If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

Main content

Formal and alternate form of the derivative

Discover the two methods for defining the derivative of a function at a specific point using limit expressions. In this video we'll explore the standard form and alternate form, enhancing our understanding of tangent line slopes and secant line approximations. Created by Sal Khan.

Want to join the conversation?

  • leaf grey style avatar for user dashpointdash
    Just a philosophical digression.

    If mathematicians until today don't know how to define a process like division by zero, isn't it impossible to assume that if a division by zero has any kind of definition, humanity can never surely estimate.

    It seems kind of strange that the financial sector, higher analysis and modelling relies on mathematical concepts which aren't rigourously defined.

    That would mean that at any given point you can estimate that a function has some
    supposed behaviour but since you can never instantaneously easily access that point parametrically you are after all only speculating.
    (77 votes)
    Default Khan Academy avatar avatar for user
    • orange juice squid orange style avatar for user Ohad
      You raised an interesting topic for discussion, and to answer that, I have to get back to one of your sentences.

      "It seems kind of strange that the financial sector, higher analysis and modelling relies on mathematical concepts which aren't rigourously defined."

      You might get the wrong impression about division by zero. Mathematicians weren't just saying "Okay, let's define a new operation and call it division- Oh wait, it doesn't work on zero- let's say it's undefined, because any sensible answer will break the rules of math". I admit it, it might have happened in the beginning- but in the last few years, mathematicians have been making extremely careful and formal definitions of math objects, including the division operation.
      Division is built as a function (of two variables), that says: given a and b, return a/b. It was built in a way that doesn't allow 0 to be the denominator, so when we say that 1/0 is not defined, we mean that division is not built to fit in zero in there.

      I'll give you an example- When I'll ask you what x to the power of a duck is, you'll think I'm crazy. Why? because exponention is not defined when talking about ducks. You say, "exponention is an operation that you put on numbers, not on animals". With the same arguement, I can tell you "division is an operation that you put on non-zero numbers as the denominator".

      Not all mathematical operation need to be defined on all numbers, you just decide what numbers do you want the operation to work with.
      (291 votes)
  • blobby green style avatar for user myr707
    I understand both forms, but what is the point of the alternative form of derivative? What can the alternative form do, which the standard form can't?
    (30 votes)
    Default Khan Academy avatar avatar for user
  • leafers ultimate style avatar for user JC
    What does derivative mean?
    (17 votes)
    Default Khan Academy avatar avatar for user
    • leafers seedling style avatar for user Y. Chen
      The derivative of a function is the measure of change in that function. Consider the parabola y=x^2.

      For negative x-values, on the left of the y-axis, the parabola is decreasing (falling down towards y=0), while for positive x-values, on the right of the y-axis, the parabola is increasing (shooting up from y=0).
      If you find the derivative of y=x^2 (it will be explained in later videos but the derivative of x^2 is 2x) and substitute in a negative x-value, you will get a negative number, meaning you have negative change i.e. the function is decreasing. On the other hand, if you substitute in a positive x-value, you will get a positive number, meaning you have positive change i.e. the function is increasing.
      (22 votes)
  • starky ultimate style avatar for user Pedro Marin
    Can´t I make something like f(x+h) -f(x)/(x+h)-x -> f(x) + f(h) - f(x)/h -> f(h)/h ?
    (6 votes)
    Default Khan Academy avatar avatar for user
    • piceratops ultimate style avatar for user Just Keith
      f(x+h) IS NOT equal to f(x) + f(h). So, no, you cannot evaluate it that way.
      Example:
      f(x) = x²
      f(x+h) = (x+h)² = x² + 2hx + h²
      while f(x)+f(h) = x² + h²

      Note: there are a few types of functions where f(x+h) happens to equal f(x) + f(h), but this is the exception to the rule.
      (20 votes)
  • ohnoes default style avatar for user Cyan Wind
    As far as I know, f'(a) is the slope of a tangent line at x = a. Does anyone have any idea how we can apply this definition in real life? In your example, please tell me what is slope and what is tangent line?
    (8 votes)
    Default Khan Academy avatar avatar for user
    • mr pants teal style avatar for user Moon Bears
      Slope can be thought as "How far you go out horizontally and how far you go up vertically to get back to the line". Another way of saying that is "Change in y over change in x." A tangent line is a line that intersects the graph at exactly one point. Slope is often seen as a rate of change, so differentials and derivatives are seen as a rate of change. Physics is interested often times how fast a particle is moving in which ever way, so you set up a coordinate system to find position, then if you differentiate position you obtain velocity. But let's say you want to know how fast your velocity is changing - that's acceleration. You can graph velocity versus time and if you create a tangent line to one point that would be your instantaneous acceleration or your acceleration at that instant.
      (5 votes)
  • purple pi purple style avatar for user Arthur Alikulov
    If we take a look at any of these two forms for derivative, we always will get 0/0. So it looks like any derivative is undefined. Could someone explain me this moment?
    (3 votes)
    Default Khan Academy avatar avatar for user
    • leaf green style avatar for user kubleeka
      If you substitute h directly into the expressions, then yes, you get 0/0. But this does not necessarily mean that the limit does not exist, it may just mean we have to work a little more to evaluate it. For example, consider the function f(x)=x. The derivative is
      lim [(x+h)-x]/h
      h->0

      Substituting h immediately gives 0/0, but if we first do some simplifying algebra, we get
      lim [x+h-x]/h
      h->0

      lim [h]/h
      h->0

      lim 1
      h->0

      1

      So the derivative of f(x)=x is simply 1.
      (13 votes)
  • leaf green style avatar for user Umar khan
    at shouldnt x approach zero
    then and only,'I think ' we might get closer to zero
    PLEASE correct if i am wrong
    I am talking about the slope at point 'a' on the second graph
    (3 votes)
    Default Khan Academy avatar avatar for user
  • spunky sam green style avatar for user Zion J
    Hey, isn't a tangent line just a secant line with two points in the same place? Before I learned calculus here on Khan Academy, I taught myself to approach a tangent line with a secant line whose points get closer and closer and closer in the same way you approach an undefined point on a function with a limit.
    (4 votes)
    Default Khan Academy avatar avatar for user
    • leafers seed style avatar for user Travis Bartholome
      That's a good way to look at it, and it's actually the basis of the typical definitions of the derivative. Geometrically, the derivative is the limit of the secant slope as the two points of the secant line get closer to each other, just like you described. I think Sal has a video on that entitled "Tangent slope as the limit of secant slope," or something to that effect.
      (5 votes)
  • piceratops ultimate style avatar for user Matthew K.
    When you draw the tangent line to point a at the beginning of the video in both graphs, would it be possible to find take another random point on that line and find the slope of the tangent line? Since the tangent line is straight, wouldnt that meant the slope is the same at every point? Thank you.
    (5 votes)
    Default Khan Academy avatar avatar for user
    • female robot grace style avatar for user tyersome
      Good question – if you could draw the tangent line "nearly perfectly" then that would work. There are, however, a number of drawbacks (so to speak) of this strategy. I encourage you to try your proposed method on a few equations at a couple of different points and compare the answers you get with the actual answers.

      Some shortcomings to a graphical strategy:
      1) Tedious.
      2) Error prone.
      3) In situation where you need an equation it would be impractical (as well as exceedingly inaccurate) to make estimates from graphs.
      – For many applications of derivatives (e.g. optimization), we need an equation not just the derivative at a single point.
      – Sometimes we even want the second derivative and so an equation for the derivative is doubly necessary.
      (3 votes)
  • spunky sam blue style avatar for user onkar.kulkarni
    at Sal cancels the x's and the equation remaining is:

    the limit (as x approaches 0) of ( f(x+h) - f(x) )/h

    As x approaches 0, doesn't the whole equation approach infinity? If this is true, then why isn't the derivative of every function infinity?
    (3 votes)
    Default Khan Academy avatar avatar for user
    • starky ultimate style avatar for user KLaudano
      The limit is when h approaches 0, not x. When h approaches 0, f(x+h) approaches f(x) and the numerator, f(x+h)-f(x), approaches 0. The numerator and denominator both approach 0, so the limit is 0/0. 0/0 is one of the inderterminate answers, so it is not necessarily infinity. The answer depends on the function we put in the limit.
      (4 votes)

Video transcript

So let's think about how we could find the slope of the tangent line to this curve right over here, so what I have drawn in red, at the point x equals a. And we've already seen this with the definition of the derivative. We could try to find a general function that gives us the slope of the tangent line at any point. So let's say we have some arbitrary point. Let me define some arbitrary point x right over here. Then this would be the point x comma f of x. And then we could take some x plus h. So let's say that this right over here is the point x plus h. And so this point would be x plus h, f of x plus h. We can find the slope of the secant line that goes between these two points. So that would be your change in your vertical, which would be f of x plus h minus f of x, over the change in the horizontal, which would be x plus h minus x. And these two x's cancel. So this would be the slope of this secant line. And then if we want to find the slope of the tangent line at x, we would just take the limit of this expression as h approaches 0. As h approaches 0, this point moves towards x. And that slope of the secant line between these two is going to approximate the slope of the tangent line at x. And so this right over here, this we would say is equal to f prime of x. This is still a function of x. You give me an arbitrary x where the derivative is defined. I'm going to plug it into this, whatever this ends up being. It might be some nice, clean algebraic expression. Then I'm going to give you a number. So for example, if you wanted to find-- you could calculate this somehow. Or you could even leave it in this form. And then if you wanted f prime of a, you would just substitute a into your function definition. And you would say, well, that's going to be the limit as h approaches 0 of-- every place you see an x, replace it with an a. f of-- I'll stay in this color for now-- blank plus h minus f of blank, all of that over h. And I left those blanks so I could write the a in red. Notice, every place where I had an x before, it's now an a. So this is the derivative evaluated at a. So this is one way to find the slope of the tangent line when x equals a. Another way-- and this is often used as the alternate form of the derivative-- would be to do it directly. So this is the point a comma f of a. Let's just take another arbitrary point someplace. So let's say this is the value x. This point right over here on the function would be x comma f of x. And so what's the slope of the secant line between these two points? Well, it would be change in the vertical, which would be f of x minus f of a, over change in the horizontal, over x minus a. Actually, let me do that in that purple color. Over x minus a. Now, how could we get a better and better approximation for the slope of the tangent line here? Well, we could take the limit as x approaches a. As x gets closer and closer and closer to a, the secant line slope is going to better and better and better approximate the slope of the tangent line, this tangent line that I have in red here. So we would want to take the limit as x approaches a here. Either way, we're doing the exact same thing. We have an expression for the slope of a secant line. And then we're bringing those x values of those points closer and closer together. So the slopes of those secant lines better and better and better approximate that slope of the tangent line. And at the limit, it does become the slope of the tangent line. That is the definition of the derivative. So this is the more standard definition of a derivative. It would give you your derivative as a function of x. And then you can then input your particular value of x. Or you could use the alternate form of the derivative. If you know that, hey, look, I'm just looking to find the derivative exactly at a. I don't need a general function of f. Then you could do this. But they're doing the same thing.