Chapter VIII · The convergence of infinite series and infinite integrals
Conditionally convergent series
Series that converge only through cancellation; why both their halves diverge, why comparison tests fail for them, and how rearranging their terms can produce any sum at all.
Why this matters
Some series converge although the series of their moduli does not. The standard example is , which converges (to , as Chapter IX will show), while diverges. §193 and §194 explain what is special about such series: their convergence rests entirely on cancellation, so the familiar tools stop working. You cannot compare them with other series, you cannot pick out some of their terms, and, most strikingly, you cannot reorder them without risking a different sum.
The idea
Recall from the previous lesson that every series is the difference of its positive part and its negative part, . For the positive part is and the negative part is . Both diverge. The series converges only because the order of the terms makes the two infinite supplies cancel each other step by step.
Now imagine you are free to choose the order. You have an unlimited supply of positive terms and an unlimited supply of negative terms, both shrinking to zero. To steer the partial sums towards any number : add positive terms until you are above , then negative terms until you are below it, and repeat. Because both supplies are unlimited, you never get stuck. Because the terms shrink, the overshoots shrink.
Try this. With “chase a target” selected, set the target to , then to , then to , and watch the partial sums of the rearranged series. Try the other two series in the menu as well, and finally build a rearrangement yourself with “by hand”.
- after a positive term
- after a negative term
- original sum 0.69315
- your target
What to notice
For and the partial sums zigzag around any target, closer and closer, and both kinds of term keep being used: every term is used exactly once, so this is a genuine rearrangement with sum . For the absolutely convergent series the same rule gets stuck, because each half has a finite total.
Precise statements
A series is conditionally convergent if it converges but diverges.
Since is a series of positive terms, “diverges” here means . Every convergent real series is therefore either absolutely or conditionally convergent.
If is conditionally convergent, then its positive part and its negative part both diverge to . Its sum is the limit of , a difference of two quantities that both tend to infinity.
The consequences that Hardy draws in §193 and §194:
- Selections fail. A series made of some of the terms need not converge (the positive terms alone diverge).
- Comparison fails. Knowing and that converges conditionally tells us nothing about . For example converges, while diverges, although each of its terms is numerically smaller. The proof of comparison for positive series needs the terms of the larger series to be positive, and the conclusion only helps when is finite. Any test for conditional convergence must therefore look at how the terms cancel, not just at their sizes.
- Rearrangements fail. The proof of §192 used the convergence of and separately, and now we know both diverge. The theorem itself fails too, as Riemann showed.
If is a conditionally convergent real series and is any real number, the terms of can be rearranged to give a series that converges to . They can also be rearranged so that the new series diverges to , or to .
Hardy only mentions this theorem (referring to Bromwich for the proof), but the proof is just the game above made precise.
Proofs
Rule out the other two possibilities: both halves converging, and exactly one converging.
They cannot both converge
If and both converged, then so would , contradicting conditional convergence.
They cannot converge one at a time
Suppose converges to while diverges. Then , and , so , contradicting the convergence of . The case with and exchanged is the same with signs reversed.
Conclude
Both are series of positive terms, so “not convergent” means “divergent to ”.
List the positive terms and the sizes of the negative terms in their original order. Both lists tend to (because ), and both have infinite sums. Build the new series greedily: positive terms while the partial sum is at most , negative terms while it exceeds .
Each stage ends
Suppose at some stage the partial sum is and we start adding the next unused positive terms. The unused still have an infinite sum (removing finitely many terms from a divergent positive series leaves it divergent), so after finitely many of them the partial sum exceeds . Similarly, a stage of negative terms ends after finitely many terms.
Every term is used
Since each stage ends, the stages alternate for ever, so infinitely many positive and infinitely many negative terms are taken, always in their listed order. Hence every and every is used exactly once. (Terms equal to , if any, can be slipped in one at a time after each stage.) The new series is a rearrangement of .
The overshoots tend to zero
When a positive stage ends, the partial sum has just crossed upwards with the term , so it lies in . During the following negative stage it decreases but stays above until the last step, which leaves it in . So at every moment after the stage ending with (or ), the distance from is at most the size of the term that ended the latest stage.
Conclude
The terms that end the stages are and with , and . Hence the partial sums of the rearranged series tend to .
To diverge to instead, aim at a moving target: take positive terms until the sum exceeds , then one negative term, then positive terms until the sum exceeds , then one negative term, and so on (exercise 3 below).
Rearrangements need not be clever to change the sum. Hardy’s example takes two positive terms of for every negative one: .
Try this. Choose “fixed pattern” with positive and negative term per block, and look at 5000 terms. Then try and , and and . Finally switch the series to with the pattern and .
- after a positive term
- after a negative term
- original sum 0.69315
- limit of this pattern
What to notice
For the alternating harmonic series the pattern positive, negative converges to (a classical formula, quoted here; the exercise below proves the case and ): about for and , which is , and about for and . For the square-root series the pattern and does not converge at all: its partial sums climb to infinity.
Worked examples and exercises
Let . Show that the rearrangement converges to , which equals .
Hint
Hint
Solution
The sum of the first even reciprocals is , and so the sum of the first odd reciprocals is . The first terms of the new series contain the first positive terms and the first negative ones:
Write with . Then
Also and differ from by at most two terms, which tend to , so the whole series converges to . By the same method , so the new sum is . (Hardy reaches the same answer by writing as a sum close to .)
Show that diverges to , although it is a rearrangement of the convergent series .
Hint
Hint
Solution
The -th block satisfies
Since diverges, the partial sums tend to . The partial sums and exceed (they add positive terms), so the whole sequence of partial sums tends to .
Describe a rearrangement of that diverges to , and prove that it does.
Hint
Solution
Take positive terms (in order) until the partial sum exceeds , then the next negative term; then positive terms until it exceeds , then the next negative term; and so on. Each stage ends because the unused positive terms have an infinite sum. Every term is used: there are infinitely many stages, each using at least one positive and exactly one negative term. After the -th stage the sum exceeds (it exceeded before the single negative term, which is at most ), and during stage it only increases. So from the end of stage on, every partial sum exceeds , and the partial sums tend to .
Give a convergent series and a series with for every , such that diverges. Explain which step of the comparison test breaks.
Hint
Solution
Take and : then , converges, and diverges. (Hardy’s version uses , which is even strictly smaller.) The comparison test bounds the partial sums of by those of ; that is useful only when is finite, that is, when converges absolutely. The sizes of the terms of a conditionally convergent series are too large to bound anything.
Try this. Choose , then , and increase .
- positive part
- negative part
- the series
- its sum
What to notice
In each case the blue and ochre curves (the positive and negative parts) rise without bound, while the purple curve, their difference, settles: the theorem “both halves diverge” in a picture.
Check yourself
converges but diverges. What can you say about the series of positive terms ?
If both halves converged, would converge; if only one did, the partial sums of would tend to . A series of positive terms cannot oscillate.
If converges and for every , then converges.
False unless converges absolutely. Take and .
To four decimal places, what is the sum of ?
, half as much again as the sum of the original order.
The terms of can be rearranged so that the new series oscillates between and .
True. Steer the partial sums alternately above and below , using the greedy method with targets . Both supplies of terms are inexhaustible, so every stage ends and every term is used; the lim sup is and the lim inf is .
Hardy’s language
There are no comparison tests for convergence of conditionally convergent series.
Summary
- A conditionally convergent series converges, but its series of moduli diverges.
- Its positive part and negative part both diverge to : the sum exists only through cancellation in the given order.
- So comparison tests, selections and rearrangements all fail for it.
- Riemann: a conditionally convergent series can be rearranged to converge to any number, or to diverge to ; for example sums to .