This is completely for my own curiosity.

Does $y = \sin(n)$ have rational solutions for $n$, an integer number of radians.

I know that this is strange because usually integers are only used in degrees. I know that there will be convergent sub series of $n$'s to any rational number in $[-1,1]$. Would we ever get an algebraic integer? It would have something like $$ e^\left(n i\right)=e^\left(\frac{n}{\pi}\pi i\right) $$

I suspect that it will never have a rational solution but I don't know how to prove it.


As @lulu states in the comments of the question, the Lindermann-Weierstrass Theorem is a theorem that goes along these lines:

If I have $\sin(n)$ where $n$ is algebraic e.g. it is a solution to a polynomial with rational coefficients, then $\sin(n)$ is transcendental.

And, furthermore, transcendental numbers are not rational.

However, I can tell you that $\sin(\pi n)$ is algebraic, but usually not rational.

We know this because $\sin(\pi n)$ can be rewritten as a polynomial $P_n(\sin(\pi))$ where the degree of this polynomial is $n$ and the coefficients are all whole numbers.

For example:

$$\sin(\pi 0)=0$$

$$\sin(\pi 1)=0$$


And so on. But, through the polynomial method, we also know that if $n$ is rational, then the result is algebraic.

$$\sin(\pi \frac12)=1$$



And that is indeed algebraic. (It's just very tedious to calculate.)


My mistake, we can only get algebraic solutions to $\sin(\pi n)$ if $n=\frac pq$ where $q=2^a3^b$ and $a,b$ are whole numbers.

This is a direct result of the inability to find the inverse of a polynomial of degree $n>4$.


I messed up with the last example of $\cos(\pi\frac72)$, it is rather tedious and prone to mistakes. But you can apply trig identities ($\cos(\pi\frac62+\pi\frac12)$) seven times.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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