Chapter V · Limits of functions of a continuous variable. Continuous and discontinuous functions
Continuous functions are bounded and attain their bounds
On a closed interval a continuous function is bounded and takes its greatest and least values; three proofs, the oscillation of a function in an interval, and why each hypothesis is needed.
Why this matters
Does a function have a largest value? For on the answer is yes; for on it is no, even though the values come as close to as we like. In §§103–105 Hardy proves that a function continuous on a closed interval is bounded and attains its upper and lower bounds. This is the theorem behind every “find the maximum” problem, behind Rolle’s theorem and the mean value theorem in Chapter VI, and (with the oscillation introduced here) behind the existence of the integral in Chapter VII.
The idea
For any function defined on , its values form a set of numbers, and we can ask for that set’s upper bound (the least number that no value exceeds) and lower bound , as in §80 for sequences. Two things can go wrong.
- The values may be unbounded, as for on : no exists at all.
- may exist but not be attained: on comes arbitrarily close to (just left of ) but is never equal to , since .
Hardy’s theorems say that neither can happen to a continuous function on a closed interval. The difference , the oscillation of on the interval, will measure how much varies there.
Try this. Drag the two purple handles (or use the sliders) to choose an interval , and read off , and the oscillation. Do it first for the continuous function. Then choose and make the window contain the integer or . Then choose and make the window contain .
- y = φ(x)
- the interval [c, d] (drag its ends)
- from m(c, d) up to M(c, d)
- a bound that is not attained
What to notice
For the continuous function both bounds are always attained, at the green dots. For , as soon as the window contains an integer to the right of , the upper bound becomes and is not attained: the hollow orange dot marks the place it is approached. For , every window containing , however small, has , and oscillation : continuity fails at in a way no shrinking can cure.
Precise statements
Let be defined on . It is bounded above if there is a with for all in . Then it has an upper bound : no value of exceeds , but every number less than is exceeded by some value. Bounded below and the lower bound are defined in the same way, and is bounded if it is bounded above and below.
If is continuous throughout , then it is bounded on .
If is continuous throughout , with upper and lower bounds and , then there are points and in with and .
Together with the intermediate value theorem, this says that the image of under a continuous is exactly the closed interval .
To compare intervals, Hardy writes the bounds of on as and .
For bounded on , the oscillation of in is
Let .
- is the greater of and , and is the lesser of and .
- increases and decreases as increases; so increases with .
- .
Proofs
Hardy proves Theorem 1 with a section, just as he proved the intermediate value theorem: look for the first point where “bounded so far” fails, and show there is none.
Put in if is bounded on , and in otherwise. We show that is empty. If not, is a section with a number ; but continuity at makes bounded on both sides of , which pushes further right. Contradiction.
contains points to the right of
By continuity at (from the right), there is an interval on which . So is bounded on for every .
Suppose is not empty
If is bounded on it is bounded on for . So every member of is less than every member of , and is a section with some number , .
Continuity at gives boundedness near
There is with on (only if ). We may take .
Contradiction
is in , so is bounded on ; it is also bounded on ; hence on . So is in , although it is to the right of and should be in . (If , the same argument shows is bounded on , so and is empty after all.) Hence is empty and is bounded on .
Hardy’s proof of Theorem 2 is short and indirect: it turns “does not attain ” into “is unbounded”, and applies Theorem 1.
If never equals , then is a positive continuous function whose values come arbitrarily close to , so its reciprocal is continuous but unbounded on . That contradicts Theorem 1.
gets arbitrarily small
By the definition of , for every there is an in with , that is .
Suppose for every
Then everywhere on , and is continuous on (a quotient of continuous functions whose denominator never vanishes).
is unbounded
At the points of the first step, . As is arbitrary, has no upper bound, contradicting Theorem 1 applied to .
Conclude
So for some . Applying this to , whose upper bound is , gives a point where .
Hardy calls this proof indirect and, “in view of the great importance of the theorem”, gives two more in §105. One uses a section again: let be the boundary between the points with and those with . Then every interval round has upper bound , so takes values as close to as we please near , and continuity gives . The other is repeated bisection, and it shows clearly where continuity is used.
Try this. Press “Halve” and watch which half is kept: the one whose own upper bound is still . Jump to step 20 and read off and . Then switch to on and do the same.
- y = φ(x)
- the upper bound M
- current interval: its upper bound is still M
What to notice
For the continuous function the intervals close on the maximum point and . For the construction runs exactly the same way, every interval has upper bound , and the intervals close on ; but . The bisection only finds a point near which comes close to . Continuity at that point is what turns “close to nearby” into “equal to here”.
Use property (1) to write and in terms of the two halves, and use the one point that the halves share to link them.
Name the four bounds
Let be and in some order, and be and . By (1), and .
The shared point gives
belongs to both intervals, so (it is at most both upper bounds) and (at least both lower bounds). Hence .
Add
, since the two oscillations on the right are and , whose sum equals .
Worked examples and exercises
Let for and . Show that has neither an upper nor a lower bound on . If instead for and , show that the lower bound on is (attained) and there is no upper bound. Which hypothesis of the theorems fails?
Hint
Solution
. and for every positive integer , so no bounds above or below.
. Every value is and , so the lower bound is and is attained. But is unbounded.
In both cases is defined everywhere on the closed interval but is discontinuous at (an infinity), so Theorem 1 does not apply.
Give a function continuous on the open interval that is unbounded there, and one that is bounded but attains neither bound. Why do the proofs above fail?
Hint
Solution
is continuous on and unbounded. is continuous and bounded on , with bounds and , attained nowhere in the open interval.
In the proof of Theorem 1 the first step needs continuity at the end point to start the class , and the section argument needs continuity at even when . On there is no point or to be continuous at; the “bad” point has been removed from the interval. That is why the theorems are always stated for a closed interval .
Show that has lower bound and upper bound on , that it equals at and , but that it never equals .
Hint
Solution
On , , and at it is . So its values are exactly . No value exceeds , and any number is exceeded (by the value at ), so the upper bound is ; it is not a value. The lower bound is attained at and at . The function is discontinuous at , so this does not contradict Theorem 2.
Let for irrational , and when is a fraction in lowest terms with . Show that on any interval with , has lower bound and no upper bound.
Hint
Solution
Every value is and irrational points give , so the lower bound is (attained). Fix . A fraction in with has , so for each there are at most choices of ; hence only finitely many fractions in have denominator at most . But contains infinitely many rationals, so it contains one, in lowest terms, with , where . As is arbitrary, is unbounded above on , however short the interval. (This is discontinuous everywhere.)
Check yourself
Which function is guaranteed by Theorem 2 to attain a greatest value?
Theorem 2 needs continuity on a closed interval. is continuous everywhere and is closed. on is unbounded (the interval is not closed at ); on attains neither bound; is discontinuous at and and never reaches its upper bound .
What is the oscillation of in the interval ?
On , has least value (at ) and greatest value (at ). . Note that the least value is not at an end point.
If is bounded on , then attains its upper bound somewhere in .
False: on is bounded and never reaches its upper bound . Boundedness gives ; continuity is what makes a value.
For on with , compare with .
and : equality. For a function that goes up and comes back down, such as on , the inequality is strict: . Property (3) never allows to be the larger.
Hardy’s language
The proof just given is indirect, and it may be well, in view of the great importance of the theorem, to indicate alternative lines of proof.
Summary
- A function continuous on a closed interval is bounded (Theorem 1) and takes its upper and lower bounds as values (Theorem 2).
- Both hypotheses are needed: on is unbounded, on never reaches .
- Proofs: a section (“bounded so far” cannot stop), an indirect argument via , and bisection towards a point where comes close to , which continuity turns into .
- The oscillation grows with the interval and satisfies .