Calculate (and argue) the critical points of an exponential function

Click For Summary
The discussion focuses on the differentiation and domain of the function x^x, which is expressed as e^(x ln x). It is established that the domain is (0, +∞) because x^x is not defined for non-positive values, particularly due to issues with negative bases and fractional exponents. The participants debate the value of 0^0, with arguments supporting its definition as 1 for continuity purposes, while also acknowledging that this definition can vary based on the context of the functions involved. The limit of x^x as x approaches 0 from the positive side is shown to equal 1, reinforcing the argument for defining 0^0 as 1. The conversation highlights the complexities of defining functions and limits in calculus, particularly regarding continuity and the behavior of functions at critical points.
mcastillo356
Gold Member
Messages
642
Reaction score
350
Homework Statement
Calculate the critical point of ##y=x^x##
Relevant Equations
Chain Rule, Product Rule
Hi everybody

We can't differentiate ##x^x## neither like a power function nor an exponential function. But ##x^x=e^{x\mbox{ln}x}##. So

##\dfrac{d}{dx}x^x=\dfrac{d}{dx}e^{x\mbox{ln}x}=x^x(\mbox{ln}x+1)##

And here comes the doubt: prove the domain of ##x^x## is ##(0, +\infty)##

Why is only positive?; why it will never be 0?

So ##(\mbox{ln}x+1)=0\Rightarrow{x=1/e}##

It's a solved exercise at Calculus, by R.A.A. It's the only question not solved (the domain, I mean).
 
Physics news on Phys.org
You can define ##0^0## if you want, but fractional powers of negative numbers are undefined, or require complex numbers. E.g. ##-0.5^{-0.5}## is not a real number.
 
  • Informative
Likes mcastillo356
Can ##0^0=1##? The book says: ##x^x## is never 0. This is the graph:

geogebra-export.png

##0^x=0## on one hand; ##x^0=1##. This is all I can provide. :frown:
 
##0^0## is given a value per definition. If you want continuity of ##x \longrightarrow x^0## at ##x=0## then you have to set ##0^0=1,## and if you want continuity of ##x\longrightarrow 0^x## at ##x=0## then you have to set ##0^0=0.##

My personal opinion is, that only ##0^0=1## is a valid definition because exponentiation is an abbreviation for multiplication and all multiplications over an empty set of indexes equals the neutral element of multiplication, which is ##1.##
$$
\sum_{i\in \{\}} a_i = 0 \quad ;\quad \prod_{i\in \{\}} a_i = 1
$$
 
Last edited:
  • Informative
Likes mcastillo356
The next step would be to prove ##\displaystyle\lim_{x \to{0+}}{x^x}=f(0)=1##? I am just wondering.
 
mcastillo356 said:
The next step would be to prove ##\displaystyle\lim_{x \to{0+}}{x^x}=f(0)=1##? I am just wondering.
Yes. E.g. we get for ##x>0##
$$
\lim_{n \to \infty}\left(\dfrac{1}{n}\right)^{1/n}= \lim_{n \to \infty}\left(1-\dfrac{\log n}{n}+O(n^{-2})\right)=1.
$$
But if you decouple base and exponent, say ##f(x,y)=x^y,## then all depends on which path you approach ##(0,0).## That is where the question: "Why isn't ##0^0=0## ?" comes from. ##x\longmapsto x^x ## and ##(x,y)\longmapsto x^y## are different functions with different properties. And even ##f(x,y)## depends on its domain: real positive numbers, all real numbers and complex values, complex numbers and complex values. These are questions that should be answered before we discuss ##0^0.## The answer depends on them.
 
  • Like
Likes mcastillo356
##\displaystyle\lim_{x \to{0^+}}{e^{x\ln(x)}}=e^{\lim_{x \to{0^+}}{x\ln(x)}}=e^{\lim_{x \to{0^+}}{\frac{\ln(x)}{1/x}}}##, and, by L'Hopital, ##\displaystyle e^{\lim_{x \to{0^+}}{-x}}=1##.

Is this equal to ##f(0)##? Don't know. ##x^x=e^{x\ln(x)}##. Does this alow me to say "no", because ##\ln 0## is nonsense?.
 
mcastillo356 said:
##\displaystyle\lim_{x \to{0^+}}{e^{x\ln(x)}}=e^{\lim_{x \to{0^+}}{x\ln(x)}}=e^{\lim_{x \to{0^+}}{\frac{\ln(x)}{1/x}}}##, and, by L'Hopital, ##\displaystyle e^{\lim_{x \to{0^+}}{-x}}=1##.
This is correct.
mcastillo356 said:
Is this equal to ##f(0)##? Don't know. ##x^x=e^{x\ln(x)}##. Does this alow me to say "no", because ##\ln 0## is nonsense?.
What do you mean? "No" as an answer to what? And where do you have ##\log 0## if you use de L'Hôpital?
You could as well use a Taylor expansion of ##\log## and observe that ##y=x## goes faster to zero than ##y=\log x ## does so that the product goes to zero. Taylor will give you a quantitative measure for "faster than".
 
Last edited:
  • Informative
Likes mcastillo356
Another try:smile:

In the exercise I'm given a function ##f(x)=x^x##. We have ##x^x:=e^{x\ln x}##, which is well defined for ##x>0##, so the domain is ##(0,\infty)##.

Now, we can define another function ##[0,\infty)\rightarrow{\mathbb R}## such that

$$g(x):=\begin{cases}{f(x)}&\text{if}& x\in{(0,\infty)}\\1 & \text{if}& x=0\end{cases}$$

Where we assume ##0^0:=1##
 
  • #10
You can define ##g(x)## whichever you want. The question is: which properties do you require? If you do it like this, then we get a continuous extension of ##f(x)##.
 
  • Love
Likes mcastillo356
  • #11
Hi, fresh_42

fresh_42 said:
The question is: which properties do you require?
No idea. I'm afraid I let myself struggle with maths with lots of naivety. Sooo...PF is a great, great, help, and a place where I feel at home, an island of give-and-take knowledge.

But I think I generate expectations I'm not able to tackle...

Conclusion: Thanks for helping me one more time. Much love. Forgive me if this post is improper. I felt I had to contextualize the thread.
 
  • Like
Likes berkeman and fresh_42

Similar threads

  • · Replies 14 ·
Replies
14
Views
2K
Replies
7
Views
2K
  • · Replies 27 ·
Replies
27
Views
2K
  • · Replies 4 ·
Replies
4
Views
3K
Replies
8
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 14 ·
Replies
14
Views
3K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 14 ·
Replies
14
Views
1K
Replies
5
Views
2K