Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have the second edition of Spivak. Consider

enter image description here

Can someone tell me why he considers $2n|a_{n-1}| \dots$? Later he shows everything is squeezed between -1/2 and 1/2 and he gets the desired result. I am perplexed as to why $2n$? He could have just done it wit $n$ and it would work out.

Also he never considered the case for $|x| < 1$, what happens then?

EDIT: On another matter, why does he write $|x| > 1, 2n|a_{n-1}| \dots$ instead of just writing $|x| > \max \{1, 2n|a_{n-1}|, \dots \}$?

EDIT2:What is more interesting is that he assumes that $\dfrac{|a_{n-k}|}{|x|} < \dfrac{|a_{n-k}|}{2n|a_{n-k}|}$. Is there a reason why he it will not be just $\dfrac{|a_{n-k}|}{|x|} < |a_{n-k}|$ and why does he even consider $|x| > 1$, why not $|x| > 0$?

share|cite|improve this question
The factor $2$ doesn't seem necessary, given the sharp inequalities involved (I mean $<$ instead of $\le$). As for $|x|<1$, he does not need to look at that because all he's interested in is showing that the polynomial takes both positive and negative values somewhere. And he finds those for large $|x|$. – Harald Hanche-Olsen May 4 '13 at 19:22
Oh so he is considering a root, not $n$ number (different $n$ in my context) of root(s) that happens to lie $x > 1$ – Hawk May 4 '13 at 19:24

The point is not just to ensure that $|1+\sum_{k=1}^{n}a_{n-k}/x^k|$ is positive but that it is actually bounded away from zero, independently of $x$. Actually, the factor $2$ is overkill, but it makes the analysis quick and easy. Harald's comment answers your second question.

Edit: (In response to the OP's query) Yes, it's sufficient to take $x\ge n|a_k|+\varepsilon\;(k=1,\dots,n)$, along with $x>1$, where $\varepsilon>0$ is independent of $x$. Then the multiplying factor $|1+\sum_{k=1}^{n}a_{n-k}/x^k|$ exceeds a constant, which can be written in terms of $\varepsilon$ and the $a_k$, which can be seen to be positive. But why bother when you can just write $\frac12$?

share|cite|improve this answer
Is my reply correct? – Hawk May 4 '13 at 19:43
@sizz: See my edit. – John Bentin May 4 '13 at 21:30

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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