When we talk about Taylor series, we say it's around point $x_0$. It's in the Taylor series formula:

$$f(x) = f(x_0) + (x-x_0)f'(x_0) + \frac{(x-x_0)^2f''(x_0)}{2} + \frac{(x-x_0)^3f'''(x_0)}{6} + + ...$$

What I don't understand what is this point $x_0$?

We're given Taylor formulas for $\sin(x)$, $\cos(x)$, etc, and they don't have this point $x_0$ in the formula? It's somehow chosen to be $0$ so it disappears, but what does that mean?

Doesn't that mean the Taylor series works only around $0$? But then we somehow can use the formula for any $x$, even $100$ or $100,000$, and it still gives the right result?

What happens if we choose $x_0$ to be $1$? or $25$? That would change the Taylor formula for the primitive functions but would it still work for any $x$?

I'm very confused.


In Taylor's formula, we have to choose an initial point $x_0$, in which the sum actually converges to the function. In any other point the sum may converge or diverge, depending on the function.

For example, by evaluating Taylor's formula for $\sin(x)$, $\cos(x)$ and also $e^x$, it doesn't matter which point $x_0$ you choose, and the sum will converge everywhere to the function (if I am not mistaken, there is a theorem that if the sum converges at $x$, it converges to the value of the function at $x$).

However, in other functions (like $\ln(1+x)$), the sum may not converge always; Taylor's formula for $\ln(1+x)$ converges $\Leftrightarrow$ $|x|<1$ (and also $x=1$, if I remember correctly).

To sum up, you may choose any point $x_0$ you want; but it is usually convinient to choose it as $0$ (because it is easier to calculate the derivatives).

| cite | improve this answer | |
  • 2
    $\begingroup$ Even if a Taylor series converges everywhere, the limit can be different from the given function. $\endgroup$ – Martin R Jan 22 '15 at 18:24
  • $\begingroup$ Very interesting. $\endgroup$ – bodacydo Jan 22 '15 at 18:30
  • $\begingroup$ However if the Taylor series around some $x_0$ converges to the function everywhere on $\mathbb R$, then the Taylor series around any other point will also converge to the function everywhere. $\endgroup$ – hmakholm left over Monica Jan 22 '15 at 18:33
  • $\begingroup$ Weird and interesting as f*ck. Thanks. $\endgroup$ – bodacydo Jan 23 '15 at 7:34

You can develop a Taylor series of a function at any point of the domain/interval where it is nice enough (continuously n-differentiable...). The formulae well known for sine cosine exp are at $x_0=0$ but you could using the general formula get a Taylor expansion of sine at $x_0=\pi$ for example or any other real by the way.

| cite | improve this answer | |
  • $\begingroup$ I'm sorry, I don't understand your answer. You didn't answer the questions that I've. $\endgroup$ – bodacydo Jan 22 '15 at 18:28
  • $\begingroup$ I meant that the Taylor formula is the general one you wrote down involving derivatives of $f$ and factorials. You could develop sine for example anywhere and this would give a different expansion from the well known expansion at zero $x-x^3/3+\ldots$ $\endgroup$ – marwalix Jan 22 '15 at 18:32
  • $\begingroup$ Thanks. But what does it mean to have a different expansion? Will it still produce the same values for sin(x)? Or they will be shifted by $x_0$? Or does it mean the same function can have many different expansions that all produce the same values? $\endgroup$ – bodacydo Jan 22 '15 at 18:36
  • $\begingroup$ The function is unchanged it is the expansion meaning the formula that is different. Value at a given point is unique but two different formulaes give that value $\endgroup$ – marwalix Jan 22 '15 at 18:44

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.