In my analysis class, I was given the following definition of the Lebesgue integral:


$O^b:=\{ \phi:[a,b]\to\mathbb R \mid \exists \text{ sequence of step functions $s_1 \leq s_2 \leq \ ... \ \rightarrow \phi$ a.e.}\}$

$O^a:=\{ \phi:[a,b]\to\mathbb R \mid \exists \text{ sequence of step functions $s_1 \geq s_2 \geq \ ... \ \rightarrow \phi$ a.e.}\}$

On $O^b$ define $\int^b_a \phi:= \lim \int^b_a s_n \leq +\infty$

On $O^a$ define $\int^b_a \phi:= \lim \int^b_a s_n \geq -\infty$

$L$-$\overline{\int^b_a}f:= \inf \{ \int^b_a \phi \mid f\leq \phi \in O^b \} $

$L$-$\underline{\int^b_a}f:= \sup \{ \int^b_a \phi \mid f\geq \phi \in O^a \} $

$O^b$ and $O^a$ stand for One-sided$^{\text{below}}$ and One-sided$^{\text{above}}.$

If $L$-$\underline{\int^b_a}f = L$-$\overline{\int^b_a}f$, we call it the Lebesgue integral of $f$.

In order to get a better understanding of what's going on, I tried to look up the definition in a textbook or on the internet but I couldn't find anything (not even close to the above). The problem is that we didn't have any measure theory so far: only the zero set (hence the "a.e." above). And all texts seem to define the Lebesgue integral after introducing measure etc.

I would highly appreciate a reference to a material which introduces the Lebesgue integral in the above way.

  • 2
    $\begingroup$ Have you tried asking your prof? Surely he/she didn't just make up a new formulation of the theory out of the blue just to confuse everybody. $\endgroup$ – Jeff Jun 29 '13 at 0:21
  • $\begingroup$ @Jeff: Yes, we did ask him. He gave us a reference to one book yet it turned out to be the same as the rest: first a chapter on measure theory, then the Lebesgue integral. $\endgroup$ – Leo Jun 29 '13 at 0:28
  • 1
    $\begingroup$ Ah, I guess it is not necessary to include it here. Looks clumsy, and it is clear we're talking about Lebesgue. $\endgroup$ – Pedro Tamaroff Jun 29 '13 at 0:38
  • 1
    $\begingroup$ I don't think this gives the Lebesgue integral unless by "step function" you mean simple function. $\endgroup$ – nullUser Jun 29 '13 at 0:48
  • 1
    $\begingroup$ Unless $f$ is bounded this doesn't work. Take $f(x) = x^{-1/2}$ on $[0,1]$. Any step function above $f$ has integral $+\infty$ but of course $\int_0^1f < \infty$. $\endgroup$ – nullUser Jun 29 '13 at 0:50

Look in Pesin, I.N. Classical and modern integration theories. (English) Probability and Mathematical Satistics. 8. New York-London: Academic Press. XVIII, 195 p. (1970). MSC 2010: 26-01 28-01 26A36 26A39 26A42 28Cxx 26-03 28-03 Sorry, I can't add a comment.


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.