Q
stringlengths
4
3.96k
A
stringlengths
1
3k
Result
stringclasses
4 values
Lemma 3.1.35 Let \( X \) be a metrizable space and \( B \subseteq X \) Borel. Then \( {\chi }_{B} : X \rightarrow \mathbb{R} \) is Baire.
Proof. Let\n\n\[ \mathcal{B} = \left\{ {B \subseteq X : {\chi }_{B}\text{ is Baire }}\right\} .\n\]\n\n(a) Let \( U \) be open in \( X \) . Write \( U = \mathop{\bigcup }\limits_{n}{F}_{n} \), where the \( {F}_{n} \) ’s are closed and \( {F}_{n} \subseteq {F}_{n + 1} \) . By 2.1.18, there is a continuous function \( {f}_{n} : X \rightarrow \left\lbrack {0,1}\right\rbrack \) identically equal to 1 on \( {F}_{n} \) and equal to 0 on \( X \smallsetminus U \) . Then the sequence \( \left( {f}_{n}\right) \) converges pointwise to \( {\chi }_{U} \) . Thus, \( U \in \mathcal{B} \) .\n\n(b) Let \( {B}_{0},{B}_{1},{B}_{2},\ldots \) be pairwise disjoint and belong to \( \mathcal{B} \) . Set\n\n\[ {f}_{n} = \mathop{\sum }\limits_{{i \leq n}}{\chi }_{{B}_{i}} \]\n\nBy our hypothesis and 3.1.34(ii), \( {f}_{n} \) is Baire. Since \( \left( {f}_{n}\right) \) converges pointwise to the characteristic function of \( \mathop{\bigcup }\limits_{n}{B}_{n} \), we see that \( \mathop{\bigcup }\limits_{n}{B}_{n} \in \mathcal{B} \) .\n\n(c) Let \( {B}_{0},{B}_{1},{B}_{2},\ldots \) belong to \( \mathcal{B} \) . Put\n\n\[ {f}_{n} = \mathop{\min }\limits_{{i \leq n}}{\chi }_{{B}_{i}} \]\n\nBy our hypothesis and 3.1.34, \( {f}_{n} \) is Baire. As \( \left( {f}_{n}\right) \) converges pointwise to the characteristic function of \( \mathop{\bigcap }\limits_{n}{B}_{n} \), it follows that \( \mathop{\bigcap }\limits_{n}{B}_{n} \in \mathcal{B} \) .\n\nThe result now follows from 3.1.11.
Yes
Theorem 3.1.36 (Lebesgue - Hausdorff theorem) Every real-valued Borel function defined on a metrizable space is Baire.
Proof. By 3.1.35 the characteristic function of every Borel set is Baire. Hence, by 3.1.34(ii), every simple Borel function is Baire. Now the result follows from 3.1.28.
Yes
Lemma 3.2.1 Let \( \\left( {X,\\mathcal{T}}\\right) \) be a (zero-dimensional, second countable) metrizable space and \( \\left( {B}_{n}\\right) \) a sequence of Borel subsets of \( X \) . Then there is a (respectively zero-dimensional, second countable) metrizable topology \( {\\mathcal{T}}^{\\prime } \) such that \( \\mathcal{T} \\subseteq {\\mathcal{T}}^{\\prime } \\subseteq {\\mathcal{B}}_{X} \) and each \( {B}_{n} \\in {\\mathcal{T}}^{\\prime } \) .
Proof. Define \( f : X \\rightarrow X \\times \\mathcal{C} \) by\n\n\[ f\\left( x\\right) = \\left( {x,{\\chi }_{{B}_{0}}\\left( x\\right) ,{\\chi }_{{B}_{1}}\\left( x\\right) ,{\\chi }_{{B}_{2}}\\left( x\\right) ,\\ldots }\\right) .\n\]\n\nThis map is clearly one-to-one. Let\n\n\[ {\\mathcal{T}}^{\\prime } = \\left\{ {{f}^{-1}\\left( U\\right) : U\\text{ open in }X \\times \\mathcal{C}}\\right\}\n\]\n\nAs \( \\left( {X,{\\mathcal{T}}^{\\prime }}\\right) \) is homeomorphic to a subset of \( X \\times \\mathcal{C} \), it is metrizable. Further, if \( X \) is zero-dimensional (separable), so is \( \\left( {X,{\\mathcal{T}}^{\\prime }}\\right) \) .\n\nLet \( U \\subseteq X \) be open with respect to the original topology \( \\mathcal{T} \) . Then\n\n\[ U = {f}^{-1}\\left( {\\{ \\left( {x,\\alpha }\\right) \\in X \\times \\mathcal{C} : x \\in U\\} }\\right)\n\]\n\nand hence belongs to \( {\\mathcal{T}}^{\\prime } \) . Thus, \( {\\mathcal{T}}^{\\prime } \) is finer that \( \\mathcal{T} \) . By 3.1.29, \( f \) is Borel measurable. Therefore, \( {\\mathcal{T}}^{\\prime } \\subseteq {\\mathcal{B}}_{X} \) . It remains to show that each \( {B}_{n} \\in {\\mathcal{T}}^{\\prime } \) . Let\n\n\[ {V}_{n} = \\{ \\left( {x,\\alpha }\\right) \\in X \\times \\mathcal{C} : \\alpha \\left( n\\right) = 1\\} .\n\]\n\nThen \( {V}_{n} \) is open in \( X \\times \\mathcal{C} \) . Since \( {B}_{n} = {f}^{-1}\\left( {V}_{n}\\right) ,{B}_{n} \\in {\\mathcal{T}}^{\\prime } \) .
Yes
Proposition 3.2.3 Let \( \left( {X,\mathcal{T}}\right) \) be a metrizable space, \( A \subseteq X, Y \) Polish, and \( f : A \rightarrow Y \) any Borel map. Then\n\n(i) there is a finer metrizable topology \( {\mathcal{T}}^{\prime } \) on \( X \) generating the same Borel \( \sigma \) -algebra such that \( f : A \rightarrow Y \) is continuous with respect to the new topology \( {\mathcal{T}}^{\prime } \), and\n\n(ii) the map \( f : A \rightarrow Y \) admits a Borel extension \( g : X \rightarrow Y \) .
Proof. Fix a countable base \( \left( {U}_{n}\right) \) for \( Y \) . Let \( n \in \mathbb{N} \) . As \( {f}^{-1}\left( {U}_{n}\right) \) is Borel in \( A \), there is a Borel set \( {B}_{n} \) in \( X \) such that \( {f}^{-1}\left( {U}_{n}\right) = A \cap {B}_{n} \) . Take \( {\mathcal{T}}^{\prime } \) as in 3.2.1. This answers (i).\n\nTo prove (ii), take \( {\mathcal{T}}^{\prime } \) as above. By 2.2.3, there is a \( {G}_{\delta } \) set \( C \supseteq A \) and a continuous extension \( {g}^{\prime } : C \rightarrow Y \) of \( f \) . Here we are assuming that \( X \) is equipped with the finer topology \( {\mathcal{T}}^{\prime } \) . As \( \mathcal{T} \) and \( {\mathcal{T}}^{\prime } \) generate the same \( \sigma \) -algebra \( {\mathcal{B}}_{X}, C \) is Borel in \( X \) and \( {g}^{\prime } \) measurable relative to the original topology \( \mathcal{T} \) . Extend \( {g}^{\prime } \) to the whole space \( X \) by defining it to be a constant on \( X \smallsetminus C \) .
Yes
Theorem 3.2.4 Suppose \( \left( {X,\mathcal{T}}\right) \) is a Polish space. Then for every Borel set \( B \) in \( X \) there is a finer Polish topology \( {\mathcal{T}}_{B} \) on \( X \) such that \( B \) is clopen with respect to \( {\mathcal{T}}_{B} \) and \( \sigma \left( \mathcal{T}\right) = \sigma \left( {\mathcal{T}}_{B}\right) \) .
Proof of 3.2.4. Let \( \mathcal{B} \) be the class of all Borel subsets \( B \) of \( X \) such that there is a finer Polish topology \( {\mathcal{T}}_{B} \) generating \( {\mathcal{B}}_{X} \) and making \( B \) clopen.\n\nBy Observation 1, \( \mathcal{B} \) contains all closed sets, and it is clearly closed under complementation.\n\nTo show \( \mathcal{B} = {\mathcal{B}}_{X} \), we need to prove only that \( \mathcal{B} \) is closed under countable unions. Let \( {B}_{n} \) belong to \( \mathcal{B} \) and \( B = \bigcup {B}_{n} \) . Let \( {\mathcal{T}}_{n} \) be a finer Polish topology on \( X \) making \( {B}_{n} \) clopen and generating the same Borel \( \sigma \) -algebra. Then \( B \in {\mathcal{T}}_{\infty } \), where \( {\mathcal{T}}_{\infty } \) is the topology generated by \( \bigcup {\mathcal{T}}_{n} \) . By Observation \( 2,{\mathcal{T}}_{\infty } \) is Polish. Take \( {\mathcal{T}}_{B} \) to be the topology generated by \( {\mathcal{T}}_{\infty }\bigcup \left\{ {B}^{c}\right\} \) . By Observation \( 1,{\mathcal{T}}_{B} \) is Polish.
Yes
Every uncountable Borel subset of a Polish space contains a homeomorph of the Cantor set. In particular, it is of cardinality \( \mathfrak{c} \) .
Proof. Let \( \left( {X,\mathcal{T}}\right) \) be Polish and \( B \) an uncountable Borel subset of \( X \) . By 3.2.4, let \( {\mathcal{T}}^{\prime } \) be a finer Polish topology on \( X \) making \( B \) closed. By 2.6.3, \( \left( {B,{\mathcal{T}}^{\prime } \mid B}\right) \) contains a homeomorph of the Cantor set, say \( K \) . By 2.3.9, \( {\mathcal{T}}^{\prime }\left| {K = \mathcal{T}}\right| K \), and the result follows.
Yes
Theorem 3.2.7 Every uncountable Borel subset of a Polish space contains a homeomorph of the Cantor set. In particular, it is of cardinality \( \mathfrak{c} \) .
Proof. Let \( \left( {X,\mathcal{T}}\right) \) be Polish and \( B \) an uncountable Borel subset of \( X \) . By 3.2.4, let \( {\mathcal{T}}^{\prime } \) be a finer Polish topology on \( X \) making \( B \) closed. By 2.6.3, \( \left( {B,{\mathcal{T}}^{\prime } \mid B}\right) \) contains a homeomorph of the Cantor set, say \( K \) . By 2.3.9, \( {\mathcal{T}}^{\prime }\left| {K = \mathcal{T}}\right| K \), and the result follows.
Yes
Example 3.2.8 By 2.6.4, there are exactly \( \mathfrak{c} \) uncountable closed subsets of \( \mathbb{R} \) . Let \( \left\{ {{C}_{\alpha } : \alpha < \mathfrak{c}}\right\} \) be an enumeration of these. We shall get distinct points \( {x}_{\alpha },{y}_{\alpha },\alpha < \mathfrak{c} \), such that \( {x}_{\alpha },{y}_{\alpha } \in {C}_{\alpha } \) . Then the set \( A = \left\{ {{x}_{\alpha } : \alpha < \mathfrak{c}}\right\} \) is easily seen to be a Bernstein set.
To define the \( {x}_{\alpha } \) ’s and \( {y}_{\alpha } \) ’s, we proceed by transfinite induction. Choose \( {x}_{0},{y}_{0} \in {C}_{0} \) with \( {x}_{0} \neq {y}_{0} \) . Let \( \alpha < \mathfrak{c} \) . Suppose \( {x}_{\beta },{y}_{\beta } \) has been chosen for all \( \beta < \alpha \) . Let \( D = \left\{ {{x}_{\beta } : \beta < \alpha }\right\} \bigcup \left\{ {{y}_{\beta } : \beta < \alpha }\right\} \) . Note that \( \left| D\right| = \left| \alpha \right| + \left| \alpha \right| < \mathfrak{c} \) . As \( \left| {C}_{\alpha }\right| = \mathfrak{c} \), we choose distinct points \( {x}_{\alpha },{y}_{\alpha } \) in \( {C}_{\alpha } \smallsetminus D \) .
Yes
Example 3.3.1 The closed unit interval \( I = \left\lbrack {0,1}\right\rbrack \) and the Cantor set \( \mathcal{C} \) are Borel isomorphic.
Proof. Let \( D \) be the set of all dyadic rationals in \( I \) and \( E \subset \mathcal{C} \) the set of all sequences of 0 's and 1's that are eventually constant. Define \( f : \mathcal{C} \smallsetminus E \rightarrow I \smallsetminus D \) by\n\n\[ f\left( {{\epsilon }_{0},{\epsilon }_{1},{\epsilon }_{2},\ldots }\right) = \mathop{\sum }\limits_{{n \in \mathbb{N}}}{\epsilon }_{n}/{2}^{n + 1}. \]\n\nIt is easy to check that \( f \) is a homeomorphism from \( \mathcal{C} \smallsetminus E \) onto \( I \smallsetminus D \) . Since both \( D \) and \( E \) are countably infinite, there is a bijection \( g : E \rightarrow D \) . The function \( h : I \rightarrow \mathcal{C} \) obtained by piecing \( f \) and \( g \) together is clearly a Borel isomorphism from \( I \) onto \( \mathcal{C} \) .
Yes
Proposition 3.3.2 Suppose \( \\left( {X,\\mathcal{A}}\\right) \) is a measurable space with \( \\mathcal{A} \) countably generated. Then there is a subset \( Z \) of \( \\mathcal{C} \) and a bimeasurable map \( g : X \\rightarrow \) \( Z \) such that for any \( x, y \) in \( X, g\\left( x\\right) = g\\left( y\\right) \) if and only if \( x \) and \( y \) belong to the same atom of \( \\mathcal{A} \) .
Proof. Let \( \\mathcal{G} = \\left\\{ {{A}_{n} : n \\in \\mathbb{N}}\\right\\} \) be a countable generator of \( \\mathcal{A} \) . Define \( g : X \\rightarrow \\mathcal{C} \) by\n\n\[ g\\left( x\\right) = \\left( {{\\chi }_{{A}_{0}}\\left( x\\right) ,{\\chi }_{{A}_{1}}\\left( x\\right) ,{\\chi }_{{A}_{2}}\\left( x\\right) ,\\ldots }\\right) .\n\]\n\nTake \( Z = g\\left( X\\right) \) . By 3.1.29, \( g \) is measurable. Also note that for any two \( x, y \) in \( X, g\\left( x\\right) = g\\left( y\\right) \) if and only if \( x \) and \( y \) belong to the same \( {A}_{i} \) ’s. Recall that the atoms of \( \\mathcal{A} \) are precisely the sets of the form \( \\mathop{\\bigcap }\\limits_{n}{A}^{\\epsilon \\left( n\\right) } \) , \( \\left( {\\epsilon \\left( 0\\right) ,\\epsilon \\left( 1\\right) ,\\ldots }\\right) \\in \\mathcal{C} \) . (See the proof of 3.1.15.) It follows that \( g\\left( x\\right) = g\\left( y\\right) \) if and only if \( x \) and \( y \) belong to the same atom of \( \\mathcal{A} \) . As\n\n\[ g\\left( {A}_{n}\\right) = Z\\bigcap \\{ \\alpha \\in \\mathcal{C} : \\alpha \\left( n\\right) = 1\\}\n\]\n\n it is Borel in \( Z \) . Now observe that\n\n\[ \\mathcal{B} = \\{ B \\in \\mathcal{A} : g\\left( B\\right) \\text{ is Borel in }Z\\}\n\]\n\n is a \( \\sigma \) -algebra containing \( {A}_{n} \) for all \( n \) . So, \( {g}^{-1} \) is also measurable.
Yes
Proposition 3.3.4 Let \( \\left( {X,\\mathcal{A}}\\right) \) be a measurable space, \( Y \) a Polish space, \( A \\subseteq X \), and \( f : A \\rightarrow Y \) a measurable map. Then \( f \) admits a measurable extension to \( X \) .
Proof. Fix a countable base \( \\left( {U}_{n}\\right) \) for \( Y \) . For every \( n \), choose \( {B}_{n} \\in \\mathcal{A} \) such that \( {f}^{-1}\\left( {U}_{n}\\right) = {B}_{n}\\bigcap A \) . Without loss of generality, we assume that \( \\mathcal{A} = \\sigma \\left( \\left( {B}_{n}\\right) \\right) \) . By 3.3.2, get a metrizable space \( Z \) and a bimeasurable map \( g : X \\rightarrow Z \) such that for any \( x,{x}^{\\prime } \\in X, g\\left( x\\right) = g\\left( {x}^{\\prime }\\right) \) if and only if \( x \) and \( {x}^{\\prime } \) belong to the same atom of \( \\sigma \\left( \\left( {B}_{n}\\right) \\right) \) . Hence, for \( x,{x}^{\\prime } \\in A, f\\left( x\\right) = f\\left( {x}^{\\prime }\\right) \) if and only if \( g\\left( x\\right) = g\\left( {x}^{\\prime }\\right) \) . Set \( B = g\\left( A\\right) \) and define \( h : B \\rightarrow Y \) by\n\n\[ h\\left( z\\right) = f\\left( x\\right) \]\n\nwhere \( x \\in A \) is such that \( g\\left( x\\right) = z \) . It is easy to see that \( h \) is well-defined and \( {h}^{-1}\\left( {U}_{n}\\right) = g\\left( {B}_{n}\\right) \\cap B \) . Hence, \( h \) is Borel. By 3.2.3, there is a Borel extension \( {h}^{\\prime } : Z \\rightarrow Y \) of \( h \) . The composition \( {h}^{\\prime } \\circ g \) is clearly a measurable extension of \( f \) to \( X \) .
Yes
Proposition 3.3.6 Let \( X \) and \( Y \) be measurable spaces and \( f : X \rightarrow Y \) , \( g : Y \rightarrow X \) one-to-one, bimeasurable maps. Then \( X \) and \( Y \) are isomorphic.
Proof. As \( f \) and \( g \) are bimeasurable, the set \( E \) described in the proof of the Schröder - Bernstein theorem (1.2.3) is measurable. So the bijection \( h : X \rightarrow Y \) obtained there is bimeasurable.
Yes
Proposition 3.3.7 Let \( X \) be a second countable metrizable space. Then the following statements are equivalent.\n\n(i) \( X \) is standard Borel.\n\n(ii) \( X \) is Borel in its completion \( \widehat{X} \) .\n\n(iii) \( X \) is homeomorphic to a Borel subset of a Polish space.
Proof. Clearly, (ii) implies (iii), and (i) follows from (iii). We show that (i) implies (ii).\n\nLet \( X \) be standard Borel. Then, there is a Polish space \( Z \), a Borel subset \( Y \) of \( Z \), and a Borel isomorphism \( f : X \rightarrow Y \) . By 3.3.5, there is a Borel isomorphism \( g : {X}^{\prime } \rightarrow {Y}^{\prime } \) extending \( f \) between Borel subsets \( {X}^{\prime } \) and \( {Y}^{\prime } \) of \( \widehat{X} \) (the completion of \( X \) ) and \( Z \) respectively. Since \( X = {g}^{-1}\left( Y\right) \), it is Borel in \( {X}^{\prime } \) and hence in \( \widehat{X} \) .
Yes
Theorem 3.3.10 The Effros Borel space of a Polish space is standard Borel.
Proof. Let \( Y \) be a compact metric space containing \( X \) as a dense subspace. By 2.2.7, \( X \) is a \( {G}_{\delta } \) set in \( Y \) . Write \( X = \bigcap {U}_{n},{U}_{n} \) open in \( Y \) . Let \( \left( {V}_{n}\right) \) be a countable base for \( Y \) . Now consider\n\n\[ \mathcal{Z} = \{ \operatorname{cl}\left( F\right) \in F\left( Y\right) : F \in F\left( X\right) \}\]\n\nwhere closure is relative to \( Y \) .\n\nNote that \( \mathcal{Z} \subseteq K\left( Y\right) \) and\n\n\[ K \in \mathcal{Z} \Leftrightarrow K\bigcap X\text{ is dense in }K \]\n\nThe result will be proved if we show the following.\n\n(i) The map \( F \rightarrow \operatorname{cl}\left( F\right) \) from \( \left( {F\left( X\right) ,\mathcal{E}\left( X\right) }\right) \) onto \( \mathcal{Z} \) is an isomorphism, and\n\n(ii) \( \mathcal{Z} \) is a \( {G}_{\delta } \) set in \( K\left( Y\right) \) .\n\nClearly, \( F \rightarrow \operatorname{cl}\left( F\right) \) is one-to-one on \( F\left( X\right) \) . Further, for any \( F \in F\left( X\right) \) and any \( U \) open in \( Y \) ,\n\n\[ \operatorname{cl}\left( F\right) \bigcap U \neq \varnothing \Leftrightarrow F\bigcap \left( {U\bigcap X}\right) \neq \varnothing .\n\nHence, (i) follows.\n\nWe now prove (ii). We have\n\n\[ K \in \mathcal{Z} \Leftrightarrow K \cap \mathop{\bigcap }\limits_{n}{U}_{n}\text{ is dense in }K \]\n\nTherefore, by the Baire category theorem,\n\n\[ K \in \mathcal{Z} \Leftrightarrow \forall n\left( {K\bigcap {U}_{n}\text{ is dense in }K}\right) \]\n\n\[ \Leftrightarrow \;\forall n\forall m\left( {K\bigcap {V}_{m} \neq \varnothing \Rightarrow K\bigcap {V}_{m}\bigcap {U}_{n} \neq \varnothing }\right) .\n\nThus,\n\n\[ Z = \mathop{\bigcap }\limits_{n}\mathop{\bigcap }\limits_{m}\left\{ {K \in F\left( Y\right) : K\bigcap {V}_{m} = \varnothing \text{ or }K\bigcap {V}_{m}\bigcap {U}_{n} \neq \varnothing }\right\} \]\n\nand the result follows.
Yes
Theorem 3.3.13 (The Borel isomorphism theorem) Any two uncountable standard Borel spaces are Borel isomorphic.
Proof of 3.3.13. Let \( B \) be an uncountable standard Borel space. Without loss of generality, we assume that \( B \) is a Borel subset of some Polish space. By 3.3.14, there is a bimeasurable bijection from \( B \) into \( \mathcal{C} \). By 3.2.7, \( B \) contains a homeomorph of the Cantor set. By 3.3.6, \( B \) is Borel isomorphic to \( \mathcal{C} \), and the proof is complete.
Yes
Lemma 3.3.14 Every standard Borel space \( B \) is Borel isomorphic to a Borel subset of \( \mathcal{C} \) .
Proof. By 3.3.1, \( I \) and \( \mathcal{C} \) are Borel isomorphic. Therefore, the Hilbert cube \( {I}^{\mathbb{N}} \) and \( {\mathcal{C}}^{\mathbb{N}} \) are isomorphic. But \( {\mathcal{C}}^{\mathbb{N}} \) is homeomorphic to \( \mathcal{C} \) . Thus, the Hilbert cube and the Cantor set are Borel isomorphic. By 2.1.32, every standard Borel space is isomorphic to a Borel subset of the Hilbert cube, and hence of \( \mathcal{C} \) .
Yes
Proposition 3.3.15 For every Borel subset \( B \) of a Polish space \( X \), there is a Polish space \( Z \) and a continuous bijection from \( Z \) onto \( B \) .
Proof. Let \( \mathcal{B} \) be the set of all \( B \subseteq X \) such that there is a continuous bijection from a Polish space \( Z \) onto \( B \) . We show that \( \mathcal{B} = {\mathcal{B}}_{X} \) . Since every open subset of \( X \) is Polish,(2.2.1), open sets belong to \( \mathcal{B} \) . By 3.1.11, it is sufficient to show that \( \mathcal{B} \) is closed under countable intersections and countable disjoint unions. Let \( {B}_{0},{B}_{1},{B}_{2},\ldots \) belong to \( \mathcal{B} \) . Fix Polish spaces \( {Z}_{0},{Z}_{1},\ldots \) and continuous bijections \( {g}_{i} : {Z}_{i} \rightarrow {B}_{i} \) . Let\n\n\[ Z = \left\{ {z \in \prod {Z}_{i} : {g}_{0}\left( {z}_{0}\right) = {g}_{1}\left( {z}_{1}\right) = \cdots }\right\} .\n\]\n\nThen \( Z \) is closed in \( \mathop{\prod }\limits_{i}{Z}_{i} \) . Therefore, \( Z \) is Polish. Define \( g : Z \rightarrow X \) by\n\n\[ g\left( z\right) = {g}_{0}\left( {z}_{0}\right) \]\n\nThen \( g \) is a one-to-one, continuous map from \( Z \) onto \( \bigcap {B}_{i} \) . Thus, \( \mathcal{B} \) is closed under countable intersections.\n\nLet us next assume that \( {B}_{0},{B}_{1},\ldots \) are pairwise disjoint. Choose \( {g}_{i},{Z}_{i} \) as before. Take \( Z = \bigoplus {Z}_{i} \), the direct sum of the \( {Z}_{i} \) ’s. Define \( g : Z \rightarrow X \) by\n\n\[ g\left( z\right) = {g}_{i}\left( z\right) \text{ if }z \in {Z}_{i} \]\n\nThen \( Z \) is a Polish space, and \( g \) is a one-to-one, continuous map from \( Z \) onto \( \bigcup {B}_{i} \) . So, \( \mathcal{B} \) is also closed under countable disjoint unions.
Yes
Theorem 3.3.17 Every Borel subset of a Polish space is a continuous image of \( {\mathbb{N}}^{\mathbb{N}} \) and a one-to-one, continuous image of a closed subset of \( {\mathbb{N}}^{\mathbb{N}} \) .
Proof. The result follows directly from 3.3.15, 2.6.9, and 2.6.13.
No
Theorem 3.3.18 For every infinite Borel subset \( X \) of a Polish space, \( \left| {\mathcal{B}}_{X}\right| = \mathfrak{c} \)
Proof. Without loss of generality, we assume that \( X \) is uncountable. Since \( X \) contains a countable infinite set, \( \left| {\mathcal{B}}_{X}\right| \geq \mathfrak{c} \) . By 2.6.6, the cardinality of the set of continuous maps from \( {\mathbb{N}}^{\mathbb{N}} \) to \( X \) is \( \mathfrak{c} \) . Therefore, by 3.3.17, \( \left| {\mathcal{B}}_{X}\right| \leq \mathfrak{c} \) . The result follows from the Schröder - Bernstein Theorem.
Yes
Theorem 3.3.22 (Ramsey - Mackey theorem) Suppose \( \left( {X,\mathcal{B}}\right) \) is a standard Borel space and \( f : X \rightarrow X \) a Borel isomorphism. Then there is a Polish topology \( \mathcal{T} \) on \( X \) generating \( \mathcal{B} \) and making \( f \) a homeomorphism.
Proof. If \( X \) is countable, we equip \( X \) with the disrete topology, and the result follows. So, we assume that \( X \) is uncountable. By the Borel isomorphism theorem, there is a Polish topology \( {\mathcal{T}}_{0} \) generating \( \mathcal{B} \) . Suppose for some \( n \in \mathbb{N} \), a Polish topology \( {\mathcal{T}}_{n} \) generating \( \mathcal{B} \) has been defined. Let \( \left\{ {{B}_{i}^{n} : i \in \mathbb{N}}\right\} \) be a countable base for \( \left( {X,{\mathcal{T}}_{n}}\right) \) . Consider\n\n\[ \mathcal{D} = \left\{ {f\left( {B}_{i}^{n}\right) : i \in \mathbb{N}}\right\} \bigcup \left\{ {{f}^{-1}\left( {B}_{i}^{n}\right) : i \in \mathbb{N}}\right\} .\n\]\n\nBy 3.2.5, there is a Polish topology \( {\mathcal{T}}_{n + 1} \) finer than \( {\mathcal{T}}_{n} \) making each element of \( \mathcal{D} \) open. Now take \( \mathcal{T} \) to be the topology generated by \( \bigcup {\mathcal{T}}_{n} \) . By Observation 2 following 3.2.4, \( \mathcal{T} \) is Polish. A routine argument now completes the proof.
Yes
Lemma 3.4.5 Let \( \\left( {X,\\mathcal{B}}\\right) \) be a measurable space and \( \\mathcal{A} \) an algebra such that \( \\sigma \\left( \\mathcal{A}\\right) = \\mathcal{B} \) . Suppose \( {\\mu }_{1} \) and \( {\\mu }_{2} \) are finite measures on \( \\left( {X,\\mathcal{B}}\\right) \) such that \( {\\mu }_{1}\\left( A\\right) = {\\mu }_{2}\\left( A\\right) \) for every \( A \\in \\mathcal{A} \) . Then \( {\\mu }_{1}\\left( A\\right) = {\\mu }_{2}\\left( A\\right) \) for every \( A \\in \\mathcal{B} \) .
Proof. Let\n\n\[ \n\\mathcal{M} = \\left\\{ {A \\in \\mathcal{B} : {\\mu }_{1}\\left( A\\right) = {\\mu }_{2}\\left( A\\right) }\\right\\} \n\]\n\nBy our hypothesis \( \\mathcal{A} \\subseteq \\mathcal{M} \) . By (iii) and (iv) above, \( \\mathcal{M} \) is a monotone class. The result follows from 3.1.14.
Yes
Example 3.4.4 Let \( X \) be a nonempty set. For \( A \subseteq X \), let \( \mu \left( A\right) \) denote the number of elements in \( A \) . ( \( \mu \left( A\right) \) is \( \infty \) if \( A \) is infinite.) Then \( \mu \) is a measure on \( \mathcal{P}\left( X\right) \), called the counting measure.
Let \( \left( {X,\mathcal{A},\mu }\right) \) be a measure space. The following are easy to check.\n\n(i) \( \mu \) is monotone: If \( A \) and \( B \) are measurable sets with \( A \subseteq B \), then \( \mu \left( A\right) \leq \mu \left( B\right) . \)\n\n(ii) \( \mu \) is countably subadditive: For any sequence \( \left( {A}_{n}\right) \) of measurable sets,\n\n\[ \mu \left( {\mathop{\bigcup }\limits_{n}{A}_{n}}\right) \leq \mathop{\sum }\limits_{0}^{\infty }\mu \left( {A}_{n}\right) \]\n\n(iii) If the \( {A}_{n} \) ’s are measurable and nondecreasing, then\n\n\[ \mu \left( {\mathop{\bigcup }\limits_{n}{A}_{n}}\right) = \lim \mu \left( {A}_{n}\right) \]\n\n(iv) If \( \mu \) is finite and \( \left( {A}_{n}\right) \) a nonincreasing sequence of measurable sets, then\n\n\[ \mu \left( {\mathop{\bigcap }\limits_{n}{A}_{n}}\right) = \lim \mu \left( {A}_{n}\right) \]
Yes
Lemma 3.4.5 Let \( \left( {X,\mathcal{B}}\right) \) be a measurable space and \( \mathcal{A} \) an algebra such that \( \sigma \left( \mathcal{A}\right) = \mathcal{B} \) . Suppose \( {\mu }_{1} \) and \( {\mu }_{2} \) are finite measures on \( \left( {X,\mathcal{B}}\right) \) such that \( {\mu }_{1}\left( A\right) = {\mu }_{2}\left( A\right) \) for every \( A \in \mathcal{A} \) . Then \( {\mu }_{1}\left( A\right) = {\mu }_{2}\left( A\right) \) for every \( A \in \mathcal{B} \) .
Proof. Let\n\n\[ \mathcal{M} = \left\{ {A \in \mathcal{B} : {\mu }_{1}\left( A\right) = {\mu }_{2}\left( A\right) }\right\} \]\n\nBy our hypothesis \( \mathcal{A} \subseteq \mathcal{M} \) . By (iii) and (iv) above, \( \mathcal{M} \) is a monotone class. The result follows from 3.1.14.
No
Example 3.4.7 Let \( \mathcal{A} \) be the algebra on \( \mathbb{R} \) consisting of finite disjoint unions of nondegenerate intervals (3.1.4). For any interval \( I \), let \( \left| I\right| \) denote the length of \( I \) . Let \( {I}_{0},{I}_{1},\ldots ,{I}_{n} \) be pairwise disjoint intervals and \( A = \) \( \mathop{\bigcup }\limits_{{k = 0}}^{n}{I}_{k} \) . Set\n\n\[ \lambda \left( A\right) = \mathop{\sum }\limits_{{k = 0}}^{n}\left| {I}_{k}\right| \]\n\nThen \( \lambda \) is a \( \sigma \) -finite measure on \( \mathcal{A} \) .
By 3.4.6, there is a unique measure on \( \sigma \left( \mathcal{A}\right) = {\mathcal{B}}_{\mathbb{R}} \) extending \( \lambda \) . We call this measure the Lebesgue measure on \( \mathbb{R} \) and denote it by \( \lambda \) itself.
No
Lemma 3.4.14 Let \( X \) be a metrizable space and \( \mu \) a finite measure on \( X \) . Then \( \mu \) is regular; i.e., for every Borel set \( B \) ,
Proof. Consider the class \( \mathcal{D} \) of all sets \( B \) satisfying the above conditions. We show that \( \mathcal{D} = {\mathcal{B}}_{X} \) . Let \( B \) be closed. Therefore, it is a \( {G}_{\delta } \) set. Write \( B = \mathop{\bigcap }\limits_{n}{U}_{n} \), the \( {U}_{n} \) ’s open and nonincreasing. Since \( \mu \) is finite, \n\n\[ \mu \left( B\right) = \inf \mu \left( {U}_{n}\right) = \lim \mu \left( {U}_{n}\right) . \]\n\nThus every closed set has the above property. \( \mathcal{D} \) is clearly closed under complementation.\n\nNow let \( {B}_{0},{B}_{1},{B}_{2},\ldots \) belong to \( \mathcal{D} \), and \( B = \mathop{\bigcup }\limits_{n}{B}_{n} \) . Fix \( \epsilon > 0 \) . Choose \( N \) such that \( \mu \left( {B \smallsetminus \mathop{\bigcup }\limits_{{i < N}}{B}_{i}}\right) < \epsilon /2 \) . For each \( 0 \leq i \leq N \), there is a closed set \( {F}_{i} \subseteq {B}_{i} \) such that \( \mu \left( {{B}_{i} \smallsetminus {F}_{i}}\right) < \epsilon /\left( {2\left( {N + 1}\right) }\right) \) . It is easy to check that \( \mu \left( {B \smallsetminus \mathop{\bigcup }\limits_{{i < N}}{F}_{i}}\right) < \epsilon \)\n\nTo show the other equality, choose closed sets \( {F}_{i} \subseteq {B}_{i}^{c} \) such that \( \mu \left( {B}_{i}^{c}\right. \smallsetminus \) \( \left. {F}_{i}\right) < \epsilon /{2}^{i + 1} \) . As \( {B}^{c} \smallsetminus \bigcap {F}_{i} \subseteq \bigcup \left( {{B}_{i}^{c} \smallsetminus {F}_{i}}\right) \), it follows that \( \mu \left( {{B}^{c} \smallsetminus \bigcap {F}_{i}}\right) < \epsilon \) . Take \( U = {\left( \bigcap {F}_{i}\right) }^{c} \) . Then \( U \) is an open set containing \( B \) such that \( \mu \left( {U \smallsetminus B}\right) < \) \( \epsilon \) . It follows that \( \mathcal{D} \) is closed under countable unions too. The result follows.
Yes
Theorem 3.4.17 If \( E \subseteq \mathbb{R} \) is a Lebesgue measurable set of positive Lebesgue measure, then the set\n\n\[ E - E = \{ x - y : x, y \in E\} \]\n\n is a neighborhood of 0 .
Proof. By 3.4.16 (ii), the function \( f\left( x\right) = \lambda \left( {E\bigcap \left( {E + x}\right) }\right), x \in \mathbb{R} \), is continuous. Since \( f\left( 0\right) = \lambda \left( E\right) > 0 \), there is a nonempty open interval \( \left( {-a, a}\right) \) such that \( f\left( x\right) > 0 \) for every \( x \in \left( {-a, a}\right) \) . In particular, \( E \cap (E + \) \( x) \neq \varnothing \) for every \( x \in \left( {-a, a}\right) \) . It follows that \( \left( {-a, a}\right) \subseteq E - E \) .
Yes
Example 3.4.18 Let \( G \) be the additive group \( \mathbb{R} \) of real numbers, \( \mathbb{Q} \) the subgroup of rationals, and \( \mathbf{\Pi } \) the partition of \( \mathbb{R} \) consisting of all the cosets of \( \mathbb{Q} \) . The partition \( \mathbf{\Pi } \) is known as the Vitali partition. By \( \mathbf{{AC}} \), there exists a set \( S \) intersecting each coset in exactly one point. We claim that \( S \) is not Lebesgue measurable.
Suppose not. Two cases arise. Either \( \lambda \left( S\right) = 0 \) or \( \lambda \left( S\right) > 0 \) . Assume first that \( \lambda \left( S\right) = 0 \) . Then, as \( \mathbb{R} = \mathop{\bigcup }\limits_{{r \in \mathbb{O}}}\left( {r + S}\right) \) , \( \lambda \left( \mathbb{R}\right) = 0 \), which is a contradiction. Now, let \( \lambda \left( S\right) > 0 \) . By 3.4.17, \( S - S \) contains a nonempty open interval. Hence, there are distinct points \( x, y \) in \( S \) such that \( x - y \) is rational. We have arrived at a contradiction again.
Yes
Theorem 3.4.19 Let \( X \) be a Polish space, \( \mu \) a finite Borel measure on \( X \) , and \( \epsilon > 0 \) . Then there is a compact set \( K \) such that \( \mu \left( {X \smallsetminus K}\right) < \epsilon \) .
Proof. Fix a compatible complete metric \( d \leq 1 \) on \( X \) . Take a regular system \( \left\{ {{F}_{s} : s \in {\mathbb{N}}^{ < \mathbb{N}}}\right\} \) of nonempty closed sets such that\n\n(i) \( {F}_{e} = X \) ,\n\n(ii) \( {F}_{s} = \mathop{\bigcup }\limits_{n}{F}_{s \cap n} \), and\n\n(iii) diameter \( \left( {F}_{s}\right) \leq 1/{2}^{\left| s\right| } \) .\n\nTo see that such a system exists, we proceed by induction on \( \left| s\right| \) . Suppose \( {F}_{s} \) has been defined. Since \( X \) is second countable, there is a sequence \( \left( {U}_{n}\right) \) of open sets of diameter \( \leq {2}^{-\left| s\right| } \) covering \( {F}_{s} \), and further, \( {F}_{s}\bigcap {U}_{n} \neq \varnothing \) for all \( n \) . Take \( {F}_{s \cap n} = \operatorname{cl}\left( {{F}_{s} \cap {U}_{n}}\right) \) .\n\nBy an easy induction, we now define positive integers \( {n}_{0},{n}_{1},{n}_{2},\ldots \) such that the following conditions hold: for every \( s = \left( {{m}_{0},{m}_{1},\ldots ,{m}_{k - 1}}\right) \) with \( {m}_{i} \leq {n}_{i} \)\n\n\[ \mu \left( {{F}_{s} \smallsetminus \mathop{\bigcup }\limits_{{j \leq {n}_{k}}}{F}_{s \cap j}}\right) < \frac{\epsilon }{{2}^{k + 1} \cdot {n}_{0}\ldots \cdot {n}_{k - 1}}. \]\n\nSet\n\n\[ K = \mathop{\bigcap }\limits_{k}\mathop{\bigcup }\limits_{s}{F}_{s} \]\n\nwhere the union varies over all \( s = \left( {{m}_{0},{m}_{1},\ldots ,{m}_{k - 1}}\right) \) with \( {m}_{i} \leq {n}_{i} \) . It is easy to check that \( K \) is closed and totally bounded and hence compact. Further, \( \mu \left( {X \smallsetminus K}\right) < \epsilon \) .
Yes
Theorem 3.4.20 Let \( \left( {X,\mathcal{T}}\right) \) be a Polish space and \( \mu \) a finite Borel measure on \( X \) . Then for every Borel subset \( B \) of \( X \) and every \( \epsilon > 0 \), there is a compact \( K \subseteq B \) such that \( \mu \left( {B \smallsetminus K}\right) < \epsilon \) .
Proof. By 3.2.4, there is a Polish topology \( {\mathcal{T}}_{B} \) on \( X \) finer than \( \mathcal{T} \) generating the same Borel \( \sigma \) -algebra such that \( B \) is clopen with respect to \( {\mathcal{T}}_{B} \) . By 3.4.19, there is a compact set \( K \) relative to \( {\mathcal{T}}_{B} \) contained in \( B \) such that \( \mu \left( {B \smallsetminus K}\right) < \epsilon \) . Since \( K \) is compact with respect to the original topology too, the result follows.
Yes
Theorem 3.4.23 (The isomorphism theorem for measure spaces) If \( \mu \) is a continuous probability on a standard Borel space \( X \), then there is a Borel isomorphism \( h : X \rightarrow I \) such that for every Borel subset \( B \) of \( I,\lambda \left( B\right) = \) \( \mu \left( {{h}^{-1}\left( B\right) }\right) \) .
Proof. By the Borel isomorphism theorem (3.3.13), we can assume that \( X = I \) . Let \( F : I \rightarrow I \) be the distribution function of \( \mu \) . So, \( F \) is a continuous, nondecreasing map with \( F\left( 0\right) = 0 \) and \( F\left( 1\right) = 1 \) . Let\n\n\[ N = \left\{ {y \in I : {F}^{-1}\left( {\{ y\} }\right) }\right. \text{contains more than one point}\} \text{.} \]\n\nSince \( F \) is monotone, \( N \) is countable. If \( N \) is empty, take \( h = F \) . Otherwise, we take an uncountable Borel set \( M \subset I \smallsetminus N \) of Lebesgue measure 0, e.g., \( \mathcal{C} \smallsetminus N \) . So, \( \mu \left( {{F}^{-1}\left( M\right) }\right) = 0 \) . Put \( Q = M \cup N \) and \( P = {F}^{-1}\left( Q\right) \) . Both \( P \) and \( Q \) are uncountable Borel sets with \( \mu \left( P\right) = \lambda \left( Q\right) = 0 \) . Fix a Borel isomorphism \( g : P \rightarrow Q \) . Define\n\n\[ h\left( x\right) = \left\{ \begin{array}{lll} g\left( x\right) & \text{ if } & x \in P, \\ F\left( x\right) & \text{ if } & x \in I \smallsetminus P. \end{array}\right. \]\n\nThe map \( h \) has the desired properties.
Yes
Proposition 3.4.24 Let \( X, Y \), and \( P \) be as above. Then for every \( A \in \) \( \mathcal{A} \otimes {\mathcal{B}}_{Y} \), the map \( x \rightarrow P\left( {x,{A}_{x}}\right) \) is measurable.
Proof. Let\n\n\[ \mathcal{B} = \left\{ {A \in \mathcal{A}\bigotimes {\mathcal{B}}_{Y} : \text{ the map }x \rightarrow P\left( {x,{A}_{x}}\right) \text{ is measurable }}\right\} .\n\]\n\nIt is obvious that \( \mathcal{B} \) contains all the measurable rectangles and is closed under finite disjoint unions. Clearly, \( \mathcal{B} \) is a monotone class. As finite disjoint unions of measurable rectangles form an algebra generating \( \mathcal{A} \otimes {\mathcal{B}}_{Y} \), the result follows from 3.1.14.
Yes
Proposition 3.5.1 The collection \( \mathcal{D} \) of all subsets of a topological space \( X \) having the Baire property forms a \( \sigma \) -algebra.
Proof. Closure under countable unions: Let \( {A}_{0},{A}_{1},{A}_{2},\ldots \) belong to \( \mathcal{D} \) . Take open sets \( {U}_{0},{U}_{1},{U}_{2},\ldots \) such that \( {A}_{n}\Delta {U}_{n} \) is meager for each \( n \) . Since\n\n\[ \left( {\mathop{\bigcup }\limits_{n}{A}_{n}}\right) \Delta \left( {\mathop{\bigcup }\limits_{n}{U}_{n}}\right) \subseteq \mathop{\bigcup }\limits_{n}\left( {{A}_{n}\Delta {U}_{n}}\right) \]\n\nand the union of a sequence of meager sets is meager, \( \mathop{\bigcup }\limits_{n}{A}_{n} \in \mathcal{D} \) .\n\nClosure under complementation: Let \( A \in \mathcal{D} \) and let \( U \) be an open set such that \( {A\Delta U} \) is meager. We have\n\n\[ \left( {X \smallsetminus A}\right) \Delta \operatorname{int}\left( {X \smallsetminus U}\right) \]\n\n\[ \subseteq \;\left( {\left( {X \smallsetminus A}\right) \Delta \left( {X \smallsetminus U}\right) }\right) \bigcup \left( {\left( {X \smallsetminus U}\right) \smallsetminus \operatorname{Int}\left( {X \smallsetminus U}\right) }\right) .\n\n\[ \text{As}\left( {X \smallsetminus A}\right) \Delta \left( {X \smallsetminus U}\right) = {A\Delta U} \]\n\n\[ \left( {X \smallsetminus A}\right) \Delta \operatorname{int}\left( {X \smallsetminus U}\right) \subseteq \left( {A\Delta U}\right) \bigcup \left( {\left( {X \smallsetminus U}\right) \smallsetminus \operatorname{Int}\left( {X \smallsetminus U}\right) }\right) .\n\nSince for any closed set \( F, F \smallsetminus \operatorname{int}\left( F\right) \) is nowhere dense, \( \left( {X \smallsetminus A}\right) \Delta \operatorname{int}(X \smallsetminus \)\n\n\( U) \) is meager.\n\nThe result follows.\n\nThe \( \sigma \) -algebra \( \mathcal{D} \) defined above is called the Baire \( \sigma \) -algebra of \( X \) .
Yes
Proposition 3.5.6 Let \( X \) be a second countable Baire space and \( \left( {U}_{n}\right) \) a countable base for \( X \) . Let \( U \) be an open set in \( X \) .\n\n(i) For every sequence \( \left( {A}_{n}\right) \) of subsets of \( X,\bigcap {A}_{n} \) is comeager in \( U \) if and only if \( {A}_{n} \) is comeager in \( U \) for each \( n \) .\n\n(ii) Let \( A \subseteq X \) be a nonmeager set with \( {BP} \) . Then \( A \) is comeager in \( {U}_{n} \) for some \( n \) .\n\n(iii) A set \( A \) with \( {BP} \) is comeager if and only if \( A \) is nonmeager in each \( {U}_{n} \) .
Proof. Suppose \( \bigcap {A}_{n} \) is comeager in \( U \) . Then clearly each of \( {A}_{n} \) is comeager in \( U \) . Conversely, if each of \( {A}_{n} \) is comeager in \( U \), then \( U \smallsetminus {A}_{n} \) is meager in \( U \) for all \( n \) . So, \( \mathop{\bigcup }\limits_{n}\left( {U \smallsetminus {A}_{n}}\right) = U \smallsetminus \mathop{\bigcap }\limits_{n}{A}_{n} \) is meager in \( U \) . Thus we have proved (i).\n\nTo prove (ii), take \( A \) with BP. Write \( A = {V\Delta I}, V \) open, \( I \) meager. If \( A \) is nonmeager, \( V \) must be nonempty. Then \( A \) is comeager in every \( {U}_{n} \) contained in \( V \) .\n\nWe now prove (iii). Let \( A \) be comeager. Then trivially \( {U}_{n} \smallsetminus A \) is meager for all \( n \) . As \( {U}_{n} \) is open, it follows that \( {U}_{n} \smallsetminus A \) is meager in \( {U}_{n} \) . Since \( X \) is a Baire space, this implies that \( A \) is nonmeager in \( {U}_{n} \) . Conversely, let \( A \) be not comeager; i.e., \( {A}^{c} \) is not meager. So, there is \( {U}_{n} \) such that \( {A}^{c} \) is comeager in \( {U}_{n} \) ; i.e., \( A \) is meager in \( {U}_{n} \) .
Yes
Proposition 3.5.7 A topological group is Baire if and only if it is of second category in itself.
Proof. The \
No
Proposition 3.5.8 Let \( Y \) be a second countable topological space and \( f \) : \( X \rightarrow Y \) Baire measurable. Then there is a comeager set \( A \) in \( X \) such that \( f \mid A \) is continuous.
Proof. Take a countable base \( \left( {V}_{n}\right) \) for \( Y \) . Since \( f \) is Baire measurable, for each \( n \) there is a meager set \( {I}_{n} \) in \( X \) such that \( {f}^{-1}\left( {V}_{n}\right) \Delta {I}_{n} \) is open. Let \( I = \mathop{\bigcup }\limits_{n}{I}_{n} \) . Plainly, \( f \mid \left( {X \smallsetminus I}\right) \) is continuous.
Yes
Proposition 3.5.9 Let \( G \) be a completely metrizable group and \( H \) a second countable group. Then every Baire measurable homomorphism \( \varphi : G \rightarrow H \) is continuous. In particular, every Borel homomorphism \( \varphi : G \rightarrow H \) is continuous.
Proof. By 3.5.8, there is a meager set \( I \) in \( G \) such that \( \varphi \mid \left( {G \smallsetminus I}\right) \) is continuous. Now take any sequence \( \left( {g}_{k}\right) \) in \( G \) converging to an element \( g \) . Let\n\n\[ J = \left( {{g}^{-1} \cdot I}\right) \bigcup \mathop{\bigcup }\limits_{k}\left( {{g}_{k}^{-1} \cdot I}\right) . \]\n\nBy 2.4.7, \( J \) is meager. Since \( G \) is completely metrizable, it is of second category in itself by 2.5.6. In particular, \( J \neq G \) . Take any \( h \in G \smallsetminus J \) . Then, \( {g}_{k} \cdot h, g \cdot h \) are all in \( G \smallsetminus I \) . Further, \( {g}_{k} \cdot h \rightarrow g \cdot h \) as \( k \rightarrow \infty \) . Since \( \varphi \mid \left( {G \smallsetminus I}\right) \) is continuous, \( \varphi \left( {{g}_{k} \cdot h}\right) \rightarrow \varphi \left( {g \cdot h}\right) \) ; i.e., \( \varphi \left( {g}_{k}\right) \cdot \varphi \left( h\right) \rightarrow \varphi \left( g\right) \cdot \varphi \left( h\right) \) . Multiplying by \( {\left( \varphi \left( h\right) \right) }^{-1} \) from the right, we have \( \varphi \left( {g}_{k}\right) \rightarrow \varphi \left( g\right) \) .
Yes
Theorem 3.5.12 (Pettis theorem) Let \( G \) be a Baire topological group and \( H \) a nonmeager subset with \( {BP} \) . Then there is a neighborhood \( V \) of the identity contained in \( {\mathrm{H}}^{-1}\mathrm{H} \) .
Proof. Since \( H \) is nonmeager with BP, there is a nonempty open set \( U \) such that \( {H\Delta U} \) is meager. Let \( g \in U \) . Choose a neighborhood \( V \) of the identity such that \( {gV}{V}^{-1} \subseteq U \) . We show that for every \( h \in V, H \cap {Hh} \) is nonmeager, in particular, nonempty. It will then follow that \( V \subseteq {H}^{-1}H \) , and the proof will be complete.\n\nLet \( h \in H \) . Note that\n\n\[ \left( {U\bigcap {Uh}}\right) \Delta \left( {H\bigcap {Hh}}\right) \subseteq \left( {U\Delta H}\right) \bigcup \left( {\left( {U\Delta H}\right) h}\right) .\n\]\n\n\( \left( *\right) \)\n\nSo, \( \left( {U\bigcap {Uh}}\right) \Delta \left( {H\bigcap {Hh}}\right) \) is meager. As \( {gV} \subseteq U\bigcap {Uh} \) and \( G \) is Baire, \( U \cap {Uh} \) is nonmeager. Therefore, \( H \cap {Hh} \) is nonmeager by \( \left( \star \right) \) .
Yes
Corollary 3.5.13 Every nonmeager Borel subgroup \( H \) of a Polish group \( G \) is clopen.
Proof. Let \( H \) be a Borel subgroup of \( G \) that is not meager. By 3.5.12, \( H \) contains a neighborhood of the identity. Hence, \( H \) is open. Since \( {H}^{c} \) is the union of the remaining cosets of \( H \), which are all open, it is open too.
Yes
Lemma 3.5.14 Let \( X \) be a Baire space and \( Y \) second countable. Suppose \( A \subseteq X \times Y \) is a closed, nowhere dense set. Then\n\n\[ \left\{ {x \in X : {A}_{x}}\right. \text{is nowhere dense}\} \]\n\nis a dense \( {G}_{\delta } \) set.
Proof. Take any \( A \subseteq X \times Y \), closed and nowhere dense. Fix a countable base \( \left( {V}_{n}\right) \) for \( Y \) . Let \( U = {A}^{c} \) . Then \( U \) is dense and open. Let\n\n\[ {W}_{n} = \left\{ {x \in X : {U}_{x}\bigcap {V}_{n} \neq \varnothing }\right\} .\n\nAs\n\n\[ {W}_{n} = {\pi }_{X}\left( {U\bigcap \left( {X \times {V}_{n}}\right) }\right) \]\n\nit is open. Also, \( {W}_{n} \) is dense. Suppose not. Then \( \left( {X \smallsetminus \operatorname{cl}\left( {W}_{n}\right) }\right) \times {V}_{n} \) is a nonempty open set disjoint from \( U \) . As \( U \) is dense, this is a contradiction.\n\nSince for any \( x \in X \) ,\n\n\( {A}_{x} \) is nowhere dense \( \Leftrightarrow {U}_{x} \) is dense, it follows that\n\n\[ \left\{ {x \in X : {A}_{x}}\right. \text{is nowhere dense}\} = \mathop{\bigcap }\limits_{n}{W}_{n}\text{.} \]\n\nSince \( X \) is a Baire space, the result follows.
Yes
Lemma 3.5.15 Let \( X \) be a Baire space, \( Y \) second countable, and suppose \( A \subseteq X \times Y \) has BP. The following statements are equivalent.\n\n(i) \( A \) is meager.\n\n(ii) \( \left\{ {x \in X : {A}_{x}}\right. \) is meager \( \} \) is comeager.
Proof. (ii) follows from (i) by 3.5.14. Now assume that \( A \) is nonmeager. Since \( A \) has BP, there exist nonempty open sets \( U \) and \( V \) in \( X \) and \( Y \) respectively such that \( A \) is comeager in \( U \times V \) . Therefore, from what we have just proved, \( {A}^{*V} \) is comeager in \( U \) . Since \( U \) is nonmeager, \( {A}^{*V} \) is nonmeager. In particular, \( {A}^{\Delta X} \) is not meager; i.e.,(ii) is false.
Yes
Proposition 3.5.18 Let \( \\left( {X,\\mathcal{A}}\\right) \) be a measurable space and \( Y \) a Polish space. For every \( A \\in \\mathcal{A}\\bigotimes {\\mathcal{B}}_{Y} \) and \( U \) open in \( Y \), the sets \( {A}^{\\Delta U},{A}^{*U} \), and \( \\left\\{ {x \\in X : {A}_{x}}\\right. \) is meager in \( \\left. U\\right\\} \) are in \( \\mathcal{A} \) .
Proof. Fix a countable base \( \\left( {U}_{n}\\right) \) for \( Y \). Step 1. Let \[ \\mathcal{B} = \\left\\{ {A \\subseteq X \\times Y : {A}^{\\Delta U} \\in \\mathcal{A}\\text{ for all open }U}\\right\\} . \] We show that \( \\mathcal{A} \\otimes {\\mathcal{B}}_{Y} \\subseteq \\mathcal{B} \). Let \( A = B \\times V, B \\in \\mathcal{A} \), and \( V \) open in \( Y \). Then \( {A}^{\\Delta U} \) equals \( B \) if \( U \\cap V \\neq \\varnothing \). Otherwise it is empty. Hence, \( A \\in \\mathcal{B} \). Our proof will be complete if we show that \( \\mathcal{B} \) is closed under countable unions and complementation. For every sequence \( \\left( {A}_{n}\\right) \) of subsets \( X \\times Y \), \[ {\\left( \\mathop{\\bigcup }\\limits_{n}{A}_{n}\\right) }^{\\Delta U} = \\mathop{\\bigcup }\\limits_{n}{A}_{n}^{\\Delta U} \] So, \( \\mathcal{B} \) is closed under countable unions. Let \( A \\in \\mathcal{B} \) and \( U \) open in \( Y \). Let \( x \\in X \). We have \( {\\left( {A}^{c}\\right) }_{x} \) is meager in \( U \\Leftrightarrow {A}_{x} \) is comeager in \( U \) \[ \\Leftrightarrow \\forall {U}_{n} \\subseteq U\\left( {A}_{x}\\right. \\text{is nonmeager in}\\left. {U}_{n}\\right) \\text{.} \] Therefore, \[ {\\left( {A}^{c}\\right) }^{\\Delta U} = {\\left( \\mathop{\\bigcap }\\limits_{{{U}_{n} \\subseteq U}}{A}^{\\Delta {U}_{n}}\\right) }^{c} \] Hence, \( {A}^{c} \\in \\mathcal{B} \). Step 2. Let \( A \\in \\mathcal{A}\\bigotimes {\\mathcal{B}}_{Y} \) and \( U \) be open in \( Y \). Then \[ {A}^{*U} = \\mathop{\\bigcap }\\limits_{{{U}_{n} \\subseteq U}}{A}^{\\Delta {U}_{n}} \] Therefore, \( {A}^{*U} \\in \\mathcal{A} \) by step 1. The remaining part of the result follows easily.
Yes
Every \( \sigma \) -finite complete measure space is Marczewski complete.
We prove this now. Let \( \left( {X,\mathcal{B},\mu }\right) \) be a \( \sigma \) -finite complete measure space. First assume that \( {\mu }^{ * }\left( A\right) < \infty \) . Take \( \widehat{A} \) to be a measurable set containing \( A \) with \( {\mu }^{ * }\left( A\right) = \mu \left( \widehat{A}\right) \) . In the general case, write \( A = \bigcup {A}_{n} \) such that \( {\mu }^{ * }\left( {A}_{n}\right) < \infty \) . Since \( \mu \) is \( \sigma \) -finite, this is possible. Take \( \widehat{A} = \bigcup {\widehat{A}}_{n} \) .
Yes
Example 3.5.21 Let \( X \) be a topological space and \( A \subseteq X \) . Take \( {A}^{ * } \) to be the union of all open sets \( U \) such that \( A \) is comeager in \( U \) . We first show that \( {A}^{ * } \smallsetminus A \) is meager.
Let \( \mathcal{U} \) be a maximal family of pairwise disjoint open sets \( U \) such that \( A \) is comeager in \( U \) . Let \( W = \bigcup \mathcal{U} \) . By the maximality of \( \mathcal{U},{A}^{ * } \subseteq \operatorname{cl}\left( W\right) \) . By the Banach category theorem, \( A \) is comeager in \( W \) . Now note that\n\n\[ \n{A}^{ * } \smallsetminus A \subseteq \left( {{A}^{ * } \smallsetminus W}\right) \bigcup \left( {W \smallsetminus A}\right) \subseteq \left( {\operatorname{cl}\left( W\right) \smallsetminus W}\right) \bigcup \left( {W \smallsetminus A}\right) .\n\]\n\nThis shows that \( {A}^{ * } \smallsetminus A \) is meager. Let \( B \) be any meager \( {F}_{\sigma } \) set containing \( {A}^{ * } \smallsetminus A \) . Take \( \widehat{A} = {A}^{ * } \cup B \) .
Yes
Theorem 3.5.22 (Marczewski) If \( \left( {X,\mathcal{B}}\right) \) is a measurable space with \( \mathcal{B} \) Marczewski complete, then \( \mathcal{B} \) is closed under the Souslin operation.
Proof. Let \( \left\{ {{B}_{s} : s \in {\mathbb{N}}^{ < \mathbb{N}}}\right\} \) be a system of sets in \( \mathcal{B} \) . We have to show that \( B = \mathcal{A}\left( \left\{ {B}_{s}\right\} \right) \in \mathcal{B} \) . Without loss of generality we assume that the system \( \left\{ {B}_{s}\right\} \) is regular. For \( s \in {\mathbb{N}}^{ < \mathbb{N}} \), let\n\n\[ \n{B}^{s} = \mathop{\bigcup }\limits_{{\{ \alpha : s \prec \alpha \} }}\mathop{\bigcap }\limits_{n}{B}_{\alpha \mid n} \subseteq {B}_{s} \n\]\n\nNote that \( {B}^{e} = B \) and \( {B}^{s} = \mathop{\bigcup }\limits_{n}{B}^{s \uparrow n} \) for all \( s \) . For each \( s \in {\mathbb{N}}^{ < \mathbb{N}} \), choose a minimal \( \mathcal{B} \) -cover \( {\widehat{B}}^{s} \) of \( {B}^{s} \) . Since \( {B}^{s} \subseteq {B}_{s} \), by replacing \( {\widehat{B}}^{s} \) by \( {B}_{s}\bigcap {\widehat{B}}^{s} \) we may assume that \( {\widehat{B}}^{s} \subseteq {B}_{s} \) . Further, by replacing \( {\widehat{B}}^{s} \) by \( \mathop{\bigcap }\limits_{{t \preccurlyeq s}}{\widehat{B}}^{t} \), we can assume that \( \left\{ {{\widehat{B}}^{s} : s \in {\mathbb{N}}^{ < \mathbb{N}}}\right\} \) is regular. Let\n\n\[ \n{C}_{s} = {\widehat{B}}^{s} \smallsetminus \mathop{\bigcup }\limits_{n}{\widehat{B}}^{s}{}^{rn}. \n\]\n\nSince \( {B}^{s} = \mathop{\bigcup }\limits_{n}{B}^{{s}^{ \frown }n} \subseteq \mathop{\bigcup }\limits_{n}{\widehat{B}}^{{s}^{ \frown }n} \), every subset of \( {C}_{s} \) is in \( \mathcal{B} \) . Let \( C = \mathop{\bigcup }\limits_{s}{C}_{s} \) .\n\nClaim: \( {\widehat{B}}^{e} \smallsetminus C \subseteq B \) .\n\nAssuming the claim, we complete the proof as follows. Since \( {\widehat{B}}^{e} \smallsetminus B \subseteq C \) and since every subset of \( C \) is in \( \mathcal{B} \), it follows that \( {\widehat{B}}^{e} \smallsetminus B \in \mathcal{B} \) . As \( B = \) \( {\widehat{B}}^{e} \smallsetminus \left( {{\widehat{B}}^{e} \smallsetminus B}\right) \), it belongs to \( \mathcal{B} \) .\n\nProof of the claim. Let \( x \in {\widehat{B}}^{e} \smallsetminus C \) . Since \( x \notin C, x \notin {C}_{e} \) . Since \( x \in {\widehat{B}}^{e} \), there is \( \alpha \left( 0\right) \in \mathbb{N} \) such that \( x \in {\widehat{B}}^{\alpha \left( 0\right) } \) . Suppose \( n > 0 \) and \( \alpha \left( 0\right) ,\alpha \left( 1\right) ,\ldots ,\alpha \left( {n - 1}\right) \) have been defined such that \( x \in {\widehat{B}}^{s} \), where \( s = \) \( \left( {\alpha \left( 0\right) ,\alpha \left( 1\right) ,\ldots ,\alpha \left( {n - 1}\right) }\right) \) . Since \( x \notin {C}_{s} \), there is \( \alpha \left( n\right) \in \mathbb{N} \) such that \( x \in \) \( {\widehat{B}}^{s \land \alpha \left( n\right) } \) . Since \( {\widehat{B}}^{\alpha \mid n} \subset {B}_{\alpha \mid n} \) for all \( n \), we conclude that \( x \in B \) .
Yes
Proposition 3.6.1 (i) For every \( 1 \leq \alpha < {\omega }_{1} \), \[ {\mathbf{\sum }}_{\alpha }^{0},{\mathbf{\Pi }}_{\alpha }^{0} \subseteq {\mathbf{\Delta }}_{\alpha + 1}^{0} \]
Proof. Since every closed (open) set in a metrizable space is a \( {G}_{\delta } \) set (respectively an \( {F}_{\sigma } \) set),(i) is true for \( \alpha = 1 \) . A simple transfinite induction argument completes the proof of (i) for all \( \alpha \) .
No
Proposition 3.6.3 Every set of additive class \( \alpha > 2 \) is a countable disjoint union of multiplicative class \( < \alpha \) sets.
Proof. Let \( A \) be a set of additive class \( \alpha > 2 \) . Write \( A = \bigcup {A}_{n} \), where \( {A}_{n} \) is of multiplicative class less than \( \alpha \) . Let \( {B}_{n} = {\left( \mathop{\bigcup }\limits_{{i < n}}{A}_{i}\right) }^{c} \) . Then \( {B}_{n} \) is of additive class \( < \alpha \) . Write \( {B}_{n} = \mathop{\bigcup }\limits_{k}{B}_{k}^{n} \), where the \( {B}_{k}^{n} \) ’s are pairwise disjoint ambiguous class \( < \alpha \) sets. This is possible since \( \alpha > 2 \) . We have\n\n\[ A = {A}_{0}\bigcup \left( {{A}_{1} \cap {B}_{0}}\right) \bigcup \left( {{A}_{2} \cap {B}_{1}}\right) \cup \cdots \]\n\n\[ = \;{A}_{0}\bigcup \mathop{\bigcup }\limits_{{n \geq 1}}\mathop{\bigcup }\limits_{k}\left( {{A}_{n}\bigcap {B}_{k}^{n - 1}}\right) ,\]\n\nand the result follows.
Yes
Proposition 3.6.3 Every set of additive class \( \alpha > 2 \) is a countable disjoint union of multiplicative class \( < \alpha \) sets.
Proof. Let \( A \) be a set of additive class \( \alpha > 2 \) . Write \( A = \bigcup {A}_{n} \), where \( {A}_{n} \) is of multiplicative class less than \( \alpha \) . Let \( {B}_{n} = {\left( \mathop{\bigcup }\limits_{{i < n}}{A}_{i}\right) }^{c} \) . Then \( {B}_{n} \) is of additive class \( < \alpha \) . Write \( {B}_{n} = \mathop{\bigcup }\limits_{k}{B}_{k}^{n} \), where the \( {B}_{k}^{n} \) ’s are pairwise disjoint ambiguous class \( < \alpha \) sets. This is possible since \( \alpha > 2 \) . We have\n\n\[ A = {A}_{0}\bigcup \left( {{A}_{1} \cap {B}_{0}}\right) \bigcup \left( {{A}_{2} \cap {B}_{1}}\right) \cup \cdots \]\n\n\[ = \;{A}_{0}\bigcup \mathop{\bigcup }\limits_{{n \geq 1}}\mathop{\bigcup }\limits_{k}\left( {{A}_{n}\bigcap {B}_{k}^{n - 1}}\right) ,\]\n\nand the result follows.
Yes
Theorem 3.6.6 Let \( 1 \leq \alpha < {\omega }_{1} \) and \( {\mathbf{\Gamma }}_{\alpha } \) the pointclass of \( {\mathbf{\Pi }}_{\alpha }^{0} \) or \( {\mathbf{\sum }}_{\alpha }^{0} \) sets. For every second countable metrizable space \( Y \), there exists a \( U \in {\mathbf{\Gamma }}_{\alpha }\left( {{\mathbb{N}}^{\mathbb{N}} \times Y}}\right) \) such that \[ A \in {\mathbf{\Gamma }}_{\alpha }\left( Y\right) \Leftrightarrow \left( {\exists x \in {\mathbb{N}}^{\mathbb{N}}}\right) \left( {A = {U}_{x}}\right) . \] We call such a set \( U \) universal for \( {\mathbf{\Gamma }}_{\alpha }\left( Y\right) \) .
Proof. We proceed by induction on \( \alpha \) . Let \( \left( {V}_{n}\right) \) be a countable base for the topology of \( Y \) with at least one \( {V}_{n} \) empty. Define \( U \subseteq {\mathbb{N}}^{\mathbb{N}} \times Y \) by \[ \left( {x, y}\right) \in U \Leftrightarrow y \in \mathop{\bigcup }\limits_{n}{V}_{x\left( n\right) } \] Evidently, \( A \) is open in \( Y \) if and only if \( A = {U}_{x} \) for some \( x \) . It remains to show that \( U \) is open. Let \( \left( {{x}_{0},{y}_{0}}\right) \in U \) . Then there is an \( n \) such that \( {y}_{0} \in {V}_{{x}_{0}\left( n\right) } \) . Then \[ \left( {{x}_{0},{y}_{0}}\right) \in \left\{ {x \in {\mathbb{N}}^{\mathbb{N}} : x\left( n\right) = {x}_{0}\left( n\right) }\right\} \times {V}_{{x}_{0}\left( n\right) } \subseteq U. \] Thus \( U \) is open. Let \( W = {U}^{c} \), where \( U \subseteq {\mathbb{N}}^{\mathbb{N}} \times Y \) is universal for open sets. Clearly, \( W \) is universal for closed sets. The result for \( \alpha = 1 \) is proved. Suppose \( \alpha > 1 \) and the result has been proved for all \( \beta < \alpha \) . Case 1: \( \alpha \) is a limit ordinal Fix a sequence of countable ordinals \( \left( {\alpha }_{n}\right) ,1 < {\alpha }_{n} < \alpha \), such that \( \alpha = \sup {\alpha }_{n} \) . Let \( {U}_{n} \) be universal for multiplicative class \( {\alpha }_{n}, n \in \mathbb{N} \) . For \( x \in {\mathbb{N}}^{\mathbb{N}} \) and \( n \in \mathbb{N} \), define \( {x}_{n} \in {\mathbb{N}}^{\mathbb{N}} \) by \[ {x}_{n}\left( m\right) = x\left( {{2}^{n}\left( {{2m} + 1}\right) - 1}\right) . \] \( \left( *\right) \) For each \( n, x \rightarrow {x}_{n} \) is a continuous function. Define \( U \subseteq {\mathbb{N}}^{\mathbb{N}} \times Y \) by \[ \left( {x, y}\right) \in U \Leftrightarrow \left( {\exists n}\right) \left( {\left( {{x}_{n}, y}\right) \in {U}_{n}}\right) \] It is routine to check that \( U \) is universal for \( {\mathbf{\sum }}_{\alpha }^{0}\left( Y\right) \) . Case 2: \( \alpha = \beta + 1 \), a successor ordinal Fix a universal \( {\mathbf{\Pi }}_{\beta }^{0} \) set \( P \subseteq {\mathbb{N}}^{\mathbb{N}} \times Y \) . Define \( U \subseteq {\mathbb{N}}^{\mathbb{N}} \times Y \) by \[ \left( {x, y}\right) \in U \Leftrightarrow \left( {\exists n}\right) \left( {\left( {{x}_{n}, y}\right) \in P}\right) , \] where \( {x}_{n} \) is as defined in \( \left( \star \right) \) . Clearly, \( U \) is universal for \( {\mathbf{\sum }}_{\alpha }^{0}\left( Y\right) \) . Having defined a universal \( {\mathbf{\sum }}_{\alpha }^{0} \) set \( U \subseteq {\mathbb{N}}^{\mathbb{N}} \times Y \), note that \( {U}^{c} \) is universal for \( {\Pi }_{\alpha }^{0}\left( Y\right) \) .
Yes
Theorem 3.6.7 Let \( 1 \leq \alpha < {\omega }_{1} \) and \( {\mathbf{\Gamma }}_{\alpha } \) the pointclass of additive or multiplicative class \( \alpha \) sets. Then for every uncountable Polish space \( X \) , there is a \( U \in {\mathbf{\Gamma }}_{\alpha }\left( {X \times X}\right) \) universal for \( {\mathbf{\Gamma }}_{\alpha }\left( X\right) \) .
Proof. Since \( X \) is uncountable Polish, it has a subset, say \( Y \), homeomorphic to \( {\mathbb{N}}^{\mathbb{N}} \) . By 3.6.6, there is \( U \subseteq Y \times X \) universal for \( {\mathbf{\Gamma }}_{\alpha }\left( X\right) \) . By 3.6.4(iii), \( V \cap \left( {Y \times X}\right) = U \) for some \( V \in {\mathbf{\Gamma }}_{\alpha }\left( {X \times X}\right) \) . The set \( V \) is universal for \( {\mathbf{\Gamma }}_{\alpha }\left( X\right) \) .
Yes
Corollary 3.6.8 Let \( X \) be any uncountable Polish space and \( 1 \leq \alpha < {\omega }_{1} \) . Then there exists an additive class \( \alpha \) set that is not of multiplicative class \( \alpha \) .
Proof. Let \( U \subseteq X \times X \) be universal for \( {\mathbf{\sum }}_{\alpha }^{0}\left( X\right) \) . Take\n\n\[ A = \{ x \in X : \left( {x, x}\right) \in U\} . \]\n\nSince \( {\mathbf{\sum }}_{\alpha }^{0} \) is closed under continuous preimages, \( A \) is of additive class \( \alpha \) . We claim that \( A \) is not of multiplicative class \( \alpha \) . To the contrary, suppose \( A \) is of multiplicative class \( \alpha \) . Choose \( {x}_{0} \in X \) such that \( {A}^{c} = {U}_{{x}_{0}} \) . Then\n\n\[ {x}_{0} \in {A}^{c} \Leftrightarrow \left( {{x}_{0},{x}_{0}}\right) \in U \Leftrightarrow {x}_{0} \in A. \]\n\nThis is a contradiction.
Yes
Proposition 3.6.9 Let a pointclass \( \mathbf{\Delta } \) be closed under complementation and continuous preimages. Then for no Polish space \( X \) is there a set in \( \mathbf{\Delta }\left( {X \times X}\right) \) universal for \( \mathbf{\Delta }\left( X\right) \) .
Proof. Suppose there is a Polish space \( X \) and a \( U \in \mathbf{\Delta }\left( {X \times X}\right) \) universal for \( \mathbf{\Delta }\left( X\right) \) . Take\n\n\[ A = \{ x \in X : \left( {x, x}\right) \in U\} . \]\n\nSince \( \mathbf{\Delta } \) is closed under continuous preimages, \( A \in \mathbf{\Delta } \) . As \( \mathbf{\Delta } \) is closed under complementation, \( {A}^{c} \in \mathbf{\Delta } \) . Let \( {A}^{c} = {U}_{{x}_{0}} \) for some \( {x}_{0} \in X \) . Then\n\n\[ {x}_{0} \in {A}^{c} \Leftrightarrow \left( {{x}_{0},{x}_{0}}\right) \in U \Leftrightarrow {x}_{0} \in A. \]\n\nThis is a contradiction.
Yes
Theorem 3.6.10 (Reduction theorem for additive classes) Let \( X \) be a metrizable space and \( 1 < \alpha < {\omega }_{1} \) . Suppose \( \left( {A}_{n}\right) \) is a sequence of additive class \( \alpha \) sets in \( X \) . Then there exist \( {B}_{n} \subseteq {A}_{n} \) such that\n\n(a) The \( {B}_{n} \) ’s are pairwise disjoint sets of additive class \( \alpha \), and\n\n(b) \( \mathop{\bigcup }\limits_{n}{A}_{n} = \mathop{\bigcup }\limits_{n}{B}_{n} \) .\n\n(See Figure 3.1.) Consequently the \( {B}_{n} \) ’s are of ambiguous class \( \alpha \) if \( \mathop{\bigcup }\limits_{n}{A}_{n} \) is so.\n\nThe result is also true for \( \alpha = 1 \) if \( X \) is zero-dimensional and second countable.
Proof. Write\n\n\[ \n{A}_{n} = \mathop{\bigcup }\limits_{m}{C}_{nm} \n\]\n\n\( \left( *\right) \)\n\nwhere the \( {C}_{nm} \) ’s are of ambiguous class \( \alpha \) . If \( \alpha > 1 \), this is always possible. If \( \alpha = 1 \), it is possible if \( X \) is zero-dimensional and second countable (3.6.1). Enumerate \( \left\{ {{C}_{nm} : n, m \in \mathbb{N}}\right\} \) in a single sequence, say \( \left( {D}_{i}\right) \) . Let\n\n\[ \n{E}_{i} = {D}_{i} \smallsetminus \mathop{\bigcup }\limits_{{j < i}}{D}_{j} \n\]\n\nTake\n\n\[ \n{B}_{n} = \bigcup \left\{ {{E}_{i} : {E}_{i} \subseteq {A}_{n}\& \left( {\forall m < n}\right) \left( {{E}_{i} \nsubseteq {A}_{m}}\right) }\right\} .\n\]
Yes
Theorem 3.6.11 (Separation theorem for multiplicative classes) Let \( X \) be metrizable and \( 1 < \alpha < {\omega }_{1} \) . Then for every sequence \( \left( {A}_{n}\right) \) of multiplicative class \( \alpha \) sets with \( \bigcap {A}_{n} = \varnothing \), there exist ambiguous class \( \alpha \) sets \( {B}_{n} \supseteq {A}_{n} \) with \( \bigcap {B}_{n} = \varnothing \) .
Proof. By 3.6.10, there exist pairwise disjoint additive class \( \alpha \) sets \( {C}_{n} \subseteq \) \( {A}_{n}^{c} \) such that \( \mathop{\bigcup }\limits_{n}{C}_{n} = \mathop{\bigcup }\limits_{n}{A}_{n}^{c} = X \) . Obviously, the \( {C}_{n} \) ’s are of ambiguous class \( \alpha \) . Take \( {B}_{n} = {C}_{n}^{c} \) .
Yes
Proposition 3.6.13 Let \( X \) be metrizable and \( 2 < \alpha < {\omega }_{1} \) . Suppose \( A \in \) \( {\mathbf{\Delta }}_{\alpha }^{0}\left( X\right) \) . Then there is a sequence \( \left( {A}_{n}\right) \) of ambiguous class \( < \alpha \) sets such that \( A = \lim {A}_{n} \) .
Proof. We write\n\n\[ A = \mathop{\bigcup }\limits_{n}{C}_{n} = \mathop{\bigcap }\limits_{n}{D}_{n} \]\n\nwhere the \( {C}_{n} \) ’s are multiplicative class \( < \alpha \) sets, the \( {D}_{n} \) ’s are additive class \( < \alpha \) sets, \( {C}_{n} \subseteq {C}_{n + 1} \), and \( {D}_{n + 1} \subseteq {D}_{n} \) . By 3.6.11, there is a set \( {A}_{n} \) of ambiguous class \( < \alpha \) such that\n\n\[ {C}_{n} \subseteq {A}_{n} \subseteq {D}_{n} \]\n\nThen \( A = \lim {A}_{n} \) as we now show. Let \( x \in \lim \sup {A}_{n} \) . Thus, \( x \in {A}_{n} \) for infinitely many \( n \) . Then \( x \in {D}_{n} \) for infinitely many \( n \) and hence for all \( n \) . Therefore,\n\n\[ \lim \sup {A}_{n} \subseteq A\text{.} \]\n\n(1)\n\nNow let \( x \in A \) . Then \( x \in {C}_{n} \) for all but finitely many \( n \) . Since \( {C}_{n} \subseteq {A}_{n} \) for all \( n \), \n\n\[ A \subseteq \liminf {A}_{n}\text{.} \]\n\n(2)\n\nThe result follows from (1) and (2).
Yes
Proposition 3.6.14 Let \( 2 < \alpha < {\omega }_{1} \) and \( X \) an uncountable Polish space. There exists a sequence \( {A}_{n} \) in \( {\mathbf{\Pi }}_{\alpha }^{0}\left( X\right) \) with \( \limsup {A}_{n} = \varnothing \) such that there does not exist \( {B}_{n} \supseteq {A}_{n} \) in \( {\mathbf{\sum }}_{\alpha }^{0}\left( X\right) \) with \( \lim \sup {B}_{n} = \varnothing \) .
Proof. Take \( A \in {\mathbf{\sum }}_{\alpha + 1}^{0}\left( X\right) \smallsetminus {\mathbf{\Pi }}_{\alpha + 1}^{0}\left( X\right) \) . Such a set exists by 3.6.8. By 3.6.3, we can find disjoint sets \( {A}_{n} \in {\mathbf{\Pi }}_{\alpha }^{0}\left( X\right) \) with union \( A \) . Quite trivially, \( \lim \sup {A}_{n} = \varnothing \) . Suppose there exist \( {B}_{n} \supseteq {A}_{n} \) in \( {\mathbf{\sum }}_{\alpha }^{0}\left( X\right) \) with \( \lim \sup {B}_{n} = \) \( \varnothing \) . We shall get a contradiction.\n\nBy 3.6.11, there is a set \( {C}_{n} \in {\mathbf{\Delta }}_{\alpha }^{0}\left( X\right) \) such that \( {A}_{n} \subseteq {C}_{n} \subseteq {B}_{n} \) . Note that \( \lim \sup {C}_{n} = \varnothing \) . As the sets \( {A}_{n} \) are in \( {\mathbf{\Delta }}_{\alpha + 1}^{0}\left( X\right) \), by 3.6.13 there are sets \( {A}_{n}^{k} \in {\mathbf{\Delta }}_{\alpha }^{0}\left( X\right) \) such that \( {A}_{n} = \mathop{\lim }\limits_{k}{A}_{n}^{k} \) . Now define\n\n\[ \n{D}_{k} = \left( {{A}_{1}^{k}\bigcap {C}_{1}}\right) \bigcup \left( {{A}_{2}^{k}\bigcap {C}_{2}}\right) \bigcup \cdots \bigcup \left( {{A}_{k}^{k}\bigcap {C}_{k}}\right) .\n\]\n\nThen \( {D}_{k} \in {\mathbf{\Delta }}_{\alpha }^{0}\left( X\right) \) . It is now fairly easy to check that \( \lim \sup {D}_{k} \subseteq A \subseteq \) \( \liminf {D}_{k} \), so \( A = \lim {D}_{k} \) . This implies that \( A \in {\mathbf{\Delta }}_{\alpha + 1}^{0}\left( X\right) \), and we have arrived at a contradiction.
Yes
Theorem 3.6.15 Suppose \( X, Y \) are metrizable spaces with \( Y \) second countable and \( 2 < \alpha < {\omega }_{1} \) . Then for every Borel function \( f : X \rightarrow Y \) of class \( \alpha \), there is a sequence \( \left( {f}_{n}\right) \) of Borel maps from \( X \) to \( Y \) of class \( < \alpha \) such that \( {f}_{n} \rightarrow f \) pointwise.
Proof of 3.6.15. Let \( d \) be a totally bounded compatible metric on \( Y \) . Such a metric exists by 2.1.32 and 2.3.12. By 3.6.16, there is a sequence \( \left( {g}_{m}\right) \) of class \( \alpha \) functions, with range finite, converging to \( f \) uniformly. Without any loss of generality, we assume that for all \( x \)
No
Lemma 3.6.16 Suppose \( Y \) is totally bounded. Then every \( f : X \rightarrow Y \) of class \( \alpha ,\alpha > 1 \), is the limit of a uniformly convergent sequence of class \( \alpha \) functions \( {f}_{n} : X \rightarrow Y \) of finite range.
Proof. Take any \( \epsilon > 0 \) . We shall obtain a function \( g : X \rightarrow Y \) of class \( \alpha \) such that the range of \( g \) is finite and \( d\left( {g\left( x\right), f\left( x\right) }\right) < \epsilon \) for all \( x \) . Let \( \left\{ {{y}_{1},{y}_{2},\ldots ,{y}_{n}}\right\} \) be an \( \epsilon \) -net in \( Y \) . Set\n\n\[ \n{A}_{i} = {f}^{-1}\left( {B\left( {{y}_{i},\epsilon }\right) }\right) \n\]\n\nThe sets \( {A}_{1},{A}_{2},\ldots ,{A}_{n} \) are of additive class \( \alpha \) with union \( X \) . By 3.6.10, there are pairwise disjoint ambiguous class \( \alpha \) sets \( {B}_{1},{B}_{2},\ldots ,{B}_{n} \) such that\n\n\[ \n{B}_{1} \subseteq {A}_{1},{B}_{2} \subseteq {A}_{2},\ldots ,{B}_{n} \subseteq {A}_{n} \n\]\n\nand\n\n\[ \n\bigcup {B}_{i} = \bigcup {A}_{i} = X \n\]\n\nDefine \( g : X \rightarrow Y \) by\n\n\[ \ng\left( x\right) = {y}_{i}\text{ if }x \in {B}_{i} \n\]\n\nThen \( d\left( {f\left( x\right), g\left( x\right) }\right) < \epsilon \) for all \( x \) .
Yes
Lemma 3.6.17 Let \( f : X \rightarrow Y \) be of class \( \alpha > 2 \) with range contained in a finite set \( E = \left\{ {{y}_{1},{y}_{2},\ldots ,{y}_{n}}\right\} \) . Then \( f \) is the limit of a sequence of functions of class \( < \alpha \) with values in \( E \) .
Proof. Let \( {A}_{i} = {f}^{-1}\left( {y}_{i}\right), i = 1,2,\ldots, n \) . Then \( {A}_{1},{A}_{2},\ldots ,{A}_{n} \) are pairwise disjoint, ambiguous class \( \alpha \) sets with union \( X \) . By 3.6.13, for each \( i \) there is a sequence \( \left( {A}_{im}\right) \) of sets of ambiguous class \( < \alpha \) such that \( {A}_{i} = \) \( \mathop{\lim }\limits_{m}{A}_{im} \) . Fix \( m \) . Let\n\n\[ \n{B}_{1}^{m} = {A}_{1m},{B}_{2}^{m} = {A}_{2m} \smallsetminus {A}_{1m},\ldots ,{B}_{n}^{m} = {A}_{nm} \smallsetminus \mathop{\bigcup }\limits_{{j < n}}{A}_{jm} \n\]\n\nand\n\n\[ \n{B}_{n + 1}^{m} = X \smallsetminus \mathop{\bigcup }\limits_{{j \leq n}}{A}_{jm} \n\]\n\nEvidently, the sets \( {B}_{1}^{m},{B}_{2}^{m},\ldots ,{B}_{n + 1}^{m} \) are pairwise disjoint and of ambiguous class less than \( \alpha \) with union \( X \) . So there is a function \( {f}_{m} : X \rightarrow Y \) of class \( < \alpha \) satisfying\n\n\[ \n{f}_{m}\left( x\right) = {y}_{i},\text{ if }x \in {B}_{i}^{m},\;1 \leq i \leq n. \n\]\n\nWe claim that \( {f}_{m}\left( {x}_{0}\right) \rightarrow f\left( {x}_{0}\right) \) for all \( {x}_{0} \in X \) . Assume that \( {x}_{0} \in {A}_{i} \) . So, \( f\left( {x}_{0}\right) = {y}_{i} \) . Since \( {x}_{0} \notin \lim \mathop{\sup }\limits_{m}{A}_{jm} \) for all \( j \neq i \), there is an integer \( M \) such that \( {x}_{0} \notin {A}_{jm} \) for \( m > M \) and \( j \neq i \) . Since \( {x}_{0} \in \lim \mathop{\inf }\limits_{m}{A}_{im} \), we can further assume that \( {x}_{0} \in {A}_{im} \) for all \( m > M \) . Thus, \( {f}_{m}\left( {x}_{0}\right) = {y}_{i} \) for all \( m > M \) . Hence, \( {f}_{m} \rightarrow f \) pointwise.
Yes
Proposition 4.1.1 Let \( X \) be a Polish space and \( A \subseteq X \) . The following statements are equivalent.\n\n(i) \( A \) is analytic.\n\n(ii) There is a Polish space \( Y \) and a Borel set \( B \subseteq X \times Y \) whose projection is \( A \) .\n\n(iii) There is a continuous map \( f : {\mathbb{N}}^{\mathbb{N}} \rightarrow X \) whose range is \( A \) .\n\n(iv) There is a closed subset \( C \) of \( X \times {\mathbb{N}}^{\mathbb{N}} \) whose projection is \( A \) .\n\n(v) For every uncountable Polish space \( Y \) there is a \( {G}_{\delta } \) set \( B \) in \( X \times Y \) whose projection is \( A \) .
Proof. (i) trivially implies (ii).\n\nLet \( Y \) be a Polish space and \( B \) a Borel subset of \( X \times Y \) such that \( {\pi }_{X}\left( B\right) = A \), where \( {\pi }_{X} : X \times Y \rightarrow X \) is the projection map. By 3.3.17, there is a continuous map \( g \) from \( {\mathbb{N}}^{\mathbb{N}} \) onto \( B \) . Take \( f = {\pi }_{X} \circ g \) . Since the range of \( f \) is \( A \) ,(ii) implies (iii).\n\nSince the graph of a continuous map \( f : {\mathbb{N}}^{\mathbb{N}} \rightarrow X \) is a closed subset of \( {\mathbb{N}}^{\mathbb{N}} \times X \) with projection \( A \) ,(iii) implies (iv).\n\nBy 2.6.5, every uncountable Polish space \( Y \) contains a homeomorph of \( {\mathbb{N}}^{\mathbb{N}} \), which is necessarily a \( {G}_{\delta } \) set in \( Y \) . Therefore,(iv) implies (v).\n\n(i) trivially follows from (v).
Yes
Proposition 4.1.2 (i) The pointclass \( {\mathbf{\sum }}_{1}^{1} \) is closed under countable unions, countable intersections and Borel preimages. Consequently, \( {\mathbf{\Pi }}_{1}^{1} \) is closed under these operations.
Proof. We first prove (i).\n\nClosure under Borel preimages: Let \( X \) and \( Z \) be Polish spaces, \( A \subseteq X \) analytic, and \( f : Z \rightarrow X \) a Borel map. Choose a Borel subset \( B \) of \( X \times X \) whose projection is \( A \) . Let\n\n\[ C = \{ \left( {z, x}\right) \in Z \times X : \left( {f\left( z\right), x}\right) \in B\} . \]\n\nThe set \( C \) is Borel, and \( {\pi }_{X}\left( C\right) = {f}^{-1}\left( A\right) \) . So \( {f}^{-1}\left( A\right) \) is analytic.\n\nClosure under countable unions and countable intersections: Let \( {A}_{0},{A}_{1},{A}_{2},\ldots \) be analytic subsets of \( X \) . By 4.1.1, there are Borel subsets \( {B}_{0},{B}_{1},{B}_{2},\ldots \) of \( X \times {\mathbb{N}}^{\mathbb{N}} \) whose projections are \( {A}_{0},{A}_{1},{A}_{2},\ldots \) respectively. Take\n\n\[ C = \left\{ {\left( {x,\alpha }\right) \in X \times {\mathbb{N}}^{\mathbb{N}} : \left( {x,{\alpha }^{ * }}\right) \in {B}_{\alpha \left( 0\right) }}\right\} \]\n\nand\n\n\[ D = \left\{ {\left( {x,\alpha }\right) \in X \times {\mathbb{N}}^{\mathbb{N}} : \left( {x,{f}_{i}\left( \alpha \right) }\right) \in {B}_{i}\text{ for every }i}\right\} , \]\n\nwhere \( {\alpha }^{ * }\left( i\right) = \alpha \left( {i + 1}\right) \) and \( \left( {{f}_{0},{f}_{1},{f}_{2},\ldots }\right) : {\mathbb{N}}^{\mathbb{N}} \rightarrow {\left( {\mathbb{N}}^{\mathbb{N}}\right) }^{\mathbb{N}} \) is a continuous surjection. Note that the map \( \alpha \rightarrow {\alpha }^{ * } \) is also continuous. Hence, the sets \( C \) and \( D \) are Borel with projections \( \mathop{\bigcup }\limits_{i}{A}_{i} \) and \( \mathop{\bigcap }\limits_{i}{A}_{i} \) respectively. We have shown that \( {\mathbf{\sum }}_{1}^{1} \) is closed under countable unions and countable intersections. The closure properties of \( {\mathbf{\Pi }}_{1}^{1} \) follow.
Yes
Theorem 4.1.4 For every Polish space \( X \), there is an analytic set \( U \subseteq \) \( {\mathbb{N}}^{\mathbb{N}} \times X \) such that \( A \subseteq X \) is analytic if and only if \( A = {U}_{\alpha } \) for some \( \alpha \) ; i.e., \( U \) is universal for \( {\mathbf{\sum }}_{1}^{1}\left( X\right) \) .
Proof. Let \( C \subseteq {\mathbb{N}}^{\mathbb{N}} \times \left( {X \times {\mathbb{N}}^{\mathbb{N}}}\right) \) be a universal closed set. The existence of such a set is shown in 3.6.6. Let\n\n\[ U = \left\{ {\left( {\alpha, x}\right) \in {\mathbb{N}}^{\mathbb{N}} \times X : \left( {\alpha, x,\beta }\right) \in C\text{ for some }\beta }\right\} .\n\]\n\nAs \( U = {\exists }^{{\mathbb{N}}^{\mathbb{N}}}C \), it follows that \( U \in {\mathbf{\sum }}_{1}^{1} \) . Let \( A \subseteq X \) be \( {\mathbf{\sum }}_{1}^{1} \) . Choose a closed set \( F \subseteq X \times {\mathbb{N}}^{\mathbb{N}} \) whose projection is \( A \) (4.1.1). Let \( \alpha \in {\mathbb{N}}^{\mathbb{N}} \) be such that \( F = {C}_{\alpha } \) . Then \( A = {U}_{\alpha } \) .
Yes
Theorem 4.1.5 Let \( X \) be an uncountable Polish space.\n\n(i) There is an analytic set \( U \subseteq X \times X \) such that for every analytic set \( A \subseteq X \), there is an \( x \in X \) with \( A = {U}_{x} \).\n\n(ii) There is a subset of \( X \) that is analytic but not Borel.
Proof. (i) Since \( X \) is uncountable Polish, it contains a homeomorph of \( {\mathbb{N}}^{\mathbb{N}} \), say \( Y \) (2.6.5). The set \( Y \) is a \( {G}_{\delta } \) set in \( X \) (2.2.7). Take \( U \subseteq Y \times X \) as in 4.1.4.\n\n(ii) Let\n\n\[ A = \{ x \in X : \left( {x, x}\right) \in U\} .\n\]\n\nSince \( {\mathbf{\sum }}_{1}^{1} \) is closed under continuous preimages, \( A \in {\mathbf{\sum }}_{1}^{1} \) . We claim that \( A \) is not coanalytic and hence not Borel. Suppose not. Then \( {A}^{c} \) analytic. Take an \( {x}_{0} \in X \) such that \( {A}^{c} = {U}_{{x}_{0}} \) . Then\n\n\[ {x}_{0} \in A \Leftrightarrow \left( {{x}_{0},{x}_{0}}\right) \in U \Leftrightarrow {x}_{0} \in {A}^{c}.\n\]\n\nWe have arrived at a contradiction.
Yes
Proposition 4.1.7 Let \( n \) be a positive integer.\n\n(i) The pointclasses \( {\mathbf{\sum }}_{n}^{1} \) and \( {\mathbf{\Pi }}_{n}^{1} \) are closed under countable unions, countable intersections and Borel preimages.\n\n(ii) \( {\mathbf{\Delta }}_{n}^{1} \) is a \( \sigma \) -algebra.\n\n(iii) The pointclass \( {\mathbf{\sum }}_{n}^{1} \) is closed under projections \( {\exists }^{Y} \), and \( {\mathbf{\Pi }}_{n}^{1} \) is closed under coprojections \( {\forall }^{Y}, Y \) Polish.
Proof. Clearly, (ii) follows from (i). So, we prove (i) and (iii) only. We proceed by induction on \( n \) . Let \( n > 1 \) and \( {\mathbf{\Pi }}_{n - 1}^{1} \) and \( {\mathbf{\sum }}_{n - 1}^{1} \) have all the closure properties stated in (i) and (iii). The arguments contained in the proof of 4.1.2 show that \( {\mathbf{\sum }}_{n}^{1} \) also has the stated closure properties. Since \( {\mathbf{\Pi }}_{n}^{1} = \neg {\mathbf{\sum }}_{n}^{1} \), the remaining part of the result follows.
No
Proposition 4.1.9 For every \( n \geq 1 \) ,\n\n\[ \n{\mathbf{\sum }}_{n}^{1}\bigcup {\mathbf{\Pi }}_{n}^{1} \subseteq {\mathbf{\Delta }}_{n + 1}^{1} \n\]
Proof. We prove the result by induction on \( n \) . Let \( X \) be a Polish space and \( A \subseteq X \) analytic. As \( {\mathbf{\Delta }}_{1}^{1} \subseteq {\mathbf{\Pi }}_{1}^{1} \), it follows that \( {\mathbf{\sum }}_{1}^{1} \subseteq {\mathbf{\sum }}_{2}^{1} \) . Since \( {\mathbf{\sum }}_{1}^{1} \) is closed under continuous preimages, the set \( C = A \times X \) is analytic. Since\n\n\[ \nA = {\forall }^{X}C, \n\]\n\n\( A \) is in \( {\mathbf{\Pi }}_{2}^{1} \) . Hence \( A \in {\mathbf{\Delta }}_{2}^{1} \) . The rest of the result now follows fairly easily by induction.
Yes
Lemma 4.1.10 Let \( n \geq 1,\mathbf{\Gamma } \) either \( {\mathbf{\sum }}_{n}^{1} \) or \( {\mathbf{\Pi }}_{n}^{1} \), and \( X \) a Polish space. There is a \( U \subseteq {\mathbb{N}}^{\mathbb{N}} \times X \) in \( \mathbf{\Gamma } \) such that \( A \subseteq X \) is in \( \mathbf{\Gamma } \) if and only if \( A = {U}_{\alpha } \) for some \( \alpha \) ; i.e., \( U \) is universal for \( \mathbf{\Gamma }\left( X\right) \) .
Proof. The result is proved by induction. Suppose \( U \subseteq {\mathbb{N}}^{\mathbb{N}} \times X \) is universal for \( {\mathbf{\sum }}_{1}^{1}\left( X\right) \) . Then \( {U}^{c} \) is universal for \( {\mathbf{\Pi }}_{1}^{1}\left( X\right) \) . Let \( C \subseteq {\mathbb{N}}^{\mathbb{N}} \times (X \times \) \( {\mathbb{N}}^{\mathbb{N}} \) ) be universal for \( {\mathbf{\Pi }}_{n}^{1}\left( {X \times {\mathbb{N}}^{\mathbb{N}}}\right) \) . As in 4.1.4, we see that \( {\exists }^{{\mathbb{N}}^{\mathbb{N}}}C \) is universal for \( {\mathbf{\sum }}_{n + 1}^{1}\left( X\right) \), and its complement is universal for \( {\mathbf{\Pi }}_{n + 1}^{1}\left( X\right) \) .
Yes
Theorem 4.1.11 Let \( X \) be an uncountable Polish space and \( n \geq 1 \) .\n\n(i) There is a set \( U \in {\mathbf{\sum }}_{n}^{1}\left( {X \times X}\right) \) such that for every \( A \in {\mathbf{\sum }}_{n}^{1}\left( X\right) \), there is an \( x \) with \( A = {U}_{x} \) .\n\n(ii) There is a subset of \( X \) that is in \( {\mathbf{\sum }}_{n}^{1}\left( X\right) \) but not in \( {\mathbf{\Pi }}_{n}^{1}\left( X\right) \) .
Proof. The result is proved in exactly the same way as 4.1.5.
No
Theorem 4.1.13 Let \( X \) be a Polish space, \( d \) a compatible complete metric on \( X \), and \( A \subseteq X \). The following statements are equivalent.\n\n(i) \( A \) is analytic.\n\n(ii) There is a regular system \( \left\{ {{F}_{s} : s \in {\mathbb{N}}^{ < \mathbb{N}}}\right\} \) of closed subsets of \( X \) such that for every \( \alpha \in {\mathbb{N}}^{\mathbb{N}} \) diameter \( \left( {F}_{\alpha \mid n}\right) \rightarrow 0 \) and \( A = \mathcal{A}\left( \left\{ {F}_{s}\right\} \right) \).\n\n(iii) There is a system \( \left\{ {{F}_{s} : s \in {\mathbb{N}}^{ < \mathbb{N}}}\right\} \) of closed subsets of \( X \) such that \( A = \mathcal{A}\left( \left\{ {F}_{s}\right\} \right) \)
Proof. (ii) implies (iii) is obvious.\n\n(iii) \( \Rightarrow \) (i): Let \( \left\{ {F}_{s}\right\} \) be a system of closed sets in \( X \) such that\n\n\[ A = \mathcal{A}\left( \left\{ {F}_{s}\right\} \right) \]\n\ni.e.,\n\n\[ x \in A \Leftrightarrow \exists \alpha \forall n\left( {x \in {F}_{\alpha \mid n}}\right) .\n\nLet\n\n\[ C = \left\{ {\left( {x,\alpha }\right) \in X \times {\mathbb{N}}^{\mathbb{N}} : \forall n\left( {x \in {F}_{\alpha \mid n}}\right) }\right\} .\n\nAs\n\n\[ C = \mathop{\bigcap }\limits_{n}\mathop{\bigcup }\limits_{{\{ s : \left| s\right| = n\} }}\left( {{F}_{s} \times \sum \left( s\right) }\right) \]\n\n\( C \) is closed. Since \( A \) is the projection of \( C \), it is analytic.\n\n(i) \( \Rightarrow \) (ii): Let \( A \subseteq X \) be analytic. By 4.1.1, there is a continuous map \( f : {\mathbb{N}}^{\mathbb{N}} \rightarrow X \) whose range is \( A \). Take\n\n\[ {F}_{s} = \operatorname{cl}\left( {f\left( {\sum \left( s\right) }\right) }\right) .\n\nClearly, the system of closed sets \( \left\{ {{F}_{s} : s \in {\mathbb{N}}^{ < \mathbb{N}}}\right\} \) is regular. Since \( f \) is continuous, diameter \( \left( {F}_{\alpha \mid n}\right) \) converges to 0 as \( n \rightarrow \infty \).\n\nLet \( x = f\left( \alpha \right) \in A \). Then for all \( n, x \in {F}_{\alpha \mid n} \). Thus \( A \subseteq \mathcal{A}\left( \left\{ {F}_{s}\right\} \right) \).\n\nTo show the reverse inclusion, take any \( x \in \mathcal{A}\left( \left\{ {F}_{s}\right\} \right) \). Let\n\n\[ x \in {F}_{\alpha \mid n} = \operatorname{cl}\left( {f\left( {\sum \left( {\alpha \mid n}\right) }\right) }\right) \]\n\nfor all \( n \). Choose \( {\alpha }_{n} \in \sum \left( {\alpha \mid n}\right) \) such that \( d\left( {x, f\left( {\alpha }_{n}\right) }\right) < {2}^{-n} \). So, \( f\left( {\alpha }_{n}\right) \rightarrow x \). Since \( {\alpha }_{n} \rightarrow \alpha \) and \( f \) is continuous, \( f\left( {\alpha }_{n}\right) \rightarrow f\left( \alpha \right) \). Hence, \( x \in A \), and the result follows.
Yes
Theorem 4.1.14 The pointclass \( {\mathbf{\sum }}_{1}^{1} \) is closed under the Souslin operation.
Proof. By 1.13.1, the Souslin operation is idempotent; i.e., for any family \( \mathcal{F} \) of sets \( \mathcal{A}\left( {\mathcal{A}\left( \mathcal{F}\right) }\right) = \mathcal{A}\left( \mathcal{F}\right) \) . Since \( {\mathbf{\sum }}_{1}^{1} = \mathcal{A}\left( \mathcal{F}\right) \), where \( \mathcal{F} \) is the family of closed sets, the result follows.
Yes
Proposition 4.1.20 Let \( A \subseteq {\mathbb{N}}^{\mathbb{N}} \) . The following statements are equivalent.\n\n(i) \( A \) is coanalytic.\n\n(ii) There is a tree \( T \) on \( \mathbb{N} \times \mathbb{N} \) such that\n\n\[ \alpha \in A \Leftrightarrow T\left\lbrack \alpha \right\rbrack \text{ is well-founded } \]\n\n\[ \Leftrightarrow \;T\left\lbrack \alpha \right\rbrack \text{ is well-ordered with respect to }{ \leq }_{KB}\text{. } \]
Proof. Let \( A \subseteq {\mathbb{N}}^{\mathbb{N}} \) be a coanalytic set. Then \( {A}^{c} \) is analytic. Let \( C \) be a closed set in \( {\mathbb{N}}^{\mathbb{N}} \times {\mathbb{N}}^{\mathbb{N}} \) such that \( {\pi }_{1}\left( C\right) = {A}^{c} \), where \( {\pi }_{1} : {\mathbb{N}}^{\mathbb{N}} \times {\mathbb{N}}^{\mathbb{N}} \rightarrow {\mathbb{N}}^{\mathbb{N}} \) is the projection onto the first coordinate space. The existence of such a set follows from 4.1.1. By 2.2.13, there is a tree \( T \) on \( \mathbb{N} \times \mathbb{N} \) such that \( \left\lbrack T\right\rbrack = C \) . Now note that\n\n\[ \alpha \in {A}^{c} \Leftrightarrow \exists \beta \left( {\left( {\alpha ,\beta }\right) \in \left\lbrack T\right\rbrack }\right) \]\n\n\[ \Leftrightarrow \;\exists \beta \left( {\beta \in \left\lbrack {T\left( \alpha \right) }\right\rbrack }\right) \]\n\n\[ \Leftrightarrow T\left\lbrack \alpha \right\rbrack \text{is not well-founded} \]\n\nThus (ii) follows from (i).\n\n(ii) \( \Rightarrow \) (i): Let \( A \subseteq {\mathbb{N}}^{\mathbb{N}} \) satisfy (ii). Then \( {A}^{c} \) is the projection of \( \left\lbrack T\right\rbrack \), and so \( A \) is coanalytic.
Yes
Example 4.1.21 Let \( g : \mathbb{R} \times \mathbb{R} \rightarrow \mathbb{R} \) be a Borel function. Define \[ f\left( x\right) = \mathop{\sup }\limits_{y}g\left( {x, y}\right) ,\;x \in X. \] Assume that \( f\left( x\right) < \infty \) for all \( x \) . The function \( f \) need not be Borel. To see this, take an analytic set \( A \subseteq \mathbb{R} \) that is not Borel. Suppose \( B \subseteq \mathbb{R} \times \mathbb{R} \) is a Borel set whose projection is \( A \) . Take \( g = {\chi }_{B} \).
Proof. Let \( v : \mathbb{R} \rightarrow \mathbb{R} \) be a Borel function such that \( v\left( x\right) \leq f\left( x\right) \) for all \( x \) . For \( n \in \mathbb{Z} \), let \[ {B}_{n} = \{ x \in \mathbb{R} : n \leq v\left( x\right) < n + 1\} . \] Fix an enumeration \( \left\{ {{r}_{m} : m \in \mathbb{N}}\right\} \) of the set of all rational numbers. Let \[ A = \{ \left( {x, y}\right) : f\left( x\right) > y\} . \] Since \[ A = \mathop{\bigcup }\limits_{m}\left\{ {\left( {x, y}\right) \in \mathbb{R} \times \mathbb{R} : f\left( x\right) > {r}_{m} > y}\right\} \] and \( f \) is an A-function, \( A \) is analytic. By 4.1.1, there is a Borel set \( B \subseteq \) \( \left( {\mathbb{R} \times \mathbb{R}}\right) \times \mathbb{R} \) whose projection is \( A \) . Define \( h : {\mathbb{R}}^{3} \rightarrow \mathbb{R} \) by \[ h\left( {x, y, z}\right) = \left\{ \begin{array}{ll} y & \text{ if }\left( {x, y, z}\right) \in B, \\ n & \text{ if }x \in {B}_{n}\& \left( {x, y, z}\right) \in {\mathbb{R}}^{3} \smallsetminus B. \end{array}\right. \] The function \( h \) is Borel, and \[ f\left( x\right) = \mathop{\sup }\limits_{\left( y, z\right) }h\left( {x, y, z}\right) . \] Let \( u : \mathbb{R} \rightarrow {\mathbb{R}}^{2} \) be a Borel isomorphism. Such a map exists by the Borel isomorphism theorem. Define \( g \) by \[ g\left( {x, y}\right) = h\left( {x, u\left( y\right) }\right) . \]
Yes
Proposition 4.1.22 (H. Sarbadhikari [99]) For every A-function \( f \) : \( \mathbb{R} \rightarrow \mathbb{R} \) dominating a Borel function there is a Borel \( g : \mathbb{R} \times \mathbb{R} \rightarrow \mathbb{R} \) such that \( f\left( x\right) = \mathop{\sup }\limits_{y}g\left( {x, y}\right) \) .
Proof. Let \( v : \mathbb{R} \rightarrow \mathbb{R} \) be a Borel function such that \( v\left( x\right) \leq f\left( x\right) \) for all \( x \) . For \( n \in \mathbb{Z} \), let\n\n\[ \n{B}_{n} = \{ x \in \mathbb{R} : n \leq v\left( x\right) < n + 1\} .\n\] \n\nFix an enumeration \( \left\{ {{r}_{m} : m \in \mathbb{N}}\right\} \) of the set of all rational numbers. Let\n\n\[ \nA = \{ \left( {x, y}\right) : f\left( x\right) > y\} .\n\] \n\nSince\n\n\[ \nA = \mathop{\bigcup }\limits_{m}\left\{ {\left( {x, y}\right) \in \mathbb{R} \times \mathbb{R} : f\left( x\right) > {r}_{m} > y}\right\} \n\] \n\nand \( f \) is an A-function, \( A \) is analytic. By 4.1.1, there is a Borel set \( B \subseteq \) \( \left( {\mathbb{R} \times \mathbb{R}}\right) \times \mathbb{R} \) whose projection is \( A \) . Define \( h : {\mathbb{R}}^{3} \rightarrow \mathbb{R} \) by\n\n\[ \nh\left( {x, y, z}\right) = \left\{ \begin{array}{ll} y & \text{ if }\left( {x, y, z}\right) \in B, \\ n & \text{ if }x \in {B}_{n}\& \left( {x, y, z}\right) \in {\mathbb{R}}^{3} \smallsetminus B. \end{array}\right. \n\] \n\nThe function \( h \) is Borel, and\n\n\[ \nf\left( x\right) = \mathop{\sup }\limits_{\left( y, z\right) }h\left( {x, y, z}\right) .\n\] \n\nLet \( u : \mathbb{R} \rightarrow {\mathbb{R}}^{2} \) be a Borel isomorphism. Such a map exists by the Borel isomorphism theorem. Define \( g \) by\n\n\[ \ng\left( {x, y}\right) = h\left( {x, u\left( y\right) }\right) .\n\]
Yes
We show that \( {WF} \) is \( {\mathbf{\Pi }}_{1}^{1} \) -complete.
Observe that\n\n\[ T \in {WF} \Leftrightarrow T \in \operatorname{Tr}\& \forall \beta \exists n\left( {T\left( {\beta \mid n}\right) = 0}\right) .\n\]\n\nTherefore, \( {WF} = {\forall }^{{\mathbb{N}}^{\mathbb{N}}}E \), where\n\n\[ E = \left\{ {\left( {T,\beta }\right) \in {2}^{{\mathbb{N}}^{ < \mathbb{N}}} \times {\mathbb{N}}^{\mathbb{N}} : T \in \operatorname{Tr}\& \exists n\left( {T\left( {\beta \mid n}\right) = 0}\right) }\right\} .\n\]\n\nIt is quite easy to see that the set \( E \) is Borel. Hence, \( {WF} \) is coanalytic.\n\nNow take any coanalytic set \( C \) in \( {\mathbb{N}}^{\mathbb{N}} \) . By 4.1.20, there is a tree \( T \) on \( \mathbb{N} \times \mathbb{N} \) such that\n\n\[ \alpha \in C \Leftrightarrow T\left\lbrack \alpha \right\rbrack \text{is well-founded. }\n\]\n\nDefine \( f : {\mathbb{N}}^{\mathbb{N}} \rightarrow \operatorname{Tr} \) by\n\n\[ f\left( \alpha \right) = T\left\lbrack \alpha \right\rbrack \n\]\n\nthe section of \( T \) at \( \alpha \) . The map \( f \) is continuous: Take any \( s \in {\mathbb{N}}^{ < \mathbb{N}} \) and note that\n\n\[ f\left( \alpha \right) \left( s\right) = 1 \Leftrightarrow T\left( {\alpha \left| \right| s \mid, s}\right) = 1.\n\]\n\nThus \( {\pi }_{s} \circ f \) is continuous for all \( s \), and so \( f \) is continuous.\n\nAs \( C = {f}^{-1}\left( {WF}\right) \), by the Borel isomorphism theorem it follows that \( {WF} \) is \( {\mathbf{\Pi }}_{1}^{1} \) -complete.
Yes
We now show that \( {WO} \) is \( {\mathbf{\Pi }}_{1}^{1} \) -complete. It is sufficient to show that there is a continuous map \( R : {Tr} \rightarrow {2}^{\mathbb{N} \times \mathbb{N}} \) such that \( {WF} = {R}^{-1}\left( {WO}\right) \) .
Fix a bijection \( u : \mathbb{N} \rightarrow {\mathbb{N}}^{ < \mathbb{N}} \) . To each \( T \in {Tr} \), associate a binary relation \( R\left( T\right) \) on \( \mathbb{N} \) as follows:\n\n\[ {kR}\left( T\right) l \Leftrightarrow \left( {u\left( k\right), u\left( l\right) \notin T\& k \leq l}\right) \]\n\n\[ \vee \left( {u\left( k\right) \in T\& u\left( l\right) \notin T}\right) \]\n\n\[ \vee \left( {u\left( k\right), u\left( l\right) \in T\& \;u\left( k\right) { \leq }_{KB}u\left( l\right) }\right) \]\n\nIt is easy to check that \( T \rightarrow R\left( T\right) \) is a continuous map from \( {Tr} \) to \( {2}^{\mathbb{N} \times \mathbb{N}} \) . Since a tree \( T \) on \( \mathbb{N} \) is well-founded if and only if \( { \leq }_{KB} \) is a well-order on \( T \) (1.10.10.), \( {WF} = {R}^{-1}\left( {WO}\right) \) .
Yes
Proposition 4.2.5 Let \( X \) be an uncountable Polish space. Then\n\n\[ U\left( X\right) = \{ K \in K\left( X\right) : K\text{ is uncountable }\} \]\n\nis \( {\mathbf{\sum }}_{1}^{1} \) -complete.
Proof. We first show that \( U\left( X\right) \in {\mathbf{\sum }}_{1}^{1} \) . Let \( P\left( X\right) \) denote the set of all nonempty perfect subsets of \( X \) . Then \( P\left( X\right) \) is Borel in \( K\left( X\right) \) . To see this, take a countable base \( \left( {V}_{n}\right) \) for \( X \) . We have\n\n\( K \) is perfect \( \Leftrightarrow \forall n\left( {K\bigcap {V}_{n} \neq \varnothing }\right.\)\n\n\[ \Rightarrow \exists k\exists l\left( {{V}_{k},{V}_{l} \subseteq {V}_{n}}\right. \]\n\n\[ \text{&}{V}_{k} \cap {V}_{l} = \varnothing \text{&}K \cap {V}_{k}, K \cap {V}_{l} \neq \varnothing )\text{).} \]\n\nSo,\n\n\[ P\left( X\right) = \mathop{\bigcap }\limits_{n}\left\lbrack {{A}_{n}^{c}\bigcup \mathop{\bigcup }\limits_{{\left( {k, l}\right) \in {S}_{n}}}\left( {{A}_{k}\bigcap {A}_{l}}\right) }\right\rbrack \]\n\nwhere\n\n\[ {A}_{n} = \left\{ {K \in K\left( X\right) : K\bigcap {V}_{n} \neq \varnothing }\right\} \]\n\nand\n\n\[ {S}_{n} = \left\{ {\left( {k, l}\right) : {V}_{k} \subseteq {V}_{n}\& {V}_{l} \subseteq {V}_{n}\& {V}_{k}\bigcap {V}_{l} = \varnothing }\right\} . \]\n\nHence, \( P\left( X\right) \) is Borel. Let \( K \in K\left( X\right) \) . By 2.6.3,\n\n\( K \) is uncountable \( \Leftrightarrow \left( {\exists P \in K\left( X\right) }\right) \left( {P \in P\left( X\right) \& P \subseteq K}\right) \) .\n\nBy 2.4.11, the set\n\n\[ \{ \left( {K, L}\right) \in K\left( X\right) \times K\left( X\right) : K \subseteq L\} \]\n\nis closed. Hence, \( U\left( X\right) \in {\mathbf{\sum }}_{1}^{1} \) .\n\nIt remains to show that \( U\left( X\right) \) is \( {\mathbf{\sum }}_{1}^{1} \) -complete. Since every uncountable Polish space contains a \( {G}_{\delta } \) set homeomorphic to \( {\mathbb{N}}^{\mathbb{N}} \), it is sufficient to prove the result for \( X = {\mathbb{N}}^{\mathbb{N}} \) . Let \( N \) be as in 4.2.3. Define \( f : {\mathbb{N}}^{\mathbb{N}} \rightarrow K\left( {\mathbb{N}}^{\mathbb{N}}\right) \) by\n\n\[ f\left( \alpha \right) = \left\{ {\beta \in {\mathbb{N}}^{\mathbb{N}} : \beta \leq \alpha \text{ pointwise }}\right\} . \]\n\nThen \( f \) is continuous. Further,\n\n\[ \alpha \in N \Leftrightarrow f\left( \alpha \right) \text{is uncountable.} \]\n\nNow consider the map \( g : K\left( {K\left( {\mathbb{N}}^{\mathbb{N}}\right) }\right) \rightarrow K\left( {\mathbb{N}}^{\mathbb{N}}\right) \) defined by\n\n\[ g\left( \mathcal{K}\right) = \bigcup \mathcal{K},\;\mathcal{K} \in K\left( {K\left( {\mathbb{N}}^{\mathbb{N}}\right) }\right) . \]\n\nThe map \( g \) is continuous (2.4.11). Define\n\n\[ h\left( K\right) = g\left( {f\left( K\right) }\right) ,\;K \in K\left( {\mathbb{N}}^{\mathbb{N}}\right) . \]\n\nThe map \( h \) is continuous, and\n\n\[ I{F}^{ * } = {h}^{-1}\left( \left\{ {K \in K\left( {\mathbb{N}}^{\mathbb{N}}\right) : K\text{ is uncountable }}\right\} \right) . \]\n\nThe result follows from 4.2.3.
Yes
Theorem 4.3.4 (B. V. Rao[95]) Let \( X \) be an uncountable Polish space and \( U \subseteq X \times X \) universal analytic. Then\n\n\[ U \notin \mathcal{P}\left( X\right) \bigotimes \mathcal{B} \]\n\nwhere \( \mathcal{B} \) is as in 4.3.3.
Proof. Suppose \( U \in \mathcal{P}\left( X\right) \otimes \mathcal{B} \) . We shall get a contradiction. From 3.1.7, there are \( {C}_{0},{C}_{1},{C}_{2},\ldots \subseteq X \) and \( {D}_{0},{D}_{1},{D}_{2},\ldots \) in \( \mathcal{B} \) such that\n\n\( U \in \sigma \left( \left\{ {{C}_{i} \times {D}_{i} : i \in \mathbb{N}}\right\} \right) \) . Let \( Y \) be an uncountable Borel subset of \( X \) such that each \( {D}_{i}\bigcap Y \) is Borel. In particular, every section \( {\left( U\bigcap \left( X \times Y\right) \right) }_{x} \) , \( x \in X \), is Borel. Let \( E \) be an analytic non-Borel set contained in \( Y \) . Since \( U \) is universal,\n\n\[ E = {U}_{{x}_{0}} = {\left( U\bigcap \left( X \times Y\right) \right) }_{{x}_{0}} \]\n\nfor some \( {x}_{0} \in X \) . We have arrived at a contradiction.
Yes
Theorem 4.3.5 Every uncountable analytic set contains a homeomorph of the Cantor set and hence is of cardinality \( \mathfrak{c} \) .
Proof. Let \( X \) be a Polish space and \( f : {\mathbb{N}}^{\mathbb{N}} \rightarrow X \) a continuous map whose range is uncountable. We first show that there is a Cantor scheme \( \left\{ {{F}_{s} : s \in {2}^{ < \mathbb{N}}}\right\} \) of closed subsets of \( {\mathbb{N}}^{\mathbb{N}} \) such that whenever \( \left| s\right| = \left| t\right| \) and \( s \neq t, f\left( {F}_{s}\right) \cap f\left( {F}_{t}\right) = \varnothing \) . Since the range of \( f \) is uncountable, we get an uncountable \( Z \subseteq {\mathbb{N}}^{\mathbb{N}} \) such that \( f \mid Z \) is one-to-one. By the Cantor - Bendixson theorem (2.6.2), we can further assume that \( Z \) is dense-in-itself. Take a compatible complete metric \( d < 1 \) on \( {\mathbb{N}}^{\mathbb{N}} \) . We define a system \( \left\{ {{U}_{s} : s \in {2}^{ < \mathbb{N}}}\right\} \) of nonempty open subsets of \( {\mathbb{N}}^{\mathbb{N}} \) satisfying the following conditions: (i) \( \operatorname{diameter}\left( {U}_{s}\right) < {2}^{-\left| s\right| } \) ; (ii) \( {U}_{s} \cap Z \neq \varnothing \) ; (iii) \( \operatorname{cl}\left( {U}_{s \cap \epsilon }\right) \subseteq {U}_{s},\epsilon = 0,1 \) ; and (iv) whenever \( \left| s\right| = \left| t\right| \) and \( s \neq t, f\left( {\operatorname{cl}\left( {U}_{s}\right) }\right) \cap f\left( {\operatorname{cl}\left( {U}_{t}\right) }\right) = \varnothing \) . In particular, \( \operatorname{cl}\left( {U}_{s}\right) \cap \operatorname{cl}\left( {U}_{t}\right) = \varnothing . \) We define such a system by induction on \( \left| s\right| \) . Take \( {U}_{e} = X \) . Suppose \( {U}_{s} \) has been defined for some \( s \) . Since \( Z \) is dense-in-itself and \( {U}_{s} \) open, \( {U}_{s} \cap Z \) has at least two distinct points, say \( {x}_{0},{x}_{1} \) . Then \( f\left( {x}_{0}\right) \neq f\left( {x}_{1}\right) \) . Let \( {W}_{0} \) and \( {W}_{1} \) be disjoint open sets containing \( f\left( {x}_{0}\right) \) and \( f\left( {x}_{1}\right) \) respectively. Since \( f \) is continuous, there are open sets \( {U}_{s}{ \hat{} }_{0} \) and \( {U}_{s}{ \hat{} }_{1} \) satisfying the following conditions: (b) \( \operatorname{diameter}\left( {U}_{s \uparrow \epsilon }\right) < \frac{1}{{2}^{\left| s\right| + 1}} \) ; and (c) \( f\left( {\operatorname{cl}\left( {U}_{s \cap \epsilon }\right) }\right) \subseteq {W}_{\epsilon },\epsilon = 0 \) or 1 . (d) In particular, \( f\left( {\operatorname{cl}\left( {U}_{{s}^{ \frown }0}\right) }\right) \cap f\left( {\operatorname{cl}\left( {U}_{{s}^{ \frown }1}\right) }\right) = \varnothing \) . Put \( {F}_{s} = \operatorname{cl}\left( {U}_{s}\right) \) . Let \( C = \mathcal{A}\left( \left\{ {F}_{s}\right\} \right) \) . Then \( C \) is homeomorphic to the Cantor set, and \( f \mid C \), being one-to-one and continuous, is an embedding.
Yes
Proposition 4.3.7 Let \( X \) be a Polish space and \( A \subseteq X \) . The following statements are equivalent.\n\n(i) \( A \) is analytic.\n\n(ii) There is a closed set \( C \subseteq X \times {\mathbb{N}}^{\mathbb{N}} \) such that\n\n\[ A = \left\{ {x \in X : {C}_{x}}\right. \text{is uncountable}\} \text{.} \]\n\n(iii) There is a Polish space \( Y \) and an analytic set \( B \subseteq X \times Y \) such that\n\n\[ A = \left\{ {x \in X : {B}_{x}}\right. \text{is uncountable}\} \text{.} \]
Proof. (i) \( \Rightarrow \) (ii): Let \( f : {\mathbb{N}}^{\mathbb{N}} \rightarrow X \) be a continuous map with range \( A \) and \( {\pi }_{1} : {\mathbb{N}}^{\mathbb{N}} \times {\mathbb{N}}^{\mathbb{N}} \rightarrow {\mathbb{N}}^{\mathbb{N}} \) the projection map. Note that \( {\pi }_{1} \) is continuous and \( {\pi }_{1}^{-1}\left( \alpha \right) \) uncountable for all \( \alpha \) . Since \( {\mathbb{N}}^{\mathbb{N}} \times {\mathbb{N}}^{\mathbb{N}} \) is homeomorphic to \( {\mathbb{N}}^{\mathbb{N}} \) , this shows that there is a continuous map \( h : {\mathbb{N}}^{\mathbb{N}} \rightarrow {\mathbb{N}}^{\mathbb{N}} \) such that \( {h}^{-1}\left( \alpha \right) \) is uncountable for all \( \alpha \) . Take \( C = \operatorname{graph}\left( {f \circ h}\right) \) .\n\n(iii) is a special case of (ii).\n\n(iii) \( \Rightarrow \) (i): By (4.3.6), we have the following: Let \( P, Q \) be Polish spaces and \( f : P \rightarrow Q \) a continuous map. The range of \( f \) is uncountable if and only if there is a countable dense-in-itself subset \( Z \) of \( P \) such that \( f \mid Z \) is one-to-one.\n\nNote also that the set\n\n\[ D = \left\{ {\left( {x}_{n}\right) \in {\left( {\mathbb{N}}^{\mathbb{N}}\right) }^{\mathbb{N}} : \left\{ {{x}_{n} : n \in \mathbb{N}}\right\} \text{ is dense-in-itself }}\right\} \]\n\nis a \( {G}_{\delta } \) set in \( {\left( {\mathbb{N}}^{\mathbb{N}}\right) }^{\mathbb{N}} \) .\n\nNow let \( X, Y \) and \( B \) be as in (iii). Let \( f : {\mathbb{N}}^{\mathbb{N}} \rightarrow X \times Y \) be a continuous map with range \( B \) . By (a),\n\n\[ {B}_{x}\text{ is uncountable } \Leftrightarrow \left( {\exists \left( {z}_{n}\right) \in D}\right) \left( {\forall i\forall j\left( {i \neq j \Rightarrow f\left( {z}_{i}\right) \neq f\left( {z}_{j}\right) }\right. }\right) ,\n\n\left. {\& \forall k\left( {{\pi }_{X}\left( {f\left( {z}_{k}\right) }\right) = x}\right) }\right) ,\n\nwhere \( {\pi }_{X} : X \times Y \rightarrow X \) is the projection map. The result follows from (b).
Yes
Theorem 4.3.8 (S. Simpson [79]) Let \( X \) be an analytic subset of a Polish space, \( Y \) a metrizable space, and \( f : X \rightarrow Y \) a Borel map. Then \( f\left( X\right) \) is separable.
Proof. Without any loss of generality, we assume that \( X \) is Polish and \( Y = f\left( X\right) \) . Suppose \( Y \) is not separable. Then there is an uncountable closed discrete subspace \( Z \) of \( Y \) . As \( \left| X\right| = \mathfrak{c},\left| Y\right| \leq \mathfrak{c} \), and hence \( \left| Z\right| \leq \mathfrak{c} \) . Let \( {X}^{\prime } = {f}^{-1}\left( Z\right) \) . Note that \( {X}^{\prime } \) is Borel. Now take any \( A \subseteq \mathbb{R} \) of the same cardinality as \( Z \) that does not contain any uncountable closed set. We have proved the existence of such a set in 3.2.8. Let \( g \) be any one-to-one map from \( Z \) onto \( A \) . Since \( Z \) is discrete, \( g \) is continuous. Clearly, \( g \circ f \) is Borel. As \( A = g\left( {f\left( {X}^{\prime }\right) }\right), A \) is an uncountable analytic set not containing a perfect set. This contradicts 4.3.5.
Yes
Corollary 4.3.9 Every Borel homomorphism \( \varphi : G \rightarrow H \) from a completely metrizable group \( G \) to a metrizable group \( H \) is continuous.
Proof. Let \( \left( {g}_{n}\right) \) be a sequence in \( G \) converging to \( g \) . Replacing \( G \) by the closed subgroup generated by \( \left\{ {{g}_{n} : n \in \mathbb{N}}\right\} \), we assume that \( G \) is Polish. By 4.3.8, \( \varphi \left( G\right) \) is separable. The result follows from 3.5.9.
No
Proposition 4.3.10 (i) Every countable set of reals has strong measure zero.
Proof. (i) and (ii) are immediate consequences of the definition. We prove (iii) now. Let \( \left( {A}_{n}\right) \) be a sequence of strong measure zero sets. Take any sequence \( \left( {a}_{n}\right) \) of positive real numbers. Choose pairwise disjoint infinite subsets \( {I}_{0},{I}_{1},{I}_{2},\ldots \) of \( \mathbb{N} \) whose union is \( \mathbb{N} \) . For each \( n \) choose open intervals \( {I}_{m}^{n}, m \in {I}_{n} \), such that \( \left| {I}_{m}^{n}\right| \leq {a}_{m} \) and \( {A}_{n} \subseteq \mathop{\bigcup }\limits_{{m \in {I}_{n}}}{I}_{m}^{n} \) . Note that\n\n\[ \mathop{\bigcup }\limits_{n}{A}_{n} \subseteq \mathop{\bigcup }\limits_{{n \in \mathbb{N}}}\mathop{\bigcup }\limits_{{m \in {I}_{n}}}{I}_{m}^{n} \]\n\nThe proof of (iii) is clearly seen now.
No
Proposition 4.3.11 Let \( A \subseteq \left\lbrack {0,1}\right\rbrack \) be a strong measure zero set and \( f \) : \( \left\lbrack {0,1}\right\rbrack \rightarrow \mathbb{R} \) a continuous map. Then the set \( f\left( A\right) \) has strong measure zero.
Proof. Let \( \left( {a}_{n}\right) \) be any sequence of positive real numbers. We have to show that there exist open intervals \( {J}_{n}, n \in \mathbb{N} \), such that \( \left| {J}_{n}\right| \leq {a}_{n} \) and \( f\left( A\right) \subseteq \mathop{\bigcup }\limits_{n}{J}_{n} \) . Since \( f \) is uniformly continuous, for each \( n \) there is a positive real number \( {b}_{n} \) such that whenever \( X \subseteq \left\lbrack {0,1}\right\rbrack \) is of diameter at most \( {b}_{n} \) , the diameter of \( f\left( X\right) \) is at most \( {a}_{n} \) . Since \( A \) has strong measure zero, there are open intervals \( {I}_{n}, n \in \mathbb{N} \), such that \( \left| {I}_{n}\right| \leq {b}_{n} \) and \( A \subseteq \mathop{\bigcup }\limits_{n}{I}_{n} \) . Take \( {J}_{n} = f\left( {I}_{n}\right) \) .
Yes
Theorem 4.3.17 Every coanalytic set is a union of \( {\aleph }_{1} \) Borel sets.
Proof. Let \( X \) be Polish and \( C \subseteq X \) coanalytic. By the Borel isomorphism theorem (3.3.13), without any loss of generality we may assume that \( X = {\mathbb{N}}^{\mathbb{N}} \) . By 4.1.20, there is a tree \( T \) on \( \mathbb{N} \times \mathbb{N} \) such that\n\n\[ \alpha \in C \Leftrightarrow T\left\lbrack \alpha \right\rbrack \text{is well-founded.} \]\n\nSo,\n\n\[ \alpha \in C \Leftrightarrow {\rho }_{T\left\lbrack \alpha \right\rbrack }\left( e\right) < {\omega }_{1} \]\n\nTherefore,\n\n\[ C = \mathop{\bigcup }\limits_{{\xi < {\omega }_{1}}}{C}_{e}^{\xi } \]\n\nwhere the \( {C}_{e}^{\xi } \) are as in 4.3.16.\n\nThe sets \( {C}_{e}^{\xi },\xi < {\omega }_{1} \), defined in the above proof are called the constituents of \( C \) . Since \( \mathbf{{CH}} \) holds for Borel sets, we now have the following result.\n\nTheorem 4.3.18 A coanalytic set is either countable or of cardinality \( {\aleph }_{1} \) or \( \mathfrak{c} \) .
Yes
Lemma 4.4.2 Suppose \( E = \mathop{\bigcup }\limits_{n}{E}_{n} \) cannot be separated from \( F = \mathop{\bigcup }\limits_{m}{F}_{m} \) by a Borel set. Then there exist \( m, n \) such that \( {E}_{n} \) cannot be separated from \( {F}_{m} \) by a Borel set.
Proof. Suppose for every \( m, n \) there is a Borel set \( {C}_{mn} \) such that\n\n\[ \n{E}_{n} \subseteq {C}_{mn}\text{ and }{F}_{m}\bigcap {C}_{mn} = \varnothing .\n\]\n\nIt is fairly easy to check that the Borel set\n\n\[ \nC = \mathop{\bigcup }\limits_{n}\mathop{\bigcap }\limits_{m}{C}_{mn}\n\]\nseparates \( E \) from \( F \) .
Yes
Theorem 4.4.3 (Souslin) A subset A of a Polish space \( X \) is Borel if and only if it is both analytic and coanalytic; i.e., \( {\mathbf{\Delta }}_{1}^{1}\left( X\right) = {\mathcal{B}}_{X} \) .
Proof. The \
No
Proposition 4.4.4 Suppose \( {A}_{0},{A}_{1},\ldots \) are pairwise disjoint analytic subsets of a Polish space \( X \) . Then there exist pairwise disjoint Borel sets \( {B}_{0},{B}_{1},\ldots \) such that \( {B}_{n} \supseteq {A}_{n} \) for all \( n \) .
Proof. By 4.4.1, for each \( n \) there is a Borel set \( {C}_{n} \) such that\n\n\[ \n{A}_{n} \subseteq {C}_{n}\text{ and }{C}_{n} \cap \mathop{\bigcup }\limits_{{m \neq n}}{A}_{m} = \varnothing .\n\]\n\nTake\n\n\[ \n{B}_{n} = {C}_{n} \cap \mathop{\bigcap }\limits_{{m \neq n}}\left( {X \smallsetminus {C}_{m}}\right)\n\]
Yes
Theorem 4.4.5 Let \( E \subseteq X \times X \) be an analytic equivalence relation on a Polish space \( X \) . Suppose \( A \) and \( B \) are disjoint analytic subsets of \( X \) . Assume that \( B \) is invariant with respect to \( E \) (i.e., \( B \) is a union of \( E \) - equivalence classes). Then there is an \( E \) -invariant Borel set \( C \) separating \( A \) from \( B \) .
Proof. First we note the following. Let \( D \) be an analytic subset of \( X \) and \( {D}^{ * } \) the smallest invariant set containing \( D \) . Since\n\n\[ \n{D}^{ * } = {\pi }_{X}\left( {E\bigcap \left( {D \times X}\right) }\right) \n\]\n\nwhere \( {\pi }_{X} : X \times X \rightarrow X \) is the projection to the second coordinate space, \( {D}^{ * } \) is analytic.\n\nWe show that there is a sequence \( \left( {A}_{n}\right) \) of invariant analytic sets and a sequence \( \left( {B}_{n}\right) \) of Borel sets such that\n\n(i) \( A \subseteq {A}_{0} \),\n\n(ii) \( {A}_{n} \subseteq {B}_{n} \subseteq {A}_{n + 1} \), and\n\n(iii) \( B \cap {B}_{n} = \varnothing \).\n\nTake \( {A}_{0} = {A}^{ * } \) . Since \( B \) is invariant, \( {A}_{0} \cap B = \varnothing \) . By 4.4.1, let \( {B}_{0} \) be a Borel set containing \( {A}_{0} \) and disjoint from \( B \) . Suppose \( {A}_{i},{B}_{i},0 \leq i \leq n \) , satisfying (i),(ii), and iii) have been defined. Put \( {A}_{n + 1} = {B}_{n}^{ * } \) . Since \( B \) is invariant, \( {A}_{n + 1} \cap B = \varnothing \) . By 4.4.1, let \( {B}_{n + 1} \) be a Borel set containing \( {A}_{n + 1} \) and disjoint from \( B \) .\n\nHaving defined \( \left( {A}_{n}\right) ,\left( {B}_{n}\right) \), let \( C = \mathop{\bigcup }\limits_{n}{B}_{n} \) . Clearly, \( C \) is a Borel set containing \( A \) and disjoint from \( B \) . Since \( C = \mathop{\bigcup }\limits_{n}{A}_{n} \), it is also invariant.
Yes
Proposition 4.5.1 Let \( A \) be an analytic subset of a Polish space, \( Y \) a Polish space, and \( f : A \rightarrow Y \) a one-to-one Borel map. Then \( f : A \rightarrow \) \( f\left( A\right) \) is a Borel isomorphism.
Proof. Let \( B \subseteq A \) be Borel in \( A \) . We need to show that \( f\left( B\right) \) is Borel in \( f\left( A\right) \) . As both \( B \) and \( C = A \smallsetminus B \) are analytic and \( f \) Borel, \( f\left( B\right) \) and \( f\left( C\right) \) are analytic. Since \( f \) is one-to-one, these two sets are disjoint. So, by 4.4.1, there is a Borel set \( D \subseteq Y \) such that \( f\left( B\right) \subseteq D \) and \( f\left( C\right) \cap D = \varnothing \) . Since \( f\left( B\right) = D \cap f\left( A\right) \), the result follows.
Yes
Theorem 4.5.2 Let \( X, Y \) be Polish spaces, \( A \subseteq X \) analytic, and \( f : A \rightarrow Y \) any map. The following statements are equivalent\n\n(i) \( f \) is Borel measurable.\n\n(ii) \( \operatorname{graph}\left( f\right) \) is Borel in \( A \times Y \). \n\n(iii) \( \operatorname{graph}\left( f\right) \) is analytic.
Proof. We only need to show that (iii) implies (i). The other implications are quite easy to see. Let \( U \) be an open set in \( Y \). As\n\n\[ \n{f}^{-1}\left( U\right) = {\pi }_{X}\left( {\operatorname{graph}\left( f\right) \bigcap \left( {X \times U}\right) }\right) , \n\]\n\nwhere \( {\pi }_{X} : X \times Y \rightarrow X \) is the projection map, it is analytic. Similarly, \( {f}^{-1}\left( {U}^{c}\right) \) is analytic. By 4.4.1, there is a Borel set \( B \subseteq X \) such that\n\n\[ \n{f}^{-1}\left( U\right) \subseteq B\text{ and }B\bigcap {f}^{-1}\left( {U}^{c}\right) = \varnothing . \n\]\n\nSince \( {f}^{-1}\left( U\right) = B \cap A \), it is Borel in \( A \), and the result follows.
No
Corollary 4.5.5 Let \( X \) be a standard Borel space and \( Y \) a metrizable space. Suppose there is a one-to-one Borel map \( f \) from \( X \) onto \( Y \) . Then \( Y \) is standard Borel and \( f \) a Borel isomorphism.
Proof. By 4.3.8, \( Y \) is separable. The result follows from 4.5.4.
No
Theorem 4.5.7 (Blackwell - Mackey theorem, [13]) Let \( X \) be an analytic subset of a Polish space and \( \mathcal{A} \) a countably generated sub \( \sigma \) -algebra of the Borel \( \sigma \) -algebra \( {\mathcal{B}}_{X} \) . Let \( B \subseteq X \) be a Borel set that is a union of atoms of \( \mathcal{A} \) . Then \( B \in \mathcal{A} \) .
Proof. Let \( \left\{ {{B}_{n} : n \in \mathbb{N}}\right\} \) be a countable generator of \( \mathcal{A} \) . Consider the map \( f : X \rightarrow {2}^{\mathbb{N}} \) defined by\n\n\[ f\left( x\right) = \left( {{\chi }_{{B}_{0}}\left( x\right) ,{\chi }_{{B}_{1}}\left( x\right) ,\ldots }\right) ,\;x \in X. \]\n\nThen \( \mathcal{A} = {f}^{-1}\left( {\mathcal{B}}_{{2}^{\mathbb{N}}}\right) \) . In particular, \( f : X \rightarrow {2}^{\mathbb{N}} \) is Borel measurable. So, \( f\left( B\right) \) and \( f\left( {B}^{c}\right) \) are disjoint analytic subsets of \( {2}^{\mathbb{N}} \) . By 4.4.1, there is a Borel set \( C \) containing \( f\left( B\right) \) and disjoint from \( f\left( {B}^{c}\right) \) . Clearly, \( B = {f}^{-1}\left( C\right) \), and so it belongs to \( \mathcal{A} \) .
Yes
Theorem 4.6.1 (The generalized first separation theorem, Novikov[90]) Let \( \\left( {A}_{n}\\right) \) be a sequence of analytic subsets of a Polish space \( X \) such that \( \\bigcap {A}_{n} = \\varnothing \) . Then there exist Borel sets \( {B}_{n} \\supseteq {A}_{n} \) such that \( \\bigcap {B}_{n} = \\varnothing \) .
Proof of 4.6.1. (Mokobodzki [86]) Let \( \\left( {A}_{n}\\right) \) be a sequence of analytic sets that is not Borel separated and such that \( \\mathop{\\bigcap }\\limits_{n}{A}_{n} = \\varnothing \) . For each \( n \), fix a continuous surjection \( {f}_{n} : {\\mathbb{N}}^{\\mathbb{N}} \\rightarrow {A}_{n} \) . We get a sequence \( {\\alpha }_{0},{\\alpha }_{1},\\ldots \) in \( {\\mathbb{N}}^{\\mathbb{N}} \) such that for every \( k > 0 \) the sequence\n\n\[ \n{f}_{0}\\left( {\\sum \\left( {{\\alpha }_{0} \\mid k}\\right) }\\right) ,{f}_{1}\\left( {\\sum \\left( {{\\alpha }_{1} \\mid \\left( {k - 1}\\right) }\\right) }\\right) ,\\ldots ,{f}_{k - 1}\\left( {\\sum \\left( {{\\alpha }_{k - 1} \\mid 1}\\right) }\\right) ,{A}_{k},{A}_{k + 1},\\ldots \n\]\n\nis not Borel separated.\n\nTo see that such a sequence exists we proceed by induction. Write \( {A}_{0} = \\mathop{\\bigcup }\\limits_{n}{f}_{0}\\left( {\\sum \\left( n\\right) }\\right) \) . By 4.6.2, there exists \( {\\alpha }_{0}\\left( 0\\right) \\in \\mathbb{N} \) such that the sequence \( {f}_{0}\\left( {\\sum \\left( {{\\alpha }_{0}\\left( 0\\right) }\\right) }\\right) ,{A}_{1},{A}_{2},\\ldots \) is not Borel separated. Write \( {f}_{0}\\left( {\\sum \\left( {{\\alpha }_{0}\\left( 0\\right) }\\right) }\\right) = \\mathop{\\bigcup }\\limits_{m}{f}_{0}\\left( {\\sum \\left( {{\\alpha }_{0}\\left( 0\\right) m}\\right) }\\right) \) and \( {A}_{1} = \\mathop{\\bigcup }\\limits_{n}{f}_{1}\\left( {\\sum \\left( n\\right) }\\right) \) . Apply 4.6.2 again to get \( {\\alpha }_{0}\\left( 1\\right) ,{\\alpha }_{1}\\left( 0\\right) \\in \\mathbb{N} \) such that the sequence \( {f}_{0}\\left( {\\sum \\left( {{\\alpha }_{0}\\left( 0\\right) {\\alpha }_{0}\\left( 1\\right) }\\right) }\\right) ,{f}_{1}\\left( {\\sum \\left( {{\\alpha }_{1}\\left( 0\\right) }\\right) }\\right) ,{A}_{2},{A}_{2},\\ldots \n\n
Yes
Lemma 4.6.2 Let \( \left( {E}_{n}\right) \) be a sequence of subsets of \( X, k \in \mathbb{N} \), and \( {E}_{i} = \) \( \mathop{\bigcup }\limits_{n}{E}_{in} \) for \( i \leq k \) . Suppose \( \left( {E}_{n}\right) \) is not Borel separated. Then there exist \( {n}_{0},{n}_{1},\ldots ,{n}_{k} \) such that the sequence \( {E}_{0{n}_{0}},{E}_{1{n}_{1}},\ldots ,{E}_{k{n}_{k}},{E}_{k + 1},{E}_{k + 2},\ldots \) is not Borel separated.
Proof. We prove the result by induction on \( k \) .\n\nInitial step: \( k = 0 \) . Suppose the result is not true. Hence, for every \( n \) , there is a sequence \( {\left( {B}_{in}\right) }_{i \in \mathbb{N}} \) of Borel sets such that\n\n(i) \( \mathop{\bigcap }\limits_{i}{B}_{in} = \varnothing \) ,\n\n(ii) \( {B}_{0n} \supseteq {E}_{0n} \), and\n\n(iii) \( {B}_{in} \supseteq {E}_{i} \) for all \( i \) .\n\nLet\n\n\[ {B}_{i} = \mathop{\bigcup }\limits_{n}{B}_{in}\;\text{ if }\;i = 0, \]\n\n\[ = \mathop{\bigcap }\limits_{n}{B}_{in}\;\text{if}\;i > 0\text{.} \]\n\nThen \( {B}_{i} \supseteq {E}_{i} \), the \( {B}_{i} \) ’s are Borel and \( \bigcap {B}_{i} = \varnothing \) . This contradicts the hypothesis that \( \left( {E}_{n}\right) \) is not Borel separated, and we have proved the result for \( k = 0 \) .\n\nInductive step. Suppose \( k > 0 \) and the result is true for all integers less than \( k \) . By the induction hypothesis, there are integers \( {n}_{0},{n}_{1},\ldots ,{n}_{k - 1} \) such that \( {E}_{0{n}_{0}},{E}_{1{n}_{1}},\ldots ,{E}_{k - 1{n}_{k - 1}},{E}_{k},{E}_{k + 1},\ldots \) is not Borel separated. By the initial step, there is an \( {n}_{k} \) such that \( {E}_{0{n}_{0}},{E}_{1{n}_{1}},\ldots ,{E}_{k{n}_{k}},{E}_{k + 1},{E}_{k + 2},\ldots \) is not Borel separated.
Yes
Theorem 4.6.5 (Weak reduction principle for coanalytic sets) Let \( {C}_{0},{C}_{1},{C}_{2},\ldots \) be a sequence of coanalytic subsets of a Polish space such that \( \bigcup {C}_{n} \) is Borel. Then there exist pairwise disjoint Borel sets \( {B}_{n} \subseteq {C}_{n} \) such that \( \bigcup {B}_{n} = \bigcup {C}_{n} \) .
Proof. Let \( {A}_{n} = X \smallsetminus {C}_{n} \), where \( X = \mathop{\bigcup }\limits_{n}{C}_{n} \) . Then \( \left( {A}_{n}\right) \) is a sequence of analytic sets such that \( \mathop{\bigcap }\limits_{n}{A}_{n} = \varnothing \) . By 4.6.1, there exist Borel sets \( {D}_{n} \supseteq {A}_{n} \) such that \( \mathop{\bigcap }\limits_{n}{D}_{n} = \varnothing \) . Take\n\n\[ \n{B}_{n} = {B}_{n}^{\prime } \smallsetminus \mathop{\bigcup }\limits_{{m < n}}{B}_{m}^{\prime }\n\]\n\nwhere \( {B}_{n}^{\prime } = X \smallsetminus {D}_{n} \) .
Yes
Theorem 4.7.1 (Saint Raymond[97]) Let \( {A}_{0} \) and \( {A}_{1} \) be disjoint analytic subsets of \( X \times Y \) with the sections \( {\left( {A}_{0}\right) }_{x}, x \in X \), closed in \( Y \). Then there is a sequence \( \left( {B}_{n}\right) \) of Borel subsets of \( X \) such that \[ {A}_{1} \subseteq \mathop{\bigcup }\limits_{n}\left( {{B}_{n} \times {V}_{n}}\right) \text{ and }{A}_{0}\bigcap \mathop{\bigcup }\limits_{n}\left( {{B}_{n} \times {V}_{n}}\right) = \varnothing . \]
Proof. By 4.4.1, there is a Borel set containing \( {A}_{1} \) and disjoint from \( {A}_{0} \). So, without any loss of generality, we assume that \( {A}_{1} \) is Borel. For each \( n \), let \[ {C}_{n} = \left\{ {x \in X : {V}_{n} \subseteq {\left( {A}_{0}\right) }_{x}^{c}}\right\} \] Then \( {C}_{n} \) is coanalytic and \[ {\left( {A}_{0}\right) }^{c} = \mathop{\bigcup }\limits_{n}\left( {{C}_{n} \times {V}_{n}}\right) \] Note that \( \left( {\left( {{C}_{n} \times {V}_{n}}\right) \bigcap {A}_{1}}\right) \) is a sequence of coanalytic sets whose union is Borel. Hence, by 4.6.5, there exist Borel sets \( {D}_{n} \subseteq \left( {{C}_{n} \times {V}_{n}}\right) \bigcap {A}_{1} \) such that \[ \bigcup {D}_{n} = \mathop{\bigcup }\limits_{n}\left( {{A}_{1}\bigcap \left( {{C}_{n} \times {V}_{n}}\right) }\right) = {A}_{1}. \] By 4.4.1, there exist Borel sets \( {B}_{n} \) such that \[ {\pi }_{X}\left( {D}_{n}\right) \subseteq {B}_{n} \subseteq {C}_{n} \] where \( {\pi }_{X} : X \times Y \rightarrow X \) is the projection map. It is now fairly easy to see that \( \left( {B}_{n}\right) \) satisfies \( \left( \star \right) \).
Yes
Theorem 4.7.2 (Kunugui, Novikov) Suppose \( B \subseteq X \times Y \) is any Borel set with sections \( {B}_{x} \) open, \( x \in X \) . Then there is a sequence \( \left( {B}_{n}\right) \) of Borel subsets of \( X \) such that\n\n\[ B = \bigcup \left( {{B}_{n} \times {V}_{n}}\right) \]
Proof. Apply 4.7.1 to \( {A}_{0} = {B}^{c} \) and \( {A}_{1} = B \) .
No
Corollary 4.7.4 Suppose \( B \subseteq X \times Y \) is a Borel set with the sections \( {B}_{x} \) closed. Then there is a Polish topology \( \mathcal{T} \) finer than the given topology on \( X \) generating the same Borel \( \sigma \) -algebra such that \( B \) is closed relative to the product topology on \( X \times Y, X \) being equipped with the new topology \( \mathcal{T} \) .
Proof. By 4.7.2, write\n\n\[ \n{B}^{c} = \mathop{\bigcup }\limits_{n}\left( {{B}_{n} \times {V}_{n}}\right) \n\]\n\nthe \( {B}_{n} \) ’s Borel. By 3.2.5, take a finer Polish topology \( \mathcal{T} \) on \( X \) generating the same Borel \( \sigma \) -algebra such that \( {B}_{n} \) is \( \mathcal{T} \) -open.
Yes
Example 4.7.9 (H. Sarbadhikari) Let \( A \subseteq \left\lbrack {0,1}\right\rbrack \) be an analytic non-Borel set and \( E \subseteq \left\lbrack {0,1}\right\rbrack \times {\mathbb{N}}^{\mathbb{N}} \) a closed set whose projection is \( A \) . Set \( B = \) \( E\bigcup \left( {\left( {\left\lbrack {0,1}\right\rbrack \smallsetminus A}\right) \times {\mathbb{N}}^{\mathbb{N}}}\right) \) and \( f : B \rightarrow \left\lbrack {0,1}\right\rbrack \) the characteristic function of \( E \) . We claim that there is no Borel extension \( F : \left\lbrack {0,1}\right\rbrack \times {\mathbb{N}}^{\mathbb{N}} \rightarrow \left\lbrack {0,1}\right\rbrack \) of \( f \) such that \( y \rightarrow F\left( {x, y}\right) \) is continuous.
Suppose not. Consider \( C = {F}^{-1}(\left( {0,1\rbrack }\right) \) . Then \( C \) is a Borel set with sections \( {C}_{x} \) open and whose projection is \( A \) . Hence \( A \) is Borel. (See the paragraph below.) We have arrived at a contradiction.
Yes
Theorem 4.7.11 (Novikov) Let \( X \) and \( Y \) be Polish spaces and \( B \) a Borel subset of \( X \times Y \) with sections \( {B}_{x} \) compact. Then \( {\pi }_{X}\left( B\right) \) is Borel in \( X \) .
Proof. (Srivastava) Since every Polish space is homeomorphic to a \( {G}_{\delta } \) subset of the Hilbert cube \( \mathbb{H} \), without any loss of generality, we assume that \( Y \) is a compact metric space. Note that the sections \( {B}_{x} \) are closed in \( Y \) . By 4.7.4, there is a finer Polish topology on \( X \) generating the same Borel \( \sigma \) -algebra and making \( B \) closed in \( X \times Y \) . Hence, by 2.3.24, \( {\pi }_{X}\left( B\right) \) is closed in \( X, X \) being equipped with the new topology. But the Borel structure of \( X \) is the same with respect to both the topologies. The result follows.
Yes
Corollary 4.7.12 Let \( X, Y \) be Polish spaces with \( {Y\sigma } \) -compact (equivalently, locally compact). Then the projection of every Borel set \( B \) in \( X \times Y \) with \( x \) -sections closed in \( Y \) is Borel.
Proof. Write \( Y = \mathop{\bigcup }\limits_{n}{Y}_{n},{Y}_{n} \) compact. Then\n\n\[ \n{\pi }_{X}\left( B\right) = \mathop{\bigcup }\limits_{n}{\pi }_{X}\left( {B\bigcap \left( {X \times {Y}_{n}}\right) }\right) \n\] \n\nNow apply 4.7.11.
No
Theorem 4.8.1 Let \( \left( {G, \cdot }\right) \) be a Polish group and \( H \) a closed subgroup. Suppose \( E = \left\{ {\left( {x, y}\right) : x \cdot {y}^{-1} \in H}\right\} \) ; i.e., \( E \) is the equivalence relation induced by the right cosets. Then the \( \sigma \) -algebra of invariant Borel sets is countably generated.
Proof. Let \( \left\{ {{U}_{n} : n \in \mathbb{N}}\right\} \) be a countable base for the topology of \( G \) . Put\n\n\[ \n{B}_{n} = \mathop{\bigcup }\limits_{{y \in H}}y \cdot {U}_{n} \n\]\n\nSo, the \( {B}_{n} \) ’s are Borel (in fact, open). We show that \( \left\{ {{B}_{n} : n \in \mathbb{N}}\right\} \) generates \( \mathcal{B} \) .\n\nLet \( {H}_{1} \) and \( {H}_{2} \) be two distinct cosets. Since \( H \) is closed, \( {H}_{1} \) and \( {H}_{2} \) are closed. Since they are disjoint, there is a basic open set \( {U}_{n} \) such that \( {U}_{n}\bigcap {H}_{1} \neq \varnothing \) and \( {U}_{n}\bigcap {H}_{2} = \varnothing \) . Then \( {H}_{1} \subset {B}_{n} \) and \( {B}_{n}\bigcap {H}_{2} = \varnothing \) . It follows that the right cosets are precisely the atoms of \( \sigma \left( \left\{ {{B}_{n} : n \in \mathbb{N}}\right\} \right) \) . The result now follows from 4.5.7.
Yes