So I have been studying the convergence of Taylor series from Tom M Apostol. There is a theorem which states the sufficient condition for convergence of Taylor series. Quoting the theorem we've

Assume $f$ is infinitely differentiable in an open interval $I = (a - r, a + r)$, and assume that there is a positive constant $A$ such that $$ |f^{(n)}(x)| \le A^n$$ for $n = 1, 2, 3, \dots$ and every $x$ in $I$. Then the Taylor's series generated by $f$ at $a$ converges to $f(x)$ for each $x$ in $I$.

My question is how to arrive at $|f^{(n)}(x)| \le A^n$ deductively ? I tried looking it up on the internet but haven't found anything that solves my issue.


  • $\begingroup$ Do you mean how to deduce that such a bound would be a sufficient condition for convergence? It is a strange question asking how to deduce a hypothesis. Hypotheses are starting points. One could guess, or motivate a hypothesis. In the case of necessary and sufficient conditions one could deduce it, but this one is only a sufficient condition. $\endgroup$ – logarithm May 13 '19 at 13:13

The assertion $$ |f^n(x)| \le A^n $$ is a hypothesis in this theorem. You can't "arrive at it deductively" in general. When it happens to be true for any particular function $f$ then you can conclude something about the Taylor series for that function.

For example, the $n$th derivatives of $\sin$ are $\sin$ and $\cos$. Their absolute values are always at most $1$. This theorem then implies that the Taylor series for $\sin$ converges everywhere to the value of the function.

| cite | improve this answer | |

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.