Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

In Calculus by Spivak (1994), the author states in Chapter 18 p. 341 that $$\exp(x+y) = \exp(x)\exp(y)$$ implies $$\exp(x) = [\exp(1)]^x$$ He refers to the discussion in the beginning of the chapter where we define a function $f(x + y) = f(x)f(y)$; with $f(1) = 10$, it follows that $f(x) = [f(1)]^x$. But I don't get this either. Can anyone please explain this? Many thanks!

share|improve this question
As a first step: try writing out the argument Spivak gave in your question, and try writing it in your own words. –  J. M. Feb 5 '12 at 16:08
Either $f(x+y)=f(x)+f(y)$ or $f(x)=[f(1)]^x$ is incorrect. One of the two should be changed to $f(x+y)=f(x)f(y)$ or $f(x)=[f(1)]x$ respectively. –  Adam Feb 5 '12 at 16:40

4 Answers 4

up vote 7 down vote accepted

I think you should assume $x$ is an integer (since $a^x$ is defined using $\exp$ if $x$ is a positive real). You can write $\exp(x) = \exp(1+1+1+1+\cdots+1)$ (there is $x$ times « 1 » in the last term).

Using the property of $\exp$, you find now $\exp(x) = \exp(1)\exp(1)\cdots\exp(1) = (\exp(1))^n$

share|improve this answer
Note that $a^x = \exp( x \log(a))$ and is defined for all complex $x$ and $a$, s.t. $a \not= 0$, where the principal branch of $\log$ is taken. –  Sasha Feb 5 '12 at 16:50

As Paul Pichaureau has mentioned, it is easy to see that the formula holds for $n$ natural. You can also easily go further: what should $\exp(-n)$ be? Well, $\exp(0) = 1 = \exp(-n+n) = \exp(-n) \exp(n)$, so $\exp(-n) = 1/\exp(n)$. Furthermore, it is easily seen that $\exp(nx) = \exp(x)^{n}$, so that $$\exp(1) = \exp(n\cdot 1/n) = \exp(1/n)^n,$$ and hence $$\exp(1/n) = \exp(1)^{1/n}$$ by the usual rules for powers. Combining these, we see that $$\exp(p/q) = \exp(1)^{p/q},$$ so the formula holds for any rational number $x$. I think this is about as far as you can get using only that formula (in addition to the requirement that $\exp(0) = 1$). If we, in addition, assume that $\exp$ is continuous on $\mathbb{R}$, then we can choose a sequence $(r_{i})_{i =1}^{\infty}$ of rational numbers such that $\lim_{n \to \infty} r_{i} = x$, and compute $$ \lim_{n\to \infty} \exp(1)^{r_n} = \lim_{n \to \infty} \exp(r_{n}) = \exp(\lim_{n \to \infty} r_{n}) = \exp(x).$$ The problem now is what $\exp(1)^x$ is supposed to be. If we already know that the power function $x \mapsto \exp(1)^x$ is continuous, then certainly the right hand side is equal to $\exp(1)^x$, but as Paul mentions, the power functions are often defined in terms of the exponential function. However, I don't think you have to do it that way (you could define rational powers as usual, and prove that a continuous extension to the reals exists), so it all depends on what Spivak has done prior to that chapter (you also didn't specify what $x$ is, so it conceivable that he assumes it is rational, in which case it doesn't matter much).

share|improve this answer
Instead of continuity, you could get by with monotonicity. –  Michael Hardy Feb 6 '12 at 2:41
@michaelhardy: interesting! –  Martin Wanvik Feb 6 '12 at 2:54
excellent answer Martin, thorough and comprehensive without abstruosity –  David Holden Dec 23 '13 at 16:22

The $n$-th $\ (n\in{\mathbb N}_{\geq1}$) power of a real number $a>0$ is defined in elementary terms as product of $n$ factors equal to $a$. It is only reasonable to define $a^0:=1$ and $a^n:=1/a^{-n}$ when $n<0$. One then has $$a^{m+n}=a^m\cdot a^n\ ,\qquad a^{m\ n}=\bigl(a^m\bigr)^n \qquad(*)$$ for all $m$, $n\in{\mathbb Z}$. The aim now is to extend the definition of an $x$-th power of $a$ from $x\in{\mathbb Z}$ to $x\in{\mathbb Q}$ and $x\in{\mathbb R}$ such that the law $(*)$ is preserved.

The first step is easy: For $p\in{\mathbb Z}$ and $q\in{\mathbb N}_{\geq1}$ put $$a^{p/q}\ :=\ \root q \of {a^p}\ .$$ Using the "rules of algebra" one then can verify that $(*)$ is true for arbitrary $x$, $y\in{\mathbb Q}$ instead of $m$, $n\in{\mathbb Z}$. That is as far as one can get using the algebraic notions of $n$-th power and $q$-th root.

Now analysis provides a wonderful (in particular, continuous) function $x\mapsto\exp (x)$ with the property that $\exp(x+y)=\exp(x)\cdot\exp(y)$ for all $x$, $y\in{\mathbb C}$. Put $\exp(1)=:e\doteq 2.718$. Then it follows by an easy induction that in fact $$\exp\Bigl({p\over q}\Bigr)\ =\ e^{p/q}$$ for all ${p\over q} \in{\mathbb Q}$. As $\exp$ is continuous it is only natural to define arbitrary real powers of the number $e$ by $$e^x\ :=\ \exp(x)\qquad(x\in{\mathbb R})\ .$$ Via the logarithm function (which I won't explain here) we can define arbitrary real powers not only of $e$, but of an arbitrary number $a>0$: $$a^x\ :=\ \exp\bigl(x\log(a)\bigr)\qquad (x\in{\mathbb R})\ .$$ Using the functional equations of $\exp$ and $\log$ one then easily proves that the rule $(*)$ remains true for arbitrary $x$, $y\in{\mathbb R}$ instead of $m$, $n\in{\mathbb Z}$.

share|improve this answer

I like (as I have done here before) to start with a functional equation and derive properties of the function.

If $f(x+y) = f(x) f(y)$ and $f$ is differentiable (and non-zero somewhere), $f(0) = 1$ and $f(x+h)-f(x) = f(x)f(h)-f(x) =f(x)(f(h)-1) =f(x)(f(h)-f(0)) $ so $$(f(x+h)-f(x))/h = f(x)(f(h)-f(0))/h.$$ Letting $h \to 0$, $f'(x) = f'(0) f(x)$. From this, $f(x) = \exp(f'(0)x)$.

This also works for $\ln$, $\arctan$, and $\sin$ and $\cos$.

Functional equations are fun!

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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