Nth Root of N
During my son’s math lesson today we got on the subject of ![\sqrt[n]n](http://nealabq.com/blog/wp-content/ql-cache/quicklatex.com-d65fbdb0addac84526184adf1d999070_l3.png "Rendered by QuickLaTeX.com")
, which I prefer to write as 
. We made a table of a few obvious values and limits:
![\[ \begin{tabular}{|r|c|} \hline $n$ & $n^{1/n}$\\ \hline $0\leftarrow$ & $0\leftarrow$ \\ 1 & 1 \\ 2 & $\sqrt{2} \approx 1.414$ \\ 3 & $\sqrt[3]{3} \approx 1.442$ \\ 4 & $\sqrt{2} \approx 1.414$ \\ $\rightarrow \infty$ & $\rightarrow 1$ \\ \hline \end{tabular}\]](http://nealabq.com/blog/wp-content/ql-cache/quicklatex.com-5833bf7173749e7ffbea0a7cf086167f_l3.png "Rendered by QuickLaTeX.com")
, peak somewhere in 
, and asymptotically drop to 1 after that. So the maximum is probably between 2 and 4, and I asked my son how we could find it. And he suggested we find where the curve is flat and the derivative is zero.
And that’s when I realized I’d forgotten know how to find 
. My son is only 11 and we’ve only touched on calculus, so he couldn’t help. I figure the first step is to express it as 
, but I don’t know where to go from there.
So we turn to the internet, and I read Wikipedia’s entry for the nth root algorithm, which doesn’t answer my question but is nice to read because it’s a little gem. And then we get to WolframAlpha, which tells us everything we want to know. See for yourself, the WolframAlpha page for , which says the derivative is 
, which is zero when 
or 
or 
.
So the maximum of is ![\sqrt[e]e\approx 1.4447](http://nealabq.com/blog/wp-content/ql-cache/quicklatex.com-8fd997b3d491aa3fbc0fe242b9354e0e_l3.png "Rendered by QuickLaTeX.com")
. Not very surprising, but nice to have it confirmed. (When I described this problem to my neighbor he said the max would be at 
as soon as he heard it was between 2 and 4.)
But I still don’t know how to calculate that derivative, except by asking WolframAlpha. Which brings up the question, is the internet making me stupid by making things easy? Or is it making me smarter by taking care of drudgery? I’d like to say it makes me smarter, but I know that, once in a while, what appears to be drudgery at first turns out to be important, providing unexpected insight.
Powered by QuickLaTeX.com” and QuickLaTeX as a WordPress plugin.
Comments
11 Responses to “Nth Root of N”
Hello!
Thank you for using QuickLaTeX!
One small tip – you can place latex code inside ${-}$!..${-}$ tags if you want formula to be centered on the page horizontally (use it without “{-}” characters). QuickLaTeX will do that automatically. Like here:
$$!
S=\sqrt{p(p-a)(p-b)(p-c)}
$$
Thanks Pavel, great advice. I just edited the above post and centered the table. Works like a charm.
In wolframalpha, just go down to the derivative section where it shows the solution and click on ‘show steps’ and then wolframalpha will show you how to use the chain rule
Thanks Pat, I didn’t notice the ‘show steps’ link before. The derivation is very clear.
My friend and I were arguing about this! Now I know that I was right. lol! Thanks for making me sure!
Sent from my Android phone
Hi Neal,
Another essentially equivalent way to make this derivation is to use the chain rule on the logarithm first: if f=f(x) is any positive function with a derivative, then d/dx(log f)=(1/f)f’(x). Thus f’(x)=f(x)*d/dx(log f).
Even though this is basically equivalent to what they do at wolframalpha, this can sometimes be easier than using the chain rule: for expressions like x^{something} this method is basically a product-rule-type derivative instead of a chain-rule-type derivative. Sometimes one of these types is easier to work with than the other. It’s also a great exercise for young people, your son, say, to familiarize themselves with the logarithm, which is one of the most ubiquitous functions in all of mathematics. One might also ask questions like, what is the limit as x goes to zero of x^x or x^{1/x}? Since one has no direct way to compute such a limit easily, one uses your idea that x=e^{log x}, that one may move a limit inside the argument of a continuous function, and that one can use L’Hospital’s rule to compute the limit of expressions like xlogx and the like. Anyway the moral is that it’s good to keep the logarithm at arm’s reach when you’re doing calculus problems!
Regards,
Peter
Hayabusa, glad to help!
Peter, wow, excellent comment. I’ve read it 4 times so far, and I think I get it now. Thanks for taking the time.
L’auteur de ce jour dans notre magasin pour acheter un latexcatsuits pièce.
You might enjoy what I call the Whammo Function, W(x), so called for its behavior for x 0. On the reals you can define it using logarithms, as usual.
I ran across a problem in my calculus book,
which asks to find the supremum of “nth root of n”
for natural values. I think the upper bound in this case should be 3^(1/3),am i right ?
Great internet site! I am caring this!! Will come back once more.
Im bookmarking your feeds also