 61$, the final triple is $(226{,}153{,}980, \; 1{,}766{,}319{,}049, \; 1)$ — Bhāskara's published answer.

Every composition step in that chain is bhāvanā. The chakravala contribution is the heuristic for choosing $m$. Bhāvanā is the operation; chakravala is the strategy that uses the operation.

Both are necessary. Brahmagupta gave the operation in 628 CE; Bhāskara II added the strategy in 1150 CE, five centuries later.

The European route to the same machinery

European mathematicians did not encounter Pell-type equations systematically until **Pierre de Fermat** in the 1650s. Fermat posed $N = 61$ as a challenge to William Brouncker and John Wallis in 1657, asking for integer solutions. Brouncker eventually produced an answer using continued fractions; Fermat himself claimed (in his usual margin-style) to have a general method but left no proof.

The full European solution came in two stages:

- **Leonhard Euler**, 1764, *Vollständige Anleitung zur Algebra*. Euler develops a multiplicative identity for Pell-like equations that is essentially the algebraic content of bhāvanā, though presented in continued-fraction language rather than as a discrete composition law. Euler does not cite Indian sources; the work appears to be independent rediscovery.

- **Joseph-Louis Lagrange**, 1768, *Solution d'un problème d'arithmétique*. Lagrange proves the general theorem that Pell's equation $N x^2 + 1 = y^2$ has infinitely many solutions for any non-square $N$, and gives a complete algorithm for generating them. The algorithm combines continued-fraction convergent expansion with what amounts to bhāvanā composition in disguise.

Euler-Lagrange (1764-1768) is the European "first complete solution" to Pell's equation. It is approximately 1,140 years after Brahmagupta's bhāvanā and approximately 620 years after Bhāskara II's chakravala. The naming convention "Pell's equation" — attached to the 17th-c. English mathematician John Pell, who never worked on the problem — is a Euler-era misattribution that stuck.

What this leaves us with

A single Sanskrit verse from 628 CE, defining a binary operation on the solution space of $N x^2 + k = y^2$ equations. The operation:

- closes the solution set under composition (additives multiply), - generates infinitely many new solutions from any two seeds, - is *exactly* the structure needed to recursively drive Pell-type equations to additive 1, - becomes the engine of Bhāskara II's chakravala 522 years later, - gets independently rediscovered by Euler-Lagrange ~1,140 years later, in continued-fraction clothing.

The bhāvanā rule has the algebraic shape of a *group operation* on a particular solution space (modern formulation: solutions of $y^2 - N x^2 = k$ as norms of elements of $\mathbb{Z}[\sqrt{N}]$, with bhāvanā the multiplication of those elements). The Sanskrit tradition didn't have the language of group theory, but it had the operation that group theory would later formalise — written down, in verse, in the 7th century, and used continuously thereafter.

When Bhāskara II solved $61 x^2 + 1 = y^2$ in 1150 CE, the algebra that let him do it was three numbered Sanskrit verses old. Five centuries between the operation and its most spectacular application. Five and a half centuries more, and the European tradition got to the same place from a different direction.

---

Sources

- [Algebra, with Arithmetic and Mensuration, from the Sanscrit of Brahmegupta and Bhāscara, H. T. Colebrooke trans., 1817](https://archive.org/details/1817-henry-colebrooke-algebra-with-arithmetic-and-mensuration-from-the-sanskrit-of-br) — *Brāhmasphuṭasiddhānta* XVIII §76-77 (bhāvanā) cited above. - Plofker, K. (2009). *Mathematics in India*. Princeton University Press, ch. 5. — bhāvanā in the Sanskrit number-theoretic tradition. - Lagrange, J. L. (1768). "Solution d'un problème d'arithmétique." *Mélanges de Turin* IV. — the European treatment.

Related claims

- [Bhāskara II's chakravala for Pell's equation](/c/e9b99c06-981f-5f4d-8628-0357df5417f0) — the algorithm bhāvanā made possible, 522 years later. - [Brahmagupta's cyclic-quadrilateral formula](/c/5c1b33ac-50a8-5b98-86d5-feb2ac2109f9) — same author, same treatise, a different chapter (XII vs XVIII) showing Brahmagupta's range from geometry to number theory.