The Basel Problem: The Infinite Sum That Hides π in Plain Sight

In 1650, an Italian mathematician named Pietro Mengoli posed a question so simple a child could understand it, and so hard that it humiliated the greatest minds in Europe for the next 85 years. Add up the reciprocals of the perfect squares — 1 + 1/4 + 1/9 + 1/16 + 1/25 + ... — forever. The terms get smaller and smaller, so the sum obviously settles down to some finite number. But what number?

Jacob Bernoulli, one of the most formidable mathematicians of his generation, tackled it in 1689 and failed. He published his attempt anyway, ending with an almost plaintive request: if anyone could find what had escaped his efforts, please let him know. For nearly half a century, nobody could. Leibniz couldn't do it. Jacob's brother Johann Bernoulli couldn't do it. It became known as the Basel problem, named for the Swiss city that was home to the Bernoulli dynasty, and it sat there, unsolved, mocking the mathematical establishment.

Then in 1734, a 28-year-old working in St. Petersburg named Leonhard Euler — who had himself studied under Johann Bernoulli back in Basel — found the answer, and it was so strange that it seemed almost like a joke: the sum equals π²/6.

Stop and sit with that for a second. On the left side of the equation you have nothing but whole numbers, squared and inverted, added together one by one, with no circles, no triangles, no curves anywhere in sight. On the right side you have π, the ratio that governs every circle in the universe. Why would a sum built entirely out of integers care about circles at all? That question — and the audacious, not-quite-rigorous way Euler answered it — is one of the great "wait, what?" moments in the history of mathematics.

The Concept

The Basel problem asks for the exact value of the infinite series:

1/1² + 1/2² + 1/3² + 1/4² + 1/5² + ... = 1 + 0.25 + 0.111... + 0.0625 + 0.04 + ...

If you just start adding these terms on a calculator, you'll notice the sum creeps upward very slowly: after 10 terms you're around 1.5498, after 100 terms around 1.6350, after 1,000 terms around 1.6439. It's clearly converging to something, but the digits crawl toward their final destination so sluggishly that brute-force addition is a terrible way to find the answer — you'd need millions of terms just to nail down a handful of decimal places.

Euler's result says the true value is π²/6, which works out to approximately 1.644934... That's a genuinely bizarre thing to discover. Pi shows up when you relate a circle's circumference to its diameter. It shows up in trigonometry, in the area of a circle, in the normal distribution's bell curve. It is not supposed to show up when you're just adding up fractions made of whole numbers.

Euler's original approach, roughly, was this: he knew the Taylor series for sin(x)/x, an infinite polynomial-like expression. He also noticed that sin(x)/x has roots (places where it equals zero) at x = ±π, ±2π, ±3π, and so on — everywhere x is a nonzero multiple of π. For an ordinary polynomial, you can factor it completely if you know all its roots; Euler made the bold leap that the same trick should work for this infinite "polynomial," factoring sin(x)/x into an infinite product of terms like (1 − x²/π²)(1 − x²/4π²)(1 − x²/9π²)... When he then compared that infinite product to the original Taylor series term by term — specifically, matching up the coefficient of x² on both sides — the π²/6 fell right out.

It was a brilliant piece of mathematical intuition, but at the time it wasn't actually rigorous. Nobody had proven that factoring an infinite series into an infinite product was valid the way it is for finite polynomials. Euler knew this was a weak point, and so did his contemporaries — some were openly skeptical even after his numbers checked out against direct calculation. It took until 1741 for Euler himself to produce a proof that mathematicians of the era accepted as solid. Today we have many different rigorous proofs of the same result, including slick ones using Fourier series and double integrals, but Euler's original leap of intuition is still the one everybody tells first, because it's the one that makes you gasp.

Why It Matters

The reason this problem isn't just a historical curiosity is that it turned out to be the simplest visible tip of an enormous iceberg: the Riemann zeta function.

Define ζ(s) = 1/1^s + 1/2^s + 1/3^s + ... for any number s greater than 1. The Basel problem is just the special case s = 2, so in this notation, Euler proved ζ(2) = π²/6. Euler didn't stop there — he went on to find exact values for ζ(4) = π⁴/90, ζ(6) = π⁶/945, and every other even integer, always landing on some rational multiple of a power of π. (The odd cases, like ζ(3), remain stubbornly mysterious to this day — we don't even know whether ζ(3) can be written as a simple fraction times a power of π, though we do know it's irrational.)

A century later, Bernhard Riemann extended this same function to complex numbers — numbers with both a real and an imaginary part — and discovered that its behavior encodes deep secrets about the distribution of prime numbers. The locations where this extended zeta function equals zero are the subject of the Riemann Hypothesis, widely considered the single most important unsolved problem in mathematics, with a million-dollar Clay Institute prize attached to it. In a very real sense, the question a 17th-century Italian monk-turned-mathematician scribbled down about adding fractions is a direct ancestor of the question that, if answered, would reshape our understanding of how prime numbers are scattered among the integers.

There's also a wonderfully concrete, almost magical consequence of the Basel problem that has nothing to do with primes: if you pick two random positive integers, the probability that they share no common factor — that they are "coprime" — approaches exactly 6/π², or about 60.8%, as the range of numbers you're picking from grows large. Since 6/π² is just the reciprocal of Euler's π²/6, this means that if you ever wanted to estimate π experimentally, you could do it by repeatedly picking random pairs of integers and checking whether they're coprime, then taking the reciprocal of that fraction and multiplying by 6. It's a strange, roundabout way to "measure" π using nothing but whole numbers and a coin-flip-style random process, and it works precisely because of the Basel problem.

The Details

It helps to actually see why squaring matters here. If instead of squares you use the plain reciprocals — 1 + 1/2 + 1/3 + 1/4 + ... — you get the harmonic series, which diverges: it grows without bound, just excruciatingly slowly (it takes roughly e^100 terms to pass 100). The difference between that series and the Basel sum is the exponent. Squaring the denominators makes each term shrink fast enough that the total sum is trapped — bounded above by a geometric-style argument — while the harmonic series' terms shrink just slowly enough to let the sum escape to infinity. The Basel problem lives exactly on the boundary where things start converging, which is part of why it was such a natural and tempting target.

Picture the terms geometrically: imagine stacking rectangles of width 1 and heights 1, 1/4, 1/9, 1/16, and so on, side by side. The combined area of an infinite number of these rectangles is finite, and Euler showed that finite area is a precise, specific multiple of π² — specifically one-sixth of it. If you wanted a visual intuition for why a circle constant sneaks into a sum of straight-edged rectangles, one way to think about it is that the sine function itself is fundamentally about circular motion — it's the height of a point moving around a unit circle — so any identity built from the zeros of sin(x) is, underneath the hood, an identity about circles. Euler's sum, in disguise, is a statement about where the sine wave crosses zero.

It's also worth appreciating just how slowly the raw series converges, because it explains why nobody could just "compute their way" to the answer. To get 10 correct decimal digits of π²/6 by brute-force summation, you'd need on the order of 10 billion terms. Euler, famously, also derived accelerated methods for approximating such series far faster — he used one to compute the sum to several decimal places by hand before he had the exact closed-form answer, which is how he first became confident the answer involved π² in the first place: he recognized the decimal expansion 1.644934... as suspiciously close to π²/6 and then went looking for the proof that it was exact.

Takeaways

The Basel problem — find the exact sum of 1/1² + 1/2² + 1/3² + ... — stumped Europe's best mathematicians, including Jacob Bernoulli, for 85 years after Pietro Mengoli posed it in 1650.
Leonhard Euler solved it in 1734 at age 28, showing the sum equals π²/6 — a startling appearance of a circle constant inside a sum of nothing but whole numbers.
Euler's original proof used an audacious, not-yet-rigorous trick: treating the infinite Taylor series for sin(x)/x as if it were a polynomial that could be factored using its known zeros at every multiple of π.
The result is the first visible instance of the Riemann zeta function, ζ(2), whose extension to complex numbers underlies the still-unsolved Riemann Hypothesis and its deep connections to the distribution of primes.
A striking byproduct: the probability that two random positive integers are coprime is exactly 6/π², linking this 17th-century puzzle to probability theory and even to experimental ways of estimating π.

Resources: - Basel problem — Wikipedia - Euler's Solution to the Basel Problem — ScienceBlogs