Absolute convergence and the two certificates for alternating harmonic

On this page
  1. Goal and scope
  2. Givens and symbols
  3. Rules and sources
  4. Main derivation
  5. Substeps and lemmas
  6. Reverse audits
  7. Error controls and negative controls
  8. Result and limitations
  9. Verification and source binding

← Guide index · Original MKO

derivation-mko-absolute-convergence-of-series-01 · ocme-new142-reader-adoption-20261021

Goal and scope

Distinguish absolute convergence of a negative-ratio geometric series from conditional convergence of alternating harmonic.

Exact geometric tails and harmonic blocks; conditional convergence needs both signed convergence and absolute divergence.

Givens and symbols

  • algebra
  • epsilon
  • monotone_complete
  • definition_here
  • G1 — a_n=(-1/2)^n for n≥0; b_n=(-1)^{n+1}/n for n≥1.
  • B_N — alternating-harmonic partial sum
  • B — proved existing limit

Rules and sources

  • mko_definition · cited — Geometric series use |r|<1 for absolute convergence. The alternating harmonic example combines an alternating-remainder certificate of convergence with harmonic divergence of the absolute series. Bounded terms, terms tending to zero, or the absolute value of a partial sum do not establish absolute convergence. · mko-absolute-convergence-of-series
  • algebra · prerequisite — Use real-field identities; division requires a nonzero denominator and multiplication by a positive number preserves inequalities.
  • epsilon · prerequisite — Convergence requires for every positive epsilon an integer N controlling every n≥N by a strict error bound; use the Archimedean property of the reals.
  • monotone_complete · prerequisite — Every bounded monotone real sequence converges; partial sums of a nonnegative series are nondecreasing.
  • definition_here · prerequisite — Absolute convergence means convergence of the termwise-absolute-value series; conditional convergence means signed convergence without absolute convergence.

Main derivation

S01 · equivalent

Before

|a_n|=2^{-n}

After

\sum_{n=0}^{N}|a_n|=2-2^{-N}

Conditions: none

Rule: algebra · Uses: G1

Leaf basis: algebra

The finite geometric subtraction identity computes every absolute partial sum.

S02 · implies

Before

2^{-N}\le1/(N+1)\to0

After

\sum_{n=0}^{\infty}|a_n|=2

Conditions: none

Rule: epsilon · Uses: S01

Leaf basis: epsilon

Induction proves 2^N≥N+1, so the whole remainder tail vanishes and the a series converges absolutely.

S03 · implies

Before

B_{2m}=\sum_{n=1}^{2m}b_n

After

B_{2(m+1)}-B_{2m}=1/(2m+1)-1/(2m+2)>0

Conditions: none

Rule: algebra · Uses: G1

Leaf basis: algebra

Even partial sums increase because each paired difference is positive.

S04 · implies

Before

B_{2m+1}=B_{2m}+1/(2m+1)

After

0\le B_{2m}<B_{2m+1}\le1

Conditions: none

Rule: algebra · Uses: G1, S03

Leaf basis: algebra

Odd partial sums start at 1 and decrease by a positive difference; the two parity sequences bound each other.

S05 · implies

Before

B_{2m}\uparrow,\ B_{2m+1}\downarrow,\ B_{2m+1}-B_{2m}=1/(2m+1)\to0

After

B_N\to B,\quad |B-B_N|\le1/(N+1)

Conditions: none

Rule: monotone_complete · Uses: S03, S04

Leaf basis: monotone_complete

Both bounded monotone parity subsequences have limits; their vanishing gap makes the limits equal. The common limit lies between adjacent parity sums, giving the alternating remainder bound.

S06 · implies

Before

\sum_{n=2^{j-1}+1}^{2^j}|b_n|

After

\sum_{n=2^{j-1}+1}^{2^j}1/n\ge2^{j-1}/2^j=1/2

Conditions: j\ge1

Rule: algebra · Uses: G1

Leaf basis: algebra

Each dyadic block has 2^{j-1} terms, each at least 1/2^j.

S07 · implies

Before

\sum_{n=1}^{2^m}|b_n|\ge1+m/2

After

\sum_{n=1}^{\infty}|b_n|\text{ diverges}

Conditions: none

Rule: algebra · Uses: S06

Leaf basis: algebra

The number of blocks m is arbitrary, making absolute partial sums unbounded.

S08 · definition

Before

\sum a_n\text{ absolutely convergent};\quad\sum b_n\text{ convergent but not absolutely}

After

\sum b_n\text{ conditionally convergent}

Conditions: none

Rule: definition_here · Uses: S02, S05, S07

Leaf basis: definition_here

Both obligations for conditional convergence have independent proofs; absolute convergence implies signed convergence through the Cauchy tail triangle bound.

Substeps and lemmas

No substeps in this guide.

Reverse audits

  • S08 · premise_audit · S02, S05, S07 — Bounded terms, vanishing terms and alternation alone have not been promoted to certificates.
  • S07 · premise_audit · S06 — Bounded signed partial sums do not imply bounded absolute partial sums.
  • S06 · premise_audit · G1 — The lower bound holds for every integer j≥1, not observed blocks only.
  • S05 · premise_audit · S03, S04 — The bound follows from bracketing, without substituting an unknown closed form for a proof.
  • S04 · premise_audit · G1, S03 — Alternation is supported by decreasing 1/n magnitudes, not arbitrary signs.
  • S03 · premise_audit · G1 — Use m≥1 or initialize B_0=0 consistently.
  • S02 · premise_audit · S01 — Absolute convergence is not inferred merely from alternating signs.
  • S01 · inverse_equivalence · G1 — Take absolute values termwise, not only after summing signed terms.

Error controls and negative controls

Two certificates

Conditional convergence needs signed convergence and absolute divergence.

Whole tails, not signs

Use full monotonicity and block formulas for every term.

  • Alternation guarantees absolute convergence. → reject — The series of termwise absolute values of the alternating harmonic series was proved divergent.
  • |sum b_n| equals sum |b_n|. → reject — Already |1-1/2|=1/2 while 1+1/2=3/2.
  • Terms tending to zero suffice for absolute convergence. → reject — 1/n tends to zero while block lower bounds force divergence.

Result and limitations

The (-1/2)^n series is absolutely convergent; alternating harmonic is conditionally convergent.

    Verification and source binding

    1 declared checks for this guide; only the listed expressions are covered.

    • source commit: b1b2f1dcf8c50d67a93fe52d95f8eef7b3abba9f
    • adopted source SHA-256: b6518fa48ca2c7fd98dd67f837b61fe5b8969690967db8d9f312d7b22b2fbbff
    • original source SHA-256: 37c8ac964ebcc44af448259b5824d6a85f22b8a877f01fcf355200915fc5e31f
    • Lean: NOT_RUN · independent AI: SCOPED_LEAD_AND_BILINGUAL_CONTEXT_REVIEW

    Same-source full JSON (formulas, ASTs, controls and revision log)