Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

I want to prove the following.

We have a function $f: \mathbb{Z} \to\mathbb{R}$ s.t.

(1) $f(mn) = f(m)f(n)$

(2) $f(m+n) \leq f(m) + f(n)$

(3) $0 \leq f(x) \leq 1$

then $f(m+n) \leq \max\big(f(m), f(n)\big)$?

I tried to use $f(m+n)^k$ and use (1)-(3), but I could not show the inequality.

Please help me.

share|cite|improve this question
Maybe of use: The function $f(n) = \frac{1}{|n|}$ satisfies all your conditions. – Lieven Nov 17 '12 at 15:39

Firstly, note that $f(0)=f(0)^2$ and $f(1)=f(1)^2$, i.e. $f(0),f(1)=0$ or $1$. If $f(0)=1$, then $1=f(0)=f(n)f(0)=f(n)$ for every $n\in\mathbb{Z}$; if $f(1)=0$, then $0=f(n)f(1)=f(n)$ for every $n\in\mathbb{Z}$. For both cases, the conclusion is trivially true. Therefore, we only need to focus on the case $f(0)=0$ and $f(1)=1$.

Note that by (1) and (3), we have (4): if $m\mid n$, then $f(n)\le f(m)$. Now given any pair of prime numbers $p\ne q$, according to Fermat's little theorem, $q\mid (p^{r(q-1)}-1)$ for every $r\ge 1$. Therefore, by (1), (2) and (4):

$$1=f(1)\le f(p^{r(q-1)})+f(1-p^{r(q-1)})\le f(p)^r+f(q).$$ If $0\le f(p)<1$, letting $r\to\infty$, we can conclude that $f(q)=1$. Since $p\ne q$ are arbitrary, it implies that either for each prime number $p$, $f(p)=1$; or there exists a unique prime number $p$, $0\le f(p)<1$ and for every prime number $q\ne p$, $f(q)=1$.

Therefore, either for every $n\ne 0$, $f(n)=1$; or there exists a prime number $p$, such that $0\le f(p)<1$, and if $n=p^rm$ with $p \nmid m$, then $f(n)=f(p)^r$. The conclusion follows from this immediately.

share|cite|improve this answer

This is not a finished solution, but maybe a stepping stone?

From 1) we get $f(m\cdot 1)=f(m)f(1)$, so either $f(1)=1$ or $f(n)=0$ for all $n$.

Assume $f(1)=1$, then 2) gives $f(n)\leq f(1+...+1)=f(1)+...+f(1)\leq n$.

In particular, $f(0)\leq f(0)+f(0)$ so $f(0)\leq 0$. Condition 3) then forces $f(0)=0$.

We also have $1=f(1)=f(-1)f(-1)$ giving $f(-1)=1$ and $f(-n)=f(n)$ for all $n$.

Now, $f(m)^2+f(n)^2\geq f(m^2-n^2)=f(m+n)f(m-n)$. Assuming $f(m-n)\neq 0$,

$f(m+n)\leq \frac{f(m)^2+f(n)^2}{f(m-n)}$.

Maybe this can be massaged into what you are after?

share|cite|improve this answer
From your third line. Condition 2) is an inequality, not an equality. – Mahtsolver11 Nov 17 '12 at 14:46
@espen180 wouldn't f(n) be bounded by n instead of equal to n? – TenaliRaman Nov 17 '12 at 14:48
Yes, you are right. I misread the question. – Espen Nielsen Nov 17 '12 at 14:51

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.