Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I came across the following exercise in Spivak: Suppose that $f(x)$ is a differentiable function and suppose that $f'(x)$ is increasing. Show that every tangent line of $f(x)$ intersects the graph only once.

My intuition is as follows: Consider the tangent line $t(x)$ for any point $a$. Clearly, we have $t'(a) = f'(a)$. Since $f'(x)$ is an increasing function whereas $t'(x)$ is constant, the two graphs can never intersect again for $x\geq a$ because $f(x)$ "grows faster" than $t(x)$.

How do make this argument precise?

share|cite|improve this question
up vote 4 down vote accepted

Consider the tangent $l$ to $f(x)$ at $x=a$. This has gradient $f'(a)$. Let $c>a$. Then by the mean value theorem, there is some $a<b<c$ such that $f'(b)(c-a)=f(c)-f(a)$. This means $f(c)=f(a)+f'(b)(c-a)>f(a)+f'(a)(c-a)$ which is the value of $l$ at $c$. Do the same for the case $c<a$.

share|cite|improve this answer
Why is there a strict inequality? We don't assume that $f'(x)$ is strictly increasing. – somebody Oct 30 '12 at 19:05

Use the integral. We have, for any $x > a$: \[ f(x) = f(a) + \int_a^x f'(\xi)\, d\xi > t(a) + \int_{a}^{\frac{a+x}2} t'(\xi)\,d\xi = t(x) \] (The latter holds as $f' \ge t'$ and $f' > t'$ on $[\frac{a+x}2, x]$. Argue analogously for $x < a$.

share|cite|improve this answer
Elite approach. +1 – Babak S. Oct 30 '12 at 14:38

Assume that for some $c$, the tangent at $c$ cut the graph two times : at $c$ and at $d$. The slope of this line is $f'(c)$. However, according to MVT, there exists some $a \in (c,d)$ such that $f'(a)=\frac{f(d)-f(c)}{d-c}=f'(c)$. Since $a \neq c$, this contradicts the fact that $f'(x)$ is increasing.

share|cite|improve this answer
But $f'(x)$ might not strictly increasing, so it can be that $f'(c)=f'(a)$ even if $a<c$, right? – somebody Oct 30 '12 at 15:33

Since $f'$ is increasing, the function $f$ is strictly convex, which means that its graph always lies "below the line segment joining any two points of the graph".

Now assume that a tangent line $t$ at $f$ in the point $(x_1,f(x_1))$ intersects the graph of $f$ twice, at $(x_1,f(x_1))$ and $(x_2,f(x_2))$. Since the graph of $f$ has to lie below the two points $(x_1,f(x_1))$ and $(x_2,f(x_2))$, the value of $(f-t)'$ cannot be zero at $x$, which is a contradiction, since $t$ should be a tangent at $f$ in $x_1$, i.e. $f'(x_1) = t'(x_1)$.

share|cite|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.