How do you prove that ln(a^x) = xln(a) and a^x = e^xln(a) without using exponent rules?

Click For Summary
SUMMARY

The discussion centers on proving the identities ln(a^x) = xln(a) and a^x = e^(xln(a)) without relying on exponent rules. Participants emphasize defining a^x for real numbers using continuity and limits, particularly through the lens of rational numbers and their density in real numbers. The proof involves using the properties of logarithms and the relationship between exponential and logarithmic functions, specifically leveraging the definitions of e^x and ln(a) through power series and functional equations.

PREREQUISITES
  • Understanding of limits and continuity in calculus
  • Familiarity with logarithmic and exponential functions
  • Knowledge of power series and their applications
  • Basic concepts of algebraic identities involving exponents
NEXT STEPS
  • Study the properties of logarithms, particularly the product and power rules
  • Explore the definitions and properties of power series for e^x and ln(x)
  • Learn about the continuity of functions and its implications in calculus
  • Investigate the density of rational numbers in real numbers and its significance in proofs
USEFUL FOR

Mathematicians, calculus students, educators, and anyone interested in the foundational proofs of logarithmic and exponential identities.

logicgate
Messages
13
Reaction score
2
Homework Statement
The proof must contain all real numbers x, not just for integers, it must be true for rational and irrational numbers.
Relevant Equations
ln(a^x) = xln(a), a^x = e^xln(a)
In the book "Calculus by Michael Spivak" it says that a^x = e^xln(a) is a definition. And I am not convinced to accept this as true without a proof.
 
Physics news on Phys.org
You want to define ##a^x## for any real number ##x##. I suppose we can assume ##a\geq 0## for otherwise, we would get into trouble already with rational values of ##x.## I know what ##a^z## for ##z\in \mathbb{Z}## is since we can define it recursively. That allows us to define ##a^{1/n}=\sqrt[n]{a}## since an ##n## long product of them yields ##1.## We get therefore the definition for all ##x\in \mathbb{Q}.##

Now, it gets complicated. The real numbers are limits of certain sequences of rational numbers. Say we have ##x=\displaystyle{\lim_{n \to \infty}}x_n## and we require that ##x \mapsto a^x## is continuous. We need this condition so that we can define ##a^x= \displaystyle{\lim_{n \to \infty}}a^{x_n}.## The requirement makes sense if we look at the drawings of ##x_n \mapsto a^{x_n}.## However, it is an artificial requirement in the sense that we can only get from the rational to the reals if we accept a topological homomorphism, a continuous function.

Thus, we have defined ##a^x## by algebraic and topological means, that mimic the process from natural numbers over integers and rational numbers to real numbers.

So far so good. Now, what are ##x\mapsto e^x## and ##a\mapsto \ln(a)##? You have to define them before we prove the statement. For example, their definition by functional equations is possibly less helpful than their definition by power series or limits. So what are they to you? In order to prove something, we must agree on what we are allowed to use.
 
  • Like
Likes   Reactions: e_jane
use power series
 
Let us use that ##e^{\ln(x)}=x## or prove it with your power series. With that, we only have to show that ##x\cdot \ln a = \ln a^x## and the second equation follows. We can also assume that ##x>0## since the negative case ...
\begin{align*}
\ln a^{-x}&=\ln \dfrac{1}{a^x}=\ln \left(\dfrac{1}{a}\right)^x =x\ln \left(\dfrac{1}{a}\right)\\
&=x\cdot \left(\ln 1 -\ln a\right)=x\cdot(0-\ln a)=-x\ln a
\end{align*}
... follows from the positive case and ##x=0## is trivial. I would suggest that you use
$$\ln a^x=\lim _{h\to 0}\int _{1}^{a^x}{\frac {1}{t^{1-h}}}\,\mathrm {d} t$$
as definition and substitute ##s=t^{1/x}.##
 
How did you arrive to the conclusion that ln(1/a)^x = xln(1/a) ? This is exactly what I want to prove.
 
I did not arrive there. I proposed possible paths. You could prove it for rationals and use continuity and the fact that ##\mathbb{Q}\subseteq \mathbb{R}## is dense, or you can use the integral definition and the substitution I suggested.

However, you should do some of the steps on your own.
 
  • Like
Likes   Reactions: e_jane and PeroK
By definition:
$$b^n = \underbrace{b \times b\ ... \times b}_{n \text{ times}}$$
Therefore we get our first identity:
$$b^{n+m} = \underbrace{b \times b\ ... \times b}_{n \text{ times}}\times\underbrace{b \times b\ ... \times b}_{m \text{ times}}$$
$$b^{n+m} = b^n b^m$$
Similarly, we can get our second identity:
$$\left({b^n}\right)^m = \underbrace{\underbrace{b \times b\ ... \times b}_{n \text{ times}}\times\underbrace{b \times b\ ... \times b}_{n \text{ times}}\ ... \times \underbrace{b \times b\ ... \times b}_{n \text{ times}}}_{m \text{ times}}$$
$$\left({b^n}\right)^m = b^{nm}$$
About how this applies to rational and irrational numbers you can read the Wikipedia page about exponentiation starting here.

If you accept that and we define that ##y = \log_b x## if ##b^y = x##, then by our first identity:
$$b^{y_1}b^{y_2} = b^{y_1 + y_2}$$
$$x_1x_2 = b^{y_1 + y_2}$$
$$\log_b(x_1x_2) = y_1+y_2$$
$$\log_b(x_1x_2) = \log_b x_1+\log_b x_2$$
From there, if ##x_1 = x_2 = a##, then:
$$\log_b(a^2) = \log_b a+\log_b a = 2 \log_b a$$
And for ##\log_b(a^3)##, we get:
$$\log_b(a^2\times a) = 2\log_b a+\log_b a = 3 \log_b a$$
Or for the general case:
$$\log_b(a^x) = x \log_b a$$
As for the other case, we used the second identity:
$$a^x = \left(a\right)^x$$
$$a^x = \left(e^{ln(a)}\right)^x$$
$$a^x = e^{xln(a)}$$
 

Similar threads

Replies
6
Views
4K
  • · Replies 5 ·
Replies
5
Views
2K
Replies
8
Views
5K
Replies
5
Views
1K
Replies
6
Views
2K
  • · Replies 4 ·
Replies
4
Views
3K
Replies
9
Views
2K
Replies
7
Views
1K
  • · Replies 4 ·
Replies
4
Views
1K
  • · Replies 4 ·
Replies
4
Views
2K