I'm trying to understand a certain way of proving that

$$(1+x)^\alpha = \sum_{n=0}^{\infty}\binom{\alpha}{n}x^n$$

for $\alpha \in \mathbb{Q}$ and $|x|<1$.

Now, my notes tackle that proof by saying that for $\alpha, \beta \in \mathbb{R}$:

$$\sum_{n=0}^{\infty}\binom{\alpha}{n}z^n \cdot \sum_{n=0}^{\infty}\binom{\beta}{n}z^n = \sum_{n=0}^{\infty}c_nz^n$$

with $$c_n =\sum_{k=0}^{\infty}\binom{\alpha}{k}\binom{\beta}{n-k}=\binom{\alpha + \beta}{n} $$

and proceeding by proving that by induction.

Now, I can't even understand why proving the second notion would prove the initial theorem, so if anybody could explain or provide further material since I can't find anything about that particular way of proof, I'd be very thankful.

  • 2
    $\begingroup$ I suppose that, when you wrote $(1+x)^\alpha=\sum_{n=0}^{\infty}\binom{n}{k}x^n$, what you meant was $(1+x)^\alpha=\sum_{n=0}^{\infty}\binom\alpha nx^n$. $\endgroup$ Aug 28 '19 at 18:38
  • $\begingroup$ Note your original series should have $\binom{\alpha}{n}$ instead of $\binom{n}{k}$. The indices in your other series are also highly suspect. $\endgroup$
    – MPW
    Aug 28 '19 at 18:39
  • $\begingroup$ $f_a(x) = \sum_{n=0}^\infty {a \choose n}x^n$ from your formula you have $f_{ma}(x) f_a(x) = f_{(m+1)a}(x)$ and this is what you need to obtain $f_{u/v}(x)^v = f_u(x) = (1+x)^u$ (where $u,v\in \Bbb{Z}_{\ge 1}$ so that $f_u(x) = (1+x)^u$ is easy to show) $\endgroup$
    – reuns
    Aug 28 '19 at 18:43
  • $\begingroup$ Sorry, just realized that I had mistakes in my notes, changed them to fit the original script. $\endgroup$
    – psyph
    Aug 28 '19 at 19:03
  • $\begingroup$ This is one of lesser known proofs of binomial theorem. The proof with all details is available in my post: paramanands.blogspot.com/2016/07/… $\endgroup$ Sep 2 '19 at 15:31

My guess is that what you wrote on the second paragraph is not meant to be a proof of the fact that $$(1+x)^\alpha=\sum_{n=0}^{\infty}\binom\alpha nx^n,\tag1$$but a proof of the fact that, if you define $(1+x)^\alpha$ by $(1)$, then$$(1+x)^\alpha(1+x)^\beta=(1+x)^{\alpha+\beta},\tag2$$since what it says there is precisely what is needed to prove that $(2)$ holds.

  • $\begingroup$ I have just edited the question to match the original script, thanks for pointing out the mistakes. And you're right, (2) is there to prove a property that is later used to prove that $\alpha$ can be rational. $\endgroup$
    – psyph
    Aug 28 '19 at 19:17
  • $\begingroup$ I'm glad I could help. $\endgroup$ Aug 28 '19 at 19:23

Here's an alternate proof, if you're interested.

One may show that $$p_n(\alpha)=\left[\frac{\partial}{\partial x}x^\alpha\right]_{x=1}=\prod_{r=1}^{n}(\alpha-r+1).$$ So then the Taylor series for $x^\alpha$ about $x=1$ is given by $$x^\alpha=\sum_{k\ge0}\frac{p_k(\alpha)}{k!}(x-1)^k.$$ We notice that $${n\choose k}=\prod_{i=1}^n\frac{n+1-i}{i}.$$ So we have $$(1+x)^\alpha=\sum_{k\ge0}{\alpha\choose k}x^k.$$


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.