Why Does 1 + 2 + 3 + 4 + … = −1/12? Ramanujan Summation Explained

Cambridge, January 1913. A parcel from distant Madras arrives on the desk of 36-year-old G. H. Hardy. The sender is an obscure 25-year-old clerk named Srinivasa Ramanujan. Inside: nine pages of formulas without proofs. Among them, a statement worthy of sending its author “to a lunatic asylum” — as Ramanujan himself had half-jokingly anticipated:

“1 + 2 + 3 + 4 + … = −1/12.”

Hardy calls for Littlewood. It is night. Trinity College. The table is covered in rows of paper and notebooks. What lies before them is not merely a set of strange equalities, but a cascade of formulas spanning number theory, elliptic functions, and q-series — all without the proofs Cambridge expects. Just pure, audacious mathematics.

The Letter That Changed Mathematics

But what to make of “1 + 2 + 3 + … = −1/12”? In the ordinary sense, the series diverges to infinity. And yet the computational traces in the letter point to a rigorous core: the Euler–Maclaurin formula, the Bernoulli numbers, careful correction terms for partial sums — in a word, regularization.

By morning, Hardy delivers his famous verdict: “This could only have been written by a mathematician of the highest class.” Ramanujan is invited to Cambridge. And so begins what we now call Ramanujan summation — a method that assigns finite values to divergent series when this assignment is consistent with deep structures of mathematical analysis.

What Is Ramanujan Summation?

In brief: it is a rigorous way to “assign” a value to a divergent series by smoothing its partial sums and applying the Euler–Maclaurin formula with Bernoulli numbers. The method is consistent with the analytic continuation of the Riemann zeta function and, across a broad class of examples, yields the same results — including −1/12.

When we try to “add” 1 + 2 + 3 + … in the ordinary sense, the partial sums race off to infinity — the series diverges. Ramanujan’s proposal: instead of chasing the limit with a running sum, proceed differently.

Step-by-Step: How the Summation Works

Step 1. Look at the partial sum function.

S(N) = ∑_{n=1}^{N} f(n), where f(n) = n.

A few values:

• S(1) = 1
• S(2) = 3
• S(3) = 6
• S(4) = 10

In general: S(N) = N(N+1)/2 = (1/2)N² + (1/2)N — parabolic growth.

Step 2. Smooth the steps with a soft weight.

Instead of a hard cutoff at N, introduce damping:

S(ε) = ∑_{n=1}^{∞} n · e^(−εn), where ε > 0.

Closed form: S(ε) = e^(−ε) / (1 − e^(−ε))².

For any fixed ε, this sum is finite.

Step 3. Expand asymptotically via Euler–Maclaurin.

Variant A — expanding S(N) for f(n) = n:

• Integral term: ∫₀ᴺ x dx = (1/2)N²
• Boundary correction: (1/2)f(N) − (1/2)f(0) = (1/2)N
• Next correction with B₂ = 1/6 yields the constant −1/12, with further terms of order O(1/N²)

Result: S(N) = (1/2)N² + (1/2)N − 1/12 + O(1/N²).

Variant B — expanding S(ε) for small ε:

S(ε) = 1/ε² − 1/12 + O(ε²).

Here 1/ε² is the divergent part, and the constant −1/12 is the “renormalized” remainder.

Step 4. Take the constant term as the renormalized value of the sum.

• From S(N): extract the constant in the asymptotic expansion: −1/12.
• From S(ε): subtract 1/ε² and take the limit ε → 0⁺ — the result is −1/12.

The compact notation is:

1 + 2 + 3 + 4 + … = −1/12 (R)

where (R) denotes “in the sense of Ramanujan summation” — a regularized value, not an ordinary sum.

The Constant Term and Its Meaning

The key idea is that the constant term — the part that neither grows with N nor vanishes like 1/N, 1/N², … — is declared to be “the sum in the sense of Ramanujan summation.” Everything else is divergent noise that the method systematically discards.

This is not a trick or a paradox. It is a precisely defined procedure with well-understood properties: it is linear, compatible with natural transformations on series, and consistent across a wide family of functions.

Why Ramanujan Summation of Divergent Series Matters

The method yields the same values as zeta-function regularization across large classes of functions. This is precisely why −1/12 appears not only in number theory, but also in calculations of quantum vacuum fluctuations and in string theory — wherever a consistent procedure for “removing infinities” is required.

When a physicist writes 1 + 2 + 3 + … = −1/12 in a research paper, they are not making an arithmetic error. They are invoking a regularization scheme — and Ramanujan summation is one of the most elegant such schemes ever devised, born from nine handwritten pages sent from Madras to Cambridge over a century ago.