MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

For any $\alpha\in [0, 1]$ and $g:[a, b]\to [0, 1]$, let us define $g^{-}:[0, 1]\to [a, b]$, by the formula $g^{-}(\alpha) = \inf (x\in [a, b] ; g(x)\ge \alpha )$.

Let $f_{n}, f : [a, b]\to [0, 1]$, $n\in \mathbb{N}$, be continuous on $[a, b]$, such that the value $1$ is attained, for $f$ and for all $f_{n}$. It is true that if $f_{n}\to f$ uniformly on $[a, b]$, then $f_{n}^{-}(\alpha)\to f^{-}(\alpha)$, almost everywhere in $[0, 1]$ ?

Thank you in advance.

Best, George

share|cite|improve this question
First you have $g$ defined on $[a,b]$, then you ask a conclusion for any $a \in [0,1]$. Not the same $a$? – Gerald Edgar Apr 18 '12 at 15:37
In addition to what Gerald Edgar asks: how do you define the inf in case the set is empty? That being said, as far as I understand the question, I doubt this is true. Say take maps from [0,1] to itself that are straight lines from some value t, say (1-1/n), at 0 to 1 at 1. Let your sequences of functions be such that t converges to 1; this will converge uniformly to constant 1. While this minus transform you consider at 1 should be 1 for all these functions it should be 0 for the limit (if I understand the def correctly). – user9072 Apr 18 '12 at 15:55
One more thought: I guess (but did not check in detail as long as the def is not completely clear) if you change gretear-equal to strict greater this would work. – user9072 Apr 18 '12 at 16:20
I apologize, $a$ in the definitions of $g^{-}$, $f_{n}^{-}$ and $f^{-}$ is in fact $\alpha$ (to be not confused with the endpoint $a$ of the interval $[a, b]$. – George Apr 18 '12 at 16:40
@quid: it seems to me that if we change the definition of f-minus to make the inequality strict, then it is still seen to be false by a similar counterexample: let f_n be 1/n on the interval [0,1/2] and a straight line on the interval [1/2,1]. For each of these functions f_n, the inf of the x with f_n(x) > 0 is 0, but their limit f is 0 on the interval [0,1/2], so the inf of the x with f(x) > 0 is 1/2. – Trevor Wilson Apr 18 '12 at 17:53

There are two problems:

  1. Replace a continuous $g:[a,b]\to [0,1]$ by its monotone non-decreasing majorant $g^{\text{mon}}$ which is the infimum of all monotone non-increasing functions $\ge g$. I did not check the details, but this operation might be continuous in the $\sup$-norm.

  2. Is taking inverse on the space of monotone non-decreasing functions $f:[a,b]\to [0,1]$ with $f(a)=0$ (?!) and $f(b)=1$ (?!) into the space of monotone non-decreasing functions $[0,1]\to [a,b]$ continuous with respect to the $\sup$-norms?

This two hints might help you to decide the question.

share|cite|improve this answer

The inf in case the set is empty is well-defined according to his assumption.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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