Chapter IX · The logarithmic, exponential and circular functions of a real variable
The exponential function
defined for every real as the inverse of ; it equals its own derivative, satisfies , and outgrows every power of .
Why this matters
We know only for rational . Rather than extend the definition of powers piece by piece, Hardy turns the logarithm round: is the number whose logarithm is . Every property of the exponential then follows from a property of the logarithm: the derivative, the addition rule, the speed of growth. This is the function at the heart of differential equations, probability and complex analysis, and §212 and §213 lay its foundation.
The idea
The logarithm is a one-to-one correspondence between the positive numbers and all real numbers : it increases strictly, and it runs from (as ) to (as ) without jumps. So every real is the logarithm of exactly one positive . Call that "". Its graph is the graph of reflected in the diagonal .
Try this. Drag along the logarithm. Watch its mirror image on the exponential curve and compare the slopes of the two tangents. Then put at and at .
- y = log x
- x = e^y, drawn as the mirror image
- mirror line
- P (drag) and the two tangents
What to notice
Reflection swaps horizontal and vertical, so the slope at is the reciprocal of the slope at : . But is precisely the height of . So at every point the exponential's slope equals its height, . At , with slope ; at , .
The speed of the exponential is the logarithm’s slowness seen in the mirror. The previous lessons showed that loses to every power of ; substituting turns ” is negligible compared with ” into ” is negligible compared with ”: the exponential beats every power of .
Try this. Set and find where finally overtakes . Then try and (widen the view).
- last crossing
- read-out point
What to notice
On these semi-logarithmic axes is a straight line and a flattening curve: the same picture as the race between and , one level up. leads until , but the line always wins. The ratio is largest at .
Precise statements
For every real , (also written ) is the unique positive number such that . Thus
is positive, continuous and strictly increasing on the whole real line; as and as ; ; and for rational , is the positive -th power of the number .
For all real and , . In particular .
For every , however large, as . More generally tends to if and to if , in each case faster than any power of .
This adds a new end to the scale of infinity: , where means and means .
Proofs
Show that maps one-to-one onto the real line; then the inverse exists, is increasing, and inherits continuity. Hardy gives a direct proof of the continuity, which we follow.
Onto
Given , the limits () and () give with . By the intermediate value theorem for some in .
One-to-one, and the inverse increases
is strictly increasing, so the is unique; and if then , since otherwise would give .
Continuity: write ,
Then . We must show that when .
Bound by
Use with . If : , which rearranges to when . If : , so . Either way for , which tends to with .
Values
because . For rational , the previous lesson showed , so by uniqueness that power is . Since is increasing and takes every positive value, it is unbounded above, so as ; and it takes values arbitrarily close to , so as .
Differentiate with respect to and invert: the derivative of an inverse function is the reciprocal of the derivative.
Difference quotients are reciprocals
With and as above (, so ): .
Let
By continuity, , so the denominator tends to . Hence .
By the chain rule, .
Translate into logarithms, where the statement becomes .
Take logarithms of the right-hand side
, by the functional equation of the logarithm.
Invert
So is the positive number whose logarithm is , which by definition is . With : .
Hardy also gives a second proof: the rule is true for rational and by the laws of indices, and both sides are continuous, so it holds in the limit along rational sequences , . This “prove it for rationals, extend by continuity” argument is used again for general powers in the next lesson.
Substitute in the theorem of the third lesson.
The known result
For every , as (the corollary with and power ).
Put
As , , and . So , that is, (the ratio is positive). For the claim is trivial.
If , with . If , .
Worked examples and exercises
Show that if for all , where is a constant, then for a constant .
Hint
Solution
. A function with zero derivative everywhere is constant (mean value theorem), so and .
Show that a differentiable solution of is either identically or of the form .
Hint
Hint
Solution
Putting gives , so or . If then for all .
Suppose . Differentiating in and then setting : . Put . By the previous exercise , and gives . (Hardy writes ; substituting into the equation shows , so the only possibilities are again and .)
Show that as .
Hint
Solution
By the mean value theorem, for some between and . So . As , and, by continuity, . The limit is . (Equivalently: it is the derivative of at .)
Show that , and are positive and strictly increasing for .
Hint
Hint
Solution
First. : and for , since increases and . So increases strictly on and is positive for .
Second. : , for .
Third. : , and
which is positive for by the second part.
Show that as for every polynomial , and that converges for every .
Hint
Solution
Each term by the theorem, so the finite sum does too.
For the integral: , so for , i.e. there. Since converges and the integrand is positive and continuous on , the integral converges by comparison (§184). (Its value is , as integration by parts shows.)
Check yourself
What is the slope of the graph of at the point where ?
The slope of equals its value, . The reciprocal is the slope of the logarithm at the mirror point .
for every real , including large negative .
True. By definition is a number in the domain of , which consists of positive numbers. Large negative correspond to close to , never or negative.
Beyond which (approximately) is for good? Use the race above, or solve .
Taking logs, means . The function is smallest at and has its last zero near . The same number appeared in the third lesson, in the race between and : putting turns one race into the other.
Which of these tends to infinity fastest as ?
Compare logarithms: , , and . The last two are the largest; and , so wins (eventually: only beyond ). Comparing logarithms is the standard way to place functions on the exponential end of the scale.
Hardy’s language
Summary
- is the positive number whose logarithm is ; its graph is the mirror image of the logarithm’s in .
- It is positive, continuous and strictly increasing from to , and agrees with for rational exponents.
- (reciprocal slopes under reflection) and (from ).
- beats every power of : substitute in ” loses to every power of ”. This gives the exponential end of the scale of infinity.