How statement and proof provenance work
The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.
- Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
- AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
- AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.
These labels describe origin, not correctness: citations and verification chips remain separate evidence.
Lebesgue Measure on Euclidean Space
1 · Prerequisites
- Areas of Elementary Plane Figures
- Binary Operations, Monoids, Groups and Subgroups
- Compactness
- Compactness in Metric Spaces
- Completeness, Completion, and Uniform Continuity
- Construction of the Natural Numbers
- Construction of the Real Numbers via Cauchy Sequences
- Construction of the Real Numbers via Dedekind Cuts
- Continuity, IVT, EVT, and Uniform Continuity
- Cosets, Index and Lagrange's Theorem
- Countability and Uncountability
- Determinants of Matrices over a Commutative Ring
- Divisibility, Euclidean Domains, Principal Ideal Domains and Unique Factorisation
- Dual Spaces, Bilinear and Quadratic Forms, and Sylvester's Law of Inertia
- Filters and Ultrafilters
- Finite Counting, Factorials and Binomial Coefficients
- Foundations of the Real Numbers for Analysis
- Fubini and Change of Variables
- Gaussian Elimination, Elementary Matrices and Reduced Row Echelon Form
- Group Actions, Orbits, Stabilisers and Cayley's Theorem
- Ideals, Quotient Rings and the Isomorphism Theorems for Rings
- Inner Product Spaces, Gram-Schmidt, Projections and Adjoints
- Limits of Real Functions
- limsup, liminf, and Subsequential Limits
- Linear Independence, Bases and Dimension
- Linear Transformations, Rank-Nullity and Quotient Spaces
- Matrices, the Matrix of a Linear Map, and Change of Basis
- Measures and Their Basic Properties
- Metric Spaces
- Monotone Functions, Discontinuities, and Continuity Sets
- Monotone Sequences, Bolzano-Weierstrass, and Cauchy Completeness
- Normal Subgroups and Quotient Groups
- Order, Zorn's Lemma, and the Axiom of Choice
- Outer Measure and the Caratheodory Extension Theorem
- Polynomial Rings, the Division Algorithm and Roots
- Relations, Functions, and Quotients
- Rings, Subrings, Integral Domains and Fields
- Rⁿ as a Normed Space; Vector-Valued Functions
- Roots, Rational Powers, and Classical Inequalities
- Separation Axioms: the Hierarchy
- Sequences and Limits
- Series: Convergence and the Nonnegative Tests
- Sigma Algebras and Borel Sets
- Simple Field Extensions and the Construction of the Complex Numbers
- Subspaces, Products, and Quotients
- Suprema and Infima
- Symmetric Groups, Cycle Decomposition and the Sign Homomorphism
- The Cantor Set, Baire Category, and Measure Zero in ℝ
- The Determinant of a Linear Operator, Cofactors and Cramer's Rule
- The Riemann Integral in Rᵐ and Jordan Content
- The Riemann Integral: Definition and Integrability
- The Topology of Euclidean Space
- The Total Derivative in ℝᵐ → ℝⁿ
- The ZFC Axioms and the Basic Set Constructions
- Topological Spaces and Continuity
- Topology of ℝ
- Vector Spaces, Linear Subspaces, Span and Direct Sums
2 · Summary
Assuming countable choice, the outer-measure and Caratheodory-extension machinery from the prerequisite page, together with the Borel sigma-algebra, Heine-Borel compactness in , the published covering notions of nullity, Jordan content, and the determinant and elementary-matrix material, gives the framework for Lebesgue measure on Euclidean space. Those dependencies are used here to pass from countable covers to measurable sets, to compare Lebesgue nullity with the earlier covering vocabulary, and to move from box computations to structural results such as regularity and invariance.
The page begins with half-open boxes and elementary sets, proves that elementary volume is a sigma-finite premeasure, and then defines Lebesgue outer measure, the Lebesgue sigma-algebra and Lebesgue measure. It next computes the measure of boxes, proves sigma-finiteness and the basic nullity results, identifies outer and inner regularity through the Littlewood characterisations and the Borel-completion description, and then turns to invariance: translation, dilation, orthogonal maps, the linear determinant formula, and finally Steinhaus with its subgroup corollary.
3 · Logical flowchart
4 · Definitions, theorems and proofs
Half-open boxes in and their volume
Definition
Fix with and let be the set of functions , writing for ( as the set of functions , and , , are metrics on it). A parameter is a function , where carries the total order of The extended real line , its order, and the arithmetic that is left undefined. For a pair of parameters set
both comparisons taken in . A half-open box is a set of this form. For a single write for the constant parameter with value , and abbreviate ; thus and is the unit cube. At , and for real , the box is the half-open interval of Intervals of : the nine order-convex forms, nondegeneracy, and length.
A box is nonempty exactly when for every . If some then no real satisfies both and , by transitivity of the order, so . Conversely suppose for every . Then in each coordinate some real satisfies : if take , which is ; if then , so take when is real and when . Assembling one such in each coordinate gives a point of ; the assembly is a definition by cases on finitely many coordinates and selects nothing.
The parameters of a nonempty box are determined by the set. Let , fix and put . Then : the inclusion is the defining condition, and for take any and replace its -th coordinate by , which leaves every other defining inequality untouched. Now exactly when has no upper bound in , and otherwise is the greatest element of ; likewise exactly when has no lower bound in , and otherwise is the greatest lower bound of in . So determines and for every , and hence determines .
Volume. For a half-open box define by
- ;
- if , with its unique parameter pair , then when or for some , and when every and every is real.
The product is the finite product of Finite sums and finite products, by recursion. In the last clause every factor is a strictly positive real, since in , so the product is a strictly positive real (Laws of finite sums and finite products, claim 6) and in particular no factor is and no product of the form is ever formed. That is what the case split buys: a box with a degenerate side is empty, not a box of volume with an infinite side. So is a total function on the half-open boxes with values in , and , .
Agreement with the published rectangle volume. For real parameters with for every , the closed rectangle of Axis-parallel rectangles in and their volume has , which is the value assigned above to . The two notions of volume therefore agree wherever both are written, and no second notion of volume is introduced.
Remarks
-
Why the upper face and not the lower one. The published For n at least one, open sets, closed sets, compact sets, open balls, boxes, rational open boxes, and rational half-open boxes generate the Borel sigma-algebra on R^n already fixes this library's half-open box as , in its Statement. The choice is a convention, but having two conventions for one phrase is not, so the published one is kept.
-
Why the parameters are allowed to be infinite. With real parameters only, every finite union of half-open boxes is a bounded set, so would not be one and the family would be closed under difference but not under complement. Admitting and as parameters makes a box, and it is what makes The elementary sets form an algebra of subsets of containing every half-open box an algebra rather than a ring.
-
Half-open, and not closed, for a second reason. Two closed rectangles that share a face are not disjoint, so a decomposition of a rectangle into closed pieces is disjoint only up to boundaries. Half-open boxes tile exactly: the cells of a coordinate grid are pairwise disjoint with union the whole box (The volume of a half-open box is the sum of the volumes of the cells of any coordinate grid subdividing it), and no boundary bookkeeping is needed anywhere below.
Half-open boxes are closed under intersection, and the complement of a half-open box is a finite disjoint union of half-open boxes
Statement
Let and let half-open boxes be as in Half-open boxes in and their volume.
- Intersection. For parameter pairs and , the extremes taken in the total order of . Consequently the intersection of the members of a finite list of half-open boxes is a half-open box, the empty list giving .
- Complement. For every parameter pair there is a finite list of pairwise disjoint half-open boxes whose union is . When the list may be taken to have members, indexed by a coordinate and a side.
Facts & Assumptions
Given: A natural number and parameter pairs , , that is, pairs of functions .
, and (Half-open boxes in and their volume).
A box is nonempty exactly when for every (Half-open boxes in and their volume).
is a totally ordered set, and the inclusion of preserves and reflects the order (The extended real line , its order, and the arithmetic that is left undefined).
Proof
For claim 1, a point lies in exactly when and for every ; the order being total, each two-element set has a greatest member and each a least member , and for a real the conjunction and says exactly while and says exactly , so the intersection is ; iterating along a list of length gives the finite case by induction on , with the empty list giving .
For claim 2 in the degenerate case, if then , a list with the single member , whose members are vacuously pairwise disjoint.
For claim 2 in the remaining case, assume , so for every , and for define two parameter pairs and by setting, in coordinates , and ; in coordinate , and ; and in coordinates , and .
Still for claim 2, every lies in one of these boxes: the set of with is a nonempty subset of , so it has a least member ; then for every , and by totality either , putting in , or , putting in , the coordinates being unconstrained in both.
Still for claim 2, each of the boxes is disjoint from , since its points satisfy or ; and two of them are disjoint from one another, because for a point of a box with index fails while a point of a box with index satisfies it, and for a common a point of both would satisfy , contradicting .
Claim 1 is step 1.1, and claim 2 is step 1.2 in the empty case and steps 2.1 and 2.2 in the nonempty case, the union of the boxes being exactly .
Elementary sets: the finite unions of half-open boxes in
Definition
Fix . A subset is an elementary set when there are a natural number and a list of half-open boxes (Half-open boxes in and their volume), that is a function on whose values are half-open boxes, with
Write for the family of all elementary subsets of .
The list is part of the data of the presentation and not of the set: one elementary set has many presentations, and nothing below reads a presentation off a set. At the union is empty, so ; at every half-open box is elementary, included. The boxes of a presentation are not required to be disjoint or nonempty.
Remarks
-
Every elementary set does admit a disjoint presentation, by a common coordinate grid built from the parameters of the given list (Every elementary set is a finite disjoint union of half-open boxes, and any finitely many boxes admit a common grid refinement). That is a theorem about , not part of the definition, and keeping it out of the definition is what lets a presentation be produced by hand wherever one is convenient.
-
is an algebra and not merely a ring of sets (The elementary sets form an algebra of subsets of containing every half-open box), and that is exactly what the infinite parameters of Half-open boxes in and their volume buy: with real parameters only, every member of would be bounded and the family would not contain .
The elementary sets form an algebra of subsets of containing every half-open box
Statement
Let . The family of elementary subsets of (Elementary sets: the finite unions of half-open boxes in ) is an algebra of subsets of (Algebras of subsets): it contains , it is closed under complement in , and it is closed under union of two members. It contains every half-open box, and it is closed under intersection of two members and under difference.
Facts & Assumptions
Given: A natural number and the family of finite unions of half-open boxes in .
A subset is an elementary set when there are a natural number and a list of half-open boxes with ; at the union is empty, so ; at every half-open box is elementary, included (Elementary sets: the finite unions of half-open boxes in ).
The intersection of the members of a finite list of half-open boxes is a half-open box, the empty list giving (Half-open boxes are closed under intersection, and the complement of a half-open box is a finite disjoint union of half-open boxes).
For every parameter pair there is a finite list of pairwise disjoint half-open boxes whose union is (Half-open boxes are closed under intersection, and the complement of a half-open box is a finite disjoint union of half-open boxes).
An algebra of subsets of is a family such that ; if , then ; and if , then (Algebras of subsets).
Proof
The empty list of boxes has union and the one-member list has union , so , every half-open box lies in , and .
If and are presentations, then concatenating the two lists into a list of length presents , so is closed under the union of two members.
With the same presentations, , each is a half-open box, and the boxes can be listed by a bijection of with the pairs , so .
The complement of a single half-open box is a finite union of half-open boxes, hence lies in .
For a presentation one has ; putting and , an induction on using step 1.1 for and steps 1.3 and 1.4 for the successor case gives for every , and .
Steps 1.1, 1.2 and 2.1 are the three clauses of [F1], so is an algebra of subsets of ; it contains every half-open box by step 1.1, is closed under binary intersection by step 1.3, and is closed under difference because .
Every elementary set is a finite disjoint union of half-open boxes, and any finitely many boxes admit a common grid refinement
Statement
Let and let be a finite list of half-open boxes in (Half-open boxes in and their volume), where . For put
a finite set with at least two members, and let be its increasing enumeration, so that , and . The cells of the grid generated by the list are the half-open boxes
Then:
- the cells are pairwise disjoint and their union is ;
- for every , a cell that meets is contained in , and is the union of the cells contained in it;
- consequently every elementary set (Elementary sets: the finite unions of half-open boxes in ) is the union of a finite list of pairwise disjoint half-open boxes.
Facts & Assumptions
Given: A natural number , a finite list of half-open boxes with parameter pairs , and the sets and cells displayed in the Statement.
A subset is an elementary set when there are a natural number and a list of half-open boxes with (Elementary sets: the finite unions of half-open boxes in ).
is a totally ordered set, and the inclusion of preserves and reflects the order; is the least and the greatest element of , and for every (The extended real line , its order, and the arithmetic that is left undefined).
Proof
Each is a finite subset of the totally ordered set containing the two distinct elements and , so it has a unique strictly increasing enumeration with , and its least and greatest members are and .
For claim 1, distinct multi-indices differ at some , say , whence and ; a common point would satisfy both and , which is impossible, so the cells are pairwise disjoint.
For claim 1 again, given and , the set contains because and omits because is not below the real , so it has a greatest member with , and then ; the multi-index so obtained puts in .
For claim 2, suppose and fix . From and it follows that , and the members of strictly below are exactly , so ; from and it follows that , and the members of strictly above are exactly , so .
For claim 2, step 1.4 gives and in every coordinate, hence ; and every point of lies in some cell by step 1.3, that cell then meeting and so contained in it, so is exactly the union of the cells contained in it.
For claim 3, let be elementary; by step 2.1 each is the union of the cells contained in it, so is the union of those cells that are contained in at least one , and by step 1.2 these finitely many cells are pairwise disjoint; listing them proves claim 3, while claims 1 and 2 are steps 1.2, 1.3 and 2.1.
The volume of a half-open box is the sum of the volumes of the cells of any coordinate grid subdividing it
Statement
Let and let be a nonempty half-open box (Half-open boxes in and their volume). Suppose that for each a strictly increasing finite list
in is given, and for a multi-index with for every put . Then the cells are nonempty half-open boxes, pairwise disjoint, with union , and
where a sum over cells is the iterated recursive sum of Grid partitions of a rectangle in , their cells, refinements and mesh, formed here in by the recursion of Series in the nonnegative extended real line.
Facts & Assumptions
Given: A natural number , a nonempty box , the lists and the cells of the Statement, and the induction principle (The principle of mathematical induction). For and a multi-index with for every , let denote the half-open box whose -th parameter pair is for and for , so that is itself and ; and let be the assertion that , a sum over no index being read as its single term.
A box is nonempty exactly when for every , and (Half-open boxes in and their volume).
For a nonempty box with parameter pair , when or for some , and when every and every is real; and (Half-open boxes in and their volume).
For sequences of reals, ; if then ; ; and (Laws of finite sums and finite products, claims 2, 3, 5 and 6).
Finite sums and finite products of a sequence of reals are defined by the recursions , and , , written and (Finite sums and finite products, by recursion).
For , when and , or and (The extended real line , its order, and the arithmetic that is left undefined).
The partial sums of a sequence in are the unique sequence with and , and finite sums use the same recursion, (Series in the nonnegative extended real line).
A sum over cells means the iterated recursive sum of Finite sums and finite products, by recursion (Grid partitions of a rectangle in , their cells, refinements and mesh).
Proof
Each cell is a nonempty box contained in , since for every , so that in every coordinate.
The cells are pairwise disjoint with union : distinct multi-indices differ at some with, say, , whence and no point can satisfy and at once; and for and the set contains and omits , so its greatest member satisfies and .
A finite sum in equals exactly when one of its terms does, and when every term is real it is the finite sum of those reals: the recursion produces once a term is and never leaves otherwise.
Let be a nonempty box all of whose parameters are real, let , let be reals, and write for the box obtained from by replacing its -th parameter pair by ; putting for and , and , the product and splitting laws give and , so scaling and telescoping give .
At the iterated sum carries no summation index, so its value is its single term and holds.
Let and assume as the induction hypothesis.
Let be a nonempty box, let , and let in with the boxes as in step 1.4; if some parameter of is infinite then , and the sum is as well, because an infinite parameter in a coordinate is shared by every nonempty , while makes and makes .
Combining the two cases, for every nonempty box , every and every strictly increasing list in one has , since either all parameters of are real, and then so are all the , or some parameter is infinite.
Each box is nonempty, by the inequalities of step 1.1 applied in coordinates and in the others, and its -th parameter pair is with the list available, so step 3.1 gives ; substituting this into the identity of step 1.6 termwise yields .
By induction holds for every , and is the displayed identity because ; together with steps 1.1 and 1.2 this is the Statement.
The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition
Statement
Let and let be an elementary set (Elementary sets: the finite unions of half-open boxes in ). If
for finite lists of pairwise disjoint half-open boxes (Half-open boxes in and their volume), then
in . Consequently there is exactly one function , the elementary volume, whose value at is the sum of the volumes of the members of any presentation of by a finite list of pairwise disjoint half-open boxes. It satisfies and for every half-open box .
Facts & Assumptions
Given: A natural number , an elementary set , and two presentations by finite lists of pairwise disjoint half-open boxes.
Every elementary set has a presentation as a finite pairwise disjoint union of half-open boxes. Applied to the concatenated list , the generated grid has pairwise disjoint cells whose union is ; for every member of the list, a cell that meets it is contained in it, and that member is the union of the cells contained in it (Every elementary set is a finite disjoint union of half-open boxes, and any finitely many boxes admit a common grid refinement).
For a nonempty box and strictly increasing lists with , the cells are nonempty pairwise disjoint boxes with union and (The volume of a half-open box is the sum of the volumes of the cells of any coordinate grid subdividing it).
, and a box is nonempty exactly when for every (Half-open boxes in and their volume).
A subset is an elementary set when there are a natural number and a list of half-open boxes with (Elementary sets: the finite unions of half-open boxes in ).
For sequences of reals, , and if then (Laws of finite sums and finite products, claims 1 and 3).
Finite sums of a sequence of reals are defined by the recursion , , written (Finite sums and finite products, by recursion).
Proof
Let be the cells of the grid generated by the concatenated list, indexed by the multi-indices with for ; they are nonempty, pairwise disjoint, cover , and each of them is either contained in or disjoint from each and each .
If one of the boxes in either decomposition has infinite volume, then both sums are and there is nothing left to prove. Indeed, by [L3] a nonempty box has infinite volume exactly when some endpoint is infinite, and such a box is unbounded. Conversely, a finite union of boxes all of whose endpoints are real is bounded: for each such box every coordinate of every point of lies between the real endpoints and , so choosing one real bound for each box and then taking the maximum over the finite list bounds the whole union. Therefore, if contains an unbounded box from one decomposition, the other decomposition cannot consist entirely of finite-volume boxes, since that would make bounded. So it remains only to treat the case in which every and every has finite volume; from now on all the volumes that appear are real numbers and the finite-sum laws of [F1] apply to them.
For each let when and otherwise; then , because for no cell is contained in it and both sides are , while for nonempty the parameters and occur among the grid points, say and with , the cells contained in are exactly those with in every coordinate, the lists subdivide so that [L2] applies, and widening each summation range from to only inserts zero terms.
A cell is contained in exactly when it is contained in exactly one , since a cell contained in meets , hence meets some and is contained in it, while a cell contained in two of the pairwise disjoint boxes would be empty; so, writing when and otherwise, one has for every .
Summing the identities of step 2.1 over and regrouping the resulting finite real sums by repeated use of the additivity law in [F1] gives . Since step 2.2 identifies the inner sum with , this is .
The right-hand side of step 3.1 is built from and the grid alone, and the same computation applied to the list , whose parameters also generate the same grid, gives for the same value; hence the two sums agree, and since every elementary set has at least one presentation by a finite list of pairwise disjoint half-open boxes, the assignment is a well-defined function on with and for a single box.
Elementary volume is finitely additive, monotone and finitely subadditive on the elementary algebra
Statement
Let , let be the elementary subsets of (Elementary sets: the finite unions of half-open boxes in ) and let be elementary volume (The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition). Let and let be a finite list in . Then:
- Finite additivity. If the are pairwise disjoint, then .
- Monotonicity. If , then .
- Finite subadditivity. .
All three hold with the value allowed, the sums being the finite sums of Series in the nonnegative extended real line.
Facts & Assumptions
Given: A natural number , the algebra , elementary volume , and elementary sets , and .
For every , there is exactly one function whose value at is the sum of the volumes of the members of any presentation of by a finite list of pairwise disjoint half-open boxes (The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition).
Every elementary set is the union of a finite list of pairwise disjoint half-open boxes (Every elementary set is a finite disjoint union of half-open boxes, and any finitely many boxes admit a common grid refinement).
is an algebra of subsets of , it contains every half-open box, and it is closed under intersection of two members and under difference (The elementary sets form an algebra of subsets of containing every half-open box).
A subset is an elementary set when there are a natural number and a list of half-open boxes with (Elementary sets: the finite unions of half-open boxes in ).
For sequences of reals, ; if then ; and if whenever then (Laws of finite sums and finite products, claims 1, 3 and 4).
Finite sums of a sequence of reals are defined by the recursion , (Finite sums and finite products, by recursion).
The partial sums of a sequence in are the unique sequence with and , and finite sums use the same recursion, (Series in the nonnegative extended real line).
For , when and , or and (The extended real line , its order, and the arithmetic that is left undefined).
Proof
A finite sum in equals exactly when one of its terms does, and otherwise is the finite sum of reals; hence such sums split over a concatenation of two lists, are monotone termwise, and satisfy for , since with all terms real these are the laws for finite sums of reals and otherwise both sides are .
For claim 1, choose for each a presentation of by a finite list of pairwise disjoint half-open boxes, finitely many instantiations of an existential statement; the concatenated list presents and its members are pairwise disjoint, boxes from different being disjoint because the are, so splitting the concatenated sum over the blocks gives .
For claim 2, is a disjoint union of two elementary sets, so claim 1 gives .
For claim 3, put ; each is elementary, the are pairwise disjoint with , and , so claim 1 and then claim 2 termwise give , which with steps 2.1 and 3.1 is the Statement.
Every elementary set is squeezed in volume between a compact subset and an elementary set whose interior contains it
Statement
Let , let be elementary volume on the elementary sets (The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition, Elementary sets: the finite unions of half-open boxes in ), and for and a real put
the translates being those of Translation of a subset of . Then:
- is an elementary set, it is determined by and alone, it contains , and every point of is an interior point of in (Interior, closure, boundary, limit point, isolated point and dense subset of a metric space, as the set of functions , and , , are metrics on it).
- For every real there is with .
- If , then for every real there are an elementary set and a compact set (Open cover, subcover, compact metric space, and compact subset of a metric space) with and .
Nothing in claim 1 or claim 2 depends on a presentation of , so the assignment and the least satisfying claim 2 are both functions of the data and involve no selection.
Facts & Assumptions
Given: A natural number , an elementary set , and a real . A presentation of is written with pairwise disjoint half-open boxes, and when these are real.
; a box is nonempty exactly when for every ; ; and for a nonempty box when or for some , and when every and every is real (Half-open boxes in and their volume).
Every elementary set is the union of a finite list of pairwise disjoint half-open boxes, and a subset is an elementary set when there are a natural number and a list of half-open boxes with (Every elementary set is a finite disjoint union of half-open boxes, and any finitely many boxes admit a common grid refinement, Elementary sets: the finite unions of half-open boxes in ).
For every , there is exactly one function whose value at is the sum of the volumes of the members of any presentation of by a finite list of pairwise disjoint half-open boxes (The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition).
Elementary volume is finitely additive on pairwise disjoint elementary sets, monotone, and finitely subadditive (Elementary volume is finitely additive, monotone and finitely subadditive on the elementary algebra).
The translate of by is (Translation of a subset of ).
and are metrics on for ( as the set of functions , and , , are metrics on it).
is an interior point of if for some , where , and a subset is open in if every has such a ball inside (Interior, closure, boundary, limit point, isolated point and dense subset of a metric space, Open ball, closed ball and sphere in a metric space, The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement).
For reals the box is a compact subset of , and a subset is compact if and only if is closed in and bounded (Heine-Borel in : with the Euclidean metric a subset of is compact if and only if it is closed and bounded, and the proof by bisection uses no choice principle; the same holds on the real line, claims 1 and 2; Axis-parallel rectangles in and their volume).
A compact subset of a metric space is closed and bounded (A compact subset of a metric space is closed and bounded).
If and are open, then is open (Arbitrary unions and finite intersections of open sets are open, open balls are open and closed balls are closed, claim 3).
is bounded if or there are and a real with (Bounded subset, diameter, distance from a point to a set, and distance between two sets in a metric space).
For every real there is a natural number with (For every in a complete ordered field there is a natural with ).
For sequences of reals, ; ; if whenever then ; and (Laws of finite sums and finite products, claims 1, 2, 4 and 6).
Finite sums and finite products of a sequence of reals are defined by the recursions , and , (Finite sums and finite products, by recursion).
For , when and , or and ; and when and , or and (The extended real line , its order, and the arithmetic that is left undefined).
Proof
For a natural number , reals and a real with for every , one has : at both products are the empty product and both sides are , and passing from to uses together with and , so the estimate follows by induction on .
For a box and one has , where is the parameter , and the two boxes have the same volume; consequently, for a real , when and , and for any presentation of , so is elementary while its definition mentions no presentation.
For and one has .
If then every nonempty box of a disjoint presentation of has all parameters real, since an infinite parameter would make its volume, and hence the sum, equal to . Put and when every box is empty. Then is a real number and every endpoint of every nonempty box has modulus at most . Every point of a nonempty box therefore has every coordinate bounded by , hence has Euclidean norm at most ; thus every box lies in the ball about the origin of radius . Therefore is bounded and so is every subset of .
For claim 1, taking gives ; and if and then satisfies for every by step 1.3, so , whence the ball of lies in and is an interior point of it.
For claim 2, if the inequality holds with ; otherwise fix a disjoint presentation, let be a real with for every nonempty and every , and let : finite subadditivity and step 1.2 give , an empty contributing and a nonempty one contributing , so step 1.1 applied with and bounds each term by and hence .
For claim 3, assume , fix a disjoint presentation with all parameters of the nonempty real by step 1.4, and define for nonempty and for empty . Let be as in step 2.2, let , and put for nonempty and otherwise, and : then , each listed closed rectangle is compact and hence closed, a finite union of closed sets is closed by complementation, is bounded because , so is compact; and in both the empty and the nonempty case, so step 1.1 applied with and , whose difference is at most , gives .
Given a real , apply [F9] to the positive real to obtain with , and put , so that satisfies and ; steps 2.2 and 3.1 then give claims 2 and 3, and step 2.1 gives claim 1.
Elementary volume is a sigma-finite premeasure on the algebra of elementary sets
Statement
Let , let be the algebra of elementary subsets of (The elementary sets form an algebra of subsets of containing every half-open box) and let be elementary volume (The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition). Then is a sigma-finite premeasure on (Premeasures on algebras of sets): ; whenever is a pairwise disjoint sequence in whose union again lies in ,
and with for every .
No choice principle is used. The one place where a textbook proof selects countably many objects is the enlargement of each , and here the enlarged set is the canonical of Every elementary set is squeezed in volume between a compact subset and an elementary set whose interior contains it with the least natural number that works, which is a definition rather than a selection. The compact inner set and the finite subcover are each a single instantiation of an existential statement.
Facts & Assumptions
Given: A natural number , the algebra with elementary volume , and a pairwise disjoint sequence in whose union lies in .
is an algebra of subsets of , it contains every half-open box, and it is closed under intersection of two members and under difference (The elementary sets form an algebra of subsets of containing every half-open box).
For every , there is exactly one function whose value at is the sum of the volumes of the members of any presentation of by a finite list of pairwise disjoint half-open boxes; it satisfies and for every half-open box (The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition).
Elementary volume is finitely additive on pairwise disjoint elementary sets, monotone, and finitely subadditive (Elementary volume is finitely additive, monotone and finitely subadditive on the elementary algebra).
is an elementary set, it is determined by and alone, it contains , and every point of is an interior point of in (Every elementary set is squeezed in volume between a compact subset and an elementary set whose interior contains it, claim 1).
For every real there is with (Every elementary set is squeezed in volume between a compact subset and an elementary set whose interior contains it, claim 2).
If , then for every real there are an elementary set and a compact set with and (Every elementary set is squeezed in volume between a compact subset and an elementary set whose interior contains it, claim 3).
For a nonempty box when or for some , and when every and every is real; ; and (Half-open boxes in and their volume).
A subset is an elementary set when there are a natural number and a list of half-open boxes with (Elementary sets: the finite unions of half-open boxes in ).
A premeasure on an algebra vanishes at the empty set and is countably additive whenever a disjoint sequence in has its union in ; it is sigma-finite if there is a sequence in with and for every (Premeasures on algebras of sets).
The nonnegative extended sum of a sequence in is , the supremum of its nondecreasing partial sums (Series in the nonnegative extended real line).
is a compact subset of if and only if for every set and every family of open subsets of with there are and indices with , or else (A subset of a metric space is open in the subspace metric exactly when it is the trace of an open set of the ambient space, and it is compact as a metric space in its own right exactly when every family of open subsets of the ambient space covering it, indexed or not, has finitely many members covering it, claim 3; Open cover, subcover, compact metric space, and compact subset of a metric space).
The interior is open, and it is the largest open subset of (Interior, closure, boundary, limit point, isolated point and dense subset of a metric space, The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement).
Every nonempty subset has a least element (The well-ordering principle).
If then ; in particular (For , , and for the series diverges).
Every complete ordered field is Archimedean: for every there is a natural number with (Every complete ordered field is Archimedean).
For sequences of reals, ; if for all then ; and if for all then , with when every (Laws of finite sums and finite products, claims 1, 4 and 6).
Finite sums and finite products of a sequence of reals are defined by the recursions , and , (Finite sums and finite products, by recursion).
For , when and , or and (The extended real line , its order, and the arithmetic that is left undefined).
Proof
is a function on the algebra with values in and , which is the first premeasure clause.
Each cube is a half-open box, hence elementary, with for and ; and , because for the Archimedean property supplies a natural above each of the finitely many reals .
For every , finite additivity gives and monotonicity gives , so every partial sum is at most and therefore , that supremum being the nonnegative extended sum.
Suppose and let be real. Fix an elementary and a compact with and ; for each let be the least natural number with , which exists because the set of such naturals is nonempty and is well ordered, and put , an open set containing . Since , compactness yields finitely many indices covering , hence a natural with , so that ; monotonicity, finite subadditivity and the geometric series then give , whence ; as was an arbitrary positive real and is finite, .
Suppose instead and put ; if then holds, and if a contradiction follows. Fix a disjoint box presentation ; some has infinite volume, hence is nonempty with or for some . Take and put when and otherwise, so that and , where for and . For a real exceeding every and every , the box with parameter pairs for and , respectively , in coordinate according as or , satisfies and has volume at least . On the other hand is elementary of finite volume and is the disjoint union of the elementary sets , so step 1.4 and monotonicity give ; taking above by the Archimedean property contradicts this.
Steps 1.3, 1.4 and 2.1 give in every case, which with steps 1.1 and 1.2 makes a sigma-finite premeasure on .
Lebesgue outer measure on
Definition
Fix . Lebesgue outer measure on is the outer set function induced by the premeasure of Elementary volume is a sigma-finite premeasure on the algebra of elementary sets on the algebra of elementary sets (Elementary sets: the finite unions of half-open boxes in ), in the sense of The outer set function induced by a premeasure:
for , the series being the nonnegative extended sum of Series in the nonnegative extended real line. The family of covering costs is nonempty, because and the sequence covers every , so the infimum is a well-determined element of . On the real line the subscript is dropped and .
The values are defined for every subset of , with no measurability hypothesis. That the resulting set function is an outer measure, and that it agrees with on , are proved in Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume; until then the name outer measure is not claimed, exactly as The outer set function induced by a premeasure stipulates.
Remarks
-
Why the covers are by elementary sets and not by boxes. Both give the same value, since an elementary set is a finite union of boxes and a countable family of finite lists reindexes to a countable family of boxes; taking elementary sets is what makes the definition an instance of the published construction, so that the Carathéodory theory applies with nothing reproved. The comparison with covers by closed, open and cubic boxes is Countable covers by closed boxes, by open boxes and by closed cubes all compute Lebesgue outer measure.
-
The definition itself spends no choice principle; it is an infimum of a nonempty subset of . Countable choice enters only when the infimum is shown to be countably subadditive, and that is recorded where it happens.
Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume
Statement
Let . Assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Then Lebesgue outer measure (Lebesgue outer measure on ) is an outer measure on (Outer measures): it vanishes at , is monotone, and is countably subadditive.
The agreement clause is a theorem of ZF and needs no choice principle: for every elementary set (Elementary sets: the finite unions of half-open boxes in ), where is elementary volume. In particular for every half-open box , and .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, the premeasure on the algebra , and its induced outer set function .
is the outer set function induced by the premeasure on the algebra of elementary sets (Lebesgue outer measure on ).
Elementary volume is a sigma-finite premeasure on (Elementary volume is a sigma-finite premeasure on the algebra of elementary sets).
Assume the Axiom of Countable Choice. The outer set function induced by a premeasure is an outer measure (Assuming countable choice, the outer set function induced by a premeasure is an outer measure).
For every , the outer measure induced by a premeasure satisfies (The induced outer measure agrees with the premeasure on the source algebra).
An outer measure on a set is a function that vanishes at the empty set, is monotone, and is countably subadditive (Outer measures).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
Elementary volume is a premeasure on the algebra of subsets of , and is by definition the outer set function it induces, so both [F1] and [F2] apply to this pair.
Under the Axiom of Countable Choice, an induced outer set function is an outer measure, which is the first assertion.
The identity on the source algebra is [F2], whose statement carries no choice hypothesis, so the agreement clause holds in ZF alone; applied to a half-open box , which is elementary, it gives , and applied to it gives .
Steps 1.1, 1.2 and 1.3 together are the Statement.
Lebesgue measurable sets, the family , and the restricted set function
Definition
Fix and let be the Lebesgue outer set function on (Lebesgue outer measure on ). A set is Lebesgue measurable when
This formula makes sense before any outer-measure theorem is invoked. Under countable choice, Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume makes an outer measure, and the displayed condition is then exactly Carathéodory measurability in the sense of Carathéodory measurable sets.
The family of Lebesgue measurable sets is written , and Lebesgue measure is the restriction
On the real line the subscript is dropped and .
The quantifier over every test set is part of the condition, and no
hypothesis on is imposed before it is tested. That
is a sigma-algebra and that is a complete
measure on it are not part of this definition: they are proved, under the Axiom
of Countable Choice, in Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume ↗, which is
recorded in this item's justified_by because it is a statement about the
objects introduced here. Until that theorem the symbols
and name a family of sets and a restricted set function, nothing more.
Remarks
-
Why the Carathéodory criterion rather than the inner-outer criterion. Lebesgue's original definition compares the outer measure of with that of its complement inside a large box, and it is available only for bounded ; the criterion above is stated for every subset at once and is what makes the published Carathéodory machinery apply verbatim. The two agree, and the equivalence with the approximation criteria is Assuming countable choice, four equivalent descriptions of a Lebesgue measurable subset of .
-
The definition is relative to and to nothing else. Changing the outer measure changes the family; the family attached to a general outer measure is written in Carathéodory measurable sets, and is the name reserved for the instance .
Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Then:
- is a sigma-algebra on and is a measure on it (Measures on sigma-algebras);
- the measure space is complete (Complete measure spaces), and every with is Lebesgue measurable with ;
- every elementary set is Lebesgue measurable and for every ; in particular for every half-open box , and .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, Lebesgue outer measure , and the family of sets Carathéodory measurable for it.
A set is Lebesgue measurable when it is Carathéodory measurable for , the family of these is , and (Lebesgue measurable sets, the family , and the restricted set function ).
Assuming countable choice, is an outer measure on , and for every elementary set (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume).
is the outer set function induced by the premeasure on the algebra of elementary sets (Lebesgue outer measure on ).
Elementary volume is a sigma-finite premeasure on (Elementary volume is a sigma-finite premeasure on the algebra of elementary sets); its value on a half-open box is the box volume (The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition, Half-open boxes in and their volume), and is elementary with infinite elementary volume (Elementary sets: the finite unions of half-open boxes in ).
For an outer measure on , the Carathéodory measurable subsets form a sigma-algebra, and the restriction of the outer measure to it is a complete measure (Carathéodory's theorem: measurable sets form a sigma-algebra carrying a complete measure).
Assume the Axiom of Countable Choice. Every member of the source algebra is Carathéodory measurable for the induced outer measure (Assuming countable choice, every source-algebra set is measurable for the induced outer measure).
Assume the Axiom of Countable Choice. If is a premeasure on an algebra of subsets of and is its induced outer set function, then and (Assuming countable choice, a premeasure extends through its induced outer measure).
Every set of outer measure zero, and every subset of it, is Carathéodory measurable and has outer measure zero (Every outer-null set is Carathéodory measurable).
A measure space is complete if every subset of every measurable -null set is measurable (Complete measure spaces); and a measure on is a function with that is countably additive on pairwise disjoint sequences (Measures on sigma-algebras).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
Under countable choice is an outer measure on , so [F1] applies to it: its Carathéodory measurable sets, which are by definition the members of , form a sigma-algebra, and the restriction of to it is a complete measure.
Since is the outer set function induced by the premeasure on , the extension theorem and the source-algebra lemma give and there; a half-open box and and are elementary, so , and .
A set with is Carathéodory measurable for , hence Lebesgue measurable, and its measure is its outer measure, namely .
Claim 1 and the completeness half of claim 2 are step 1.1, the null-set half of claim 2 is step 1.3, and claim 3 is step 1.2.
Dyadic cubes of generation in
Definition
Fix . For and a function (The integers as equivalence classes of pairs of naturals), whose values are read inside along the canonical embedding, the dyadic cube of generation and index is the half-open box (Half-open boxes in and their volume)
that is , the powers being the integer powers of Integer powers . A dyadic cube is a set of this form for some and ; its generation is and its side length is .
Every dyadic cube is nonempty, since for every , so by Half-open boxes in and their volume its parameters are determined by the set; the generation and the index are therefore determined by the cube as well. At the cubes are the translates of the unit cube by integer vectors, and .
Remarks
-
The half-open convention is what makes the family a tiling. The cubes of a fixed generation are pairwise disjoint and cover (For each generation, the dyadic cubes of that generation are pairwise disjoint and cover ), which would fail for closed cubes, whose faces overlap, and for open cubes, which miss the grid points.
-
Generations are natural numbers here, so the side lengths are at most and there is a coarsest generation. Nothing below needs cubes larger than the unit cube, and bounding the generations below is what makes the maximal-cube selection of Every open subset of is the union of a countable pairwise disjoint family of dyadic cubes work.
For each generation, the dyadic cubes of that generation are pairwise disjoint and cover
Statement
Let and let . Every lies in exactly one dyadic cube of generation (Dyadic cubes of generation in ); that is, the generation- dyadic cubes are pairwise disjoint and their union is . Each of them has volume (Half-open boxes in and their volume).
Facts & Assumptions
Given: A natural number , a natural number , and the dyadic cubes of generation .
For a nonempty box when every and every is real (Half-open boxes in and their volume).
For every real there is exactly one integer with (Integer part: for every real there is exactly one integer with ).
For and , and (Laws of integer exponents, claims 1 and 3; Integer powers ).
, and finite products are defined by the recursion , (Laws of finite sums and finite products, claim 6; Finite sums and finite products, by recursion).
Proof
For a real there is exactly one integer with : applying [F1] to gives the unique integer with , and then satisfies , while any integer with yields , so by the uniqueness in [F1] and .
Since , the condition is equivalent to , the powers satisfying .
The volume of is , the last two equalities by the recursion for finite products and the power laws.
Given , step 1.1 applied in each coordinate to the real produces exactly one integer with , so by step 1.2 the function so determined is the unique index of a generation- dyadic cube containing ; hence the generation- cubes cover and no two of them share a point.
Steps 2.1 and 1.3 are the Statement.
Two dyadic cubes are either disjoint or one contains the other
Statement
Let and let and be dyadic cubes in (Dyadic cubes of generation in ) of generations and with . If then . Consequently any two dyadic cubes are either disjoint or one of them contains the other, and two dyadic cubes of the same generation are either equal or disjoint.
Facts & Assumptions
Given: A natural number and dyadic cubes and with .
, and every dyadic cube is nonempty (Dyadic cubes of generation in ).
For and , (Laws of integer exponents, claim 3; Integer powers ).
The order relation on is a total order compatible with addition: implies (The integers form a totally ordered ring).
The canonical embedding of into is injective and preserves addition, multiplication and order, and its image is exactly the set of nonnegative integers, so every in is the image of a unique natural number (The naturals embed in the integers, The integers as equivalence classes of pairs of naturals).
For all : if and only if (Discreteness: is the immediate successor).
Proof
Put and , an integer; then and , so in coordinate the cube is cut out by and the cube by .
For integers one has , since is the image of a unique natural number, that natural is not , hence it is at least and ; consequently, for integers and , if the real conditions and hold for some real , then and , because would give and , contradicting , while would give and , contradicting .
If then step 1.2, applied in each coordinate with , , and , gives and , so the parameter interval of in coordinate is contained in that of , and hence .
For arbitrary dyadic cubes, relabel so that the generation of the first is the smaller, and step 2.1 gives the dichotomy; when the generations are equal, and the symmetric conclusion both hold, so .
Every open subset of is the union of a countable pairwise disjoint family of dyadic cubes
Statement
Let and let be open in the metric topology of (The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement, as the set of functions , and , , are metrics on it). Then there is an at most countable family of pairwise disjoint dyadic cubes (Dyadic cubes of generation in , Finite, countably infinite, countable, uncountable) with
For the family is empty. No choice principle is used: the cube attached to a point is the one of least generation that fits inside , and least is a definition.
Facts & Assumptions
Given: A natural number and an open subset .
Every lies in exactly one dyadic cube of generation (For each generation, the dyadic cubes of that generation are pairwise disjoint and cover ).
If and are dyadic cubes of generations and , then (Two dyadic cubes are either disjoint or one contains the other).
, and every dyadic cube is nonempty (Dyadic cubes of generation in , Integer powers ).
A nonempty box determines its parameter pair, since determines and for every (Half-open boxes in and their volume).
A subset is open in if for every there is a real with , where (The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement, Open ball, closed ball and sphere in a metric space).
and are metrics on for ( as the set of functions , and , , are metrics on it).
For every , and , being the canonical natural of ; and , (The finite and reverse triangle inequalities for a norm; and for every norm on satisfies and is Lipschitz, hence continuous, for , claim 3; Each is a norm on , and the induced metrics are exactly , and of the published metric-spaces page, claim 3; The -norms for rational , and ).
If then is null, that is (For the sequence is null, and for the sequence diverges to , claim 1; Limits and Cauchy sequences of reals).
Every nonempty subset has a least element (The well-ordering principle).
The set is countable and dense in ( is a countable dense subset of , and rational open boxes form a countable basis).
If and are at most countable then so is (A product of two at most countable sets is at most countable); and if is at most countable and then is at most countable (Every subset of an at most countable set is at most countable, Finite, countably infinite, countable, uncountable).
Proof
The set of all dyadic cubes is at most countable: each of its members is nonempty and so determines its parameter pair, whose entries and are rational, so the assignment of a cube to that pair is an injection of into , a countable set, and is therefore equinumerous with an at most countable subset of it.
For every there is a natural number such that the generation- dyadic cube containing is a subset of : openness supplies a real with ; since is null there is with ; and every in the generation- cube containing has in each coordinate, because and lie in one parameter interval of length , so and .
For let be the least natural number provided by step 1.2 and let be the generation- dyadic cube containing , which is unique; then , and is maximal among the dyadic cubes contained in , for if with of generation , then contains and is therefore the generation- cube containing , so by minimality, and then with gives and hence .
Put : its members are dyadic cubes contained in and every lies in one of them, so ; two members meeting each other are nested by [L2], and each being maximal in they are equal, so the members are pairwise disjoint; and is at most countable by step 1.1.
The sigma-algebra generated by the half-open boxes of is the Borel sigma-algebra
Statement
Let , let be the family of half-open boxes in (Half-open boxes in and their volume) and let be the family of elementary sets (Elementary sets: the finite unions of half-open boxes in ). With carrying its product topology, which is the metric topology of the Euclidean metric (A subset of with the product topology is compact exactly when it is closed and bounded, the product topology being the Euclidean metric topology),
(The sigma-algebra generated by a family of sets, The Borel sigma-algebra of a topological space).
Facts & Assumptions
Given: A natural number , the topology of , the family of half-open boxes and the family of elementary sets.
, with parameters in (Half-open boxes in and their volume).
A subset is an elementary set when there are a natural number and a list of half-open boxes with ; at every half-open box is elementary (Elementary sets: the finite unions of half-open boxes in ).
Every open is the union of an at most countable family of pairwise disjoint dyadic cubes (Every open subset of is the union of a countable pairwise disjoint family of dyadic cubes), and a dyadic cube is the half-open box (Dyadic cubes of generation in ).
Each of the following families generates : all open sets; and all rational half-open boxes with rational endpoints (For n at least one, open sets, closed sets, compact sets, open balls, boxes, rational open boxes, and rational half-open boxes generate the Borel sigma-algebra on R^n).
The Borel sigma-algebra of is the sigma-algebra generated by its open sets, (The Borel sigma-algebra of a topological space), and is the unique smallest sigma-algebra on containing (The sigma-algebra generated by a family of sets, Nonempty intersections of sigma-algebras are sigma-algebras, so the generated sigma-algebra exists and is minimal).
A sigma-algebra is closed under countable unions and under countable intersections (Sigma-algebras, Sigma-algebras are closed under countable intersections, differences, symmetric differences, and set limits).
The product topology on is the metric topology of (A subset of with the product topology is compact exactly when it is closed and bounded, the product topology being the Euclidean metric topology, claim 1; The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement).
A subset is open in if for every there is a real with , and a finite intersection of open sets is open (The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement, Open ball, closed ball and sphere in a metric space, Arbitrary unions and finite intersections of open sets are open, open balls are open and closed balls are closed, claim 3).
For every , , and , , where (The finite and reverse triangle inequalities for a norm; and for every norm on satisfies and is Lipschitz, hence continuous, for , claim 3; Each is a norm on , and the induced metrics are exactly , and of the published metric-spaces page, claim 3; The -norms for rational , and ; as the set of functions , and , , are metrics on it).
For every real there is a natural number with (For every in a complete ordered field there is a natural with ).
For , when and , and when and (The extended real line , its order, and the arithmetic that is left undefined).
An at most countable family may always be presented as a sequence (Finite, countably infinite, countable, uncountable).
Proof
For parameters the set is open: given , the finitely many quantities with real and with real are strictly positive, so their minimum is a positive real, or if there are none, and forces in every coordinate, hence throughout.
Every half-open box is a countable intersection of sets of the form , namely : a point of satisfies when is real and when , while a point of every satisfies and, for real , cannot have , since some is below ; a coordinate with makes both sides empty.
In the other direction every open lies in : it is the union of an at most countable family of dyadic cubes, that family may be presented as a sequence, and each dyadic cube is a half-open box.
The published generator theorem gives that same inclusion by a second route, since the rational half-open boxes with rational are half-open boxes in the sense of [L1] and already generate .
Every half-open box is therefore a Borel set, being a countable intersection of open sets, so and, consisting of finite unions of half-open boxes, also .
By steps 1.3 and 2.1 the families and lie in each other's generated sigma-algebras, so ; and gives by the same criterion.
Remarks
- The convention is the published one. For n at least one, open sets, closed sets, compact sets, open balls, boxes, rational open boxes, and rational half-open boxes generate the Borel sigma-algebra on R^n states its half-open generators as , and Half-open boxes in and their volume uses the same face. Step 1.4 is exactly the point at which a second convention would have shown up as a mismatch rather than as a silent redefinition.
Assuming countable choice, every Borel subset of is Lebesgue measurable
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Then
every Borel subset of (The Borel sigma-algebra of a topological space) is Lebesgue measurable (Lebesgue measurable sets, the family , and the restricted set function ). In particular every open set, every closed set and every countable intersection of open sets is Lebesgue measurable.
Facts & Assumptions
Given: A natural number and the Axiom of Countable Choice.
Assuming countable choice, is a sigma-algebra on and every elementary set is Lebesgue measurable (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
, where is the family of half-open boxes and the family of elementary sets (The sigma-algebra generated by the half-open boxes of is the Borel sigma-algebra).
At every half-open box is elementary (Elementary sets: the finite unions of half-open boxes in ).
is the unique smallest sigma-algebra on containing (Nonempty intersections of sigma-algebras are sigma-algebras, so the generated sigma-algebra exists and is minimal, The sigma-algebra generated by a family of sets).
The Borel sigma-algebra of is the sigma-algebra generated by its open sets (The Borel sigma-algebra of a topological space); a sigma-algebra on is an algebra of subsets closed under countable unions (Sigma-algebras).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
Under countable choice is a sigma-algebra on containing every elementary set, hence containing the family of half-open boxes.
Since is the smallest sigma-algebra containing , step 1.1 gives , and ; open sets, closed sets and countable intersections of open sets are Borel.
A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included
Statement
Let , assume the Axiom of Countable Choice (The Axiom of Countable Choice ()), and let be reals for . Write
(Axis-parallel rectangles in and their volume). Then is open and is closed, so both are Borel and Lebesgue measurable, and every set with is Lebesgue measurable with
In particular this covers the four one-dimensional face conventions in each coordinate — the open box, the closed box , the half-open box of Half-open boxes in and their volume, and every mixture of them, in any combination of coordinates — and it gives measure to all of them whenever for some . For a half-open box with infinite parameters the value is already (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, reals for , and the sets , displayed in the Statement.
Assuming countable choice, is a sigma-algebra, is a complete measure on it, every set of Lebesgue outer measure zero is Lebesgue measurable of measure zero, and for every half-open box (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
Assuming countable choice, is an outer measure on (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume), so it is monotone and countably subadditive (Outer measures, Lebesgue outer measure on ).
For a nonempty box when every and every is real, and a box is nonempty exactly when for every (Half-open boxes in and their volume).
A measure on is a function with that is countably additive on pairwise disjoint sequences (Measures on sigma-algebras).
For every real there is a natural number with (For every in a complete ordered field there is a natural with ).
; if for all then , with when every ; and finite products are defined by the recursion , (Laws of finite sums and finite products, claim 6; Finite sums and finite products, by recursion).
A subset is open in if for every there is a real with ; a subset is closed if its complement is open (The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement).
The forms and are half-open, and an interval is open when both of its written endpoints are excluded, closed when both are included (Intervals of : the nine order-convex forms, nondegeneracy, and length).
Proof
is open and is closed in , by the same coordinatewise estimate in each case, so both are Borel and hence Lebesgue measurable.
A closed rectangle with a degenerate side is Lebesgue null: let be reals with for some , and let be a positive real; the half-open box with parameter pairs for and in coordinate is nonempty, contains , and has volume where with and otherwise, so monotonicity of the outer measure gives for every positive real and hence .
The difference is contained in the union of the closed rectangles obtained from by replacing the -th side by the degenerate side or by , each of which is Lebesgue null by step 1.2, so countable subadditivity of the outer measure, applied to that finite list padded with empty sets, gives ; every subset of is therefore Lebesgue measurable of measure .
Suppose instead for some . Then , the rectangle is Lebesgue null by step 1.2, every between them is a subset of it and so is measurable of measure , and the product has the factor and is therefore as well.
Suppose first that for every . Then is a nonempty half-open box with and . For with , both and are contained in , hence are measurable of measure by step 2.1, so is measurable, and additivity on the two disjoint decompositions and gives .
Steps 3.1 and 2.2 exhaust the two cases and give the displayed value in each, and step 1.1 supplies the Borel and measurability clauses for and .
Lebesgue measure is sigma-finite, and every metrically bounded subset of has finite outer measure
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Then:
- is sigma-finite (Finite, sigma-finite, and semifinite measures): the cubes are Lebesgue measurable with , they increase with , and their union over is .
- Every bounded subset (Bounded subset, diameter, distance from a point to a set, and distance between two sets in a metric space) has ; a bounded Lebesgue measurable set therefore has finite measure, and every compact subset of is Lebesgue measurable of finite measure.
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and Lebesgue measure on .
Assuming countable choice, is a sigma-algebra, is a complete measure on it, and for every half-open box (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Every set with is Lebesgue measurable with (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included).
Assuming countable choice, is an outer measure on , hence monotone (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume, Outer measures).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
For a nonempty box when every and every is real, and (Half-open boxes in and their volume, Integer powers ).
is sigma-finite if there is a sequence in such that and for every (Finite, sigma-finite, and semifinite measures).
is bounded if or there are and a real with , where (Bounded subset, diameter, distance from a point to a set, and distance between two sets in a metric space, Open ball, closed ball and sphere in a metric space).
For every , , and , , where (The finite and reverse triangle inequalities for a norm; and for every norm on satisfies and is Lipschitz, hence continuous, for , claim 3; Each is a norm on , and the induced metrics are exactly , and of the published metric-spaces page, claim 3; The -norms for rational , and ; as the set of functions , and , , are metrics on it).
A subset is compact if and only if is closed in and bounded (Heine-Borel in : with the Euclidean metric a subset of is compact if and only if it is closed and bounded, and the proof by bisection uses no choice principle; the same holds on the real line, claim 2; The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement); and a compact subset of a metric space is closed and bounded (A compact subset of a metric space is closed and bounded).
If and , then (Measures are monotone).
Every complete ordered field is Archimedean: for every there is a natural number with (Every complete ordered field is Archimedean).
Proof
Each cube is a half-open box, hence Lebesgue measurable with , a real number; the cubes increase with ; and every lies in one of them, because the Archimedean property gives a natural above each of the finitely many reals , so their union is and is sigma-finite.
Let be bounded and nonempty, say with a positive real; every satisfies in each coordinate, so is contained in the half-open box with parameter pairs , whose volume is ; monotonicity of the outer measure therefore gives , and the empty set has outer measure .
A bounded Lebesgue measurable set has by step 1.2; and a compact is closed, hence Borel and Lebesgue measurable, and bounded, hence of finite measure.
Step 1.1 is claim 1 and steps 1.2 and 2.1 are claim 2.
Every at most countable subset of is Lebesgue null; in particular
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Every at most countable subset (Finite, countably infinite, countable, uncountable) is Lebesgue measurable with
so is a -null set (Measure-null sets and almost-everywhere statements relative to a measure). In particular every singleton is null, and on the real line the set of rational reals (The rationals embed densely in the reals) satisfies .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and an at most countable set .
Every set with is Lebesgue measurable with , and this gives measure to all of them whenever for some (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included).
Assuming countable choice, is a sigma-algebra and is a complete measure on it (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
is at most countable if it is finite or countably infinite (Finite, countably infinite, countable, uncountable); a nonempty is at most countable if and only if there is a surjection (A nonempty set is at most countable iff it is a surjective image of ).
For a measure and measurable , (Finite and countable subadditivity of measures).
: the rationals are countably infinite ( is countably infinite), and denotes the image of in under the canonical order-preserving field embedding (The rationals embed densely in the reals).
A measurable set is -null if (Measure-null sets and almost-everywhere statements relative to a measure); a sigma-algebra is closed under countable unions (Sigma-algebras).
Proof
A singleton is the closed rectangle , whose sides all satisfy , so it is Lebesgue measurable with .
The empty set is Lebesgue measurable with measure .
Let be nonempty and at most countable and fix a surjection ; then is a countable union of measurable sets, hence measurable, and countable subadditivity gives .
Steps 1.2 and 2.1 cover both cases, and is a countably infinite subset of , so .
A box with a degenerate side is Lebesgue null, and so is every coordinate hyperplane in
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Then:
- Degenerate boxes. If are reals with for some , then every set between the open box and the closed rectangle is Lebesgue measurable with .
- Coordinate hyperplanes. For and a real , the set is Lebesgue measurable with .
At the hyperplane is the singleton .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, an index and a real .
Every set with is Lebesgue measurable with , and it gives measure to all of them whenever for some (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included).
Assuming countable choice, is a sigma-algebra and is a complete measure on it (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
For a measure and measurable , (Finite and countable subadditivity of measures).
A measurable set is -null if (Measure-null sets and almost-everywhere statements relative to a measure); a sigma-algebra is closed under countable unions (Sigma-algebras).
Every complete ordered field is Archimedean: for every there is a natural number with (Every complete ordered field is Archimedean).
Proof
Claim 1 is the degenerate case of the box theorem, whose value carries the factor .
For a natural number put . This is always the closed rectangle with sides for and the degenerate side in coordinate .
Each is Lebesgue measurable of measure by claim 1.
The union is , since a point of the hyperplane has finitely many coordinates and the Archimedean property supplies a natural above each and above .
Therefore is a countable union of measurable sets, hence measurable, and countable subadditivity gives ; at the set is .
Assuming countable choice, the Lebesgue outer measure of an arbitrary subset of is the infimum of the measures of the open sets containing it
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). For every subset , measurable or not,
the infimum being taken in over a family that is nonempty because is open.
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a subset .
for every and (Lebesgue outer measure on , Elementary sets: the finite unions of half-open boxes in ).
Assuming countable choice, is an outer measure on , hence monotone and countably subadditive, and for every elementary set (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume, Outer measures).
Assuming countable choice, is a sigma-algebra and is a complete measure on it (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, every Borel subset of is Lebesgue measurable; in particular every open set is (Assuming countable choice, every Borel subset of is Lebesgue measurable).
is an elementary set determined by and alone, it contains , and every point of is an interior point of ; and for every real there is with (Every elementary set is squeezed in volume between a compact subset and an elementary set whose interior contains it, claims 1 and 2).
Elementary volume is finitely additive on pairwise disjoint elementary sets, monotone, and finitely subadditive (Elementary volume is finitely additive, monotone and finitely subadditive on the elementary algebra, The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition).
The interior is open, and an arbitrary union of open sets is open (Interior, closure, boundary, limit point, isolated point and dense subset of a metric space, The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement, Arbitrary unions and finite intersections of open sets are open, open balls are open and closed balls are closed, claim 2).
The nonnegative extended sum of a sequence in is , the supremum of its nondecreasing partial sums (Series in the nonnegative extended real line).
Every nonempty subset has a least element (The well-ordering principle).
If then ; in particular (For , , and for the series diverges).
For sequences of reals, , and if whenever then (Laws of finite sums and finite products, claims 1 and 4; Finite sums and finite products, by recursion).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
Every open is Lebesgue measurable with , so monotonicity of the outer measure gives for every open , and therefore is a lower bound of the family whose infimum is displayed; that family is nonempty since is open.
Suppose and let be a positive real; by the definition of as an infimum there is a sequence of elementary sets with and .
For each let be the least natural number with , which exists because that set of naturals is nonempty and is well ordered, and put ; each is open and contains , so is open and contains .
Countable subadditivity, monotonicity and the agreement of with on elementary sets give ; every partial sum of the last series is at most , so the series itself, being the supremum of its partial sums, is at most .
So when the infimum is at most for every positive real and hence at most ; when the infimum is at most for the same reason of triviality; with step 1.1 the infimum equals in both cases.
Every subset of has a measurable hull of the same outer measure
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Every has a set ( and subsets of a topological space, agreeing with the real-line notion) with
Such a is Borel, hence Lebesgue measurable, so it is a measurable hull of and is a regular outer measure (Measurable hulls and regular outer measures). The regularity also follows from Assuming countable choice, a premeasure-induced outer measure is regular with generated measurable hulls, which supplies a measurable hull inside ; the point added here is that the hull may be taken of the special form .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a subset .
Assuming countable choice, open and for every subset (Assuming countable choice, the Lebesgue outer measure of an arbitrary subset of is the infimum of the measures of the open sets containing it).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
Assuming countable choice, is a sigma-algebra and is a complete measure on it, and is the restriction of (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, is an outer measure on , hence monotone (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume, Outer measures, Lebesgue outer measure on ).
is a set of when there is a sequence of open subsets of with ( and subsets of a topological space, agreeing with the real-line notion).
A measurable hull of is a Carathéodory measurable set with ; the outer measure is regular when every subset has a measurable hull (Measurable hulls and regular outer measures).
Assume the Axiom of Countable Choice. An outer measure induced by a premeasure is regular, and every set has a measurable hull in (Assuming countable choice, a premeasure-induced outer measure is regular with generated measurable hulls).
The product topology on is the metric topology of (A subset of with the product topology is compact exactly when it is closed and bounded, the product topology being the Euclidean metric topology, claim 1; The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement).
For every real there is a natural number with (For every in a complete ordered field there is a natural with ).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
If , take , which is open and hence a by the constant sequence, contains , and has by monotonicity.
If , then for each the family of open sets with is nonempty, because the infimum in [L1] is not a lower bound of anything larger; countable choice selects one such for every .
Put , a set containing ; monotonicity gives for every , so .
In both cases is a countable intersection of open sets, hence Borel and Lebesgue measurable, so is a measurable hull of and is regular; the same regularity is delivered by the published theorem on premeasure-induced outer measures, with the hull taken in instead.
For a Lebesgue measurable set and every positive there is an open superset whose difference from it has outer measure below
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). For every Lebesgue measurable and every real there is an open set with
No finiteness hypothesis on is imposed; the excess is measured by the outer measure of the difference, not by a difference of measures, which is what lets the statement hold when .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, a Lebesgue measurable set , and a real .
Assuming countable choice, open and for every subset (Assuming countable choice, the Lebesgue outer measure of an arbitrary subset of is the infimum of the measures of the open sets containing it).
Assuming countable choice, is a sigma-algebra, is a complete measure on it, and is the restriction of (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, is an outer measure on , hence monotone and countably subadditive (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume, Outer measures).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
Every bounded subset has (Lebesgue measure is sigma-finite, and every metrically bounded subset of has finite outer measure).
Every set with is Lebesgue measurable with (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included), and (Half-open boxes in and their volume).
Let be a measure and let be measurable with ; then (Measure of a set difference when the smaller set has finite measure).
If then ; in particular (For , , and for the series diverges).
The nonnegative extended sum of a sequence in is , the supremum of its nondecreasing partial sums (Series in the nonnegative extended real line).
For sequences of reals, , and if whenever then (Laws of finite sums and finite products, claims 2 and 4; Finite sums and finite products, by recursion).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
Suppose first and let be a positive real. Outer regularity supplies an open with ; both and are measurable, so the difference formula gives and hence .
For put ; each is Lebesgue measurable, being an intersection and difference of measurable sets, is bounded and therefore of finite measure, and because the cubes increase to .
By step 1.1 applied to each with , the family of open with is nonempty for every , so countable choice selects such a for every ; the union is open and contains .
Since , one has , so countable subadditivity gives , whose partial sums are , so the sum is at most .
A subset of with open supersets of arbitrarily small excess is Lebesgue measurable
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Let be such that for every real there is an open with . Then there are a set ( and subsets of a topological space, agreeing with the real-line notion) and a set with
and is Lebesgue measurable (Lebesgue measurable sets, the family , and the restricted set function ).
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a set admitting open supersets of arbitrarily small outer excess.
Assuming countable choice, is a sigma-algebra, is a complete measure on it, and every with is Lebesgue measurable with (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
Assuming countable choice, is an outer measure on , hence monotone (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume, Outer measures).
is a set of when there is a sequence of open subsets of with ( and subsets of a topological space, agreeing with the real-line notion).
The product topology on is the metric topology of (A subset of with the product topology is compact exactly when it is closed and bounded, the product topology being the Euclidean metric topology, claim 1; The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement).
For every real there is a natural number with (For every in a complete ordered field there is a natural with ).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
For every the family of open with is nonempty by hypothesis, since is a positive real, so countable choice selects such a for every .
Put and ; then is a set containing , so , and for every , whence monotonicity gives for every and therefore .
is a countable intersection of open sets, hence Borel and Lebesgue measurable; has outer measure , hence is Lebesgue measurable; and is a difference of measurable sets, hence Lebesgue measurable.
Assuming countable choice, four equivalent descriptions of a Lebesgue measurable subset of
Statement
Let , assume the Axiom of Countable Choice (The Axiom of Countable Choice ()) and let . Then is Lebesgue measurable (Lebesgue measurable sets, the family , and the restricted set function ) if and only if each of the following four conditions holds, and the four are equivalent to one another.
- Open excess. For every real there is an open with .
- minus null. There are a set and a set with and ( and subsets of a topological space, agreeing with the real-line notion).
- Closed deficit. For every real there is a closed with .
- plus null. There are an set and a set with and .
Each condition is stated for sets of infinite measure as well as finite ones, which is why the excess and the deficit are measured by the outer measure of a difference rather than by a difference of measures.
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a subset .
Assuming countable choice, for every Lebesgue measurable and every real there is an open with and (For a Lebesgue measurable set and every positive there is an open superset whose difference from it has outer measure below ).
Assuming countable choice, a set admitting open supersets of arbitrarily small outer excess is with a containing it and , and is Lebesgue measurable (A subset of with open supersets of arbitrarily small excess is Lebesgue measurable).
Assuming countable choice, is a sigma-algebra, is a complete measure on it, and every with is Lebesgue measurable of measure (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
Assuming countable choice, is an outer measure on , hence monotone (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume, Outer measures).
Assuming countable choice, open and (Assuming countable choice, the Lebesgue outer measure of an arbitrary subset of is the infimum of the measures of the open sets containing it).
is a set of when there is a sequence of open subsets of with , and an set of when there is a sequence of closed subsets with ( and subsets of a topological space, agreeing with the real-line notion).
A subset is closed in if its complement is open (The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement), and the product topology on is the metric topology of (A subset of with the product topology is compact exactly when it is closed and bounded, the product topology being the Euclidean metric topology, claim 1).
For every real there is a natural number with (For every in a complete ordered field there is a natural with ).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
Measurability implies condition 1, which is the cited lemma on the open excess of a measurable set.
Condition 1 implies condition 2, which is the first clause of the cited lemma on small open excess.
Condition 2 implies measurability: is a countable intersection of open sets, hence Borel and measurable; has outer measure , hence is measurable; so is measurable.
Condition 4 implies measurability, by the same argument read for unions: is a countable union of closed sets, hence Borel and measurable, is measurable because , and is measurable.
Measurability implies condition 3: the complement is measurable, so for a real step 1.1 supplies an open with ; then is closed, , and , so .
Condition 3 implies condition 4: for each the family of closed with is nonempty, so countable choice selects such an ; then is an set with , and gives for every , hence and .
The implications of steps 1.1, 1.2 and 1.3 close the cycle between measurability and conditions 1 and 2, and those of steps 2.1, 3.1 and 1.4 close the cycle between measurability and conditions 3 and 4; so all five statements are equivalent, and outer regularity is what stands behind the open sets produced in step 1.1.
Assuming countable choice, the Lebesgue measure of a measurable set is the supremum of the measures of its compact subsets
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). For every Lebesgue measurable ,
(Open cover, subcover, compact metric space, and compact subset of a metric space), the supremum being over a nonempty family since is compact.
The choice hypothesis is inherited, not decorative. The proof runs through Assuming countable choice, four equivalent descriptions of a Lebesgue measurable subset of , which is itself stated under countable choice, so the conclusion carries the same hypothesis and says so.
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a Lebesgue measurable set .
Assuming countable choice, is Lebesgue measurable if and only if for every real there is a closed with (Assuming countable choice, four equivalent descriptions of a Lebesgue measurable subset of , condition 3).
Assuming countable choice, is a sigma-algebra and is a complete measure on it, and is the restriction of (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
Every bounded Lebesgue measurable subset of has finite measure, and every compact subset of is Lebesgue measurable of finite measure (Lebesgue measure is sigma-finite, and every metrically bounded subset of has finite outer measure).
Every set with is Lebesgue measurable with (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included), and (Half-open boxes in and their volume).
A subset is compact if and only if is closed in and bounded (Heine-Borel in : with the Euclidean metric a subset of is compact if and only if it is closed and bounded, and the proof by bisection uses no choice principle; the same holds on the real line, claim 2), and a compact subset of a metric space is closed and bounded (A compact subset of a metric space is closed and bounded, Bounded subset, diameter, distance from a point to a set, and distance between two sets in a metric space, The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement).
Closed balls are closed, for every and every (Arbitrary unions and finite intersections of open sets are open, open balls are open and closed balls are closed, claim 4), where (Open ball, closed ball and sphere in a metric space).
Let be an increasing sequence of measurable sets for a measure ; then (Continuity from below for measures).
Let be a measure and let be measurable with ; then (Measure of a set difference when the smaller set has finite measure).
If and , then (Measures are monotone).
Every complete ordered field is Archimedean: for every there is a natural number with (Every complete ordered field is Archimedean).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
Every compact is Lebesgue measurable of finite measure and satisfies by monotonicity, and the empty set is compact, so the displayed family is nonempty and its supremum is at most .
Suppose and let be real. Applying the closed-deficit condition with gives a closed with ; is Borel, hence measurable, of finite measure, and the difference formula gives , so .
The sets for are closed and bounded, hence compact subsets of , they increase with , and their union is because the Archimedean property puts every point of inside some ; continuity from below therefore gives , so some has .
Suppose instead and let be any real. The sets are measurable, bounded and hence of finite measure, they increase with and their union is , so continuity from below gives and some has ; steps 1.2 and 2.1 applied to that set of finite measure produce a compact subset of it, hence of , of measure above .
In both cases every real below is below the measure of some compact subset of , so the supremum is at least , and step 1.1 gives the reverse inequality.
is exactly the completion of the restriction of to the Borel sets
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Write for the restriction of to the Borel sigma-algebra (The Borel sigma-algebra of a topological space). Then is exactly the completion domain of (The completion domain and proposed completed set function of a measure space), and is the completed measure there. Explicitly,
and then .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, Lebesgue measure on , and the restriction of to .
Assume the Axiom of Countable Choice and let be a sigma-finite premeasure; the Carathéodory sigma-algebra of its induced outer measure is exactly the completion of under the extended measure, and the Carathéodory restriction equals the completed measure there (Assuming countable choice, the Carathéodory domain is the completion of the sigma-finite extension).
Elementary volume is a sigma-finite premeasure on the algebra of elementary sets (Elementary volume is a sigma-finite premeasure on the algebra of elementary sets, Elementary sets: the finite unions of half-open boxes in ).
is the outer set function induced by the premeasure on (Lebesgue outer measure on ), and is the family of sets Carathéodory measurable for , with its restriction (Lebesgue measurable sets, the family , and the restricted set function ).
Assuming countable choice, is a sigma-algebra and is a complete measure on it (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume), and every Borel set is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
The completion domain of is for some and with , and the completed set function is (The completion domain and proposed completed set function of a measure space).
Assume the Axiom of Countable Choice; then is a complete measure on extending , and it is the unique complete measure on that extends (Assuming countable choice, every measure space has a unique complete extension to its completion).
The Axiom of Countable Choice says that for every family of nonempty sets indexed by there is a function with domain such that for every (The Axiom of Countable Choice ()).
Proof
Elementary volume is a sigma-finite premeasure on the algebra , and is exactly the outer set function it induces, so the hypotheses of the Carathéodory-domain theorem are met with .
The sigma-algebra generated by is , and the measure that the extension theorem places on it is the restriction of , since every Borel set is Lebesgue measurable and is the restriction of .
The Carathéodory-domain theorem therefore says that , the Carathéodory sigma-algebra of , is the completion domain of on and that agrees there with the completed measure; unwinding the published description of that domain gives the displayed equivalence and the value , which is well posed because the completed measure is a measure extending and is the unique complete one.
Regularity of an outer measure and regularity of a measure with respect to open and compact sets are different conditions, both satisfied here
Assuming the Axiom of Countable Choice, the word regular is carrying two different conditions in this development, and both of them hold for Lebesgue measure. They are not variants of one statement: one is about arbitrary subsets and measurable supersets, the other about measurable sets and topologically distinguished sub- and supersets.
Regularity of an outer measure. Measurable hulls and regular outer measures calls an outer measure regular when every subset of the ambient set has a measurable hull: a Carathéodory measurable with . This mentions no topology at all, and it is a condition that fails for some outer measures. For it holds, with the hull available in the special form : Every subset of has a measurable hull of the same outer measure.
Regularity of a measure with respect to open and compact sets. Here the statements are that is the infimum of over open (Assuming countable choice, the Lebesgue outer measure of an arbitrary subset of is the infimum of the measures of the open sets containing it), and that is the supremum of over compact for measurable (Assuming countable choice, the Lebesgue measure of a measurable set is the supremum of the measures of its compact subsets). Both mention the topology essentially, and the second is restricted to measurable sets, which the first is not.
Why the distinction has to be made rather than left to context. The two conditions have different hypotheses on , different quantifiers, and different witnesses: a measurable hull is a superset with equal outer measure, while outer regularity produces supersets whose measures merely approach the outer measure and are open. The one implies the other only through an argument — here, intersecting a sequence of open supersets, which is exactly the proof of the hull. Nothing below uses the word regular without saying which of the two is meant.
Countable covers by closed boxes, by open boxes and by closed cubes all compute Lebesgue outer measure
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). For with for every write for the closed rectangle and for the open box, both of size (Axis-parallel rectangles in and their volume); a closed cube of side is a set , of size . For put
infima over countable covers of the stated kind, which exist because is covered by the rectangles , by the open boxes and by the cubes . Then
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, a subset , and the three infima displayed in the Statement.
for every and (Lebesgue outer measure on , Elementary sets: the finite unions of half-open boxes in ).
; a box is nonempty exactly when for every ; ; and for a nonempty box with real parameters (Half-open boxes in and their volume).
Assuming countable choice, for every elementary set , and is an outer measure (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume), being the elementary volume of The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition.
Assuming countable choice, open and (Assuming countable choice, the Lebesgue outer measure of an arbitrary subset of is the infimum of the measures of the open sets containing it).
Every open is the union of an at most countable family of pairwise disjoint dyadic cubes (Every open subset of is the union of a countable pairwise disjoint family of dyadic cubes), each of the form (Dyadic cubes of generation in , Integer powers ).
Assuming countable choice, is a measure on the sigma-algebra with for every half-open box (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume), so it is countably additive on pairwise disjoint measurable sequences (Measures on sigma-algebras).
The nonnegative extended sum of a sequence in is , the supremum of its nondecreasing partial sums (Series in the nonnegative extended real line).
Every nonempty subset has a least element (The well-ordering principle).
For every real there is a natural number with (For every in a complete ordered field there is a natural with ).
If then ; in particular (For , , and for the series diverges).
For sequences of reals, ; ; if whenever then ; and (Laws of finite sums and finite products, claims 1, 2, 4 and 6; Finite sums and finite products, by recursion).
An at most countable family may always be presented as a sequence (Finite, countably infinite, countable, uncountable).
Proof
For a natural number , reals and a real with for every , one has : at both products are and both sides are , and the passage from to uses with , so the estimate follows by induction on .
For real one has and , the box being empty and the product zero together when some ; moreover for every real , whose size is , and a closed cube of side is the closed rectangle of size .
, because every closed-cube cover is a closed-rectangle cover with the same terms.
, because an open-box cover gives the elementary cover whose covering cost has exactly the same terms.
: given a closed-rectangle cover and a real , let be the least natural number with , which exists by step 1.1 with a real at least bounding all and by the Archimedean property; the open boxes with cover , and each partial sum of their sizes is at most , so is at most that closed cover's total plus , for every positive real .
: the inequality is trivial when , and otherwise, given a real , outer regularity supplies an open with , the dyadic decomposition writes as a disjoint union of an at most countable family of dyadic cubes, presented as a sequence and padded with copies of if it is finite, countable additivity gives , and each is contained in the closed cube of side and size , a padding term contributing the degenerate cube of side .
The four quantities therefore satisfy , so all four are equal.
A subset of has Lebesgue outer measure zero if and only if it has measure zero in the sense of countable closed-interval covers
Statement
Assume the Axiom of Countable Choice (The Axiom of Countable Choice ()) and let . Then
measure zero being the covering notion of Measure zero (a countable cover by intervals of total length below every ) and content zero (a finite such cover): that is, if and only if for every real there are sequences and of reals with for every such that and converges with sum at most .
Facts & Assumptions
Given: The Axiom of Countable Choice, the case of Lebesgue outer measure, and a subset .
Assuming countable choice, , where is the infimum of over countable covers of by closed rectangles (Countable covers by closed boxes, by open boxes and by closed cubes all compute Lebesgue outer measure, Lebesgue outer measure on ).
has measure zero, equivalently is null, when for every real there are sequences and of reals with for every , such that and converges with sum (Measure zero (a countable cover by intervals of total length below every ) and content zero (a finite such cover), Intervals of : the nine order-convex forms, nondegeneracy, and length).
For a fixed , converges with sum if and only if for every (Measure zero (a countable cover by intervals of total length below every ) and content zero (a finite such cover)).
The nonnegative extended sum of a sequence in is , the supremum of its nondecreasing partial sums (Series in the nonnegative extended real line).
Under the standard identification , the rectangle of is the interval and its volume is its length (Axis-parallel rectangles in and their volume).
Proof
At a closed rectangle is a closed interval with and its volume is the length , so the covers admitted in are exactly the covers admitted in the published definition of measure zero.
For a sequence of nonnegative reals, the nonnegative extended sum is the supremum of the partial sums, so it is at most exactly when every partial sum is, which is exactly the condition that the real series converges with sum at most .
Hence has measure zero in the published sense if and only if for every real some admissible cover has total length at most , which says exactly that the infimum is ; and .
A subset of has Lebesgue outer measure zero if and only if it is null in the sense of countable closed-cube covers
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). For ,
nullity being the covering notion of Measure zero and content zero in by countable and finite cube covers: that is, if and only if for every real the set is covered by a sequence of closed cubes whose nonnegative volume series converges with sum at most .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a subset .
Assuming countable choice, , where is the infimum of over countable covers of by closed cubes (Countable covers by closed boxes, by open boxes and by closed cubes all compute Lebesgue outer measure, Lebesgue outer measure on ).
A closed cube is a rectangle with ; its volume is . A set is null when, for every , it is covered by a sequence of closed cubes whose nonnegative volume series converges with sum at most (Measure zero and content zero in by countable and finite cube covers, Axis-parallel rectangles in and their volume).
The nonnegative extended sum of a sequence in is , the supremum of its nondecreasing partial sums (Series in the nonnegative extended real line).
Proof
The closed cubes admitted in the published definition of nullity are exactly the sets with , with the same size as in , so the two notions quantify over the same covers with the same terms.
For a sequence of nonnegative reals, the nonnegative extended sum is the supremum of the partial sums, so the condition that the volume series converges with sum at most says exactly that this sum, taken in , is at most .
Hence is null in the published sense if and only if for every real some admissible cube cover has total volume at most , which says exactly that the infimum is ; and .
A property holding outside a set of elementary measure zero is exactly a property holding -almost everywhere
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Then:
- A subset of has measure zero in the covering sense of Measure zero (a countable cover by intervals of total length below every ) and content zero (a finite such cover) if and only if it is Lebesgue measurable with -measure ; and a subset of is null in the covering sense of Measure zero and content zero in by countable and finite cube covers if and only if it is Lebesgue measurable with -measure .
- For a property of points of , the exceptional set is null in the covering sense if and only if holds -almost everywhere (Measure-null sets and almost-everywhere statements relative to a measure).
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a property of points of with exceptional set .
Assuming countable choice, if and only if has measure zero in the covering sense (A subset of has Lebesgue outer measure zero if and only if it has measure zero in the sense of countable closed-interval covers, Measure zero (a countable cover by intervals of total length below every ) and content zero (a finite such cover)).
Assuming countable choice, if and only if is null in the covering sense (A subset of has Lebesgue outer measure zero if and only if it is null in the sense of countable closed-cube covers, Measure zero and content zero in by countable and finite cube covers).
Assuming countable choice, is a sigma-algebra, is a complete measure on it and is the restriction of , and every with is Lebesgue measurable of measure (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, is an outer measure on , hence monotone (Assuming countable choice, Lebesgue outer measure is an outer measure that restricts to elementary volume, Outer measures).
A property holds -almost everywhere if its exceptional set is contained in a measurable -null set: there is with such that holds for every (Measure-null sets and almost-everywhere statements relative to a measure).
Proof
A set with Lebesgue outer measure is Lebesgue measurable of measure , and conversely a Lebesgue measurable set of measure has outer measure , since is the restriction of .
Combining step 1.1 with the two agreement theorems gives claim 1 in both dimensions: covering nullity and Lebesgue nullity name the same class of sets.
If is null in the covering sense then , so itself is a measurable null set containing the exceptional set and holds -almost everywhere; conversely if holds -almost everywhere, with measurable and , then monotonicity gives and is null in the covering sense.
A bounded function on a closed bounded interval, or on a closed nondegenerate rectangle, is Riemann integrable exactly when its discontinuity set has Lebesgue measure zero
Statement
Assume the Axiom of Countable Choice (The Axiom of Countable Choice ()).
- Let be reals and let be bounded, with discontinuity set . Then is Riemann integrable on if and only if is Lebesgue measurable with .
- Let and let be a bounded real function on a closed nondegenerate rectangle in , with discontinuity set . Then is Riemann integrable if and only if is Lebesgue measurable with .
The choice ledger of the cited criteria is inherited, not discharged. In Lebesgue's criterion for Riemann integrability: a bounded on is Riemann integrable if and only if its set of discontinuities has measure zero the implication from integrability to nullity of uses countable choice and the converse implication is a theorem of ZF; the translation performed here rests on the construction of , which uses countable choice in both directions, so the statement above carries the hypothesis throughout.
Facts & Assumptions
Given: The Axiom of Countable Choice, a bounded real function on a closed bounded interval or on a closed nondegenerate rectangle, and its discontinuity set .
Assuming countable choice, if and only if has measure zero in the covering sense, and a set of Lebesgue outer measure zero is measurable of measure zero (A subset of has Lebesgue outer measure zero if and only if it has measure zero in the sense of countable closed-interval covers, Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, if and only if is null in the covering sense (A subset of has Lebesgue outer measure zero if and only if it is null in the sense of countable closed-cube covers).
Let be reals, let be bounded and let be its set of discontinuities; then is Riemann integrable on if and only if has measure zero (Lebesgue's criterion for Riemann integrability: a bounded on is Riemann integrable if and only if its set of discontinuities has measure zero, Measure zero (a countable cover by intervals of total length below every ) and content zero (a finite such cover)).
A bounded real function on a closed nondegenerate rectangle in , , is Riemann integrable if and only if its discontinuity set is null (Lebesgue's criterion in : a bounded function on a closed nondegenerate rectangle is Riemann integrable iff its discontinuity set is null, Measure zero and content zero in by countable and finite cube covers).
A measurable set is -null if (Measure-null sets and almost-everywhere statements relative to a measure).
Proof
On the line, " has measure zero" in the covering sense of the cited criterion is equivalent to , and a set of Lebesgue outer measure zero is Lebesgue measurable of measure zero, while conversely for a measurable says .
In , "the discontinuity set is null" in the covering sense of the cited criterion is likewise equivalent to , hence to being Lebesgue measurable with .
Substituting these equivalences into the two published criteria gives claims 1 and 2.
The published refutations separating nullity from nowhere density hold verbatim for Lebesgue measure
Assume the Axiom of Countable Choice. Two notions of smallness for subsets of are now in play: being -null, and being nowhere dense (Nowhere dense, meager (first category), residual, and second category subsets of ). Neither implies the other, and the two published refutations transfer to Lebesgue measure without a new argument, because A subset of has Lebesgue outer measure zero if and only if it has measure zero in the sense of countable closed-interval covers identifies with the covering condition of Measure zero (a countable cover by intervals of total length below every ) and content zero (a finite such cover) that those items are stated in.
Null does not imply nowhere dense. FALSE: every subset of of measure zero is nowhere dense records the false claim and its witness. Read through the agreement theorem, the witness is a -null set whose closure is all of ; the rationals of the line are one, and their nullity is also the case of Every at most countable subset of is Lebesgue null; in particular .
Nowhere dense does not imply null. FALSE: every nowhere dense subset of has measure zero records that false claim, and The Smith-Volterra-Cantor set is compact, perfect and nowhere dense, and does not have measure zero proves of the Smith–Volterra–Cantor set that it is compact, perfect and nowhere dense while no cover of it by intervals has total length below . The equality of the closed-interval cover infimum with Lebesgue outer measure in Countable covers by closed boxes, by open boxes and by closed cubes all compute Lebesgue outer measure therefore gives , so is nowhere dense and not -null. The exact value is computed on the companion page.
Why the transfer needs saying at all. The published items were written before any outer measure existed here, so they are stated as assertions about interval covers and cannot mention . Without the agreement theorem, a reader meeting both vocabularies would have two apparently unrelated notions of "measure zero" on the line; with it there is one notion, and the earlier refutations keep their force in the new vocabulary.
Lebesgue outer measure is at most Jordan outer content, and a bounded Jordan measurable set is Lebesgue measurable with Lebesgue measure equal to its Jordan content
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Write and for the Jordan outer and inner content of a bounded (Jordan inner and outer content and Jordan measurable bounded sets in ). Then:
- for every bounded , Jordan measurable or not;
- if is bounded and Jordan measurable, with Jordan content , then is Lebesgue measurable and
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a bounded set .
Assuming countable choice, , the infimum of over countable covers of by closed rectangles (Countable covers by closed boxes, by open boxes and by closed cubes all compute Lebesgue outer measure).
Assuming countable choice, if and only if is null in the covering sense of closed-cube covers (A subset of has Lebesgue outer measure zero if and only if it is null in the sense of countable closed-cube covers, Measure zero and content zero in by countable and finite cube covers).
Assuming countable choice, is a sigma-algebra, is a complete measure on it and is the restriction of , and every set of Lebesgue outer measure zero is Lebesgue measurable of measure zero (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable).
Every set with is Lebesgue measurable with , and it gives measure to all of them whenever for some (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included).
A box with a degenerate side is Lebesgue measurable of measure (A box with a degenerate side is Lebesgue null, and so is every coordinate hyperplane in ).
For bounded its Jordan outer content is the infimum of over finite axis-parallel rectangle covers of , its Jordan inner content is the supremum of the same sums over finite families of rectangles contained in whose interiors are pairwise disjoint, and the set is Jordan measurable when the contents agree (Jordan inner and outer content and Jordan measurable bounded sets in , Bounded subset, diameter, distance from a point to a set, and distance between two sets in a metric space).
A metric-bounded set is Jordan measurable if and only if its boundary is null, equivalently has content zero (A bounded set in is Jordan measurable iff its boundary is null, equivalently of content zero).
The boundary of is , the interior is open and contained in , and (Interior, closure, boundary, limit point, isolated point and dense subset of a metric space, The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement).
A measure is countably additive on pairwise disjoint measurable sequences (Measures on sigma-algebras), it is monotone (Measures are monotone), and it is finitely and countably subadditive (Finite and countable subadditivity of measures).
The nonnegative extended sum of a sequence in is the supremum of its nondecreasing partial sums (Series in the nonnegative extended real line), and for real sequences whenever throughout (Laws of finite sums and finite products, claim 4; Finite sums and finite products, by recursion).
Proof
A finite cover of by axis-parallel rectangles becomes a countable cover by closed rectangles once it is padded with copies of the degenerate rectangle , whose volume is , and the padded series has the same value, so ; taking the infimum over all finite rectangle covers gives claim 1.
Two closed rectangles and with disjoint interiors meet in a set with empty interior, and that intersection is either empty or the closed rectangle whose -th side is ; a nonempty closed rectangle with empty interior has for some , so it is Lebesgue measurable of measure .
If is bounded and Jordan measurable, then is null in the covering sense, hence and is Lebesgue measurable of measure ; is open, hence Borel and Lebesgue measurable; and because , so is Lebesgue measurable.
Let be closed rectangles contained in with pairwise disjoint interiors and put ; each is a finite union of sets of measure by step 1.2, hence of measure , so additivity on the decomposition gives , and the are pairwise disjoint measurable sets with union , so .
For bounded and Jordan measurable, step 1.3 makes measurable, step 1.1 gives , and step 2.1 with monotonicity gives for every admissible inner family, hence ; since , the two bounds force .
The Cantor set is an uncountable subset of of Lebesgue measure zero
Statement
Assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). The Cantor middle-thirds set (The Cantor middle-thirds set as the intersection of the sets obtained by removing open middle thirds) is Lebesgue measurable with
and is uncountable (Finite, countably infinite, countable, uncountable).
Facts & Assumptions
Given: The Axiom of Countable Choice and the Cantor middle-thirds set .
Assuming countable choice, if and only if has measure zero in the covering sense (A subset of has Lebesgue outer measure zero if and only if it has measure zero in the sense of countable closed-interval covers, Measure zero (a countable cover by intervals of total length below every ) and content zero (a finite such cover)).
Assuming countable choice, every with is Lebesgue measurable with (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
has content zero, and therefore measure zero, and is uncountable (The Cantor set is compact, perfect, uncountable, nowhere dense and of measure zero, and it contains no interval of positive length, so its only nonempty connected subsets are single points, claims 2 and 4; The Cantor middle-thirds set as the intersection of the sets obtained by removing open middle thirds, Finite, countably infinite, countable, uncountable).
Proof
The published theorem gives that has measure zero in the covering sense, so the agreement theorem gives .
A set of Lebesgue outer measure zero is Lebesgue measurable with measure zero, so , while the same published theorem gives that is uncountable.
Lebesgue outer measure, Lebesgue measurability and Lebesgue measure are unchanged by translation
Statement
Let , let , and let be the translate of (Translation of a subset of ). Then:
- for every subset ;
- is Lebesgue measurable if and only if is;
- for every Lebesgue measurable .
No choice principle is used. Lebesgue outer measure is defined as an infimum and the Carathéodory condition is a family of equations between its values, so all three clauses are statements about objects that exist in ZF; countable choice is needed to know that is a measure, not to know that it is translation invariant.
Facts & Assumptions
Given: A natural number , a vector , and a subset .
for every and (Lebesgue outer measure on , Series in the nonnegative extended real line).
; a box is nonempty exactly when for every ; and for a nonempty box with real parameters , the value being when a parameter is infinite (Half-open boxes in and their volume).
A subset is an elementary set when there are a natural number and a list of half-open boxes with (Elementary sets: the finite unions of half-open boxes in ), and every elementary set is the union of a finite list of pairwise disjoint half-open boxes (Every elementary set is a finite disjoint union of half-open boxes, and any finitely many boxes admit a common grid refinement).
For every , elementary volume on has value at the sum of the volumes of the members of any presentation of by a finite list of pairwise disjoint half-open boxes (The sum of the volumes of a disjoint box decomposition of an elementary set does not depend on the decomposition).
A set is Lebesgue measurable when for every , and is the restriction of to the family of these (Lebesgue measurable sets, the family , and the restricted set function , Carathéodory measurable sets).
The translate of by is ; translation by is the bijection , whose inverse is (Translation of a subset of ).
Addition of a real to an extended real is defined in every case, with when and , and when and (The extended real line , its order, and the arithmetic that is left undefined).
Proof
For a parameter pair one has , where is the parameter : a point lies in the left side exactly when satisfies , that is . The translated box is empty exactly when the original is, and has the same volume, because when both are real and an infinite parameter stays infinite.
Consequently, if is a presentation of an elementary set by pairwise disjoint half-open boxes, then is such a presentation of , so is elementary and .
A sequence of elementary sets covers if and only if the sequence covers , and the two covering costs are equal by step 2.1; the correspondence is a bijection between the two families of covers, with inverse given by translating by , so the two infima agree and .
For test sets, and , so by step 3.1 the Carathéodory identity for tested against is exactly the identity for tested against ; as ranges over all subsets so does , and therefore is Lebesgue measurable if and only if is, with in that case.
For a nonzero real , dilation by multiplies Lebesgue outer measure by , and reflection in the origin preserves it
Statement
Let , assume the Axiom of Countable Choice (The Axiom of Countable Choice ()), let be a nonzero real and write for , where . Then:
- for every subset , the product being defined in because ;
- is Lebesgue measurable if and only if is;
- for every Lebesgue measurable .
At the map is reflection in the origin and , so it preserves outer measure, measurability and measure. The value is excluded because is or and carries no information about .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, a nonzero real , and a subset .
Assuming countable choice, , the infimum of over countable covers of by closed rectangles (Countable covers by closed boxes, by open boxes and by closed cubes all compute Lebesgue outer measure, Lebesgue outer measure on ).
A set is Lebesgue measurable when for every , and is the restriction of to the family of these (Lebesgue measurable sets, the family , and the restricted set function , Carathéodory measurable sets, Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
, and finite products are defined by the recursion , (Laws of finite sums and finite products, claim 6; Finite sums and finite products, by recursion).
The defining recursion for natural powers is and (Integer powers ), and (Laws of integer exponents, claim 1).
when one of is , the other is , and both are or both are ; every product with one factor and the other is left undefined (The extended real line , its order, and the arithmetic that is left undefined).
The absolute value satisfies for and (Absolute value in an ordered field, Basic properties of the absolute value).
For positive reals, multiplication preserves order and reciprocals stay positive: if and then , and if then (Sign rules for products and monotonicity of multiplication, Inverses of positives are positive, and reciprocation reverses order, Ordered field).
Proof
For reals one has when and when , in both cases a closed rectangle whose -th side length is ; its volume is therefore .
Put . If for a nonempty , then is a lower bound of : for every real the inequality gives by [F6], while the claim is automatic when . Conversely, let be a lower bound of . If , then every element of is , hence every element of is and therefore . If is real, then by [F6], so implies for every real , and again the claim is automatic when ; thus is a lower bound of , so and therefore . Hence .
The assignment is a bijection from the countable closed-rectangle covers of onto those of , with inverse given by multiplication by . For one such cover, let be its sequence of rectangle volumes and the partial sums of in the sense of Series in the nonnegative extended real line; let be the partial sums of the transformed cover cost. By step 1.1 each transformed term is , and the shared recursion of nonnegative extended series gives for every . Therefore the transformed cover cost is by step 1.2. So step 1.2 turns the infimum of all transformed cover costs into , and [L1] then gives the same identity for .
For a test set one has and , so step 2.1 turns the Carathéodory identity for tested against into times the identity for tested against ; multiplication by the positive real is injective on , and is a bijection of the power set, so is Lebesgue measurable exactly when is, and then .
A translation-invariant Borel measure giving the unit cube measure one gives each generation- dyadic cube measure
Statement
Let and let be a measure on (Measures on sigma-algebras, The Borel sigma-algebra of a topological space) such that
Then for every dyadic cube of generation (Dyadic cubes of generation in ).
Only translates of half-open boxes are used, and those are Borel (The sigma-algebra generated by the half-open boxes of is the Borel sigma-algebra), so the invariance hypothesis is applied only where it is unambiguously meaningful.
Facts & Assumptions
Given: A natural number , a natural number , and a measure on the Borel sets of that is translation invariant and gives the unit cube measure .
Every lies in exactly one dyadic cube of generation (For each generation, the dyadic cubes of that generation are pairwise disjoint and cover ).
Every half-open box is a Borel set (The sigma-algebra generated by the half-open boxes of is the Borel sigma-algebra).
A measure on is a function with that is countably additive on pairwise disjoint sequences (Measures on sigma-algebras); padding a finite disjoint list with empty sets makes it finitely additive.
The translate of by is (Translation of a subset of ).
, where denotes the canonical natural of (Laws of finite sums and finite products, claim 2; Finite sums and finite products, by recursion).
For and , and (Laws of integer exponents, claims 1 and 3; Integer powers ).
Let ; if and whenever , then (The principle of mathematical induction).
The order on is total and compatible with addition (The integers form a totally ordered ring); the canonical embedding of into has as image exactly the nonnegative integers (The naturals embed in the integers, The integers as equivalence classes of pairs of naturals); and in exactly when (Discreteness: is the immediate successor).
Proof
A generation- dyadic cube is contained in exactly when and for every , and every point of lies in such a cube: if and is the index of the generation- cube containing , then and , so and by discreteness of ; conversely such a cube lies in because and .
Every generation- dyadic cube is a translate of , namely , and it is a half-open box, hence Borel; so all generation- cubes receive the same value under .
The indices admitted in step 1.1 are exactly the functions from to the set , and there are of them: by induction on , at there is exactly one such function and , while each function on coordinates is a function on coordinates together with one of values in the new coordinate, so the count is multiplied by and .
By steps 1.1 and 2.1 the cube is the union of a list of pairwise disjoint generation- dyadic cubes, so finite additivity and step 1.2 give ; no term can be , since then the sum would be rather than , so the common value is a real and the sum is .
Dividing by the strictly positive real gives , and step 1.2 transfers the value to every generation- dyadic cube.
A translation-invariant measure on the Borel sets of giving the unit cube measure one is the restriction of Lebesgue measure
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Let be a measure on (Measures on sigma-algebras) such that for every Borel set and every , and . Then
The hypothesis is meaningful because a translate of a Borel set is Borel, and it is satisfied by the restriction of to , so the theorem says that measure is the only one satisfying it. Finiteness on bounded sets is a consequence of the normalisation, not a further hypothesis.
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a translation-invariant measure on with .
If is a measure on the Borel sets of that is translation invariant and gives the unit cube measure , then for every dyadic cube of generation (A translation-invariant Borel measure giving the unit cube measure one gives each generation- dyadic cube measure , Dyadic cubes of generation in ).
Every open is the union of an at most countable family of pairwise disjoint dyadic cubes (Every open subset of is the union of a countable pairwise disjoint family of dyadic cubes).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable) and is a measure on with for every half-open box (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Every set with is Lebesgue measurable with (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included).
for every subset , is Lebesgue measurable if and only if is, and for measurable (Lebesgue outer measure, Lebesgue measurability and Lebesgue measure are unchanged by translation, Translation of a subset of ).
Let be a pi-system on generating , and let be measures on that agree on ; suppose there is an increasing sequence in with and for every ; then on (Measures agreeing on a generating pi-system are equal under an increasing finite-measure exhaustion from that pi-system).
A pi-system on is a nonempty family closed under binary intersections (Pi-systems).
The Borel sigma-algebra of is the sigma-algebra generated by its open sets (The Borel sigma-algebra of a topological space), and is the unique smallest sigma-algebra on containing (Nonempty intersections of sigma-algebras are sigma-algebras, so the generated sigma-algebra exists and is minimal); a sigma-algebra is closed under complements and countable unions (Sigma-algebras).
A subset is open in when every has a ball (The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement, Open ball, closed ball and sphere in a metric space); a finite intersection of open sets is open (Arbitrary unions and finite intersections of open sets are open, open balls are open and closed balls are closed, claim 3); and is a metric on ( as the set of functions , and , , are metrics on it).
A measure is countably additive on pairwise disjoint measurable sequences (Measures on sigma-algebras), and an at most countable family may be presented as a sequence (Finite, countably infinite, countable, uncountable).
Every complete ordered field is Archimedean: for every there is a natural number with (Every complete ordered field is Archimedean).
Proof
A translate of a Borel set is Borel: the family of whose translate is Borel contains every open set, since makes and hence open for open , and it is a sigma-algebra because translation commutes with complements and with countable unions; minimality of over the open sets finishes it.
The open subsets of form a pi-system generating : the family is nonempty and closed under binary intersections, and the Borel sigma-algebra is by definition the one it generates.
The restriction of to the Borel sets is a measure satisfying the two hypotheses, by translation invariance and by .
By the dyadic lemma both and give a generation- dyadic cube the value , the latter because a dyadic cube is a half-open box of that volume.
Both measures therefore agree on every open set: such a set is the union of an at most countable pairwise disjoint family of dyadic cubes, which may be presented as a sequence, and countable additivity gives the same value for the two measures.
The open cubes form an increasing sequence of open sets with union , by the Archimedean property, and by step 3.1 and the box theorem; the uniqueness theorem for a sigma-finite generating pi-system therefore gives on , and step 1.1 makes the invariance hypothesis meaningful throughout.
An invertible linear map of scales the Lebesgue measure of every Borel set by a positive constant depending only on the map
Statement
Let , assume the Axiom of Countable Choice (The Axiom of Countable Choice ()) and let be an invertible linear map (Linear map between vector spaces over the same field). Then:
- is a Borel set for every Borel set , and carries open sets to open sets;
- there is a strictly positive real , namely , with
- for invertible linear and , and .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and an invertible linear map of .
Assuming countable choice, a measure on with for every Borel and every , and with , equals on ; in particular the theorem notes that the restriction of to satisfies these hypotheses (A translation-invariant measure on the Borel sets of giving the unit cube measure one is the restriction of Lebesgue measure).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable), and is a measure on (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
Every set with is Lebesgue measurable with (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included).
Every bounded subset of has finite outer measure (Lebesgue measure is sigma-finite, and every metrically bounded subset of has finite outer measure).
For every linear there is a unique matrix such that , and there is with for every (Every Euclidean linear map has a unique matrix and satisfies for some , Linear map between vector spaces over the same field).
The translate of by is (Translation of a subset of ).
A measure on is a function with that is countably additive on pairwise disjoint sequences (Measures on sigma-algebras), and a scalar multiple is again a measure (Nonnegative scalar multiples and countable weighted sums of measures are measures, Nonnegative scalar multiples and countable weighted sums of measures).
The Borel sigma-algebra is the sigma-algebra generated by the open sets (The Borel sigma-algebra of a topological space), is the smallest sigma-algebra containing (Nonempty intersections of sigma-algebras are sigma-algebras, so the generated sigma-algebra exists and is minimal), and a sigma-algebra is closed under complements and countable unions (Sigma-algebras).
A subset is open in when every has a ball (The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement, Open ball, closed ball and sphere in a metric space), and is bounded when it is empty or lies in some ball (Bounded subset, diameter, distance from a point to a set, and distance between two sets in a metric space).
For every , , and , (The finite and reverse triangle inequalities for a norm; and for every norm on satisfies and is Lipschitz, hence continuous, for , claim 3; Each is a norm on , and the induced metrics are exactly , and of the published metric-spaces page, claim 3; The -norms for rational , and ; as the set of functions , and , , are metrics on it).
Proof
The inverse is linear, so there are reals and with and for all ; put .
carries open sets to open sets: if is open, and , then gives , so and .
The family of with Borel is a sigma-algebra, because is a bijection and so commutes with complements and with countable unions, and it contains every open set by step 2.1; minimality of over the open sets gives claim 1.
is bounded, being contained in the ball about the origin of radius , so it has finite measure; and it contains for the nonempty open box , which is open and nonempty by step 2.1, hence contains a ball and with it the open box , whose measure is a strictly positive real. So is a strictly positive real.
The assignment is well defined on by claim 1, and it is a measure: , and being injective carries a pairwise disjoint sequence to a pairwise disjoint sequence with , so countable additivity of transfers. It is translation invariant, since by linearity and is translation invariant.
By step 3.2 the scalar multiple is a measure on , it is translation invariant, and it gives the unit cube the value , so the uniqueness theorem identifies it with on the Borel sets; that is claim 2. Claim 3 follows by evaluating at the unit cube: , and the identity map gives .
A coordinate scaling and a coordinate transposition send the unit cube to a set of measure equal to the absolute value of the determinant
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Work with real matrices and identify a matrix with the linear map it defines by (Every Euclidean linear map has a unique matrix and satisfies for some ).
- Coordinate scaling. Let , let be real and let be the elementary matrix obtained from the identity by multiplying row by (Elementary matrices obtained by applying one elementary row operation to an identity matrix). Then sends to the point whose -th coordinate is and whose other coordinates are those of , the image is Lebesgue measurable, and
- Coordinate transposition. Let , let be below and let be the elementary matrix interchanging rows and . Then exchanges the -th and -th coordinates, , and
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and the elementary matrices and over .
If are real for , then any box obtained from the coordinate interval product by independently choosing for each endpoint whether it is included has Lebesgue measure (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included). In particular (Half-open boxes in and their volume).
An elementary matrix is a matrix obtained by applying one elementary row operation to the identity matrix ; there are three types: interchanges rows and ; multiplies row by ; and adds times row to the distinct row (Elementary matrices obtained by applying one elementary row operation to an identity matrix, Rectangular matrix multiplication and the identity matrix , including zero-sized shapes).
Let and let be a matrix over a commutative ring; interchanging two rows changes to , and multiplying one row by any changes it to (For every square matrix, including singular ones, a row swap negates the determinant, scaling a row by any scalar scales it, and row addition leaves it unchanged, claims 1 and 2; For , the determinant over a commutative ring by the Leibniz formula, and for a real matrix).
If is upper or lower triangular over a commutative ring, with , then (The determinant of a triangular matrix is the product of its diagonal entries).
For every linear there is a unique matrix such that (Every Euclidean linear map has a unique matrix and satisfies for some ).
The absolute value satisfies for , for and for (Absolute value in an ordered field, Basic properties of the absolute value).
Proof
The identity matrix is triangular with every diagonal entry , so ; the row-operation table applied to then gives and , hence and .
Reading off the matrix entries, sends to the point with -th coordinate and the other coordinates unchanged, and sends to the point with -th coordinate , -th coordinate and the others unchanged.
For claim 1, . When this is the half-open box with -th side ; when it is the box with -th side and all other sides . In either case [L1] gives Lebesgue measurability and measure .
For claim 2, restricts to a bijection of onto itself, since exchanging two coordinates of a point all of whose coordinates lie in again gives such a point and the map is its own inverse; hence the image is , of measure .
Steps 1.1, 2.1 and 2.2 are the two claims.
A shear sends the unit cube to a set of Lebesgue measure one
Statement
Let , assume the Axiom of Countable Choice (The Axiom of Countable Choice ()), let be below and let be real. Let be the linear map with matrix the elementary matrix obtained from the identity by adding times row to row (Elementary matrices obtained by applying one elementary row operation to an identity matrix), so that
Then is Lebesgue measurable and
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, distinct indices , a real , and the shear with matrix .
for every Lebesgue measurable and every (Lebesgue outer measure, Lebesgue measurability and Lebesgue measure are unchanged by translation, Translation of a subset of ).
An invertible linear map carries Borel sets to Borel sets and open sets to open sets (An invertible linear map of scales the Lebesgue measure of every Borel set by a positive constant depending only on the map, claim 1).
Assuming countable choice, every Borel subset of is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable), and is a complete measure on (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
For real parameters , every set between the open box and closed box is Lebesgue measurable with (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included), and (Half-open boxes in and their volume).
An elementary matrix is a matrix obtained by applying one elementary row operation to ; adds times row to the distinct row (Elementary matrices obtained by applying one elementary row operation to an identity matrix, Rectangular matrix multiplication and the identity matrix , including zero-sized shapes), and for every linear map there is a unique such matrix acting by (Every Euclidean linear map has a unique matrix and satisfies for some ).
Adding times one row to a distinct row leaves the determinant equal to (For every square matrix, including singular ones, a row swap negates the determinant, scaling a row by any scalar scales it, and row addition leaves it unchanged, claim 3; For , the determinant over a commutative ring by the Leibniz formula, and for a real matrix), and a triangular matrix has determinant the product of its diagonal entries (The determinant of a triangular matrix is the product of its diagonal entries).
For every real there is exactly one integer with (Integer part: for every real there is exactly one integer with ).
A measure is countably additive on pairwise disjoint measurable sequences, hence finitely additive after padding with empty sets (Measures on sigma-algebras).
A subset is open in when every has a ball , a subset is closed when its complement is open, and a finite intersection of open sets is open (The metric topology: a set is open when every one of its points has a ball around it inside the set; closed means open complement, Open ball, closed ball and sphere in a metric space, Arbitrary unions and finite intersections of open sets are open, open balls are open and closed balls are closed, claim 3).
For every , , and , (The finite and reverse triangle inequalities for a norm; and for every norm on satisfies and is Lipschitz, hence continuous, for , claim 3; Each is a norm on , and the induced metrics are exactly , and of the published metric-spaces page, claim 3; The -norms for rational , and ; as the set of functions , and , , are metrics on it).
The Borel sigma-algebra is the sigma-algebra generated by the open sets, and a sigma-algebra is closed under complements and countable unions (The Borel sigma-algebra of a topological space, Sigma-algebras); in particular every open and every closed subset of is Borel.
Proof
The matrix is obtained from the identity by a row addition, so and is invertible, with inverse the shear ; consequently carries Borel sets to Borel sets.
For every real there is exactly one integer with : applying the integer part to gives the unique integer with , and is the integer sought, uniqueness following the same way.
The linear functional satisfies , so for every real the set is open and is closed. Also , with closed and each open, hence Borel by [F7]; therefore each is a Borel set.
Only finitely many integers admit a point of : for one has , so and , and the integers satisfying both lie between the two integers supplied by the integer part of and of , hence form a finite consecutive list .
By step 1.2 every lies in exactly one , so the sets for in the list of step 1.4 are pairwise disjoint with union .
Define by for the unique with , where is the -th standard vector. Then takes values in , since its -th coordinate is and its other coordinates are those of .
is a bijection of onto itself. It is injective: if then for every , so and is an integer of absolute value below , hence . It is surjective: given , step 1.2 supplies the unique integer with ; setting for and gives with , so and .
The sets are pairwise disjoint, Borel and have union , because is an injective linear bijection; each , so translation invariance gives ; and by step 4.1 the sets are pairwise disjoint with union .
Finite additivity applied twice therefore gives , which with step 1.1 is the Statement.
A Lipschitz self-map of carries Lebesgue null sets to Lebesgue null sets
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Let be Lipschitz for the Euclidean metric (Lipschitz map, -Hölder map for rational , and contraction, as the set of functions , and , , are metrics on it) and let satisfy . Then is Lebesgue measurable and
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, a Lipschitz map of into itself, and a set with .
Assuming countable choice, if and only if is null in the covering sense of closed-cube covers (A subset of has Lebesgue outer measure zero if and only if it is null in the sense of countable closed-cube covers, Measure zero and content zero in by countable and finite cube covers).
Assuming countable choice, every with is Lebesgue measurable with (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
If is Lipschitz and is null, then is null (A Lipschitz map sends null sets to null sets).
is Lipschitz with constant if for all (Lipschitz map, -Hölder map for rational , and contraction).
Proof
By the agreement theorem, says exactly that is null in the covering sense of closed-cube covers.
The published theorem on Lipschitz images therefore applies and gives that is null in that same covering sense, so the agreement theorem read the other way gives ; completeness then makes Lebesgue measurable with .
Every affine hyperplane of , and hence every proper linear subspace, is Lebesgue null
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Then:
- For every with and every real , the affine hyperplane (The Euclidean inner product on ) is Lebesgue measurable with , and so is every subset of it.
- Every proper linear subspace (Linear subspace of a vector space) is Lebesgue measurable with .
At a hyperplane is the singleton and the only proper linear subspace is .
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, a nonzero , a real , and a proper linear subspace of .
Assuming countable choice, a Lipschitz self-map of carries a set of Lebesgue outer measure zero to a Lebesgue measurable set of measure zero (A Lipschitz self-map of carries Lebesgue null sets to Lebesgue null sets).
For and a real , the coordinate hyperplane is Lebesgue measurable with measure (A box with a degenerate side is Lebesgue null, and so is every coordinate hyperplane in ).
Assuming countable choice, is a complete measure on , so every subset of a measurable null set is measurable of measure (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
The Euclidean inner product of is , and it is symmetric, bilinear and positive definite, making an inner product space (The Euclidean inner product on , Finite sums and finite products, by recursion, Laws of finite sums and finite products).
For a linear subspace of an inner product space , for every , and (The orthogonal complement , Linear subspace of a vector space).
For every subspace of a finite-dimensional inner product space , (In finite dimension, and ).
For every linear there is with for every (Every Euclidean linear map has a unique matrix and satisfies for some , Linear map between vector spaces over the same field).
is Lipschitz with constant if for all (Lipschitz map, -Hölder map for rational , and contraction), and (Each is a norm on , and the induced metrics are exactly , and of the published metric-spaces page, claim 3; The -norms for rational , and ; as the set of functions , and , , are metrics on it).
Proof
Fix with and define by for and . Then carries the coordinate hyperplane onto : a point of has , and conversely a point is for the point agreeing with off the coordinate and having .
is Lipschitz: the difference equals for the linear map obtained from by deleting the constant , so for a real .
The coordinate hyperplane is Lebesgue measurable of measure , hence of outer measure , so steps 1.1 and 1.2 with the Lipschitz lemma give that is Lebesgue measurable with ; completeness then gives the same for every subset of it, which is claim 1.
If is a proper linear subspace then : otherwise . Choosing a nonzero puts inside , so claim 1 and completeness make Lebesgue measurable of measure ; at the hyperplane is the singleton and the only proper subspace is .
A linear map of sends Lebesgue measurable sets to Lebesgue measurable sets, with when is invertible and Lebesgue null when it is not
Statement
Let , assume the Axiom of Countable Choice (The Axiom of Countable Choice ()) and let be linear with matrix (Linear map between vector spaces over the same field, Every Euclidean linear map has a unique matrix and satisfies for some ).
- Invertible case. If , then is Lebesgue measurable for every Lebesgue measurable and both sides possibly ; the product is defined in because .
- Singular case. If , then is Lebesgue measurable with for every .
The singular clause is stated as nullity and not as a product. When and the expression is , which The extended real line , its order, and the arithmetic that is left undefined leaves undefined; writing the conclusion as says the same thing wherever the product is defined and remains a statement where it is not.
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, a linear map of with matrix , and a set .
An invertible linear map carries Borel sets to Borel sets, and there is a strictly positive real with for every Borel , with and (An invertible linear map of scales the Lebesgue measure of every Borel set by a positive constant depending only on the map).
For a shear satisfies (A shear sends the unit cube to a set of Lebesgue measure one).
Every proper linear subspace is Lebesgue measurable with (Every affine hyperplane of , and hence every proper linear subspace, is Lebesgue null).
A Lipschitz self-map of carries a set of Lebesgue outer measure zero to a Lebesgue measurable set of measure zero (A Lipschitz self-map of carries Lebesgue null sets to Lebesgue null sets, Lipschitz map, -Hölder map for rational , and contraction).
is Lebesgue measurable if and only if for an set and a set with (Assuming countable choice, four equivalent descriptions of a Lebesgue measurable subset of , condition 4; and subsets of a topological space, agreeing with the real-line notion).
Assuming countable choice, is a complete measure on (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume) and every Borel set is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable, Measures on sigma-algebras).
Every invertible matrix is a finite product of elementary matrices, the identity being the empty product (Every invertible finite square real matrix is a finite product of elementary matrices), and every elementary matrix is invertible (Every elementary matrix is invertible, with inverse given by the reverse elementary operation, Elementary matrices obtained by applying one elementary row operation to an identity matrix).
For and over a commutative ring, (For same-sized finite square matrices over a commutative ring, , For , the determinant over a commutative ring by the Leibniz formula, and for a real matrix, Rectangular matrix multiplication and the identity matrix , including zero-sized shapes), and a triangular matrix has determinant the product of its diagonal entries (The determinant of a triangular matrix is the product of its diagonal entries).
For every and every real matrix , is invertible if and only if (A finite square real matrix is invertible if and only if its determinant is nonzero).
For every linear there is a unique matrix with , and there is with for every (Every Euclidean linear map has a unique matrix and satisfies for some , Each is a norm on , and the induced metrics are exactly , and of the published metric-spaces page, The -norms for rational , and , as the set of functions , and , , are metrics on it).
For a linear map , (Kernel and image of a linear map), and it is a linear subspace (The kernel and image are linear subspaces, and a linear map is injective if and only if its kernel is trivial).
Every product with one factor and the other is left undefined in (The extended real line , its order, and the arithmetic that is left undefined).
Let ; if and whenever , then (The principle of mathematical induction).
Proof
Suppose . Then is a proper linear subspace of : were surjective, each standard vector would be for some , finitely many instantiations, and the matrix with would satisfy , so and , contradicting .
is Lipschitz, since for a real .
Every elementary matrix of satisfies : at the only elementary matrices are the scalings , and for the three types are the scalings, the transpositions and the shears, whose unit-cube images have the measures , and , matching in each case.
Suppose , so is invertible and factors as a finite product of elementary matrices, each invertible. Multiplicativity of and of the determinant then give by induction on , the empty product giving .
For , step 1.1 makes a proper linear subspace, hence Lebesgue null; every is a subset of it, so completeness makes Lebesgue measurable with , which is claim 2; stating it as a product would require the undefined when .
For and Lebesgue measurable, write with an set, hence Borel, and ; then , where is Borel and is Lebesgue measurable of measure by step 1.2 and the Lipschitz lemma, so is measurable. Since and , and with , both pairs differ by null sets, so and , which is claim 1; claim 2 is step 2.2.
Lebesgue measure on is invariant under every orthogonal linear map
Statement
Let , assume the Axiom of Countable Choice (The Axiom of Countable Choice ()), and let be an orthogonal operator on with the Euclidean inner product (Linear isometries, and orthogonal or unitary operators on finite-dimensional inner product spaces, The Euclidean inner product on ). Then is Lebesgue measurable for every Lebesgue measurable and
Both the orientation-preserving operators, of determinant , and those of determinant are covered, since only the absolute value of the determinant enters; nothing is asserted here about which matrices occur in either class.
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and an orthogonal operator on .
Assuming countable choice, an invertible linear with matrix sends Lebesgue measurable sets to Lebesgue measurable sets with (A linear map of sends Lebesgue measurable sets to Lebesgue measurable sets, with when is invertible and Lebesgue null when it is not).
An invertible linear isometry from a real finite-dimensional inner product space to itself is an orthogonal operator (Linear isometries, and orthogonal or unitary operators on finite-dimensional inner product spaces).
Every orthogonal or unitary operator satisfies ; over , this says (Orthogonal and unitary operators form groups, and their determinants have modulus one, For , the determinant over a commutative ring by the Leibniz formula, and for a real matrix).
For every linear there is a unique matrix such that (Every Euclidean linear map has a unique matrix and satisfies for some ).
Proof
An orthogonal operator is by definition an invertible linear map of to itself, and its matrix satisfies , so in particular .
The linear change of variables therefore applies in its invertible clause and gives for every Lebesgue measurable , with measurable.
How the Lebesgue change-of-variables formula relates to the published formula for Jordan content
Two determinant formulas are now in force, for two different set functions, and this remark says how they meet.
The published one is about Jordan content. A linear endomorphism of sends bounded Jordan sets to bounded Jordan sets and scales their content by the absolute determinant states that a linear endomorphism of with standard matrix sends every bounded Jordan set to a bounded Jordan set with , and that a singular linear image has content zero (Jordan inner and outer content and Jordan measurable bounded sets in ).
This page's is about Lebesgue measure. Assuming the Axiom of Countable Choice, A linear map of sends Lebesgue measurable sets to Lebesgue measurable sets, with when is invertible and Lebesgue null when it is not states the same identity with replaced by , for every Lebesgue measurable , bounded or not, and with the singular case stated as nullity of rather than as a product.
Where the two agree, and why that is not an accident. On a bounded Jordan set the two set functions take the same value, by Lebesgue outer measure is at most Jordan outer content, and a bounded Jordan measurable set is Lebesgue measurable with Lebesgue measure equal to its Jordan content, so on that class the two formulas are the same equation read twice. Neither implies the other: the published formula says nothing about a Lebesgue measurable set that is not Jordan measurable, and this page's formula says nothing about Jordan measurability of an image, which the published one asserts.
What the extension costs, and where it is spent. Passing from bounded Jordan sets to arbitrary Lebesgue measurable sets is not a matter of taking limits: the Lebesgue proof runs through the uniqueness of a normalised translation-invariant Borel measure, the factorisation of an invertible matrix into elementary matrices, and the fact that a Lipschitz image of a null set is null. The last of these is what carries the argument across the gap between Borel sets and the larger Lebesgue class, and it is why the change of variables holds on all of and not merely on the Borel sets.
A measurable set of positive finite measure occupies more than any prescribed proportion of some dyadic cube
Statement
Let , assume the Axiom of Countable Choice (The Axiom of Countable Choice ()), let be Lebesgue measurable with , and let be a real with . Then there is a dyadic cube (Dyadic cubes of generation in ) with
Both hypotheses on are used: positivity is what makes the strict inequality available, and finiteness is what makes the division by legitimate.
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, a Lebesgue measurable set with , and a real with .
Assuming countable choice, open and (Assuming countable choice, the Lebesgue outer measure of an arbitrary subset of is the infimum of the measures of the open sets containing it).
Every open is the union of an at most countable family of pairwise disjoint dyadic cubes (Every open subset of is the union of a countable pairwise disjoint family of dyadic cubes).
Assuming countable choice, is a sigma-algebra and is a complete measure on it and is the restriction of (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume); every Borel set, in particular every open set and every dyadic cube, is Lebesgue measurable (Assuming countable choice, every Borel subset of is Lebesgue measurable, A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included).
A measure is countably additive on pairwise disjoint measurable sequences (Measures on sigma-algebras) and monotone (Measures are monotone).
The nonnegative extended sum of a sequence in is , the supremum of its nondecreasing partial sums (Series in the nonnegative extended real line), and an at most countable family may be presented as a sequence (Finite, countably infinite, countable, uncountable).
For the product is when one factor is and the other is a nonzero real of the same sign; multiplication by a strictly positive real is therefore an order isomorphism of (The extended real line , its order, and the arithmetic that is left undefined).
Proof
Suppose, for contradiction, that for every dyadic cube .
Since and is a strictly positive real, is a real strictly above , so outer regularity supplies an open with .
Write as the union of an at most countable pairwise disjoint family of dyadic cubes; the family is nonempty because is, and presenting it as a sequence when it is infinite, or using finite additivity when it is finite, countable additivity gives and, since and the cubes are disjoint, also .
Applying the assumption of step 1.1 termwise and scaling the sum by the strictly positive real gives , which is impossible; so some dyadic cube satisfies the displayed strict inequality.
If a Lebesgue measurable subset of has positive measure, its difference set contains an open ball about the origin
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Let be Lebesgue measurable with , and put
Then there is a real with , the open Euclidean ball of centre the origin and radius (Open ball, closed ball and sphere in a metric space, as the set of functions , and , , are metrics on it).
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a Lebesgue measurable set with .
Assuming countable choice, a Lebesgue measurable with and a real with admit a dyadic cube with (A measurable set of positive finite measure occupies more than any prescribed proportion of some dyadic cube, Dyadic cubes of generation in ).
for every Lebesgue measurable and every , and is measurable exactly when is (Lebesgue outer measure, Lebesgue measurability and Lebesgue measure are unchanged by translation, Translation of a subset of ).
Every set with is Lebesgue measurable with (A box in with parameters is Lebesgue measurable of measure , whichever of its faces are included), and (For each generation, the dyadic cubes of that generation are pairwise disjoint and cover , Half-open boxes in and their volume).
Assuming countable choice, is a complete measure on , a sigma-algebra (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume), and every bounded Lebesgue measurable set has finite measure (Lebesgue measure is sigma-finite, and every metrically bounded subset of has finite outer measure).
Let be an increasing sequence of measurable sets for a measure ; then (Continuity from below for measures).
A measure is countably additive on pairwise disjoint measurable sequences, hence finitely additive (Measures on sigma-algebras), and monotone (Measures are monotone).
For and rational with , , where is the unique nonnegative -th root of (Rational powers of a positive base, Existence and uniqueness of -th roots: a unique with ), and the value does not depend on the representative (Rational powers do not depend on the representative).
If and then ; if then (Monotonicity of and of , claims 2 and 3; Integer powers ), and (Laws of integer exponents, claim 1).
For every , , and , (The finite and reverse triangle inequalities for a norm; and for every norm on satisfies and is Lipschitz, hence continuous, for , claim 3; Each is a norm on , and the induced metrics are exactly , and of the published metric-spaces page, claim 3; The -norms for rational , and ; as the set of functions , and , , are metrics on it).
Proof
The sets for are Lebesgue measurable, increase with and have union , so continuity from below gives and some has ; that set is bounded, hence of finite measure. Replacing by it shrinks , so it suffices to prove the theorem when .
Put , the unique nonnegative -th root of ; then , since would give , and is a strictly positive real with and .
Assume and apply the density lemma with : there is a dyadic cube , of some generation and side , with , since .
Let with , so that in every coordinate. Writing with , both and are contained in the half-open box with parameter pairs , whose measure is .
The two sets are Lebesgue measurable with the same measure, by translation invariance, so if they were disjoint then additivity and monotonicity inside would give , which is impossible; hence they meet, and a common point with exhibits .
Therefore , and is a strictly positive real.
A Lebesgue measurable subgroup of of positive measure is all of
Statement
Let and assume the Axiom of Countable Choice (The Axiom of Countable Choice ()). Let be a subgroup of the additive group (Subgroup, Group and abelian group) that is Lebesgue measurable with . Then
Equivalently, in the contrapositive form the sources state: a Lebesgue measurable proper subgroup of has measure zero. Nothing is asserted about subgroups that are not Lebesgue measurable.
Facts & Assumptions
Given: A natural number , the Axiom of Countable Choice, and a Lebesgue measurable subgroup of with .
Assuming countable choice, a Lebesgue measurable with has a real with (If a Lebesgue measurable subset of has positive measure, its difference set contains an open ball about the origin, Open ball, closed ball and sphere in a metric space).
Assuming countable choice, is a complete measure on the sigma-algebra (Assuming countable choice, is a sigma-algebra containing every elementary set and is a complete measure extending elementary volume).
A subset is a subgroup when , is closed under the operation, and is closed under inverses (Subgroup, Group and abelian group).
Every complete ordered field is Archimedean: for every there is a natural number with (Every complete ordered field is Archimedean); and for every real there is a natural with (For every in a complete ordered field there is a natural with ).
Let ; if and whenever , then (The principle of mathematical induction).
Proof
Since is a subgroup, and whenever , so ; conversely , and therefore .
Steinhaus applied to supplies a real with .
Let . The Archimedean property gives a natural with , so and by steps 1.1 and 1.2.
A subgroup is closed under addition, so an induction on shows for every natural , the case being ; taking gives , and as was arbitrary, .
5 · Examples, counterexamples and false statements
None yet.
Sources
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Section 1
- T. Tao, An Introduction to Measure Theory (GSM 126), Section 1.1
- John K. Hunter, Measure Theory (UC Davis lecture notes), Chapter 2
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Theorem 1.1
- John K. Hunter, Measure Theory (UC Davis lecture notes), Definition 2.2
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Definition 1.2
- John K. Hunter, Measure Theory (UC Davis lecture notes), Theorem 2.4
- T. Tao, An Introduction to Measure Theory (GSM 126), Section 1.2
- John K. Hunter, Measure Theory (UC Davis lecture notes), Definition 2.10
- John K. Hunter, Measure Theory (UC Davis lecture notes), Theorem 2.9 and Proposition 2.12
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.1.14
- T. Tao, An Introduction to Measure Theory (GSM 126), Lemma 1.2.11
- John K. Hunter, Measure Theory (UC Davis lecture notes), Proposition 2.20
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Proposition 1.4
- John K. Hunter, Measure Theory (UC Davis lecture notes), Proposition 2.21
- John K. Hunter, Measure Theory (UC Davis lecture notes), Proposition 2.7
- John K. Hunter, Measure Theory (UC Davis lecture notes), Example 2.3
- T. Tao, An Introduction to Measure Theory (GSM 126), Lemma 1.2.12
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.2.14
- John K. Hunter, Measure Theory (UC Davis lecture notes), Theorem 2.24
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.2.7
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercises 1.2.7 and 1.2.19
- John K. Hunter, Measure Theory (UC Davis lecture notes), Theorems 2.24, 2.25 and 2.27
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Theorem 1.6
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.2.15
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Theorem 1.5
- John K. Hunter, Measure Theory (UC Davis lecture notes), Theorem 2.28
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.2.8
- John K. Hunter, Measure Theory (UC Davis lecture notes), Example 2.14
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.2.9
- John K. Hunter, Measure Theory (UC Davis lecture notes), Proposition 2.16
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Theorem 2.3
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Theorem 3.4
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.2.23
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Theorem 3.1
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.2.21
- John K. Hunter, Measure Theory (UC Davis lecture notes), Proposition 2.32
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Section 3
- John K. Hunter, Measure Theory (UC Davis lecture notes), Theorem 2.33
- John K. Hunter, Measure Theory (UC Davis lecture notes), Theorem 2.31
- E. A. Carlen, Notes on Lebesgue Measure on $\mathbb{R}^n$ and $S^{n-1}$ (Rutgers Math 501), Theorem 3.2
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.6.25
- T. Tao, An Introduction to Measure Theory (GSM 126), Exercise 1.6.8
- J. Ye, L. Yu, X. Zhao, When is $A+xA=\mathbb{R}$?, Theorem 1.1
- J. Ye, L. Yu, X. Zhao, When is $A+xA=\mathbb{R}$?, Corollary 1.2