1
00:00:00,000 --> 00:00:04,285
A derivative starts with a simple question. If this curve is changing,

2
00:00:04,385 --> 00:00:07,918
how fast is it changing right here, at one exact point? It

3
00:00:08,018 --> 00:00:12,116
is the number hiding behind velocity, growth, and every tiny nudge.

4
00:00:12,216 --> 00:00:16,167
The trouble is that a single point has no direction by itself. So

5
00:00:16,267 --> 00:00:20,156
we borrow a nearby point, and ask for the slope between the two.

6
00:00:20,266 --> 00:00:24,570
That slope is honest, but rough. It measures the average change across

7
00:00:24,670 --> 00:00:27,150
a small step, a run in x and a rise in y.

8
00:00:27,266 --> 00:00:30,406
Now shrink the step. Let the two points slide closer

9
00:00:30,506 --> 00:00:33,958
and closer, until the secant line becomes a tangent line.

10
00:00:34,066 --> 00:00:38,184
The derivative is that limiting slope. Not the height of the curve,

11
00:00:38,284 --> 00:00:42,339
but the instantaneous rate at which the height is changing. If the

12
00:00:42,439 --> 00:00:45,990
curve were a road, this would be the tilt under your feet.

13
00:00:46,100 --> 00:00:49,450
Move the point, and the tangent turns with it. At every

14
00:00:49,550 --> 00:00:53,152
x value, the curve quietly carries a different local slope.

15
00:00:53,266 --> 00:00:57,626
When the tangent leans upward, the derivative is positive. The curve is

16
00:00:57,726 --> 00:01:01,206
climbing, so a small push to the right raises the output.

17
00:01:01,316 --> 00:01:04,662
When the tangent lies flat, the derivative is zero. For

18
00:01:04,762 --> 00:01:08,296
one tiny instant, the curve is neither rising nor falling.

19
00:01:08,400 --> 00:01:12,707
And when the tangent leans downward, the derivative is negative. The curve

20
00:01:12,807 --> 00:01:16,460
is descending, even if the point itself sits high on the graph.

21
00:01:16,566 --> 00:01:19,570
So imagine a new machine. Feed it an x value, and

22
00:01:19,670 --> 00:01:22,610
it returns the slope of the old curve at that x.

23
00:01:22,716 --> 00:01:27,344
Sweep across the original curve, recording each tangent slope as a height on

24
00:01:27,444 --> 00:01:31,450
a second graph. Each dot below is not a new measurement of height.

25
00:01:31,550 --> 00:01:35,432
It is a measurement of change. The slopes turn into a new curve.

26
00:01:35,533 --> 00:01:38,694
Where the original curve has a peak or valley, the new

27
00:01:38,794 --> 00:01:42,921
graph crosses zero, because the tangent there is perfectly horizontal.

28
00:01:43,033 --> 00:01:46,333
Positive regions on the derivative mean rising motion

29
00:01:46,433 --> 00:01:50,119
above. Negative regions mean falling motion. The derivative

30
00:01:50,219 --> 00:01:52,365
is a motion detector for functions.

31
00:01:52,466 --> 00:01:55,557
For this particular curve, the algebra says the same

32
00:01:55,657 --> 00:01:59,422
thing the picture says. Different notation, same slope machine.

33
00:01:59,533 --> 00:02:03,062
This is a derivative: a way to turn a curve into the story

34
00:02:03,162 --> 00:02:06,753
of how it changes. It is local motion, drawn as a function.

