Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

Construct an example of a differentiable function such that $$ \forall r \in {\Bbb Q}\quad f(r) \in {\Bbb Q}\text{ but } f'(r) \notin {\Bbb Q} $$ this example is not trivial, in a paper they prove the existence, but they don't give an explicit example, but they said that other paper as an explicit example of that, but i could not find it.

EDITED: This is the paper:

Walter Rudin, Restrictions on the Values of Derivatives, The American Mathematical Monthly, Vol. 84, No. 9 (1977), pp. 722-723, MR480908.

share|cite|improve this question
The link works for you...Its better to give the journal ref. and then links. – Tapu Oct 13 '11 at 19:10
On MathSciNet there is the possibility to create a link to the review: right below the bibliographic info there are five red links PDF | Clipboard | Journal | Article | Make Link Click on Make Link, and you'll get a nice and clickable link, looking like this: While you're at it, you can also click on "Article" (if it exists) and you'll get a stable link to the paper. Please paste both of these links in the future. – t.b. Oct 13 '11 at 19:23
Thanks t.b! it looks so better – August Oct 13 '11 at 19:24
up vote 6 down vote accepted

The problem posed is available here, but here it is in its entirety:

Hammer's problem

Here's a link to the paper you're looking for:

F. D. Hammer and William Knight, Solution to problem 5955, The American Mathematical Monthly Vol. 82, No. 4 (Apr., 1975), pp. 415-416. MR1537708.

The solution given in that paper is very simple and elegant:

Let $\tilde{g}: \left[-\frac{1}{2},\frac{1}{2}\right] \to \mathbb{R}$ be the function $\tilde{g}(x) = x(1-4x^2)$. Its $1$-periodic extension $g$ to all of $\mathbb{R}$ is a $C^1$-function that is zero at the integers and whose derivative at the integers is $1$.

Now let $$f(x) = \sum_{n=0}^{\infty} \frac{g(n!x)}{(n!)^2}.$$

It is straightforward to check that $f\in C^1$ [let $f_k \in C^1$ be the $k$th partial sum of the series. Then $f_k \to f$ pointwise and $f_{k}^\prime$ is uniformly Cauchy, so $f \in C^1$]. Now notice that for rational $x$ only finitely many summands of $f$ are non-zero, hence $f(\mathbb{Q}) \subset \mathbb{Q}$. On the other hand, for rational $x$ we have $$f'(x) - e = f'(x) - \sum_{n=1}^{\infty} \frac{1}{n!} = \text{finitely many rational terms} \in \mathbb{Q},$$ so $f'(x) \notin \mathbb{Q}$, as desired.

(Thanks to robjohn for pointing out a mistake in the previous version of this answer)

share|cite|improve this answer
$f'-e\in\mathbb{Q}$ on $\mathbb{Q}$. – robjohn Oct 13 '11 at 23:25
Thanks! Fixed... – t.b. Oct 13 '11 at 23:32
O= I Know the theorem that if the sequence of continuous functions, converges uniformly then the limit is also continuous. But there you are proving that f is $ C^1 $ what theorem is that? – August Oct 14 '11 at 2:03
@August: I don't know its name. But here's the statement and proof: Let $f$ and $g$ be continuous, $f_n \in C^1$ on an interval. If $f_n \to f$ pointwise and $f_{n}^\prime \to g$ uniformly then $f \in C^1$ and $f^\prime = g$. To see this, write $$f_{n}(x) = f_{n}(a) + \int_{a}^{x} f_{n}^\prime(t)\,dt.$$ Passing to the limit we get $$f(x) = f(a) + \int_{a}^{x} g(t)\,dt$$ where interchange of integration and limit is justified by uniform convergence. By the fundamental theorem of calculus, the derivative of the right hand side exists and is $g(x)$, so $f' = g$ is continuous – t.b. Oct 14 '11 at 8:19

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.