Lebesgue Measure is Invariant under Translations
Theorem
Let $\lambda^n$ be the $n$-dimensional Lebesgue measure on $\R^n$ equipped with the Borel $\sigma$-algebra $\map \BB {\R^n}$.
Let $\mathbf x \in \R^n$.
Then $\lambda^n$ is translation invariant; that is, for all $B \in \map \BB {\R^n}$, have:
- $\map {\lambda^n} {\mathbf x + B} = \map {\lambda^n} B$
where $\mathbf x + B$ is the set $\set {\mathbf x + \mathbf b: \mathbf b \in B}$.
Proof
Denote with $\tau_{\mathbf x}: \R^n \to \R^n$ the translation by $\mathbf x$.
From Translation in Euclidean Space is Measurable Mapping, $\tau_{\mathbf x}$ is $\map \BB {\R^n} \, / \, \map \BB {\R^n}$-measurable.
Consider the pushforward measure $\lambda^n_{\mathbf x} := \paren {\tau_{\mathbf x} }_* \lambda^n$ on $\map \BB {\R^n}$.
By Characterization of Euclidean Borel Sigma-Algebra, it follows that:
- $\map \BB {\R^n} = \map \sigma {\JJ^n_{ho} }$
where $\sigma$ denotes generated $\sigma$-algebra, and $\JJ^n_{ho}$ is the set of half-open $n$-rectangles.
Let us verify the four conditions for Uniqueness of Measures, applied to $\lambda^n$ and $\lambda^n_{\mathbf x}$.
Condition $(1)$ follows from Half-Open Rectangles Closed under Intersection.
Condition $(2)$ is achieved by the sequence of half-open $n$-rectangles given by:
- $J_k := \hointr {-k} k^n$
For condition $(3)$, let $\horectr {\mathbf a} {\mathbf b} \in \JJ^n_{ho}$ be a half-open $n$-rectangle.
Since:
- $\map {\tau_{\mathbf x}^{-1} } {\horectr {\mathbf a} {\mathbf b} } = \mathbf x + \horectr {\mathbf a} {\mathbf b} = \horectr {\mathbf {a + x} } {\mathbf {b + x} }$
we have:
\(\ds \map {\lambda^n_{\mathbf x} } {\horectr {\mathbf a} {\mathbf b} }\) | \(=\) | \(\ds \map {\lambda^n} {\horectr {\mathbf {a + x} } {\mathbf {b + x} } }\) | Definition of Pushforward Measure | |||||||||||
\(\ds \) | \(=\) | \(\ds \prod_{i \mathop = 1}^n \paren {\paren {b_i + x_i} - \paren {a_i + x_i} }\) | Definition of Lebesgue Measure | |||||||||||
\(\ds \) | \(=\) | \(\ds \prod_{i \mathop = 1}^n \paren {b_i - a_i}\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds \map {\lambda^n} {\horectr {\mathbf a} {\mathbf b} }\) | Definition of Lebesgue Measure |
Finally, since:
- $\ds \map {\lambda^n} {J_k} = \prod_{i \mathop = 1}^n \paren {k - \paren {-k} } = \paren {2 k}^n$
the last condition, $(4)$, is also satisfied.
Whence Uniqueness of Measures implies that:
- $\lambda^n_{\mathbf x} = \lambda^n$
and since for all $B \in \map \BB {\R^n}$ we have:
- $\mathbf x + B = \map {\tau_{\mathbf x}^{-1} } B$
this precisely boils down to:
- $\map {\lambda^n} {\mathbf x + B} = \map {\lambda^n} B$
$\blacksquare$
Motivation
This theorem formalizes the physical intuition that the size of an object does not depend on its position.
Sources
- 2005: René L. Schilling: Measures, Integrals and Martingales ... (previous) ... (next): $4.9 \ \text{(i)}$
- 2005: René L. Schilling: Measures, Integrals and Martingales ... (previous) ... (next): $5.8 \ \text{(i)}$