A Converging Approximation Under Debate
A student proposes approximating a quantity with the partial sums S_n = sum_{k=1}^{n} (−1)^{k+1}/k, the alternating harmonic series. A classmate argues that because the terms 1/k shrink to zero, the series must converge - and moreover that reordering the terms cannot change the total, since 'addition is commutative.' The teacher notes the series converges to ln 2 ≈ 0.6931 but that the reordering claim deserves scrutiny. A second student claims that summing the positive terms and negative terms separately gives finite totals that add to ln 2.
1. Assess the convergence of S_n and the claim that terms tending to zero is sufficient.
Read the worked answer
S_n converges by the alternating series test: terms 1/k decrease to 0, so partial sums converge (to ln 2). However, 'terms → 0' alone is not sufficient in general (harmonic series diverges despite terms → 0). The alternating sign is essential. So the classmate's conclusion is right but the stated reason is incomplete.
2. Evaluate the reordering claim and the claim about separating positive and negative terms.
Read the worked answer
The series is conditionally, not absolutely, convergent: sum 1/k diverges. By Riemann's rearrangement theorem, a conditionally convergent series can be reordered to any real sum (or diverge). So 'reordering cannot change the total' is false here. Separating positive terms (sum over odd k of 1/k) and negative terms (sum over even k of −1/k) gives two divergent subseries - their totals are not finite. The second student is wrong: the split sums diverge and do not simply add to ln 2.
3. If the student groups terms in pairs (1/1−1/2)+(1/3−1/4)+…, does the value change?
Read the worked answer
Pairing adjacent terms without reordering preserves the sum for a convergent series: each pair is a partial-sum subsequence limit, still ln 2. Pairing is not a rearrangement (order preserved), so the value is unchanged. Contrast with genuine reordering, which can change the sum.