Differentiation Explained: The Maths of Change

Look at a car's speedometer. It tells you how fast you are moving right now. But average speed is distance traveled divided by elapsed time. At one frozen instant, no time passes. So what, exactly, is that needle measuring? This question takes us into differentiation, and into a bitter fight between Isaac Newton and Gottfried Wilhelm Leibniz. By the end, you will turn a distance formula into a speed reading, find the slope of a curve, and spot a trap that catches people who can recite the rules perfectly.

In sixteen sixty six, Newton wrote about a method for studying changing quantities. He called his rates of change fluxions. Leibniz developed calculus independently, and published his first calculus paper in sixteen eighty four, before Newton published his own method. Who deserved the credit? Supporters traded accusations of copying. We will return to that fight. First, meet the idea they both helped uncover: a curve can have a slope at one point, even though the whole curve keeps bending.

Imagine a car moving forward along a straight road. Our deliberately simple model says its distance traveled, s, in meters, equals time, t, in seconds, squared. After two seconds, it has traveled four meters. After four seconds, sixteen meters. Over that interval it travels twelve meters in two seconds, an average speed of six meters per second. On the graph, join those two points. The slope is rise divided by run: twelve divided by two. That straight line is a secant. It compares two moments.

But our question is the speed at exactly two seconds. Bring the second point closer. From two to three seconds, the average is five meters per second. From two to two point one, it is four point one. From two to two point zero one, four point zero one. Notice where those numbers are heading. As the time interval shrinks toward zero, the secant slopes approach four. The limiting line is the tangent at two seconds. Its slope gives the instantaneous speed: four meters per second.

Here is the crucial move. We never divide by a zero time interval. We calculate slopes over nonzero intervals, then ask what value they approach. That is a limit. Call the shrinking input gap h. The derivative of f at x is a limit. Evaluate f at x plus h, then subtract f at x. Divide that whole difference by h. Now let h approach zero. The numerator is the change in output. The denominator is the change in input. If that limit exists, differentiation gives the instantaneous rate of change. A smaller gap reveals a local slope.

Watch the algebra for f of x equals x squared. Replace x with the whole quantity x plus h, square that quantity, then subtract x squared. Expanding leaves two x h plus h squared. Divide by h. Because h is nonzero, factor it out of the numerator and cancel, leaving two x plus h. Now take the limit as h approaches zero. The answer is two x. We have found a whole slope function. At x equals two, the slope is four. At x equals three, it is six. One formula answers every point on this curve.

You will see the derivative written as f prime of x, or as d y over d x when y equals f of x. Both mean the rate at which y changes with x. The d notation comes from Leibniz. Do not confuse the original function with its derivative. For our square function, f of three equals nine: that is the height of a point. f prime of three equals six: that is the slope there. Same input, different question. Height and steepness are not interchangeable.

Chapter three: a shortcut with a reason behind it. You do not need to expand a limit every time. For a positive integer power, the power rule says: multiply by the exponent, then reduce the exponent by one. x cubed becomes three x squared. x to the fifth becomes five x to the fourth. A constant multiplier stays. So three x cubed becomes nine x squared. A constant on its own becomes zero, because a horizontal line has zero slope. And you can differentiate a sum or difference term by term.

Try this: y equals three x cubed minus five x plus seven. First, the cubic term becomes nine x squared. Next, minus five x becomes minus five. Finally, seven becomes zero. So d y over d x equals nine x squared minus five. If the question asks for the gradient at x equals two, substitute into the derivative: nine times four minus five equals thirty one. Here is the trap: substituting into the original formula would give the point's height, not its gradient. Differentiate first, then substitute for the gradient.

Now use a derivative to write a tangent equation. Return to y equals x squared, at x equals three. The original function gives the point with coordinates three comma nine. The derivative, two x, gives the slope: six. Put both into the point slope formula: y minus nine equals six times x minus three. Simplify to y equals six x minus nine. That line matches the curve's slope at the chosen point. It is a good local approximation nearby, but it does not follow the curve forever. Zoom out and the gap grows.

Chapter four: read what the curve is doing. A positive derivative throughout an interval means the graph is rising there. A negative derivative throughout an interval means it is falling. A zero derivative means a horizontal tangent. We call that a stationary point. Picture a smooth hilltop: rising, then flat, then falling. The derivative changes from positive to negative, giving a local maximum. In a smooth valley, it changes from negative to positive, giving a local minimum. The derivative's sign tells a story that a list of coordinates can hide.

Here is a worked example. For y equals x squared minus four x plus five, the derivative is two x minus four. Set it equal to zero. That gives x equals two. Before two, the derivative is negative; after two, it is positive. Falling then rising: a minimum. Substitute two into the original function to find its height: one. The stationary point has coordinates two comma one. But zero slope alone does not guarantee a maximum or minimum. For y equals x cubed, the slope is zero at the origin, yet the graph keeps rising through it.

Your turn. Pause and differentiate y equals two x cubed plus four x minus nine. Then find its gradient at x equals one. Ready? The derivative is six x squared plus four. At one, that gives ten. The minus nine disappears because it contributes no change. Now a second quick check: if a graph is high above the axis but sloping downward, is its derivative positive or negative? Negative. Height describes where you are. The derivative describes how that height is changing.

Return to the car. Distance traveled, s, equals t squared. Its derivative is two t, in meters per second. At two seconds, the speed is four meters per second. Differentiate again and you get two meters per second squared: the acceleration in this model. More generally, differentiating position gives velocity, which includes direction. Speed is the magnitude of velocity. Our forward moving car makes them agree numerically. A modern speedometer uses sensors; it does not need to draw this graph. But instantaneous rate of change is the mathematical idea its reading represents.

And the fight? In seventeen twelve, the Royal Society issued a report favoring Newton, but Newton himself had directed the investigation. Imagine choosing yourself as referee, then declaring yourself the winner. Historians today credit Newton and Leibniz with developing calculus independently, building on earlier mathematics. Newton's language of motion and Leibniz's notation both survived. The useful inheritance is bigger than the argument: a way to connect a changing quantity with its rate of change. That is why a curve on paper and a needle on a dashboard can tell the same story.

Keep four ideas. One: a derivative is an instantaneous rate of change, found through a limit. Two: on a graph, it gives the tangent's gradient. Three: rules let you calculate derivatives efficiently. Four: use the derivative's sign to interpret change, and test stationary points before calling them maxima or minima. Useful phrases for a written answer are instantaneous rate of change, gradient of the tangent, and derivative changes sign. Use them with the actual calculation and its units. Next time you see a speedometer, you know what that single instant means.