## The Differential Mean Value Theorem

Let’s say we’ve got a function that’s continuous on the closed interval and differentiable on . We don’t even assume the function is defined outside the interval, so we can’t really set up the limit for differentiability at the endpoints, but they don’t matter much in the end.

Anyhow, if we look at the graph of we could just draw a straight line from the point to the point . The graph itself wanders away from this line and back, but the line tells us that on average we’re moving from to at a certain rate — the slope of the line. Since this is an *average* behavior, sometimes we must be going faster and sometimes slower. The differential mean value theorem says that there’s at least one point where we’re going *exactly* that fast. Geometrically, this means that the tangent line will be parallel to the secant we drew between the endpoints. In formulas we say there is a point with .

First let’s nail down a special case, called “Rolle’s theorem”. If , we’re asserting that there is some point with . Since is compact and is continuous, the extreme value theorem tells us that must take a maximum and a minimum. If these are both zero, then we’re looking at the constant function , and any point in the middle satisfies . On the other hand, if either the maximum or minimum is nonzero, then we have a local extremum at a point where is differentiable (since it’s differentiable all through the open interval). Now Fermat’s theorem tells us that since is a local extremum! Thus Rolle’s theorem is proved.

Now for the general case. Start with the function and build from it the function . On the graph, this corresponds to applying an “affine transformation” (which sends straight lines in the plane to other straight lines in the plane) to pull both and down to zero. In fact, it’s a straightforward calculation to see that . Thus Rolle’s theorem applies and we find a point with . But applying our laws of differentiation, we see that . And so , as desired.

The “trick” behind proving the Mean Value theorem is basically to write , where is the equation of the straight line connecting the points and . And this is what you did above. I know it might not sound very interesting, but I have always wondered if we can obtain another useful result (just like the mean value theorem, for instance) if we let be the equation of a parabola passing through the points and . And, do we obtain more “useful” results if is some function other than a straight line or a parabola?

(That was just a random thought.)

Comment by Vishal | January 22, 2008 |

Indeed you do, Vishal, but you ask that the function be twice-differentiable and you match another parameter: the derivative of at . Then those three parameters specify a unique parabola, which has constant second derivative. And there’s some point in between and at which the function actually has that second derivative.

There’s a whole sequence of these “generalized mean value theorems”, which can be used as Taylor series remainder approximations. Your exercise is to work out the analogues of Rolle’s theorem you guessed were around, and to see if you can prove them in general. I’ll come back to this when I’m doing Taylor series.

Comment by John Armstrong | January 22, 2008 |

I had a hunch earlier that we could indeed derive more of those “generalized mean value theorems” and somehow relate those to the Taylor series approximations. Your comment, it seems, has confirmed that hunch. I will be looking forward to your future post(s) on Taylor series.

Please keep the articles coming!

Comment by Vishal | January 22, 2008 |

This theorem is also known as the Lagrange mean value theorem.

Comment by Michael Livshits | January 23, 2008 |

It also has an interesting corollary that says that the derivative of any function that is differentiable on an interval has the mean value property, i.e. it hits all its intermediate values between any 2 of the values that it assumes, even if that derivative is not continuous.

In partcular, a simple step function H(x) that is 0 for x1 is not a derivative of any function F differentiable on (-1,1), so H(x) has no primitive on (-1,1).

Comment by Michael Livshits | January 23, 2008 |

In comment #5 H(x)=0 for x0, also known as the Heaviside function.

Comment by Michael Livshits | January 23, 2008 |

There must be a bug in the system, that messed up my posts, again (now using LaTex): for

and for .

Comment by Michael Livshits | January 23, 2008 |

There must be a bug in the system, that messed up my posts, again (now using LaTex): for

and for .

Comment by Michael Livshits | January 23, 2008 |

Michael, you’re having trouble with the < symbol. That’s not a bug in the system.

The bug in the system is that I can’t seem to log in to WordPress at all. Whether this has to do with a comment flood or not, I have no idea. Just in case, could you try in the future to think carefully about what you want to say and then say it all in one comment?

Comment by John Armstrong | January 23, 2008 |

Sorry, I got troubles with > that seems to have some special meaning on this site and also with LaTex preprocessor here that messed up my formula. I wish we could edit our entries.

I doubt that your trouble logging in is related to the number of comments on your page.

Comment by Michael Livshits | January 23, 2008 |

It’s not this site. > has a special meaning in HTML, which is sort of the basis of the web. WordPress’ processor can’t tell the difference between using it to mean “greater than” and “close a tag”.

But still, you made three separate comments before the litany of trying to fix the problem with the > sign. Slow down, take a deep breath, and decide what you want to say before saying it. This isn’t Twitter.

Comment by John Armstrong | January 23, 2008 |

Actually I wanted to say that H(x) is the Heaviside function that jumps from 0 to 1 at x=0.

Comment by Michael Livshits | January 23, 2008 |

[...] of the Mean Value Theorem So now that we have the mean value theorem what can we do with it? First off, we can tell something that seems intuitively obvious. We know [...]

Pingback by Consequences of the Mean Value Theorem « The Unapologetic Mathematician | January 24, 2008 |

[...] to get back on track. Today, we’ll see a theorem about integrals that’s similar to the Differential Mean Value Theorem. Specifically, it states that if we have a continuous function then there is some so [...]

Pingback by The Integral Mean Value Theorem « The Unapologetic Mathematician | February 12, 2008 |

[...] the Differential Mean Value Theorem tells us that in each subinterval there is some so that . We stick this into the [...]

Pingback by The Riemann-Stieltjes Integral II « The Unapologetic Mathematician | March 6, 2008 |

[...] an extension of the Fundamental Theorem of Calculus we’ll see it’s an extension of the Differential Mean Value Theorem. And this despite the most strenuous objections of my resident gadfly that the DMVT and FToC have [...]

Pingback by Taylor’s Theorem again « The Unapologetic Mathematician | October 1, 2008 |

[...] a nice technical result we may have call for from time to time: a higher-dimensional version of the differential mean value theorem. Remember that this says that if we’ve got a function continuous on the closed interval and [...]

Pingback by The Mean Value Theorem « The Unapologetic Mathematician | October 13, 2009 |

[...] we can apply the mean value theorem to [...]

Pingback by Clairaut’s Theorem « The Unapologetic Mathematician | October 15, 2009 |

[...] Taylor’s theorem. Specifically, the version of Taylor’s theorem that resembles the mean value theorem. And we’ll even use an approach like we did for the extension of that result to higher [...]

Pingback by Taylor’s Theorem « The Unapologetic Mathematician | October 20, 2009 |

[...] is some number between and that exists by the differential mean value theorem. Now to find the derivative, we take the limit of the difference [...]

Pingback by Differentiation Under the Integral Sign « The Unapologetic Mathematician | January 13, 2010 |

[...] let denote the sum of the terms and , the remainder after terms, of the series (2), so that . By Lagrange’s mean value theorem, we have , . We shall now consider . So far we have taken as an arbitrary but we shall now choose [...]

Pingback by Everywhere Continuous Non-derivable Function | My Digital NoteBook | July 7, 2011 |

[…] let denote the sum of the terms and , the remainder after terms, of the series (2), so that . By Lagrange’s mean value theorem, we have , . We shall now consider . So far we have taken as an arbitrary but we shall now choose […]

Pingback by Everywhere Continuous Non-derivable Function | MY DIGITAL NOTEBOOK | October 22, 2013 |