Kronecker’s theorem
1 Equivalent statements of Kronecker’s theorem
We shall now give two statements of Kronecker’s theorem, and prove that they are equivalent before proving that they are true.
Theorem 1.
If are real numbers that are linearly independent over , are real numbers, and and are positive real numbers, then there are integers and such that for ,
Theorem 2.
If are real numbers that are linearly independent over , are real numbers, and and are positive real numbers, then there is a real number and integers such that for ,
We now prove that the above two statements are equivalent.11 1 K. Chandrasekharan, Introduction to Analytic Number Theory, pp. 92–93, Chapter VIII, §5.
Proof.
Assume that Theorem 2 is true and let be real numbers that are linearly independent over , let be real numbers, let and let . Let with . Because are linearly independent over , so are . Using Theorem 2 with instead of , instead of , instead of , applied with
there is a real number and integers such that for ,
and
Then , and for , because ,
Thus for , we have , and for ,
proving Theorem 1.
Assume that Theorem 1 is true. The claim of Theorem 2 is immediate when . For , let be linearly independent over , let be real numbers, and let and be positive real numbers. Let , and because are linearly independent over , so are , and then
are linearly independent over . Applying Theorem 1 with and
we get that there are integers and such that for ,
Let . Then and for ,
and
On the other hand, applying Theorem 1 with and
we get that there are integers and such that for ,
and
For ,
and
Therefore for ,
which proves Theorem 2. ∎
2 Proof of Kronecker’s theorem
We now prove Theorem 2.22 2 K. Chandrasekharan, Introduction to Analytic Number Theory, pp. 93–96, Chapter VIII, §5.
Proof of Theorem 2.
Let be real numbers that are linearly independent over , let be real numbers, and let and be positive real numbers.
For real and ,
For with for , and for , let
Then for ,
Let
and let
which satisfies .
Define by
and let be a positive integer. By the multinomial theorem,
for which
and the number of terms in the above sum is . We can write as
Then
Because are linearly independent over , for it is the case that . Write and
with which
Then for each multi-index ,
(1) |
Suppose by contradiction that
Then there is some and some such that when ,
Thus for a positive integer,
But then by (1),
and then
Let , for which , and so for each positive integer it holds that
(2) |
Now,
In particular,
and because , as , contradicting (2) being true for all positive integers . This contradiction shows that in fact
and because ,
(3) |
Now let . By (3) there is some for which . For , write
It is straightforward from the definition of that
which yields
Because ,
hence
so
Furthermore,
Therefore
hence
For , denote by the distance from to the nearest integer. We check that
Thus, for each ,
We have taken .Take , i.e. , and take to be the nearest integer to , for which , proving the claim. ∎
3 Uniform distribution modulo
For let be the greatest integer , and let , called the fractional part of . For let , which belongs to the set . Let , , be a sequence in , and for let
We say that is uniformly distributed modulo if for each closed rectangle contained in ,
where is Lebesgue measure on : for , .
We have proved that if are linearly independent over , then the sequence is dense in .a It can in fact be proved that is uniformly distributed modulo .33 3 Giancarlo Travaglini, Number Theory, Fourier Analysis and Geometric Discrepancy, p. 108, Theorem 6.3.
4 Unique ergodicity
Let be a compact metric space, let be the Banach space of continuous functions , and let be the space of Borel probability measures on , with the subspace topology inherited from with the weak-* topology.44 4 This is the same as the narrow topology on . One proves that and in are equal if and only if for all . is a closed set in that is contained in the closed unit ball, and by the Banach-Alaoglu theorem that closed unit ball is compact, so is itself compact. , with the weak-* topology, is not metrizable, but it is the case that with the subspace topology inherited from is metrizable.
For a continuous map , define by
for Borel sets in . For in and , by the change of variables theorem we have
which means that , and therefore the map is continuous. We say that is -invariant if . Equivalently, is measure-preserving. We denote by the set of -invariant . The Kryloff-Bogoliouboff theorem states that is nonempty. It is immediate that is a convex subset of . Let converge to some . For we have, because is continuous,
which shows that is -invariant. Therefore is a closed set in , and we have thus established that is a nonempty compact convex set.
A measure is called ergodic if for any with it holds that or . It is proved that is ergodic if and only if is an extreme point of .55 5 Manfred Einsiedler and Thomas Ward, Ergodic Theory with a view towards Number Theory, p. 99, Theorem 4.4. The Krein-Milman theorem states that if is a nonempty compact convex set in a locally convex space, then is equal to the closed convex hull of the set of extreme points of .66 6 Walter Rudin, Functional Analysis, second ed., p. 75, Theorem 3.23. In particular this shows us that there exist extreme points of . Let be the set of extreme points of , and applying the Krein-Milman theorem with , which is a nonempty compact convex set in the locally convex space , we have that is equal to the closed convex hull . That is, is equal to the closed convex hull of the set of ergodic .
Choquet’s theorem77 7 Manfred Einsiedler and Thomas Ward, Ergodic Theory with a view towards Number Theory, p. 103, Theorem 4.8. tells us that for each there is a unique Borel probability measure on the compact metrizable space such that
and for all ,
We have established that contains at least one element. is called uniquely ergodic if is a singleton. If then is an extreme point of , hence is ergodic. If , then for , by Choquet’s theorem there is a unique Borel probability measure on satisfying and
i.e.
which means that . Therefore, is uniquely ergodic if and only if is a singleton. It can be proved that is uniquely ergodic if and only if for each there is some such that
uniformly on .88 8 Manfred Einsiedler and Thomas Ward, Ergodic Theory with a view towards Number Theory, p. 105, Theorem 4.10. This constant is equal to , where .
For a topological group and for , define , which is continuous . For a compact metrizable group, there is a unique Borel probability measure on that is -invariant for every , called the Haar measure on . Thus for each , the Haar measure belongs to , and for to be uniquely ergodic means that is the only element of . For a locally compact abelian group , let be its Pontryagin dual. The following theorem gives a condition that is equivalent to a translation being uniquely ergodic.99 9 Manfred Einsiedler and Thomas Ward, Ergodic Theory with a view towards Number Theory, p. 108, Theorem 4.14.
Theorem 4.
Let be a compact metrizable group and let . is uniquely ergodic if and only if is abelian and for all nontrivial .
Let , let , which is a compact abelian group, and let . For , ,
if and only if if and only if there is some such that . Therefore for , the set is linearly independent over if and only if for , the map , , is uniquely ergodic.