Absolute and Conditional Convergence: Concepts and Examples

A series Σ an is absolutely convergent when the series of absolute values Σ |an| converges, and conditionally convergent when Σ an converges but Σ |an| diverges. This distinction is one of the most useful ideas in the whole theory of series, because absolute convergence is a much stronger and safer property than ordinary convergence.

The distinction matters because absolutely convergent series behave like finite sums: you can rearrange their terms, group them, and split them into positive and negative parts without changing the total. Conditionally convergent series have none of these guarantees, and rearranging their terms can actually change the sum, a genuinely surprising fact.

Key Concepts

• Absolute convergence: Σ |an| converges. This always implies that Σ an itself converges.
• Conditional convergence: Σ an converges while Σ |an| diverges.
• Proof idea: since 0 ≤ an + |an| ≤ 2|an|, the comparison test shows Σ (an + |an|) converges, and subtracting the convergent Σ |an| leaves Σ an convergent.
• Rearrangement theorem (Riemann): an absolutely convergent series keeps its sum under any rearrangement of terms; a conditionally convergent series can be rearranged to converge to any real number, or to diverge.
• Classic example: the alternating harmonic series Σ (−1)n+1/n converges to ln 2, but Σ 1/n diverges, so the convergence is conditional.

Formulas

• Absolute convergence test: if Σ |an| converges (by comparison, ratio, root or integral test), then Σ an converges. No extra work on signs is needed.
• Alternating harmonic series: Σn=1∞ (−1)n+1/n = ln 2 ≈ 0.6931, conditionally convergent.
• Absolute example: Σ (−1)n/n² converges absolutely because Σ 1/n² converges (p-series with p = 2 > 1); its sum is −π²/12.
• Splitting trick: write pn = max(an, 0) and qn = max(−an, 0); for absolutely convergent series both Σ pn and Σ qn converge.

Worked Example

Classify Σ (−1)n/n², starting at n = 1, as absolutely convergent, conditionally convergent, or divergent.

1. Take absolute values: |(−1)n/n²| = 1/n².
2. The series Σ 1/n² is a p-series with p = 2 > 1, so it converges (by the integral test).
3. Since Σ |an| converges, the original series converges absolutely.
4. Because the convergence is absolute, any rearrangement of its terms still sums to the same value, −π²/12 ≈ −0.8225.

Contrast this with Σ (−1)n+1/n: it converges by the Leibniz test, but Σ 1/n diverges, so its convergence is only conditional and rearrangements can change its sum.

Common Misconceptions

• "Convergent always means absolutely convergent." The alternating harmonic series is the standard counterexample: convergent, but not absolutely.
• "You can always rearrange the terms of a convergent series." Only for absolute convergence. Riemann showed that a conditionally convergent series can be rearranged to hit any target value.
• "The ratio test detects conditional convergence." The ratio test with L < 1 proves absolute convergence; it never proves conditional convergence. Use the Leibniz test for alternating series instead.

Real-World Applications

• Rearranging terms safely in numerical software: only absolutely convergent series can be summed in any convenient order without error.
• Fourier series in engineering: absolute convergence guarantees the series represents a continuous function, which matters for physical signals.
• Probability theory: expected values of random variables with both positive and negative outcomes need absolute convergence for the expectation to be well defined.

Absolute convergence means the series of absolute values converges, which automatically forces the original series to converge and makes rearrangement safe. Conditional convergence, shown by series like the alternating harmonic series, is weaker: the series converges, but only because positive and negative terms cancel, and rearranging it can change everything.