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

Determine all functions $f \colon\mathbb{R}\to\mathbb{R}$ satisfying the following two conditions:

(a) $f(x +y) + f(x - y) = 2 f(x) f(y)$ for all $x, y\in\mathbb{R}$;

(b) $\lim\limits_{x\to\infty}f(x) = 0$.

I found this problem in IMO 1985 longlist. I have been able to figure out that there is one solution of the form $f(x) = 0$ for all $x$.

Any other function satisfying the above has to have $f(0) = 1$ and $f(x) = f(-x)$.

However I don't know how to proceed.

Any thoughts?

share|cite|improve this question
Why is it that $f(0)=1$ if $f\neq0$? – miniBill Jun 14 '12 at 9:26
@miniBill: If $f\neq 0$ choose $x$ with $f(x)\neq 0$. Then take $y=0$ and apply the cancellation law to equation (a). – Shane O Rourke Jun 14 '12 at 11:25
Understood, thanks :) – miniBill Jun 14 '12 at 15:00
up vote 12 down vote accepted

For any $x_0,y\in\mathbb{R}$, let $x=y+x_0$, then we have



$f(x_0) = 2f(y+x_0)f(y)-f(2y+x_0),\quad \forall y\in\mathbb{R}$


$\displaystyle f(x_0) = \lim_{y\to\infty}\left[f(x_0)\right] = \lim_{y\to\infty}\left[2f(y+x_0)f(y)-f(2y+x_0)\right] = 0$.

share|cite|improve this answer
This argument generalizes much more easily than mine. – Erick Wong Jun 14 '12 at 5:24
I don't understand why you can say that $f(x_0)=\lim_{y\to\infty}\left[...\right]$ – miniBill Jun 14 '12 at 9:15
@miniBill The second to last line is true for all $y$ (in $\mathbb R$), so you can take the limit of $y$ to infinity and get the same value, but we know that value is also $0$. – Mark Hurd Jun 14 '12 at 10:19
I'm still not convinced that $f(x) = g(x,y) \forall y$ implies $f(x) = \lim_{y\to\infty}g(x,y)$ – miniBill Jun 14 '12 at 10:44
@miniBill You just take the limit of both sides. You don't need anything for that - in the end both sides are equal. Then on the LHS the sequence is constant and on the RHS we may use the assumption. – Simon Markett Jun 14 '12 at 10:47

Using $x=y$, we have $f(2x) = 2f(x)^2 - 1$. Pick $X>0$ large enough so that $|f(x)| < 1/2$ for all $x \ge X$. Then $|f(2X)| > 1/2$, a contradiction.

share|cite|improve this answer
Your argument is a nice one! – user17762 Jun 14 '12 at 5:30
With $y=x$, the condition is $f(2x)+f(0)=2f(x)^2$, not what you wrote. Taking $x\to\infty$ in the correct condition, we find that $f(0)=0$. – Jonathan Sep 5 '12 at 23:21
@Jonathan Did you not notice that $f(0)=1$? – Erick Wong Sep 6 '12 at 1:10
That's not a condition of the problem. You can show that $f(0)=f(0)^2$ so either $f(0)=0$ or $f(0)=1$. The limit condition rules out the possibility of $f(0)=1$. – Jonathan Sep 6 '12 at 1:43
@Jonathan Now you're just being deliberately obtuse. You can just as easily show that $f(x) = f(x) f(0)$ so either $f$ is identically zero or $f(0)=1$. Notice the OP already deduced this. – Erick Wong Sep 6 '12 at 4:27

This is really a comment, not a solution, but I don't have that option. If one only considers the functional equation (a), then there are lots of interesting solutions, namely $\cos ax$ and $\cosh ax$. If one "cheats" and uses imaginary arguments, one can combine these in one formula. By the standard method of inserting a power series and comparing coefficients, one can show that these are the only analytic solutions. Presumably, minimal smoothness conditions on a solution would imply analyticity. However, there are pathological non-measurable solutions. Of course, these solutions are not relevant for the original question since they do not vanish at infinity.

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.