3

I have to prove the following for my math study:

Prove: $e^x$ is transcendental over the polynomials with coefficients in $\mathbb{R}$.

So far, I've done this:

It's enough to prove that if $a_n(x),\dots,a_1(x), a_0(x)$ are polynomials with coefficients in $\mathbb{R}$ such that $a_n(x)e^{nx} + a_{n-1}(x)e^{(n - 1)x} + \dots + a_1(x)e^x + a_0(x) = 0$, then $a_i(x) = 0$ $\forall i$
To prove this I assumed that $\exists i : a_i(x) \neq 0$. Then take the smallest $n$ such that $a_n(x)e^{nx} + a_{n-1}(x)e^{(n - 1)x} + ... + a_1(x)e^x + a_0(x) = 0$.

I don't know how I have to complete the proof from here. Could you explain it to me?

egreg
  • 238,574
Peter
  • 2,122

3 Answers3

3

Clearly, as $n>0$, $$ \lim_{x\to-\infty}-a_0(x)= \lim_{x\to-\infty} \bigl(a_n(x)e^{nx} + a_{n-1}(x)e^{(n - 1)x} + ... + a_1(x)e^x\bigr)= 0 $$ So…

(Note: the only needed property is $\lim_{x\to-\infty}x^he^{kx}=0$ for $h\ge0$, $k>0$.)

egreg
  • 238,574
1

You can do it, well, by using calculus.

Suppose $$\sum_{i=0}^{n}a_i(x)e^{ix}=0$$

with $n$ being smallest possible, i.e. $a_n(x)\neq 0$.

In particular, $$\lim_{x\rightarrow \infty}a_n(x) \neq 0.$$

Then it follows that for any $\alpha \in \mathbb{R},$

$$\lim_{x \rightarrow \infty} \frac{\sum_{i=0}^{n}a_i(x)e^{ix}}{e^{\alpha x}}=0.$$

Also, for positive $\alpha$ and any polynomial $a(x)$, we have $$\lim_{x\rightarrow \infty}\frac{a(x)}{e^{\alpha x}}=0.$$

Putting $\alpha=n$, one has that $$0 \neq \lim_{x \rightarrow \infty}a_n(x)=\lim_{x \rightarrow \infty} \frac{\sum_{i=0}^{n}a_i(x)e^{ix}}{e^{\alpha x}}=0,$$

a contradiction.

1

An approach not using limits. Let $f(x) = \sum_{k=0}^n a_k(x)e^{kx}$ and call $a_0$ the lowest coefficient. Then $f'(x)$ has a similar form and only the the degree of the lowest coefficient decreases, other coefficients remain of the same degree. Take the derivative repeatedly until the lowest coefficient vanishes.

The same approach also works for the difference operator $(\Delta f)(x)=f(x+1)-f(x)$.

WimC
  • 32,192
  • 2
  • 48
  • 88
  • @ WimC Ok, I do understand what you mean. So if I say that the degree of the lowest coefficient is m, then that coefficient vanishes when I take the (m +1)th derative. But what contradiction should I get from that? – Peter Apr 26 '15 at 11:45
  • I do not understand what you are trying to say in your last sentence. I am supposed to get a contradiction from this, right? – Peter Apr 26 '15 at 11:49
  • If $a_0=0$ you can factor out $e^x$ and get a combination of degree one less. – WimC Apr 26 '15 at 11:49
  • Oh and then n was not the smallest degree. So that is the contradiction, right? – Peter Apr 26 '15 at 11:50