3
$\begingroup$

Let $(X,A,\mu)$ a measure space. Assume a sequence $(f_n)$ of positive (i.e. $\geq 0$) measurable functions from $X$ to $\mathbb{R}$ converges to $f$ in $L^p$. Then is $f$ positive ?

My idea is the following : we know that a subsequence $(f_{n_k})_k$ converges pointwise almost everywhere to $f$, but each $f_{n_k}$ is also positive, so $f$ is positive almost everywhere. Is that true ?

$\endgroup$
2
  • 4
    $\begingroup$ It means $\geq 0$, I'll precise it, thanks. In French positive means non-negative, hence my confusion $\endgroup$ Commented May 22 at 16:07
  • 1
    $\begingroup$ Your idea is correct and is sufficient. $\endgroup$ Commented May 22 at 16:11

2 Answers 2

5
$\begingroup$

We have $$f_n\to f,\ f_n=|f_n|\to|f|$$ The second convergence follows from the triangle inequality $|\,|f_n|-|f|\,|\le |f_n-f|$. Hence $f=|f|\ge 0.$

$\endgroup$
1
  • $\begingroup$ Nice trick there ! $\endgroup$ Commented May 23 at 14:17
3
$\begingroup$

I assume that by positive you mean $\geq0$ (the other possible meaning leads trivially to a negative answer).

Now, consider the measurable subset $E=f^{-1}(]-\infty,0[)$.

Then for all $n$,

$\lvert\lvert f-f_n\rvert\rvert_p\geq(\int_X 1_E\lvert f-f_n\rvert^p d\mu)^\frac{1}{p}\geq(\int_X 1_E\lvert f\rvert^p d\mu)^\frac{1}{p}>0$

as $\mu(E)>0$ and $1_E\lvert f\rvert^p>0$ on $E$.

Absurd.

So $f$ is positive in the sense of $L^p$ function (positive up to a measure $0$ subset).

$\endgroup$

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.