A mathematical problem inside a ritual practice
The Śulbasūtras are Sanskrit rules for constructing sacrificial altars, associated with names including Baudhāyana and Āpastamba. They belong broadly to the first millennium BCE; precise dates for their composition are uncertain. Their geometry concerns measured lengths and transformations of area. A cord makes a straight line, a radius, or a diagonal tangible. These are construction texts, and a surviving rule is not automatically a surviving proof. 1
The setting matters. A question such as “make this shape twice as large” is ambiguous until we specify whether “large” means length or area. Doubling a square’s side quadruples its area. Producing twice the area requires a different operation. This distinction gives us a useful way into the mathematical reasoning, without pretending that an ancient manual used our symbols or classroom vocabulary.
A diagonal that doubles an area
The following is a modern explanatory reconstruction. Begin with a square of side 1, and draw its diagonal. The diagonal has length √2. A new square whose side is that diagonal therefore has area (√2)² = 2. A geometric object has supplied the length needed for an area transformation.
We can check the relationship without measuring a decimal. Divide the original square along a diagonal into two congruent right triangles. Four such triangles can be arranged around a central point so that their hypotenuses form the boundary of a larger square. Each triangle has area 1/2, so the new square has area 4 × 1/2 = 2. Its side is the diagonal we started with. This is a present-day area argument, not a quotation or an attributed ancient proof.
Combining two unequal squares leads to the same kind of question. If their sides are 3 and 4, the desired area is 9 + 16 = 25. A right triangle with perpendicular sides 3 and 4 supplies a hypotenuse of 5. The important operation is adding areas, not adding side lengths. Cord constructions for combining squares are discussed in the source overview. 1
Exact relationships, approximate measurements
A rule associated with the Baudhāyana text gives the following approximation, written here in modern notation: 2
√2 ≈ 1 + 1/3 + 1/(3 × 4) − 1/(3 × 4 × 34) = 577/408.
To assess it, we can do an exact integer calculation: 577² − 2 × 408² = 1. Consequently, (577/408)² is 2 + 1/166464, slightly greater than 2. The fraction therefore lies above √2. Its decimal value is approximately 1.414215686, compared with √2 ≈ 1.414213562. The absolute error is about 0.000002124.
This calculation tells us about the approximation, but not how its author discovered it. Several procedures can produce the same fraction. Reconstructing a plausible procedure and documenting the procedure historically are different tasks. Nor does a good approximation by itself establish that its author had proved the irrationality of √2.
Read the practice before claiming priority
These texts make a rich subject for studying construction, accuracy, and the relationship between a practical instruction and an explanation. We avoid attaching a single exact date to a layered textual tradition or using a modern theorem’s name to settle who first knew every version of it.
Continue with Āryabhaṭa’s numerical astronomy for another setting in which geometry supports calculation, or compare the problem-based teaching of the Līlāvatī.