$\exp(x)$ is usually defined in three different ways:

1) By its Taylor series: $\exp(x)=\sum_{k=0}^{\infty} \frac{x^k}{k!}$

2) By its derivative: $\exp(x)'=\exp(x)$

3) By the limit $\exp(x)=\lim_{N \rightarrow \infty} \left(1+\frac{x}{N} \right)^N$

In textbooks they mostly use the 3rd way, because $\exp(x)$ is introduced before the derivative.

But it just occured to me - we don't need this much information to define this function. There is a very simple definition which allows to recover all the other properties:


$$\lim_{x \rightarrow 0} \exp(x) = \lim_{x \rightarrow 0} (1+x)$$

Now, it's just two terms of Taylor expansion - it doesn't seem like much. Yet we can prove the following properties:

a) It follows that $\lim_{x \rightarrow 0} \exp(a x) = \lim_{x \rightarrow 0} \exp^a(x)$

$$\lim_{x \rightarrow 0} \exp^a(x) = \lim_{x \rightarrow 0} (1+x)^a = \lim_{x \rightarrow 0} (1+a x) = \lim_{x \rightarrow 0} \exp(a x)$$

b) Using this property we recover the 3rd definition

$$ \lim_{N \rightarrow \infty} \left( 1+\frac{x}{N} \right)^N=\lim_{N \rightarrow \infty} \left( \exp \left(\frac{x}{N}\right) \right)^N=\lim_{N \rightarrow \infty} \left( \exp^{\frac{1}{N}}(x) \right)^N=\exp(x) $$

Does this work, or did I make a mistake somewhere?


I was obviously wrong - there is an infinite number of functions with the property 4. So the definition needs to be:

4) For $x \rightarrow 0$ $$ \exp(x) \approx 1+x$$

For any $x$

$$\exp^a(x) = \exp(a x)$$

  • 1
    $\begingroup$ Your definition (4) is satisfied by any function $f(x)$ such that $\lim_{x\to 0} f(x) = 1$, so there is no way that you can prove $f(x)=\exp(x)$ from it... $\endgroup$ – Hans Lundmark Feb 11 '16 at 11:34
  • $\begingroup$ If $a$ is real, how do you define $\text{thing}^a$? $\endgroup$ – Martín-Blas Pérez Pinilla Feb 11 '16 at 11:51
  • $\begingroup$ As $(\exp(x))^a$ $\endgroup$ – Yuriy S Feb 11 '16 at 11:52
  • $\begingroup$ See the edit of my previous comment. $\endgroup$ – Martín-Blas Pérez Pinilla Feb 11 '16 at 11:53
  • $\begingroup$ $\approx$ doesn't have a standard meaning. However, if we define $f(x)\approx g(x)$ to mean $\lim\limits_{x\to0}\frac{f(x)-g(x)}x=0$, then $\exp(x)\approx 1+x$ is actually equivalent to $\exp'(0)=\exp(0)$. Along with the rule $\exp(a)\exp(b)=\exp(a+b)$, this would imply $\exp'(x)=\exp(x)$ (why?), which is the second definition at the top. $\endgroup$ – Akiva Weinberger Feb 11 '16 at 11:55

$$\lim_{x\to 0} (1+x)$$ is simply equal to $1$, so the equation you wrote is equal to the equation $$\lim_{x\to0} \exp(x) = 1$$

which is most certainly not enough to characterise the exponential function, for example $\cos(x)$.

As for what you proved in points (a) and (b), I don't even know what you were intending to prove so I cannot comment on that.

  • $\begingroup$ Thanks, I noticed it, see the edit $\endgroup$ – Yuriy S Feb 11 '16 at 11:35
  • $\begingroup$ I wanted to write $\exp(x) \approx 1+x$ for $x \rightarrow 0$ at first, would it be better? $\endgroup$ – Yuriy S Feb 11 '16 at 11:39
  • $\begingroup$ @YuriyS Not really, since the function $1+x$ also satisfies that condition. $\endgroup$ – 5xum Feb 11 '16 at 11:40
  • $\begingroup$ And together with the property $\exp^a(x)=\exp(ax)$? $\endgroup$ – Yuriy S Feb 11 '16 at 11:42
  • $\begingroup$ @YuriyS That might be enough, yes. But I'm not sure. Also, I don't see the benefit of such a definition. $\endgroup$ – 5xum Feb 11 '16 at 11:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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