# Thread: Definition of Natural Logarithmic Function

1. ## Definition of Natural Logarithmic Function

My book defines this function in the provided attachment. My confusion is why is this true? No where in my book does it explain it with scarcely any adequacy.

2. ## Re: Definition of Natural Logarithmic Function

It's better to define \displaystyle \begin{align*} \ln{x} \end{align*} as the inverse function of \displaystyle \begin{align*} e^x \end{align*}. Since the range of \displaystyle \begin{align*} e^x \end{align*} is \displaystyle \begin{align*} (0, \infty) \end{align*}, that means the domain of \displaystyle \begin{align*} \ln{x} \end{align*} is \displaystyle \begin{align*} (0, \infty) \end{align*}

It follows therefore, that

\displaystyle \begin{align*} y &= \ln{x} \\ e^y &= x \\ \frac{d}{dx}\left(e^y\right) &= \frac{d}{dx}(x) \\ \frac{d}{dy}\left(e^y\right)\frac{dy}{dx} &= 1 \\ e^y\,\frac{dy}{dx} &= 1 \\ \frac{dy}{dx} &= \frac{1}{e^y} \\ \frac{dy}{dx} &= \frac{1}{x} \end{align*}

It follows then that if \displaystyle \begin{align*} x > 0 \end{align*}, then \displaystyle \begin{align*} \int{\frac{1}{x}\,dx} = \ln{x} + C \end{align*}.

Therefore your definition follows from the Fundamental Theorem of Calculus.

3. ## Re: Definition of Natural Logarithmic Function

Well, Prove It, I graphed both y = ln(x) and e^y=x, and they don't appear to be inverse functions--that is, they are not symmetrical with respect to the identity function, y = x.

4. ## Re: Definition of Natural Logarithmic Function

Oh, wait, I believe I have figured it out: you wanted those two functions to equal each other, correct? So, that knowing what one is equal, we may be able to infer what the other is equal to.

5. ## Re: Definition of Natural Logarithmic Function

Originally Posted by Bashyboy
Well, Prove It, I graphed both y = ln(x) and e^y=x, and they don't appear to be inverse functions--that is, they are not symmetrical with respect to the identity function, y = x.
Of course they're not inverse functions, what you have graphed are the exact same function, as \displaystyle \begin{align*} y = \ln{x} \end{align*} is equivalent to \displaystyle \begin{align*} x = e^y \end{align*}.

What you should be graphing are \displaystyle \begin{align*} y = e^x \end{align*} and \displaystyle \begin{align*} y = \ln{x} \end{align*}.

6. ## Re: Definition of Natural Logarithmic Function

Originally Posted by Bashyboy
My book defines this function in the provided attachment. My confusion is why is this true?
Why is what true? This is just a definition, not a proposition. I am pretty sure your book later derives various properties of ln(x) from this definition.

7. ## Re: Definition of Natural Logarithmic Function

Emakarov, my confusion came about because I wondered why they were able to define it this way, what was their justification.

8. ## Re: Definition of Natural Logarithmic Function

In mathematics, one does not need a justification for definitions. One is free to define whatever they like as long as the objects are indeed "well-defined." (This may warrant a separate discussion.) The only possible flaw with a definition is that other mathematicians may not find it interesting and may, e.g., refuse to accept a conference paper about it, but this is not a properly mathematical issue.

I am not sure why in this case the authors decided to define logarithm this way. Maybe they found that proving properties of logarithm is easier this way.

9. ## Re: Definition of Natural Logarithmic Function

Yes, ln(x) can be defined as the inverse of $e^x$ but I, personally, don't agree with Prove It that that definition is "better".

For one thing, we can see immediately from this definition that ln(x) is differentiable for all for all positive x and the derivative is 1/x while finding the derivative of $e^x$ involves showing that $\lim_{h\to 0}\frac{e^h- 1}{h}= 1$ which, while possible, is not trivial.

We can prove a number of the basic properties of the logarithm directly from that definition:
Since its derivative, 1/x, is positive for all x, ln(x) is an increasing function.

To prove ln(1/x)= -ln(x), let u= 1/t in $\int_1^{1/x} dt/t$. From u= 1/t, t= 1/u so $dt= -u^{-2}du$,when t= 1, u= 1, and when t= 1/x, u= x so the integral becomes $\int_1^x (1/(1/u))(-(1/u^2)du)= -int_1^x (1/u)du= -ln(x)$.

To prove ln(xy)= ln(x)+ ln(y), let u= t/y so t=yu dt= ydu, when t=1, u= 1/y, and when t= xy, u=x so the integral becomes $\int_{1/y}^x (1/yu)(ydu)= \int_{1/y}^x (1/u)du= \int_{1/y}^1 du/u+ \int_1^x du/u= -\int_1^{1/y} du/u+ \int_1^x du/u= -ln(1/y)+ ln(x)$ and since ln(1/y)= -ln y, that is, ln(x)+ ln(y),

To prove [tex]ln(a^y)= y ln(a), if $y\ne 0$, let $u= t^{1/y}$ so that $t= u^y$ and $dt= y u^{y- 1}du$. When t= 1, u= 1 and when $t= x^y$, u= x so the integral becomes $\int_1^x (1/u^y)(y u^{y-1}du= y\int_1^x du/u= yln(x)$.
(If y= 0, then $x^y= x^0= 1$ and so $ln(x^y)= ln(1)= 0= 0(ln(x))= yln(x)$

Since ln(x) is differentiable for all x, we can apply the mean value theorem to any interval of real numbers and, in particular, to [1, 2]. The mean value theorem says that there exist c between 1 and 2 so that the mean value theorem says (ln(2)- 1)/(2- 1)= (ln(2)- 0)/1= ln(2)= 1/c. Since $1\le c\le 2$, $1/2< c< 1$. That is, ln(2)> 1/2. From that, for any X> 0, $ln(2^{2X})= 2Xln(2)> X$. That means that the ln(x) function has no upper bound and, since 1/x is positive for all positive x, $\lim_{x\to\infty} ln(x)= +\infty$. Since ln(1/x)= -ln(x), $lim_{x\to 0} ln(x)= -\infty$. That is, ln(x) maps the set of all positive real numbers, one to one, to the set of all real numbers.

That tells us that ln(x) has an inverse which we can call "exp(x)" (I am deliberately avoiding calling that function $e^x$) which maps all real numbers to the positive real numbers. And, of course, since the derivative of ln(x) is 1/x, it follows that if y= exp(x), then x= ln(y) so that dx/dy= 1/y and thus dy/dx= y= exp(x).

But I think the main thing you want is this: If y= exp(x) then x= ln(y). If $x\ne 0$, $1= (1/x)ln(y)= ln(y^{1/x})$. Therefore $y^{1/x}= exp(1)$ and so $y= exp(x)= (exp(1))^y$. If x= 0, 0= ln(y) means that y= 1= exp(1)^0. In either case, this function, exp(x), really is a some number to the x power. If we define e to be exp(1), the value of x so that ln(x)= 1, we have $exp(x)= e^x$.