Chapter IV · Limits of functions of a positive integral variable
Interpolation
Finding a formula in x that agrees with a given function of n is always possible and never unique; what is wanted is a natural one.
Why this matters
The last lesson showed that a function of always extends to functions of , in infinitely many ways. §51 (printed in Hardy’s small type) asks the sharper question that working mathematicians actually ask: is there a good formula in that takes the given values at ? This is the problem of functional interpolation. It explains why and need care, and it leads to one of the most important functions in analysis, the gamma function, which extends .
The idea
Mere existence is cheap. We can always fill in the values between the integers however we like, or even leave them undefined. So the problem only has content when we ask for a formula with good properties: simple, smooth, and agreeing with the “obvious” formula when there is one.
Even then there is never only one answer. If agrees with at every positive integer, so does for every number , because . You can see the whole family at once below.
Try this. With selected, move the slider in both directions. Then do the same for . For , also switch on the dots that show where the formula has a meaning, and move the probe to or .
- the values φ(1), …, φ(5)
- an interpolating curve
What to notice
Every curve passes through the blue dots, but between them the curves disagree as much as you like: the dots cannot choose. For the formula is useless as an interpolation. It is defined only at rationals with odd denominator, and there it is always or , for example , while .
Hardy’s examples show three typical situations.
- The formula already works for every . For or we simply take . But even here, is “almost equally obvious”.
- The formula breaks down, but can be rewritten. makes sense for integers, but does not make sense for or for irrational . Since for every integer , the function solves the problem.
- There is no formula at all. has no meaning unless is a positive integer. Here the search for a good interpolation led to a genuine discovery: the gamma function, with , which you can see in the explorer.
A fourth example sits in between: suggests , which (for ) makes sense when is rational, using the meaning of fractional powers from algebra. But with the tools available at this point of the book, has no meaning at all for irrational . Hardy promises to fill that gap later, and he does, with the exponential function in Chapter IX.
Precise statements
Let be a function of the positive integral variable . A function of the real variable interpolates if is defined at every positive integer and for The problem of functional interpolation is to find such an given by a formula, as simple and natural as possible.
The family of all solutions is easy to describe exactly.
Suppose interpolates . Then a function (defined at least at the positive integers) interpolates if and only if , where is a function with for every positive integer .
Proofs
Look at the difference of two interpolations. It must vanish at the positive integers, and any function that vanishes there may be added.
If interpolates, then vanishes at the integers
For each positive integer , . And .
Conversely, adding such an keeps the values
If for every positive integer , then . So interpolates .
There are infinitely many such
For example for any real , or , or , or any function that is at the integers and anything at all elsewhere. So an interpolation is never unique.
The theorem moves the difficulty to the right place. Uniqueness can only come from extra conditions on , chosen so that the answer is the natural one. For the natural condition is the rule , but the next exercise shows that this rule alone is not enough.
Worked examples and exercises
Using the meaning of fractional powers from algebra, show that is a real number when in lowest terms with odd, and find its value. Why does it have no real value when is even? Conclude that and agree only at the integers among such .
Hint
Solution
For odd , the equation has exactly one real solution, . So , which is for even and for odd . For even , for every real , so no real has , and has no real value.
Now compare with . Its value is only when is an integer. So for a non-integral with odd , lies strictly between and , while . The two agree only at the integers (where ). The formula is therefore no solution of the interpolation problem, and is.
Show that for any finite list there is a polynomial with for , but that no polynomial satisfies for every positive integer .
Hint
Solution
Finite lists. In with (product over ), each is a polynomial with and for the other . Hence for . (This is Lagrange’s interpolation formula.)
The whole sequence. Suppose for every . Then the polynomial vanishes at , infinitely many points. A non-zero polynomial of degree has at most roots, so is the zero polynomial and for every . But then . This contradiction shows no such polynomial exists. The infinite data is a genuinely stronger demand than any finite part of it.
Define for and for (so is defined on all , one unit interval at a time). Show that for every positive integer , and that satisfies , yet is not .
Hint
Solution
, and if then , so for all by induction. The rule holds for by construction. But , while . So . The functional equation pins the function down only once its values on one unit interval are chosen. (The modern theorem of Bohr and Mollerup says that is the only positive solution of , for which is convex.)
Explain why has a meaning for every positive rational , and compute and .
Solution
For , is the -th power of the positive -th root of the positive number , which exists by Chapter I. So and . For irrational , nothing defined so far gives a meaning.
Check yourself
If interpolates a function of , then so does .
True. for every integer , so adding does not change any value at a positive integer. Any function that vanishes at all the integers may be added.
Which function of interpolates and is defined for every real ?
for every integer . The formula has no real value at (or at any irrational ). , not . And gives at but at .
Using real roots, what is ?
is the real cube root of , which is , since . Note that , so and are different functions even where both are defined.
Hardy’s language
What is usually wanted is some formula involving x (of as simple a kind as possible) which assumes the given values for x = 1, 2, ….
Summary
- Interpolating a function of means finding a function of that takes the values at .
- Any two interpolations differ by a function that vanishes at all the integers, so there are always infinitely many (add ).
- A formula that works for integers may fail elsewhere: makes sense only at certain rationals, so is interpolated by instead.
- For the natural interpolation is the gamma function; extra conditions, not the values alone, are what single it out.