6
$\begingroup$

We shall say that two functions $f$ and $g$ defined on a set $E$ are equal almost everywhere, and write $f(x)=g(x)$ a.e $x\in E$, if the set $\{x\in E: f(x)\neq g(x)\}$ has measure zero.

I just can't wrap my brain around the fact that such functions exist! Certainly, we can take the cantor set which has measure zero but how to pick $f$ and $g$? Is there an example that I am not aware of ?

$\endgroup$
  • 5
    $\begingroup$ Any function is equal to itself almost everywhere. $\endgroup$ – David Mitra Dec 26 '12 at 3:04
  • $\begingroup$ A more interesting question is whether or not such functions "naturally" exist; defining them piecewise is easy. $\endgroup$ – MathematicsStudent1122 Sep 9 '16 at 19:53
11
$\begingroup$

Let $f$ map all real numbers to $0$ and $g$ map all irrationals to $0$ and all rationals to $1$. Then $f$ and $g$ are distinct but equal a.e.

$\endgroup$
12
$\begingroup$

$\newcommand{\R}{\Bbb R}$ There are other sets of measure $0$. For example $Z=\{1\}$ has measure $0$. Consider $f:\R\to \R$ the function defined by $$f(x)=x,$$ and $g:\R\to\R$ given by $$g(x)=\begin{cases} x &\text{if $x\neq 1$}\\ \pi &\text{if $x=1$}\end{cases}$$ Then $f=g$ a.e. since $f\neq g$ on $Z$.

$\endgroup$
4
$\begingroup$

Consider a measure zero set i.e. $E_1$ such that $\mu(E_1) = 0$. Let $$h(x) = \begin{cases} 0 & x \in E_1^c\\ \text{any value} & x \in E_1 \end{cases}$$ Now consider any function $f(x)$. Then $g(x) = f(x) + h(x)$ satisfies your criteria.

$\endgroup$
3
$\begingroup$

There are good answers already, but I think this needs a visual.

Two functions that are equal almost everywhere

$\endgroup$
2
$\begingroup$

If we let $f(x) = \lceil x \rceil$ and $g(x) = \lfloor x \rfloor + 1$ (ceiling and floor functions respectively) then $f$ and $g$ only differ on the set of integers $\mathbb{Z}$, which has measure zero.

$\endgroup$
1
$\begingroup$

Consider $\mathbb{R}$ equipped with the Borel $\sigma$-field $\mathcal{B}(\mathbb{R})$ and the Lebesgue measure $\lambda$, i.e. $\lambda$ is the measure on $\mathcal{B}(\mathbb{R})$ which satisfies $\lambda((a,b))=b-a$ for $a,b\in\mathbb{R}$, $a<b$.

Let $f:\mathbb{R}\to\mathbb{R}$ be any Borel measurable function and suppose we "change" $f$ in at most countably many points. If we call this new function $g$, then $f=g$ almost everywhere (with respect to $\lambda$). In particular, $f$ and $g$ have the same integral over any Borel set.

Let us formalize this: Let $(x_n)_{n\geq 1},(y_n)_{n\geq 1}\subseteq \mathbb{R}$ be two sequences of real numbers and define $$ g(x)= \begin{cases} y_n\quad &\text{if } x=x_n,\\ f(x)&\text{if }x\neq x_n\text{ for all }n, \end{cases} $$ i.e. we have "changed" $f$ along the sequence $(x_n)_{n\geq 1}$. Then $$ \{x\in\mathbb{R}:f(x)\neq g(x)\}\subseteq \{x_n: n\geq 1\}, $$ which has measure zero.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.