What the derivative is: slope and the rules
In the textbook: Ch. 3.1 Die Ableitung einer Funktion3.2 Differenzieren
For the café in the article “What a function is: cost, revenue, profit and demand” every extra cup added the same 2.50 € to the cost. The cost there grew evenly, along a straight line. It is not always like that.
Take a farmer who waters his field. The more water, the bigger the harvest, but every further thousand litres adds less than the one before:
| Water, thousand litres | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|
| Harvest, kg | 100 | 141 | 173 | 200 | 224 |
| Gain from the last thousand litres, kg | 41 | 32 | 27 | 24 |
A thousand litres cost the farmer 20 €, and he sells a kilo of harvest for 1 €. Right now he puts 4 thousand litres on the field. Should he water more? The fifth thousand adds 24 kg, that is 24 €, and costs 20 €: for now, yes. But the gain keeps falling and at some point drops below 20 kg. Where exactly, the table does not show: a step of a whole thousand litres is too coarse.
What you need is not the gain over a whole thousand litres but the gain at the point itself, at 4 thousand litres. The derivative (“Ableitung”) gives it. It shows how fast a quantity changes at one point, the way a speedometer shows the car's speed in this second. For the farmer's question it gives this answer:
| Water, thousand litres | Gain at this point | A thousand litres cost | Water more? |
|---|---|---|---|
| 4 | 25 kg, that is 25 € | 20 € | yes |
| 6.25 | 20 kg, that is 20 € | 20 € | stop here: beyond it the gain is below 20 kg |
In the same way the derivative tells you where profit is largest and where cost grows most slowly. This article builds the derivative from scratch, from the slope of a straight line to the second derivative. After it you can differentiate on your own:
- sums and powers;
- the exponential function and the logarithm;
- products and chains, where one function sits inside another.
Slope: from the secant to the tangent
A straight line has the same slope everywhere: how much grows when grows by 1. Working it out from two points is covered in “Straight line through two points”. A curve has its own slope at every point. To measure it, pick the point where you want the slope, take a second point on the curve at a distance from it and draw a straight line through both points, the secant (“Sekante”). Its slope
is called the difference quotient (“Differenzenquotient”). Then make smaller and smaller. The second point moves up to the first, and the secant turns into the tangent (“Tangente”), the straight line that touches the curve at the point . The number the slope approaches is the derivative :
Take at the point , where . With the second point is , with it is . The slope of the secant for smaller and smaller :
| Slope of the secant | |
|---|---|
| 5 | |
| 1 | |
| 0.1 | about 0.994 |
| 0.01 | about 0.999 |
The slopes approach 1, so . In the picture below the second point slides towards the first, and the secant settles onto the tangent with slope 1.
What the number means: if grows from 4 by a small amount, grows by almost as much. That is why the derivative is also called the rate of change (“Änderungsrate”). You do not have to work it out this way in the exam: the rules below give the derivative at once.
Your turn1 of 2
For you need the slope of the secant between the points and . What is it?
Sum, factor and number
Two cases you see at once from straight lines. has the slope 3, so and . A number without , say 7, does not change at all; its derivative is 0. Three rules follow, where and are any functions and is any number:
A sum is differentiated term by term, a constant factor stays. An example: .
Powers and roots
Sum and factor break a polynomial into single powers, but they do not give the derivative of itself. The secant finds it again. For at the point 3, the secant at gives the slope , which approaches 6. The same calculation gives 2 at the point 1 and 10 at the point 5: each time . For the points 1, 2, 3 give 3, 12, 27, that is . The exponent moves to the front as a factor and itself becomes 1 smaller. So for every power :
Together with the sum and factor rules: .
The rule also holds for fractional and negative exponents. There is no proof for them here: the rule is simply accepted. A root is written as a power: , because . That is why you first write roots and fractions as a power:
| Function | As a power | Derivative |
|---|---|---|
This also confirms the slope that the section on the secant found for at the point 4: , at the point 4 that is .
The number e, the exponential function and the logarithm
Powers cover polynomials and roots. In tasks on the growth of capital and demand, sits in the exponent. Such functions need a number of their own. The number comes from interest: if 1 € at 100 % a year is compounded more and more often, after a year the amount approaches €. The section “Continuous compounding and the number e” shows this step by step.
The function (the exponential function, “e-Funktion”) grows at every point as fast as its value is large. This is the same number : if the base is chosen so that has slope 1 at the point 0, the result is 2.71828, and the two definitions do not disagree.
The natural logarithm answers the question to which power you must raise to get : , , . On the calculator these are the keys “” and “ln”.
At the point 2, has the slope .
The product
Now the derivatives of all the single functions are known. Tasks often multiply them, say , and it is tempting to just multiply the derivatives. Check it on . By the power rule its derivative is . Multiply the derivatives of the factors and you get , which is not . The right answer comes from a sum of two products, each the derivative of one factor times the other factor: . This is the product rule (“Produktregel”): if two functions and , both with , are multiplied,
For a fraction there is the quotient rule (“Quotientenregel”) ; it is on the formula sheet printed in the task booklet, see “The formula sheet: how to find the right line”.
Chain: a function inside a function
In a product two functions stand side by side. It also happens that one sits inside the other. In the number first goes through the inner function , then the result goes through the outer one, squaring . For such a chain the chain rule holds:
You differentiate the outer function, leave inside, and multiply by the inner derivative (“innere Ableitung”) . An example: .
With the chain you get two more rules that come up all the time:
- Exponential function with an expression in the exponent: . An example: , the exponent stays up top.
- Another base : , because , so ; hence . An example: has the slope at the point 1. Here is in the exponent, so the power rule does not fit.
Your turn2 of 2
What is the derivative of at the point ?
The second derivative
All the rules above give the first derivative. You can take it once more. In the exam that is what tells a maximum from a minimum. For , the derivative is a function too and has a derivative of its own: . It is called the second derivative (“zweite Ableitung”), , and is worked out by the same rules. The first derivative says whether rises, the second says whether the slope itself grows. For the slope grows towards the right: the further right, the steeper, and . What this means for maxima and minima is explained in the article “Marginal cost and elasticity: what they mean”.
The topic in a minute
- Derivative (“Ableitung”) is the slope of the tangent to the graph at the point : if grows by 1, changes by about . Trap: the slope of the secant through two points is only an approximation.
- Sum and factor: , , a number without gives 0. Trap: has the derivative 7, not 0; only a number without gives zero.
- Power: , also for roots and fractions: , . Trap: not lowering the exponent by 1, or losing the minus in .
- Exponential function and logarithm: , , . Trap: differentiating like a power.
- Product: . Trap: .
- Chain (chain rule, “Kettenregel”): . Trap: forgetting the inner derivative .
- Second derivative is the derivative of . Trap: putting the point into too early and then differentiating a number (that gives 0).
- Formula sheet (“Formelsammlung”, the page of formulas in the exam booklet (“Angabenheft”), right after the tasks): the basic derivatives and the sum, product, quotient and chain rules are on it.
What comes next
- Values of the derivative at a point with the chain rule, products with , roots and in the exponent: the method “The derivative at a point: roots, the chain rule, x to the power x”.
- Marginal cost, elasticity, rising, maximum and minimum: “Marginal cost and elasticity: what they mean”; the methods “Maximum profit, shutdown point, order size” and “Elasticity and extrema” stand on it.
Was this article helpful?