Why Does 1+2+3+... Equal −1/12?

Pose the driving question, show the partial sums S_N = N(N+1)/2 racing toward infinity, and announce the claim that the full sum equals −1/12. Set up the tension between common sense and the formula.

• Partial sums S_N grow without bound
• Yet Ramanujan, Euler, and modern physics assign the value −1/12
• Question: in what sense can this be true?

Ask the learner to commit to a single answer before any explanation: infinity, −1/12, or 0. This records their intuition and primes them to notice which definition of 'sum' supports it.

• Commit to one interpretation
• Notice whether the chosen value comes from ordinary summation or a regularized one

A simulation that displays two series — 1 − 1 + 1 − 1 + … (Grandi's series) and 1 − 2 + 3 − 4 + … — and lets the learner try two rearrangements: Abel summation (take the average of partial sums) and Euler summation (multiply by x, let x → 1). The widget shows how each rule produces 1/2 and 1/4 respectively, with the partial sums visible as a bar chart.

• Different summation rules give different finite values
• Abel summation of Grandi's series gives 1/2
• Euler summation of 1 − 2 + 3 − 4 + … gives 1/4

Walk through Euler's manipulation: starting from ζ(s) = Σ 1/n^s, write 1 − 2 + 3 − 4 + … = (1 − 2^(1−s)) ζ(s), let s → 0, and obtain 1 − 2 + 3 − 4 + … = 1/4. Then take one more step: subtract to isolate 1 + 2 + 3 + ….

• Euler's identity links alternating sums to the zeta function
• ζ(0) = −1/2 enters the picture
• From (1 − 2 + 3 − …) − 4·(1 + 2 + 3 + …) = 1 − 2 + 3 − …, solving gives 1 + 2 + 3 + … = −1/12

A visualization3d widget that plots ζ(s) as a surface over the complex plane, marks the pole at s = 1, and traces the analytic continuation to s = −1. The learner drags a slider for s along the real axis and watches the real part of ζ(s) cross zero at s ≈ −0.295 and land exactly at −1/12 at s = −1.

• ζ(s) is defined by a sum only for Re(s) > 1
• Analytic continuation gives a unique value at every other s
• At s = −1 the analytically continued value is −1/12

State clearly what the equality does and does not claim: ordinary arithmetic is intact, the partial sums still diverge, and 'equals' here means 'has the same regularized value as.' Show that switching to other summation rules (Cesàro, Borel, Abel) for 1+2+3+… generally yields different — or still divergent — answers, except for those that match Ramanujan/analytic continuation.

• Equality is conditional on a summation method
• Cesàro and ordinary limits still diverge for 1+2+3+…
• Analytic continuation is privileged because it is the unique holomorphic extension

A simulation widget that applies Ramanujan summation to a user-chosen series — 1 + 2 + 4 + 8 + … (geometric), 1 + 1 + 1 + …, or 1 + 4 + 9 + 16 + … — and reports the regularized value ζ(0), ζ(−1), or ζ(−2). The learner sees that the same machinery returns finite answers (−1/2, −1/12, 0) for series that all 'diverge' in the ordinary sense.

• The method generalizes: ζ(−n) gives Ramanujan sums of p-series
• 1 + 1 + 1 + … regularizes to −1/2
• 1 + 4 + 9 + 16 + … regularizes to 0

Resolve the driving question: 1 + 2 + 3 + … is not a statement about ordinary arithmetic. Under analytic continuation of ζ(s) — equivalently, Ramanujan summation — the series evaluates to ζ(−1) = −1/12. The original partial sums still grow like N²/2; the finite value is the unique number that makes the zeta function a holomorphic function on ℂ.

• Partial sums diverge; the equality is regularized
• The value is ζ(−1) = −1/12 by analytic continuation
• Euler's classical rearrangement is a special case of this continuation