Q
stringlengths
4
3.96k
A
stringlengths
1
3k
Result
stringclasses
4 values
Example 2.37 (Kronecker’s Theorem). The rotation \( \left( {\mathbb{T};a}\right) \) is topologically transitive if and only if \( a \in \mathbb{T} \) is not a root of unity.
Proof. If \( {a}^{{n}_{0}} = 1 \) for some \( {n}_{0} \in \mathbb{N} \), then \( \left\{ {z \in \mathbb{T} : {z}^{{n}_{0}} = 1}\right\} \) is closed and \( \varphi \) -invariant, so by Theorem 2.36, \( \left( {\mathbb{T};a}\right) \) is not transitive.\n\nFor the converse, suppose that \( a \) is not a root of unity an...
Yes
The product of two topologically transitive systems need not be topologically transitive.
Consider \( {a}_{1} = {\mathrm{e}}^{\mathrm{i}},{a}_{2} = {\mathrm{e}}^{2\mathrm{i}} \), and the product system \( \left( {{\mathbb{T}}^{2};\left( {{a}_{1},{a}_{2}}\right) }\right) \) . Then \( M = \left\{ {\left( {x, y}\right) \in {\mathbb{T}}^{2} : {x}^{2} = y}\right\} \) is a nontrivial, closed invariant set which c...
Yes
Lemma 2.40. For \( n \in \mathbb{N} \) one has \( {\left\lbrack {A}^{n}\right\rbrack }_{ij} > 0 \) if and only if there is \( x \in F \) with \( {x}_{0} = i \) and \( {x}_{n} = j \) .
Proof. We argue by induction on \( n \in \mathbb{N} \), the case \( n = 1 \) being clear by the definition of the transition matrix \( A \) . Suppose that the claimed equivalence is proved for \( n \geq 1 \) . The inequality \( {\left\lbrack {A}^{n + 1}\right\rbrack }_{ij} > 0 \) holds if and only if there is \( m \) s...
Yes
Proposition 2.41. Let \( \left( {F;\tau }\right) \) be a subshift of order 2 and suppose that every letter occurs in some word in \( F \) . Consider the next assertions.\n\n(i) The transition matrix \( A \) of \( \left( {F;\tau }\right) \) is irreducible.\n\n(ii) \( \left( {F;\tau }\right) \) is forward transitive.\n\n...
Proof. (i) \( \Rightarrow \) (ii): Since \( F \) is metrizable, we can apply Proposition 2.35. It suffices to consider open sets \( U \) and \( V \) intersecting \( F \) that are of the form\n\n\[ U = \left\{ {x : {x}_{0} = {u}_{0},\ldots ,{x}_{n} = {u}_{n}}\right\} \text{ and }V = \left\{ {x : {x}_{0} = {v}_{0},\ldots...
Yes
Proposition 3.3. For a topological system \( \left( {K;\varphi }\right) \) the following assertions are equivalent:\n\n(i) \( \left( {K;\varphi }\right) \) is minimal.\n\n(ii) \( {\operatorname{orb}}_{ + }\left( x\right) \) is dense in \( K \) for each \( x \in K \) .\n\n(iii) \( K = \mathop{\bigcup }\limits_{{n \in {\...
Proof. This is Exercise 4.
No
Theorem 3.5. Every topological system \( \left( {K;\varphi }\right) \) has at least one minimal subsystem.
Proof. Let \( \mathcal{M} \) be the family of all nonempty closed \( \varphi \) -invariant subsets of \( K \) . Then, of course, \( K \in \mathcal{M} \), so \( \mathcal{M} \) is nonempty. Further, \( \mathcal{M} \) is ordered by set inclusion. Given a chain \( \mathcal{C} \subseteq \mathcal{M} \) the set \( C \mathrel{...
Yes
Proposition 3.6. An isometric topological system \( \left( {K;\varphi }\right) \) is minimal if and only if it is topologically transitive.
Proof. Suppose that \( {x}_{0} \in K \) has dense forward orbit and pick \( y \in K \) . By Proposition 3.3 it suffices to prove that \( {x}_{0} \in {\overline{\operatorname{orb}}}_{ + }\left( y\right) \) . Let \( \varepsilon > 0 \) be arbitrary. Then there is \( m \in {\mathbb{N}}_{0} \) such that \( d\left( {{\varphi...
Yes
Corollary 3.7 (“Structure Theorem” for Isometric Systems). An isometric system is a (possibly infinite) disjoint union of minimal subsystems.
Proof. By Remark 3.2.4, different minimal subsystems must be disjoint. Hence, the statement is equivalent to saying that every point in \( K \) is contained in a minimal system. But for any \( x \in K \) the system \( \left( {{\overline{\operatorname{orb}}}_{ + }\left( x\right) ;\varphi }\right) \) is topologically tra...
Yes
c) By a) and b) a recurrent, but not uniformly recurrent point is easy to find: Let \( k = 2 \), and enumerate the words formed from the alphabet \( \{ 0,1\} \) according to the lexicographical ordering. Now write these words into one infinite word \( x \in {\mathcal{W}}_{2}^{ + } \) :
## \(0\left| 1\right| {00}\left| {01}\right| {10}\left| {11}\right| {000}\left| {001}\right| {010}\left| {011}\right| {100}\left| {101}\right| {110}\left| {111}\right| {0000}\cdots .\) All blocks of \( x \) occur as a sub-block of later finite blocks, hence they are repeated infinitely often, and hence \( x \) is recur...
Yes
Theorem 3.11. Let \( \left( {K;\varphi }\right) \) be a topological system and \( x \in K \) . Then the following assertions are equivalent:\n\n(i) \( x \) is uniformly recurrent.\n\n(ii) \( \left( {{\overline{\operatorname{orb}}}_{ + }\left( x\right) ;\varphi }\right) \) is minimal.\n\n(iii) \( x \) is contained in a ...
Proof. The equivalence of (ii) and (iii) is evident. Suppose that (iii) holds. Then we may suppose without loss of generality that \( \left( {K;\varphi }\right) \) is minimal. Let \( U \subseteq K \) be a nonempty open neighborhood of \( x \) . By Proposition 3.3, \( \mathop{\bigcup }\limits_{{n \in {\mathbb{N}}_{0}}}{...
Yes
Proposition 3.12. a) If a topological system contains a forward transitive uniformly recurrent point, then it is minimal.
Proof. a) and b) are immediate from Theorem 3.11, and c) follows from Theorem 3.11 together with Corollary 3.7.
No
Let \( \alpha \in \lbrack 0,1) \) and \( I \subseteq \lbrack 0,1) \) be an interval containing \( \alpha \) in its interior. Then the set\n\n\[ \{ n \in \mathbb{N} : {n\alpha } - \lfloor {n\alpha }\rfloor \in I\} \]\n\nhas bounded gaps.
Proof. Consider \( \left( {\lbrack 0,1}\right) ;\alpha ) \), the translation mod 1 by \( \alpha \) (Example 2.7). It is isomorphic to the group rotation \( \left( {\mathbb{T};{\mathrm{e}}^{{2\pi }\mathrm{i}\alpha }}\right) \), so the claim follows from Proposition 3.12.
No
Proposition 3.15. Let \( \left( {K;\varphi }\right) \) be a topological system, \( G \) a compact group, and \( \left( {H;\psi }\right) \) the group extension along some \( \Phi : K \rightarrow G \) . If \( {x}_{0} \in K \) is a recurrent point, then \( \left( {{x}_{0}, g}\right) \in H \) is recurrent in \( H \) for al...
Proof. It suffices to prove the assertions for \( g = 1 \in G \) . Indeed, for every \( g \in G \) the map \( {\rho }_{g} : H \rightarrow H,{\rho }_{g}\left( {x, h}\right) = \left( {x,{hg}}\right) \), is an automorphism of \( \left( {H;\psi }\right) \), and hence maps recurrent points to recurrent points. For every \( ...
Yes
Proposition 3.16. Let \( \left( {K;\varphi }\right) \) be a topological system, \( G \) a compact group, and \( \left( {H;\psi }\right) \) the group extension along \( \Phi : K \rightarrow G \) . If \( {x}_{0} \in K \) is a uniformly recurrent point, then \( \left( {{x}_{0}, g}\right) \in H \) is uniformly recurrent in...
Proof. As before, it suffices to prove that \( \left( {{x}_{0}, h}\right) \) is uniformly recurrent for one \( h \in G \) . The set \( {\overline{\operatorname{orb}}}_{ + }\left( {x}_{0}\right) \) is minimal by Theorem 3.11, so by passing to a subsystem we can assume that \( \left( {K;\varphi }\right) \) is minimal. No...
Yes
Corollary 3.17. Let \( \alpha \in \mathbb{R} \) and \( \varepsilon > 0 \) be given. Then there exists \( n \in \mathbb{N}, m \in \mathbb{Z} \) such that\n\n\[ \left| {{n}^{2}\alpha - m}\right| \leq \varepsilon \]
Proof. Consider the topological system \( \left( {\lbrack 0,1}\right) ;\alpha ) \) from Example 2.7, and recall that, endowed with the appropriate metric and with addition modulo 1 , it is a compact group isomorphic as a topological group to \( \mathbb{T} \) . We consider a group extension similar to Example 2.22. Let\...
Yes
Proposition 3.18. Let \( p \in \mathbb{R}\left\lbrack x\right\rbrack \) be a polynomial of degree \( k \in \mathbb{N} \) with \( p\left( 0\right) = 0 \) . Then for every \( \varepsilon > 0 \) there is \( n \in \mathbb{N} \) and \( m \in \mathbb{Z} \) with\n\n\[ \left| {p\left( n\right) - m}\right| < \varepsilon \]
Proof. Start from a polynomial \( p\left( x\right) \) of degree \( k \) and define\n\n\[ {p}_{k}\left( x\right) \mathrel{\text{:=}} p\left( x\right) ,\;{p}_{k - i}\left( x\right) \mathrel{\text{:=}} {p}_{k - i + 1}\left( {x + 1}\right) - {p}_{k - i + 1}\left( x\right) \;\left( {i = 1,\ldots, k}\right) . \]\n\nThen each...
Yes
Lemma 4.1. Let \( A, B \) be disjoint closed subsets of a compact space \( K \) . Then there are disjoint open sets \( U, V \subseteq K \) with \( A \subseteq U \) and \( B \subseteq V \) ; or, equivalently, \( A \subseteq U \) and \( \bar{U} \cap B = \varnothing \) .
Proof. Let \( x \in A \) be fixed. For every \( y \in B \) there are disjoint open neighborhoods \( U\left( {x, y}\right) \) of \( x \) and \( V\left( {x, y}\right) \) of \( y \) . Finitely many of the \( V\left( {x, y}\right) \) cover \( B \) by compactness, i.e., \( B \subseteq V\left( {x,{y}_{1}}\right) \cup \cdots ...
Yes
Theorem 4.3 (Tietze). Let \( K \) be a compact space, let \( A \subseteq K \) be closed, and let \( f \in \mathrm{C}\left( A\right) \) . Then there is \( g \in \mathrm{C}\left( K\right) \) such that \( {\left. g\right| }_{A} = f \) .
Proof. The real and imaginary parts of a continuous function are continuous, hence it suffices to consider the case that \( f \) is real-valued. Then, since \( A \) is compact, \( f\left( A\right) \) is a compact subset of \( \mathbb{R} \), and by scaling and shifting we may suppose that \( f\left( A\right) \subseteq \...
Yes
Theorem 4.4 (Stone-Weierstraß). Let \( A \) be a complex conjugation invariant subalgebra of \( \mathrm{C}\left( K\right) \) containing the constant functions and separating the points of \( K \) . Then \( A \) is dense in \( \mathrm{C}\left( K\right) \) .
For the proof we note that the closure \( \bar{A} \) of \( A \) also satisfies the hypotheses of the theorem, hence we may suppose without loss of generality that \( A \) is closed. The next result is the key to the proof.
Yes
Proposition 4.5. Let \( A \) be a closed conjugation invariant subalgebra of \( \mathrm{C}\left( K\right) \) containing the constant function 1 . Then any positive function \( f \in A \) has a unique real square root \( g \in A \), i.e., \( f = {g}^{2} = \bar{g} \cdot g \) .
Proof. We prove that the square root of \( f \), defined pointwise, belongs to \( A \) . By normalizing first we can assume \( \parallel f{\parallel }_{\infty } \leq 1 \) . Recall that the binomial series\n\n\[ \n{\left( 1 + x\right) }^{\frac{1}{2}} = \mathop{\sum }\limits_{{n = 0}}^{\infty }\left( \begin{array}{l} \fr...
No
Lemma 4.6. Each compact metric space is separable.
Proof. For fixed \( m \in \mathbb{N} \) the balls \( \mathrm{B}\left( {x,\frac{1}{m}}\right), x \in K \), cover \( K \), so there is a finite set \( {F}_{m} \subseteq K \) such that\n\n\[ K \subseteq \mathop{\bigcup }\limits_{{x \in {F}_{m}}}B\left( {x,\frac{1}{m}}\right) \]\n\nThen the set \( F \mathrel{\text{:=}} \ma...
Yes
Theorem 4.7. A compact topological space \( K \) is metrizable if and only if \( \mathrm{C}\left( K\right) \) is separable.
Proof. Suppose that \( \mathrm{C}\left( K\right) \) is separable, and let \( {\left( {f}_{n}\right) }_{n \in \mathbb{N}} \) be a sequence in \( \mathrm{C}\left( K\right) \) such that \( \left\{ {{f}_{n} : n \in \mathbb{N}}\right\} \) is dense in \( \mathrm{C}\left( K\right) \) . Define\n\n\[ \Phi : K \rightarrow \Omega...
Yes
Theorem 4.8. Let \( I \subseteq \mathrm{C}\left( K\right) \) be a closed algebra ideal. Then there is a closed subset \( F \subseteq K \) such that \( I = {I}_{F} \) .
Proof. Define\n\n\[ F \mathrel{\text{:=}} \{ x \in K : f\left( x\right) = 0\text{ for all }f \in I\} = \mathop{\bigcap }\limits_{{f \in I}}\left\lbrack {f = 0}\right\rbrack .\n\]\n\nObviously, \( F \) is closed and \( I \subseteq {I}_{F} \) . Fix \( f \in {I}_{F},\varepsilon > 0 \) and define \( {F}_{\varepsilon } \mat...
Yes
Lemma 4.9. An ideal \( I \) of \( \mathrm{C}\left( K\right) \) is maximal if and only if \( I = {I}_{\{ x\} } \) for some \( x \in K \) .
Proof. It is straightforward to see that each \( {I}_{\{ x\} }, x \in K \), is a maximal ideal. Suppose conversely that \( I \) is a maximal ideal. By Theorem 4.8 it suffices to show that \( I \) is closed. Since \( \bar{I} \) is again an ideal and \( I \) is maximal, \( \bar{I} = I \) or \( \bar{I} = \mathrm{C}\left( ...
Yes
Lemma 4.10. A nonzero linear functional \( \psi : \mathrm{C}\left( K\right) \rightarrow \mathbb{C} \) is multiplicative if and only if \( \psi = {\delta }_{x} \) for some \( x \in K \) .
Proof. Let \( \gamma : \mathrm{C}\left( K\right) \rightarrow \mathbb{C} \) be nonzero and multiplicative. Then there is \( f \in \mathrm{C}\left( K\right) \) such that \( \gamma \left( f\right) = 1 \) . Hence\n\n\[ 1 = \gamma \left( f\right) = \gamma \left( {\mathbf{1}f}\right) = \gamma \left( \mathbf{1}\right) \gamma ...
Yes
Lemma 4.12. Let \( K \) be a compact space, \( \Omega \) a topological space, and let \( \varphi : \Omega \rightarrow \) \( K \) be a mapping. Then \( \varphi \) is continuous if and only if \( f \circ \varphi \) is continuous for all \( f \in \mathrm{C}\left( K\right) \) .
Proof. Clearly, if \( \varphi \) is continuous, then also \( f \circ \varphi \) is continuous for every \( f \in \mathrm{C}\left( K\right) \) . Conversely, if this condition holds, then \( {\varphi }^{-1}\left\lbrack {\left| f\right| > 0}\right\rbrack = \left\lbrack {\left| {f \circ \varphi }\right| > 0}\right\rbrack \...
Yes
Theorem 4.13. Let \( K \), L be (nonempty) compact spaces and let \( T : \mathrm{C}\left( K\right) \rightarrow \mathrm{C}\left( L\right) \) be linear. Then the following assertions are equivalent:\n\n(i) \( T \) is an algebra homomorphism.\n\n(ii) There is a continuous mapping \( \varphi : L \rightarrow K \) such that ...
Proof. Urysohn’s lemma yields that \( \varphi \) as in (ii) is uniquely determined, and it is clear from (ii) that \( \parallel T\parallel = 1 \) . For the proof of the implication (i) \( \Rightarrow \) (ii) take \( y \in L \) . Then\n\n\[ {T}^{\prime }{\delta }_{y} \mathrel{\text{:=}} {\delta }_{y} \circ T : \mathrm{C...
Yes
Lemma 4.14. Let \( K, L \) be compact spaces, and let \( \varphi : L \rightarrow K \) be continuous, with Koopman operator \( T \mathrel{\text{:=}} {T}_{\varphi } : \mathrm{C}\left( K\right) \rightarrow \mathrm{C}\left( L\right) \) . Then the following hold:\n\na) \( \varphi \) is surjective if and only if \( T \) is i...
Proof. This is Exercise 1.
No
Lemma 4.16. Let \( \Omega, K \) be topological spaces, \( K \) compact. For a mapping \( \Phi \) : \( \Omega \times K \rightarrow \mathbb{C} \) the following assertions are equivalent:\n\n(i) \( \Phi \) is continuous.\n\n(ii) For each \( x \in \Omega \) the mapping \( \widetilde{\Phi }\left( x\right) \mathrel{\text{:=}...
Proof. (i) \( \Rightarrow \) (ii): Fix \( x \in \Omega \) and \( \varepsilon > 0 \) . For each \( y \in K \) there are open sets \( {U}_{y} \subseteq \Omega \) and \( {V}_{y} \subseteq K \) with \( x \in {U}_{y}, y \in {V}_{y} \) such that\n\n\[ \left| {\Phi \left( {x, y}\right) - \Phi \left( {{x}^{\prime },{y}^{\prime...
Yes
Theorem 4.17. Let \( \Omega, K, L \) be topological spaces, and suppose that \( K \) and \( L \) are compact. Let \( \Phi : \Omega \times K \rightarrow L \) be a mapping such that for every \( x \in \Omega \) the mapping \( \Phi \left( {x, \cdot }\right) : K \rightarrow L \) is continuous. Let \( {T}_{x} : \mathrm{C}\l...
Proof. Note that \( \Phi \) is continuous if and only if for every \( f \in \mathrm{C}\left( L\right) \) the mapping \( f \circ \Phi \) is continuous. This follows from Lemma 4.12. Hence, the assertion is a direct consequence of Lemma 4.16 above.
No
Lemma 4.18. Let \( \\left( {K;\\varphi }\\right) \) be a topological system with Koopman operator \( T = {T}_{\\varphi } \) and let \( A \\subseteq K \) be a closed subset. Then \( A \) is \( \\varphi \) -invariant if and only if the ideal \( {I}_{A} \) is \( T \) -invariant.
Proof. Suppose that \( A \) is \( \\varphi \) -invariant and \( f \\in {I}_{A} \) . If \( x \\in A \), then \( \\varphi \\left( x\\right) \\in A \) and hence \( \\left( {Tf}\\right) \\left( x\\right) = f\\left( {\\varphi \\left( x\\right) }\\right) = 0 \) since \( f \) vanishes on \( A \) . Thus \( {Tf} \) vanishes on ...
Yes
Lemma 4.20. Let \( \\left( {K;\\varphi }\\right) \) be a topological system with Koopman operator \( T = {T}_{\\varphi } \) on \( \\mathrm{C}\\left( K\\right) \). If \( \\left( {K;\\varphi }\\right) \) is topologically transitive, then \( \\mathrm{{fix}}\\left( T\\right) \) is one-dimensional.
Proof. As already remarked, if \( x \\in K \) and \( f \\in \\operatorname{fix}\\left( T\\right) \), then \( f\\left( {{\\varphi }^{n}\\left( x\\right) }\\right) = \\left( {{T}^{n}f}\\right) \\left( x\\right) = \) \( f\\left( x\\right) \) is independent of \( n \\geq 0 \), and hence \( f \) is constant on \( {\\overlin...
Yes
Theorem 4.21. Let \( \left( {K;\varphi }\right) \) be a topological system with Koopman operator \( T = {T}_{\varphi } \). Then the peripheral point spectrum of \( T \) is a union of subgroups of \( \mathbb{T} \). If \( \operatorname{fix}\left( T\right) \) is one-dimensional, then the peripheral point spectrum is a gro...
Proof. Let \( \lambda \in \mathbb{T} \) be an eigenvalue of \( T \), and let \( 0 \neq f \in \mathrm{C}\left( K\right) \) be a corresponding eigenvector with \( \parallel f{\parallel }_{\infty } = 1 \). Then for each \( n \in \mathbb{N} \) we have that \( {\lambda }^{n} \) is an eigenvalue with eigenvector \( {f}^{n} \...
Yes
In order to determine this subgroup, take \( \lambda \in \mathbb{T} \) and \( \chi \in \mathrm{C}\left( G\right) \) such that \( \left| \chi \right| = \mathbf{1} \) and \( {L}_{a}\chi = {\lambda \chi } \), i.e., \( \chi \left( {ax}\right) = {\lambda \chi }\left( x\right) \) for all \( x \in G \) . Without loss of gener...
\[ \chi \left( {{a}^{n}{a}^{m}}\right) = \chi \left( {a}^{n + m}\right) = {\lambda }^{n + m} = {\lambda }^{n}{\lambda }^{m} = \chi \left( {a}^{n}\right) \chi \left( {a}^{m}\right) \] for all \( n, m \in {\mathbb{N}}_{0} \) . By continuity of \( \chi \) and since the powers of \( a \) are dense in \( G \), this implies ...
Yes
Proposition 4.26. Let \( A \) be a commutative complex Banach algebra. If \( \psi \in \Gamma \left( A\right) \) , then \( \psi \) is continuous with \( \parallel \psi \parallel \leq 1 \) .
Proof. Suppose by contradiction that there is \( a \in A \) with \( \parallel a\parallel < 1 \) and \( \psi \left( a\right) = 1 \) . The series\n\n\[ b \mathrel{\text{:=}} \mathop{\sum }\limits_{{n = 1}}^{\infty }{a}^{n} \]\n\nis absolutely convergent in \( A \) with \( {ab} + a = b \) . Since \( \psi \) is a homomorph...
Yes
Lemma 4.27. a) For every \( a \in A \) we have \( r\left( a\right) \leq \parallel a\parallel \).
Proof. a) This follows from the submultiplicativity of the norm.
No
Proposition 4.29. Let \( A \) be a complex Banach algebra and let \( a \in A \) . Then the following are true:\n\na) The spectrum \( \operatorname{Sp}\left( a\right) \) is a compact subset of \( \mathbb{C} \) and is contained in the closed ball \( \overline{\mathrm{B}}\left( {0, r\left( a\right) }\right) \).
Proof. a) If \( \lambda \in \mathbb{C} \) is such that \( \left| \lambda \right| > r\left( a\right) \), then Lemma 4.28 implies that \( \lambda \mathrm{e} - a \) is invertible, hence \( \operatorname{Sp}\left( a\right) \subseteq \overline{\mathrm{B}}\left( {0, r\left( a\right) }\right) \) . We now show that the complem...
Yes
Theorem 4.30 (Gelfand-Mazur). Let \( A \neq \{ 0\} \) be a complex Banach algebra such that every nonzero element in \( A \) is invertible. Then \( A \) is isomorphic to \( \mathbb{C} \) .
Proof. Let \( a \in A \) . Then by Proposition 4.29 there is \( {\lambda }_{a} \in \operatorname{Sp}\left( a\right) \) with \( \left| {\lambda }_{a}\right| = r\left( a\right) \) . By assumption, \( {\lambda }_{a}\mathrm{e} - a = 0 \), so \( a = {\lambda }_{a}\mathrm{e} \), hence \( A \) is one-dimensional. This proves ...
Yes
Proposition 4.31. Let \( A \) be a commutative complex Banach algebra. If \( \psi \in \Gamma \left( A\right) \) , then \( \ker \left( \psi \right) \) is a maximal ideal. Conversely, if \( I \subseteq A \) is a maximal ideal, then \( I \) is closed and there is a unique \( \psi \in \Gamma \left( A\right) \) such that \(...
Proof. Clearly, \( \ker \left( \psi \right) \) is a closed ideal. Since \( \ker \left( \psi \right) \) is of codimension one, it must be maximal.\n\nFor the second assertion let \( I \) be a maximal ideal. Consider its closure \( \bar{I} \), still an ideal. Since \( \mathrm{B}\left( {\mathrm{e},1}\right) \) consists of...
Yes
Theorem 4.32. Let \( A \) be a commutative unital Banach algebra and let \( a \in A \) . Then\n\n\[ \operatorname{Sp}\left( a\right) = \{ \psi \left( a\right) : \psi \in \Gamma \left( A\right) \} = \widehat{a}\left( {\Gamma \left( A\right) }\right) . \]
Proof. Since \( \psi \left( \mathrm{e}\right) = 1 \) for \( \psi \in \Gamma \left( A\right) \), one has \( \psi \left( {\psi \left( a\right) \mathrm{e} - a}\right) = 0 \), so \( \psi \left( a\right) \mathrm{e} - a \) cannot be invertible, i.e., \( \psi \left( a\right) \in \operatorname{Sp}\left( a\right) \) .\n\nOn the...
Yes
Lemma 4.34. Let \( A \) be a commutative \( {C}^{ * } \) -algebra. For \( a \in A \) with \( a = {a}^{ * } \) the following assertions are true:\n\na) \( r\left( a\right) = \parallel a\parallel \) .\n\nb) \( \operatorname{Sp}\left( a\right) \subseteq \mathbb{R} \) ; equivalently, \( \psi \left( a\right) \in \mathbb{R} ...
Proof. a) We have \( a{a}^{ * } = {a}^{2} \), so \( \begin{Vmatrix}{a}^{2}\end{Vmatrix} = \parallel a{\parallel }^{2} \) holds. By induction one can prove \( \begin{Vmatrix}{a}^{{2}^{n}}\end{Vmatrix} = \parallel a{\parallel }^{{2}^{n}} \) for all \( n \in {\mathbb{N}}_{0} \) . From this we obtain\n\n\[ r\left( a\right)...
Yes
Lemma 4.35. Let \( A \) be a commutative \( {C}^{ * } \) -algebra and let \( \psi \in \Gamma \left( A\right) \) . Then\n\n\[ \psi \left( {a}^{ * }\right) = \overline{\psi \left( a\right) }\;\text{ for all }a \in A, \]\n\ni.e., \( \psi \) is a \( * \) -homomorphism.
Proof. For \( a \in A \) fixed define\n\n\[ x \mathrel{\text{:=}} \frac{a + {a}^{ * }}{2}\;\text{ and }\;y \mathrel{\text{:=}} \frac{b - {b}^{ * }}{2\mathrm{i}}. \]\n\nThen we have \( x = {x}^{ * } \) and \( y = {y}^{ * } \) and \( a = x + \mathrm{i}y \) . By Lemma 4.34.b, \( \psi \left( x\right) ,\psi \left( y\right) ...
Yes
Lemma 5.3. Every finite positive Baire measure is regular in the sense that for every \( B \in \operatorname{Ba}\left( K\right) \) one has\n\n\[ \mu \left( B\right) = \sup \{ \mu \left( A\right) : A \in \mathrm{{Ba}}\left( K\right), A\text{ compact,}A \subseteq B\} \]\n\n\[ = \inf \{ \mu \left( O\right) : O \in \operat...
Proof. This is a standard argument involving Dynkin systems (Theorem B.1); for a different proof, see Bogachev (2007, II, 7.1.8).
No
Proposition 5.4. If \( \mu, v \) are regular finite positive Borel measures on \( K \), then to each Borel set \( A \in \operatorname{Bo}\left( K\right) \) there is a Baire set \( B \in \operatorname{Ba}\left( K\right) \) such that\n\n\[ \mu \left( {A\bigtriangleup B}\right) = 0 = v\left( {A\bigtriangleup B}\right) . \...
Proof. Let \( A \subseteq K \) be a Borel set. By regularity, there are open sets \( {O}_{n},{O}_{n}^{\prime } \), closed sets \( {L}_{n},{L}_{n}^{\prime } \) such that \( {L}_{n},{L}_{n}^{\prime } \subseteq A \subseteq {O}_{n},{O}_{n}^{\prime } \) and \( \mu \left( {{O}_{n} \smallsetminus {L}_{n}}\right), v\left( {{O}...
Yes
Lemma 5.5. If \( \mu, v \in \mathrm{M}\left( K\right) \) with \( {\int }_{K}f\mathrm{\;d}\mu = {\int }_{K}f\mathrm{\;d}v \) for all \( f \in \mathrm{C}\left( K\right) \), then \( \mu = v \) .
Proof. By passing to \( \mu - v \) we may suppose that \( v = 0 \) . By Exercise 8.a and standard measure theory it suffices to prove that \( \mu \left( A\right) = 0 \) for each compact \( {G}_{\delta } \) -subset \( A \) of \( K \) . Given such a set, one can find open subsets \( {O}_{n} \) of \( K \) such that \( {O}...
Yes
Theorem 5.7 (Riesz’ Representation Theorem). Let \( K \) be a compact space. Then the mapping\n\n\[ \n\mathrm{M}\left( K\right) \rightarrow \mathrm{C}{\left( K\right) }^{\prime },\;\mu \mapsto \langle \cdot ,\mu \rangle \n\]\n\nis an isometric isomorphism.
For the convenience of the reader, we have included a proof in Appendix E; see also Rudin (1987, 2.14) or Lang (1993, IX.2).
No
Proposition 5.9. Let \( 0 \leq \mu \in \mathrm{M}\left( K\right) \) . Then\n\n\[ \operatorname{supp}\left( \mu \right) = \{ x \in K : \mu \left( U\right) > 0\text{ for each open neighborhood }U\text{ of }x\} . \]
Proof. Let \( M \mathrel{\text{:=}} \operatorname{supp}\left( \mu \right) \) and let \( L \) denote the right-hand side of (5.3). Let \( x \in M \) and \( U \) be an open neighborhood of \( x \) . By Urysohn’s lemma there is \( f \in \mathrm{C}\left( K\right) \) such that \( x \in \left\lbrack {f \neq 0}\right\rbrack \...
Yes
Theorem 5.10 (Krylov-Bogoljubov). Let \( \left( {K;\varphi }\right) \) be a topological system. Then there is at least one \( \varphi \) -invariant Baire probability measure on \( K \) .
Proof. We postpone the proof of this theorem to Chapter 10, see Theorem 10.2. Note however that under the identification \( \mathrm{M}\left( K\right) = \mathrm{C}{\left( K\right) }^{\prime } \) from above, the \( \varphi \) -invariance of \( \mu \) just means that \( {T}_{\varphi }^{\prime }\left( \mu \right) = \mu \),...
No
Consider the rotation topological system \( \left( {\mathbb{T};a}\right) \) for some \( a \in \mathbb{T} \). Obviously, the normalized arc-length measure is invariant. If \( a \) is an \( {n}^{\text{th }} \) root of unity, then the convex combination of point measures
\[ \mu \mathrel{\text{:=}} \frac{1}{n}\mathop{\sum }\limits_{{j = 1}}^{n}{\delta }_{{a}^{j}} \] is another invariant probability measure, see also Exercise 5.
No
We claim that for every Borel set \( M \subseteq G/\Gamma \) \n\n\[ \n\lambda \left( {A \cap {q}^{-1}\left( M\right) }\right) = \lambda \left( {B \cap {q}^{-1}\left( M\right) }\right) .\n\]
Indeed, by the invariance of \( \lambda \) and since \( {q}^{-1}\left( M\right) h = {q}^{-1}\left( M\right) \) for all \( h \in \Gamma \) \n\n\[ \n\lambda \left( {A \cap {q}^{-1}\left( M\right) }\right) = \lambda \left( {A \cap {q}^{-1}\left( M\right) \cap {\mathbb{R}}^{3}}\right) = \lambda \left( {A \cap {q}^{-1}\left...
Yes
Theorem 5.18. Let \( G \) be a locally compact group, and let \( \Gamma \) be an unimodular, closed and cocompact subgroup of \( G \) . Then \( G \) is unimodular and there is a unique Baire probability measure \( \mathrm{m} \) on \( G/\Gamma \) that is invariant under all rotations by elements of \( G \) .
For the proof we need some auxiliary results. Integration against the measures \( {\mathrm{m}}_{G} \) and \( {\mathrm{m}}_{\Gamma } \) on \( G \) and \( \Gamma \) is denoted by \( \mathrm{d}x \) and \( \mathrm{d}y \), respectively. The modular function on \( G \) is abbreviated by \( \Delta \) . We identify continuous ...
No
Lemma 5.19. In the situation described above, the following assertions hold:\n\na) \( \Phi : {\mathrm{C}}_{\mathrm{c}}\left( G\right) \rightarrow \mathrm{C}\left( {G/\Gamma }\right) \) is linear and surjective. More precisely, there is a positive linear operator \( \Psi : \mathrm{C}\left( {G/\Gamma }\right) \rightarrow...
Proof. a) Linearity is clear. By Exercise 2.16 there is a compact subset \( K \subseteq G \) with \( {K\Gamma } = G \) . By Exercise 14.b we can find a function \( 0 \leq h \in {\mathrm{C}}_{\mathrm{c}}\left( G\right) \) with \( K \subseteq \left\lbrack {h = 1}\right\rbrack \) . Then \( \left( {\Phi h}\right) \left( x\...
No
Lemma 6.1 (Approximation). Let \( \left( {X,\sum ,\mu }\right) \) be a finite measure space and let \( \mathcal{E} \subseteq \) \( \sum \) be an algebra of subsets such that \( \sigma \left( \mathcal{E}\right) = \sum \) . Then \( \mathcal{E} \) is dense in the measure algebra, i.e., for every \( A \in \sum \) and \( \v...
Proof. This is just Lemma B. 17 from Appendix B. Its proof is standard measure theory using Dynkin systems.
No
Example 6.7. Let \( X \mathrel{\text{:=}} \{ 0,1\} \) with the trivial \( \sigma \) -algebra \( \sum \mathrel{\text{:=}} \{ \varnothing, X\} \) and the unique probability measure thereon. Consider the measure-preserving mappings \( \varphi \) and \( \psi \) on \( X \) defined by\n\n\[ \varphi \left( x\right) \mathrel{\...
We also note that the map \( \varphi \) in Example 6.7 does not have an essential inverse although its induced map \( {\varphi }^{ * } \) is invertible. Hence, the converse of Corollary 6.5 does not hold in general.
No
Proposition 6.10. Let \( \varphi ,\psi : \mathrm{X} \rightarrow \mathrm{Y} \) be measure-preserving maps between probability spaces \( \mathrm{X} \) and \( \mathrm{Y} \) such that \( {\varphi }^{ * } = {\psi }^{ * } \) . If \( \mathrm{Y} \) is a standard probability space, then \( \varphi = \psi \) almost everywhere.
Proof. Note that if \( \mathrm{Y} \) is a Borel probability space then the hypotheses of Lemma 6.9 are satisfied since there is countable collection of open balls separating the points. In the general case we can find a Borel probability space \( {\mathrm{Y}}^{\prime } \) and an essentially invertible measure-preservin...
Yes
Lemma 6.12. Let \( \left( {\mathrm{X};\varphi }\right) \) be a measure-preserving system. Then the following statements are equivalent:\n\n(i) Every \( A \in {\sum }_{\mathrm{X}} \) is recurrent.\n\n(ii) Every \( A \in {\sum }_{\mathrm{X}} \) is infinitely recurrent.\n\n(iii) For every \( \varnothing \neq A \in \sum \l...
Proof. The implication (ii) \( \Rightarrow \) (i) is evident. For the converse take \( A \in {\sum }_{\mathrm{X}} \) and apply \( {\varphi }^{ * } \) to (6.1) to obtain \( {\varphi }^{ * }A \subseteq \mathop{\bigcup }\limits_{{n \geq 2}}{\varphi }^{*n}A \) . Inserting this back into (6.1) yields\n\n\[ A \subseteq \math...
Yes
Theorem 6.13 (Poincaré). Every measure-preserving system \( \left( {\mathrm{X};\varphi }\right) \) is (infinitely) recurrent, i.e., every set \( A \in {\sum }_{\mathrm{X}} \) is infinitely recurrent.
Proof. Let \( A \in \sum \left( \mathrm{X}\right) \) be such that \( A \cap {\varphi }^{*n}A = \varnothing \) for all \( n \geq 1 \) . Thus for \( n > m \geq 0 \) we have\n\n![2ce18656-55b1-426e-a113-dde3fcb83791_119_0.jpg](images/2ce18656-55b1-426e-a113-dde3fcb83791_119_0.jpg)\n\nFig. 6.1 What happens after removing t...
Yes
Lemma 6.15. Let \( \\left( {\\mathrm{X};\\varphi }\\right) \) be a measure-preserving system, \( \\mathrm{X} = \\left( {X,\\sum ,\\mu }\\right) \), and let \( A, B \\in \\sum \) . Then, for \( n \\geq 1 \) ,
Proof. Using the \( \\varphi \) -invariance of \( \\mu \) we write\n\n\[ \n\\mu \\left( B\\right) = \\mu \\left( {{\\varphi }^{ * }B}\\right) = \\mu \\left( {A \\cap {\\varphi }^{ * }B}\\right) + \\mu \\left( {{A}^{\\mathrm{c}} \\cap {\\varphi }^{ * }B}\\right) ,\n\]\n\nand this is (6.4) when \( n = 1 \) and with \( X ...
Yes
Theorem 6.16. Let \( \\left( {\\mathrm{X};\\varphi }\\right) \) be a measure-preserving system, \( \\mathrm{X} = \\left( {X,\\sum ,\\mu }\\right) \), and let \( A \\in \\sum \) with \( \\mu \\left( A\\right) > 0 \) . Then\n\n\[ \n{\\int }_{A}{n}_{A}\\mathrm{\\;d}{\\mu }_{A} = \\frac{\\mu \\left( {\\mathop{\\bigcup }\\l...
Proof. We specialize \( B = X \) in Lemma 6.15. Note that\n\n\[ \nA \\cap \\mathop{\\bigcap }\\limits_{{j = 1}}^{{k - 1}}{\\varphi }^{*j}{A}^{\\mathrm{c}} = \\mathop{\\bigcup }\\limits_{{j = k}}^{\\infty }{A}_{j}\\;\\left( {k \\geq 1}\\right)\n\]\n\nsince \( A \) is recurrent by Poincaré’s theorem. Hence, by (6.4) in L...
Yes
Lemma 6.19. For a measure-preserving system \( \left( {\mathrm{X};\varphi }\right) \) the following statements are equivalent:\n\n(i) The measure-preserving system \( \left( {\mathrm{X};\varphi }\right) \) is ergodic.\n\n(ii) For every \( \varnothing \neq A \in \sum \left( \mathrm{X}\right) \) one has\n\n\[ \mathop{\bi...
Proof. (i) \( \Rightarrow \) (ii): For a set \( A \in {\sum }_{\mathrm{X}} \) and \( n \geq 0 \) the set\n\n\[ {A}^{\left( n\right) } \mathrel{\text{:=}} \mathop{\bigcup }\limits_{{k \geq n}}{\varphi }^{*k}A \]\n\nsatisfies \( {\varphi }^{ * }{A}^{\left( n\right) } \subseteq {A}^{\left( n\right) } \) and hence is an in...
Yes
Proposition 6.20. Let \( \left( {{\mathcal{W}}_{k}^{ + },\sum ,\mu ;\tau }\right) = B\left( {{p}_{0},\ldots ,{p}_{k - 1}}\right) \) be a Bernoulli shift. Then\n\n\[ \mathop{\lim }\limits_{{n \rightarrow \infty }}\mu \left( {{\tau }^{*n}A \cap B}\right) = \mu \left( A\right) \mu \left( B\right) \]\n\nfor all \( A, B \in...
Proof. We use the notation of Example 5.1.5. Let \( \mathcal{E} \) denote the algebra of cylinder sets on \( {\mathcal{W}}_{k}^{ + } = {L}^{{\mathbb{N}}_{0}} \) . If \( B \in \mathcal{E} \), then \( B = {B}_{0} \times \mathop{\prod }\limits_{{k \geq {n}_{0}}}L \) for some \( {n}_{0} \in \mathbb{N} \) and \( {B}_{0} \su...
No
Corollary 6.22 (Kac). Let \( \\left( {\\mathrm{X};\\varphi }\\right) \) be an ergodic measure-preserving system, \( X = \\left( {X,\\sum ,\\mu }\\right) \), and \( A \\in \\sum \) with \( \\mu \\left( A\\right) > 0 \) . Then for the expected return time to A one has\n\n\[ \n{\\int }_{A}{n}_{A}\\mathrm{\\;d}{\\mu }_{A} ...
Proof. Since the system is ergodic, the implication (i) \( \\Rightarrow \) (iii) of Lemma 6.19 shows that \( X = \\mathop{\\bigcup }\\limits_{{n \\geq 0}}{\\varphi }^{*n}A \) . Hence, the claim follows from Theorem 6.16.
No
Theorem 6.23. Let \( \left( {\mathrm{X};\varphi }\right) \) be a measure-preserving system, \( \mathrm{X} = \left( {X,\sum ,\mu }\right) \), and let \( A \in \sum \) be a set of positive measure. Then the induced transformation \( {\varphi }_{A} \) is measurable with respect to \( {\sum }_{A} \) and preserves the induc...
Proof. Take \( B \in \sum, B \subseteq A \) . Then\n\n\[ \left\lbrack {{\varphi }_{A} \in B}\right\rbrack = \mathop{\bigcup }\limits_{{n \geq 1}}{A}_{n} \cap \left\lbrack {{\varphi }^{n} \in B}\right\rbrack \]\n\nshowing that \( {\varphi }_{A} \) is indeed \( {\sum }_{A} \) -measurable. To see that \( {\varphi }_{A} \)...
Yes
Corollary 6.25. Let \( \left( {\mathrm{X};\varphi }\right) \) be an ergodic measure-preserving system such that there are sets with arbitrarily small positive measure. Then for each \( \varepsilon > 0 \) and \( n \in \mathbb{N} \) there is a Rokhlin tower \( B,{\varphi }^{ * }B,{\varphi }^{*2}B,\ldots ,{\varphi }^{*\le...
Proof of Theorem 6.24. Define\n\n\[ {A}_{0} \mathrel{\text{:=}} A,\;{A}_{k + 1} \mathrel{\text{:=}} {\varphi }^{ * }{A}_{k} \cap {A}^{\mathrm{c}}\;\left( {k \geq 0}\right) . \]\n\nThen \( {A}_{n} = {\varphi }^{*n}A \cap \mathop{\bigcap }\limits_{{j = 0}}^{{n - 1}}{\varphi }^{*j}{A}^{\mathrm{c}} \) is the set of points ...
Yes
Lemma 7.5. Let \( E, F \) be Banach lattices and let \( S : E \rightarrow F \) be a positive operator. Then the following assertions hold:\n\na) \( f \leq g\; \Rightarrow \;{Sf} \leq {Sg} \) for all \( f, g \in {E}_{\mathbb{R}} \) .
Proof. a) follows from linearity of \( S \) .
No
Theorem 7.6. Let \( \mathrm{X} \) be a measure space and let \( 1 \leq p < \infty \) . Let \( \mathcal{F} \subseteq {\mathrm{L}}_{ + }^{p}\left( \mathrm{X}\right) \) be a \( \vee \) -stable set such that\n\n\[ s \mathrel{\text{:=}} \sup \left\{ {\parallel f{\parallel }_{p} : f \in \mathcal{F}}\right\} < \infty . \]\n\n...
Proof. Take a sequence \( {\left( {f}_{n}\right) }_{n \in \mathbb{N}} \) in \( \mathcal{F} \) with \( {\begin{Vmatrix}{f}_{n}\end{Vmatrix}}_{p} \rightarrow s \) . By passing to the sequence \( {\left( {f}_{1} \vee {f}_{2} \vee \cdots \vee {f}_{n}\right) }_{n \in \mathbb{N}} \) we may suppose that \( {\left( {f}_{n}\rig...
Yes
Corollary 7.8. Let \( \mathrm{X} \) be a measure space and let \( 1 \leq p < \infty \) . Then the Banach lattice \( {\mathrm{L}}^{p}\left( {\mathrm{X};\mathbb{R}}\right) \) is order complete.
Proof. Let \( {\mathcal{F}}^{\prime } \subseteq {\mathrm{L}}^{p}\left( {\mathrm{X};\mathbb{R}}\right) \) and suppose that there exists \( F \in {\mathrm{L}}^{p}\left( {\mathrm{X};\mathbb{R}}\right) \) with \( f \leq F \) for all \( f \in {\mathcal{F}}^{\prime } \) . We may suppose without loss of generality that \( {\m...
Yes
Theorem 7.10. Let \( \mathrm{X} \) be a finite measure space and \( 1 \leq p < \infty \) . Then each closed lattice ideal \( I \subseteq {\mathrm{L}}^{p}\left( \mathrm{X}\right) \) has the form \( {I}_{A} \) for some \( A \in {\sum }_{\mathrm{X}} \) .
Proof. Let \( I \subseteq {\mathrm{L}}^{p}\left( \mathrm{X}\right) \) be a closed lattice ideal. The set\n\n\[ J \mathrel{\text{:=}} \{ f \in I : 0 \leq f \leq 1\} \]\n\nis nonempty, closed, \( \vee \) -stable and has upper bound \( \mathbf{1} \in {\mathrm{L}}^{p}\left( \mathrm{X}\right) \) (since \( {\mu }_{\mathrm{X}...
Yes
Proposition 7.12. A measure-preserving system \( \left( {\mathrm{X};\varphi }\right) \) is invertible if and only if its Koopman operator \( {T}_{\varphi } \) is invertible on \( {\mathrm{L}}^{p}\left( \mathrm{X}\right) \) for one/each \( 1 \leq p \leq \infty \) .
Proof. Fix \( 1 \leq p \leq \infty \) and abbreviate \( T \mathrel{\text{:=}} {T}_{\varphi } \) . Let \( \left( {\mathrm{X};\varphi }\right) \) be invertible, i.e., \( {\varphi }^{ * } \) is surjective (Definition 6.2). Since \( T{\mathbf{1}}_{A} = {\mathbf{1}}_{{\varphi }^{ * }A} \) for any \( A \in \sum \left( \mathr...
Yes
Consider the space \( {\mathrm{L}}^{1}\left( {\{ 0,\ldots, n - 1\} }\right) = {\mathbb{R}}^{n} \) and a positive operator \( T \) on it, identified with its \( n \times n \) -matrix. Then the irreducibility of \( T \) according to Definition 7.13 coincides with that notion introduced in Section 2.4 on page 27. Namely, ...
After a permutation of the points we may suppose that \( A = \{ k,\ldots, n - 1\} \) for \( 0 < k < n \), and this means that the representing matrix (with respect to the canonical basis) has the form: ![2ce18656-55b1-426e-a113-dde3fcb83791_142_0.jpg](images/2ce18656-55b1-426e-a113-dde3fcb83791_142_0.jpg)
Yes
Proposition 7.15. Let \( \left( {\mathrm{X};\varphi }\right) \) be a measure-preserving system with Koopman operator \( T \mathrel{\text{:=}} {T}_{\varphi } \) on \( {\mathrm{L}}^{1}\left( \mathrm{X}\right) \) . Then for \( 1 \leq p \leq \infty \) the space \( \operatorname{fix}\left( T\right) \cap {\mathrm{L}}^{p} \) ...
Proof. That \( \operatorname{fix}\left( T\right) \cap {\mathrm{L}}^{p} \) is a Banach sublattice of \( {\mathrm{L}}^{p} \) is clear from the identities\n\n\[ T\left( \bar{f}\right) = \overline{Tf} = \bar{f}\;\text{ and }\;T\left| f\right| = \left| {Tf}\right| = \left| f\right| \]\n\nfor \( f \in \operatorname{fix}\left...
No
Proposition 7.16. For \( a \in \mathbb{T} \) the rotation measure-preserving system \( \left( {\mathbb{T},\mathrm{m};a}\right) \) is ergodic if and only if \( a \) is not a root of unity.
Proof. Let \( T \mathrel{\text{:=}} {L}_{a} \) be the Koopman operator on \( {\mathrm{L}}^{2}\left( \mathbb{T}\right) \) (cf. Example 4.22) and suppose that \( f \in \operatorname{fix}\left( T\right) \) . The functions \( {\chi }_{n} : x \mapsto {x}^{n}, n \in \mathbb{Z} \), form a complete orthonormal system in \( {\m...
Yes
Proposition 7.18. Let \( \left( {\mathrm{X};\varphi }\right) \) be a measure-preserving system with Koopman operator \( {T}_{q} \mathrel{\text{:=}} {T}_{\varphi } \) on \( {\mathrm{L}}^{q}\left( \mathrm{X}\right) \left( {1 \leq q \leq \infty }\right) \). Then the following assertions hold:\n\na) \( \ker \left( {\lambda...
Proof. a) Fix \( \lambda \in \mathbb{T} \) and \( f \in {\mathrm{L}}^{q} \) with \( {Tf} = {\lambda f} \). Then \( \left| f\right| = \left| {\lambda f}\right| = \left| {Tf}\right| = T\left| f\right| \), whence \( \left| f\right| \in \operatorname{fix}\left( T\right) \). Hence, for any \( n \geq 0,{g}_{n} \mathrel{\text...
Yes
Proposition 7.19. Let \( \varphi ,\psi : \mathrm{X} \rightarrow \mathrm{Y} \) be measure-preserving mappings between standard probability spaces \( \mathrm{X},\mathrm{Y} \), and let \( {T}_{\varphi },{T}_{\psi } : {\mathrm{L}}^{1}\left( \mathrm{Y}\right) \rightarrow {\mathrm{L}}^{1}\left( \mathrm{X}\right) \) be the in...
Proof. Only one implication is not trivial. If \( {T}_{\varphi } = {T}_{\psi } \), then \( {\mathbf{1}}_{{\varphi }^{ * }A} = {T}_{\varphi }{\mathbf{1}}_{A} = {T}_{\psi }{\mathbf{1}}_{A} = \) \( {\mathbf{1}}_{{\psi }^{ * }A} \) almost everywhere for every \( A \in {\sum }_{\mathrm{Y}} \), i.e., \( {\varphi }^{ * } = {\...
No
Corollary 7.21. Let \( \left( {\mathrm{X};\varphi }\right) \) be a measure-preserving system over a standard probability space \( \mathrm{X} \) . Then the system \( \left( {\mathrm{X};\varphi }\right) \) is invertible if and only if \( \varphi \) is essentially invertible.
Proof. One implication is Corollary 6.5. For the converse, suppose that \( \left( {\mathrm{X};\varphi }\right) \) is invertible. Then its Koopman operator \( T \mathrel{\text{:=}} {T}_{\varphi } \) is invertible, by Proposition 7.12. The inverse \( {T}^{-1} \) satisfies properties 1) and 2) from Theorem 7.20, hence by ...
Yes
Theorem 7.23. Let \( E = \mathrm{C}\left( K\right) \) or \( E = {\mathrm{L}}^{\infty }\left( \mathrm{X}\right) \), and let \( A \subseteq E \) be a closed, conjugation invariant linear subspace with \( \mathbf{1} \in A \) . Then the following assertions are equivalent:\n\n(i) \( A \) is a subalgebra of \( E \) .\n\n(ii...
Proof. The proof of the first part is left as Exercise 16.
No
Theorem 7.24 (Hölder’s Inequality for Positive Operators). Let \( E = \mathrm{C}\left( K\right) \) or \( {\mathrm{L}}^{\infty }\left( \mathrm{X}\right) \), and let \( A \subseteq E \) be a \( {C}^{ * } \) -subalgebra of \( E \) . Furthermore, let \( \mathrm{Y} \) be any measure space, and let \( T : A \rightarrow {\mat...
Proof. We start with the representation\n\n\[ {x}^{1/p} = \mathop{\inf }\limits_{{t > 0}}\frac{1}{p}{t}^{-1/q}x + \frac{1}{q}{t}^{1/p} \]\n\nwhich is (7.5) with \( p \) replaced by \( \frac{1}{p} \) . For fixed \( y > 0 \), multiply this identity with \( {y}^{1/q} \) and arrive at\n\n\[ {x}^{1/p}{y}^{1/q} = \mathop{\in...
Yes
Lemma 8.2. Let \( E \) be a Banach space and let \( T : E \rightarrow E \) be a bounded linear operator on \( E \) . Then, with \( {\mathrm{A}}_{n} \mathrel{\text{:=}} {\mathrm{A}}_{n}\left\lbrack T\right\rbrack \), the following assertions hold:\n\na) If \( f \in \mathrm{{fix}}\left( T\right) \), then \( {\mathrm{A}}_...
Proof. a) is trivial, and the formulae (8.2)-(8.4) are established by simple algebraic manipulations. The remaining statements then follow from these formulae.
No
Lemma 8.3. Let \( T \) be a bounded linear operator on a Banach space \( E \) . Then\n\n\[ F \mathrel{\text{:=}} \left\{ {f \in E : {P}_{T}f \mathrel{\text{:=}} \mathop{\lim }\limits_{{n \rightarrow \infty }}{\mathrm{A}}_{n}f\text{ exists }}\right\} \]\n\nis a \( T \) -invariant subspace of \( E \) containing \( \opera...
Proof. It is clear that \( F \) is a subspace of \( E \) and \( {P}_{T} : F \rightarrow E \) is linear. By Lemma 8.2.b, \( F \) is \( T \) -invariant, \( \operatorname{ran}\left( {P}_{T}\right) \subseteq \operatorname{fix}\left( T\right) \) and \( {P}_{T}{Tf} = T{P}_{T}f = {P}_{T}f \) for all \( f \in F \) . Finally it...
Yes
Theorem 8.5. Let \( T \in \mathcal{L}\left( E\right) \) , \( E \) a Banach space. Suppose that \( \mathop{\sup }\limits_{{n \in \mathbb{N}}}\begin{Vmatrix}{\mathrm{A}}_{n}\end{Vmatrix} < \infty \) and that \( \frac{1}{n}{T}^{n}f \rightarrow 0 \) for all \( f \in E \) . Then the subspace\n\n\[ F \mathrel{\text{:=}} \lef...
Proof. By Lemma 8.3, all that remains to show is that \( F \) is closed, \( {P}_{T} \) is bounded, and \( \ker \left( {P}_{T}\right) = \overline{\operatorname{ran}}\left( {\mathrm{I} - T}\right) \) . The closedness of \( F \) and the boundedness of \( {P}_{T} \) are solely due to the uniform boundedness of the operator...
No
Theorem 8.6 (Mean Ergodic Theorem on Hilbert Spaces). Let \( H \) be a Hilbert space and let \( T \in \mathcal{L}\left( H\right) \) be a contraction, i.e., \( \parallel T\parallel \leq 1 \) . Then\n\n\[ \n{P}_{T}f \mathrel{\text{:=}} \mathop{\lim }\limits_{{n \rightarrow \infty }}\frac{1}{n}\mathop{\sum }\limits_{{j = ...
Proof. If \( T \) is a contraction, then the powers \( {T}^{n} \) and hence the Cesàro averages \( {\mathrm{A}}_{n}\left\lbrack T\right\rbrack \) are contractions, too, and \( \frac{1}{n}{T}^{n} \rightarrow 0 \) . Therefore, Theorem 8.5 can be applied and so the subspace \( F \) is closed and \( {P}_{T} : F \rightarrow...
Yes
Corollary 8.7. Let \( T \) be a contraction on a Hilbert space \( H \) . Then \( \operatorname{fix}\left( T\right) = \operatorname{fix}\left( {T}^{ * }\right) \) and \( {P}_{T} = {P}_{{T}^{ * }} \) .
Proof. Note that \( f \in \operatorname{fix}\left( {T}^{ * }\right) \) implies that \( \left( {{Tf} \mid f}\right) = \left( {f \mid {T}^{ * }f}\right) = \parallel f{\parallel }^{2} \) and hence \( {Tf} = f \) as in the proof of Theorem 8.6. By symmetry, \( \operatorname{fix}\left( T\right) = \operatorname{fix}\left( {T...
Yes
Theorem 8.8. Let \( \\left( {\\mathrm{X};\\varphi }\\right) \) be a measure-preserving system, and consider the Koopman operator \( T = {T}_{\\varphi } \) on the space \( {\\mathrm{L}}^{1} = {\\mathrm{L}}^{1}\\left( \\mathrm{X}\\right) \) . Then\n\n\[ \n{P}_{T}f \\mathrel{\\text{:=}} \\mathop{\\lim }\\limits_{{n \\righ...
Proof. Let \( f \\in {\\mathrm{L}}^{\\infty } \) . Then, by von Neumann’s theorem, the limit \( {P}_{T}f = \) \( \\mathop{\\lim }\\limits_{{n \\rightarrow \\infty }}{\\mathrm{A}}_{n}f \) exists in \( {\\mathrm{L}}^{2} \), thus a fortiori in \( {\\mathrm{L}}^{1} \) . Moreover, since \( \\left| {{T}^{n}f}\\right| \\leq \...
Yes
Theorem 8.10. Let \( \\left( {\\mathrm{X};\\varphi }\\right) \) be a measure-preserving system, \( \\mathrm{X} = \\left( {X,\\sum ,\\mu }\\right) \), with associated Koopman operator \( T \\mathrel{\\text{:=}} {T}_{\\varphi } \) on \( {\\mathrm{L}}^{1}\\left( \\mathrm{X}\\right) \), and let \( 1 \\leq p < \\infty \) . ...
Proof. The equivalence of (i)-(iv) has been proved in Proposition 7.15.\n\n(ii) \( \\Rightarrow \) (v): Let \( f \\in {\\mathrm{L}}^{1} \) . Then \( {P}_{T}f \\in \\operatorname{fix}\\left( T\\right) \), so \( {P}_{T}f = c \\cdot \\mathbf{1} \) by (ii). Integrating yields \( c = {\\int }_{\\mathrm{X}}f \)\n\n\( \\left(...
Yes
Lemma 8.13. For a row-stochastic \( k \times k \) -matrix \( S \) the following assertions are equivalent:\n\n(i) \( S \) is irreducible.\n\n(ii) There is \( m \in \mathbb{N} \) such that \( {\left( \mathrm{I} + S\right) }^{m} \) is strictly positive.\n\n(iii) \( Q = \mathop{\lim }\limits_{{n \rightarrow \infty }}\frac...
Proof. (i) \( \Rightarrow \) (ii): Simply expand \( {\left( \mathrm{I} + S\right) }^{m} = \mathop{\sum }\limits_{{j = 0}}^{m}\left( \begin{matrix} m \\ j \end{matrix}\right) {S}^{j} \) . If \( S \) is irreducible, then for large enough \( m \in \mathbb{N} \), the resulting matrix must have each entry strictly positive....
Yes
Theorem 8.14. Let \( S \) be a row-stochastic \( k \times k \) -matrix with fixed probability vector p. Then \( p \) is strictly positive and the Markov shift \( \left( {{\mathcal{W}}_{k}^{ + },\sum ,\mu \left( {S, p}\right) ;\tau }\right) \) is ergodic, if and only if \( S \) is irreducible.
Proof. Let \( {i}_{0},\ldots ,{i}_{l} \in L \) and \( {j}_{0},\ldots ,{j}_{r} \in L \) . Then for \( n \in \mathbb{N} \) we have\n\n\[ \mu \left( {\left\{ {i}_{0}\right\} \times \cdots \times \left\{ {i}_{l}\right\} \times {L}^{n - 1} \times \left\{ {j}_{0}\right\} \times \cdots \times \left\{ {j}_{r}\right\} \times \p...
Yes
Lemma 8.16. If \( E \) is a Banach space and \( T \in \mathcal{L}\left( E\right) \) is mean ergodic, then \( T \) is Cesàro bounded and \( \frac{1}{n}{T}^{n}f \rightarrow 0 \) for every \( f \in E \) .
Proof. As \( \mathop{\lim }\limits_{{n \rightarrow \infty }}{\mathrm{A}}_{n}f \) exists for every \( f \in E \) and \( E \) is a Banach space, it follows from the uniform boundedness principle (Theorem C.1) that \( \mathop{\sup }\limits_{{n \in \mathbb{N}}}\begin{Vmatrix}{\mathrm{A}}_{n}\end{Vmatrix} < \infty \) . From...
Yes
Lemma 8.17. Let \( C \) be a convex subset of a Hausdorff topological vector space, \( T : C \rightarrow C \) be a continuous affine mapping and \( g \in C \) . Then \( {Tg} = g \) if and only if there is \( f \in C \) and a subsequence \( {\left( {n}_{j}\right) }_{j \in \mathbb{N}} \) such that\n\n\[ \text{(1)}\frac{1...
Proof. If \( {Tg} = g \), then (1) and (2) hold for \( f = g \) and every subsequence \( {\left( {n}_{j}\right) }_{j \in \mathbb{N}} \) . Conversely, suppose that (1) and (2) hold for some \( f \in E \) and a subsequence \( {\left( {n}_{j}\right) }_{j \in \mathbb{N}} \) . Then\n\n\[ g - {Tg} = \left( {\mathrm{I} - T}\r...
Yes
Proposition 8.18. Let \( T \) be a Cesàro bounded operator on some Banach space \( E \) such that \( \frac{1}{n}{T}^{n}h \rightarrow 0 \) for each \( h \in E \) . Then for \( f, g \in E \) the following statements are equivalent:\n\n(i) \( {\mathrm{A}}_{n}f \rightarrow g \) in the norm of \( E \) as \( n \rightarrow \i...
Proof. The implication \( \left( \mathrm{v}\right) \Rightarrow \left( \mathrm{i}\right) \) follows from Theorem 8.5 while the implications (i) \( \Rightarrow \) (ii) \( \Rightarrow \) (iii) are trivial. If (iii) holds, then \( g \in \operatorname{fix}\left( T\right) \) by Lemma 8.17. Moreover,\n\n\[ g \in {\operatornam...
Yes
Theorem 8.20 (Mean Ergodic Operators). Let \( T \) be a Cesàro bounded operator on some Banach space \( E \) such that \( \frac{1}{n}{T}^{n}f \rightarrow 0 \) weakly for each \( f \in E \) . Further, let \( D \subseteq E \) be a dense subset of \( E \) . Then the following assertions are equivalent:\n\n(i) \( T \) is m...
Proof. The equivalence of (i)-(iv) and (vi) is immediate from Proposition 8.18 and Theorem 8.5. Suppose that (vi) holds and \( 0 \neq {f}^{\prime } \in \operatorname{fix}\left( {T}^{\prime }\right) \) . Then there is \( f \in E \) such that \( \left\langle {f,{f}^{\prime }}\right\rangle \neq 0 \) . By (vi) we can write...
No
Theorem 8.22. Every power-bounded linear operator on a reflexive Banach space is mean ergodic.
Proof. Let \( f \in E \) . Then, by power-boundedness of \( T \) the set \( \left\{ {{\mathrm{A}}_{n}f : n \in \mathbb{N}}\right\} \) is norm-bounded. Since \( E \) is reflexive, it is even relatively weakly compact, hence the sequence \( {\left( {\mathrm{A}}_{n}f\right) }_{n \in \mathbb{N}} \) has a weak cluster point...
Yes
Theorem 8.23. Let \( T : {\mathrm{L}}^{1}\left( \mathrm{X}\right) \rightarrow {\mathrm{L}}^{1}\left( \mathrm{Y}\right) \) be a Dunford-Schwartz operator. Then\n\n\[ \parallel {Tf}{\parallel }_{p} \leq \parallel f{\parallel }_{p}\;\text{ for all }f \in {\mathrm{L}}^{p} \cap {\mathrm{L}}^{1},1 \leq p \leq \infty . \]
Proof. The claim is a direct consequence of the Riesz-Thorin interpolation theorem (Folland 1999, Thm. 6.27). If \( T \) is positive, there is a more elementary proof, which we give for convenience. For \( p = 1,\infty \) there is nothing to show, so let \( 1 < p < \infty \) . Take \( f \in {\mathrm{L}}^{p} \cap {\math...
Yes
Theorem 8.24. A Dunford-Schwartz operator \( T \) on \( {\mathrm{L}}^{1}\left( \mathrm{X}\right) \) over a finite measure space \( \mathrm{X} \) is mean ergodic.
Proof. By Theorem 8.23, \( T \) restricts to a contraction on \( {\mathrm{L}}^{2} \), which is a Hilbert space. By Theorem 8.6, \( T \) is mean ergodic on \( {\mathrm{L}}^{2} \), which means that for \( f \in {\mathrm{L}}^{2} \) the limit \( \mathop{\lim }\limits_{{n \rightarrow \infty }}{\mathrm{A}}_{n}\left\lbrack T\...
Yes
Theorem 8.26. Let \( E \) be a Banach space and let \( S \in \mathcal{L}\left( E\right) \) be a power-bounded mean ergodic operator: Let \( T \) be a \( {k}^{\text{th }} \) root of \( S \), i.e., \( {T}^{k} = S \) for some \( k \in \mathbb{N} \) . Then \( T \) is also mean ergodic.
Proof. Denote by \( {P}_{S} \) the mean ergodic projection of \( S \) . Define \( P \mathrel{\text{:=}} \left( {\frac{1}{k}\mathop{\sum }\limits_{{j = 0}}^{{k - 1}}{T}^{j}}\right) {P}_{S} \) and observe that \( {Pf} \in \overline{\operatorname{conv}}\left\{ {{T}^{j}f : j \in {\mathbb{N}}_{0}}\right\} \) for all \( f \i...
Yes
Take \( E = \mathrm{c} \), the space of convergent scalar sequences, and the multiplication operator\n\n\[ M : \mathrm{c} \rightarrow \mathrm{c},\;{\left( {x}_{n}\right) }_{n \in \mathbb{N}} \mapsto {\left( {a}_{n}{x}_{n}\right) }_{n \in \mathbb{N}} \]\n\nfor some sequence \( {\left( {a}_{n}\right) }_{n \in \mathbb{N}}...
Consider now a \( {k}^{\text{th }} \) root of unity \( 1 \neq b \in \mathbb{T} \) and define \( {T}_{k} \mathrel{\text{:=}} {bM} \) . Then it is easy to see that \( \operatorname{fix}\left( {T}_{k}^{\prime }\right) = \{ 0\} \), and hence, again by Theorem \( {8.20}\left( \mathrm{v}\right) ,{T}_{k} \) is mean ergodic. I...
No
Theorem 8.28 (Kakutani). Let \( E \) be a Banach space. Then the identity operator \( {\mathrm{I}}_{E} \) is an extreme point of the closed unit ball in \( \mathcal{L}\left( E\right) \) .
Proof. Suppose that \( {\mathrm{I}}_{E} = \frac{1}{2}\left( {R + S}\right) \) with \( \parallel S\parallel ,\parallel R\parallel \leq 1 \) . Then \( {\mathrm{I}}_{{E}^{\prime }} = \frac{1}{2}\left( {{R}^{\prime } + {S}^{\prime }}\right) \) . Let \( {f}^{\prime } \) be an extreme point of the dual unit ball \( {\mathrm{...
Yes
Lemma 8.29. Let \( R, S \) be two commuting power-bounded operators, and for \( t \in \) \( \left( {0,1}\right) \) let \( T \mathrel{\text{:=}} {tR} + \left( {1 - t}\right) S \) . Then \( \operatorname{fix}\left( T\right) = \operatorname{fix}\left( R\right) \cap \operatorname{fix}\left( S\right) \) .
Proof. Only the inclusion \( \operatorname{fix}\left( T\right) \subseteq \operatorname{fix}\left( R\right) \cap \operatorname{fix}\left( S\right) \) is not obvious. Endow \( E \) with an equivalent norm \( \parallel f{\parallel }_{1} \mathrel{\text{:=}} \sup \left\{ {\begin{Vmatrix}{{R}^{n}{S}^{m}f}\end{Vmatrix} : n, m...
Yes
Theorem 8.30. Let \( {T}_{1},{T}_{2},\ldots ,{T}_{m} \) be commuting power-bounded, mean ergodic operators. Then every convex combination\n\n\[ T \mathrel{\text{:=}} \mathop{\sum }\limits_{{j = 1}}^{m}{t}_{j}{T}_{j} \]\n\nwith all \( {t}_{j} > 0 \), is mean ergodic. Denoting by \( {P}_{j} \) the mean ergodic projection...
Proof. It suffices to prove the statement for the case of \( m = 2 \), the general case can then be established by induction. So let \( S = {T}_{1}, R = {T}_{2} \), let \( 0 < t < 1 \), and let \( T \mathrel{\text{:=}} {tR} + \left( {1 - t}\right) S \) . By Lemma 8.29 we have \( \operatorname{fix}\left( T\right) = \ope...
Yes
Theorem 8.32 (Contraction Semigroups on Hilbert Spaces). Let \( \mathcal{T} \) be a semigroup of linear contractions on a Hilbert space \( H \), and let \( P \in \mathcal{L}\left( H\right) \) be the orthogonal projection onto \( \operatorname{fix}\left( \mathcal{T}\right) \) . Then \( \mathcal{T} \) is mean ergodic wit...
Proof. Let \( T \in \mathcal{T} \) . Then \( {TP} = P \) by definition of \( P \), and \( \operatorname{fix}\left( T\right) = \operatorname{fix}\left( {T}^{ * }\right) \) by Corollary 8.7. Hence, \( {T}^{ * }P = P \), and taking adjoints yields \( {PT} = P \) .\n\nBy Theorem D.2, the closed convex set \( C \mathrel{\te...
Yes
Theorem 8.34. Let \( E \) be a strictly convex Banach space \( E \) which has a strictly convex dual space. Then each relatively weakly compact contraction semigroup on \( E \) is mean ergodic.
Proof. We only sketch the proof. Let \( \mathcal{T} \) be a relatively weakly compact semigroup of contractions on \( E \), and let \( f \in E \) . Then by weak compactness, \( \overline{\operatorname{conv}}\left( {\mathcal{T}f}\right) \) contains an element \( g \) with minimal norm. This element is unique because of ...
No