Skip to main content

Derivative Definitions

For f(x)=x2f(x)=x^2, moving xx from 33 to 44 changes the output from 99 to 1616, so the average rate of change is (16−9)/(4−3)=7(16-9)/(4-3)=7. Moving instead from 33 to 3.13.1 changes the output from 99 to 9.619.61, giving an average rate of (9.61−9)/0.1=6.1(9.61-9)/0.1=6.1.

Write the nonzero input change as hh, keeping the starting point at a=3a=3. Dividing the output change by hh gives:

f(3+h)−f(3)h=(3+h)2−9h=6+h(h≠0).\frac{f(3+h)-f(3)}h =\frac{(3+h)^2-9}h=6+h\quad(h\ne0).

This ratio is the difference quotient, the average rate of change between the two points.

Input change hhDifference quotient 6+h6+h
1177
0.10.16.16.1
0.010.016.016.01
−0.1-0.15.95.9

The slopes approach 66 as hh approaches zero from either side. At h=0h=0, the quotient is undefined; the derivative comes from the limit, not from substituting zero into the quotient.

In the square-function view below, P=(3,9)P=(3,9) stays fixed while Q=(3+h,(3+h)2)Q=(3+h,(3+h)^2) moves with hh; for h≠0h\ne0, the line through them is a secant whose slope is the difference quotient.

xPQ
CurveSecant PQTangent at P
P = (3.00, 9.00) · Q = (4.00, 16.00)Difference quotient: 7.00The secant slope is 6 + h. As Q approaches P from either side, the slope approaches 6.

Formal Definition​

For a function y=f(x)y=f(x), the derivative is the instantaneous rate of change defined by

f′(x)=lim⁡h→0f(x+h)−f(x)h,f'(x) = \lim_{{h \to 0}} \frac{f(x + h) - f(x)}{h},

provided the limit is a finite real number. The ordinary two-sided derivative requires ff to be defined on an open interval around the point and the quotients from both sides to approach the same value. At a domain endpoint, specify a one-sided derivative. The limit definition and examples develop this distinction between a function value and a derivative.

For the square function, f′(3)=6f'(3)=6. More generally, at any real aa,

f(a+h)−f(a)h=a2+2ah+h2−a2h=2a+h(h≠0),\frac{f(a+h)-f(a)}h =\frac{a^2+2ah+h^2-a^2}h=2a+h\quad(h\ne0),

so f′(a)=2af'(a)=2a. Cancellation is valid for nonzero hh; no division by zero is performed.

From a difference quotient to a local approximation​

Slope of the Tangent Line​

The difference quotient is the slope of a secant through (a,f(a))(a,f(a)) and (a+h,f(a+h))(a+h,f(a+h)). Its finite limit m=f′(a)m=f'(a) is the slope of the tangent at (a,f(a))(a,f(a)).

Equation of the Tangent Line​

With this slope and the point (a,f(a))(a,f(a)), the tangent line is

y=f(a)+f′(a)(x−a).y = f(a) + f'(a)(x - a).

At a=3a=3 for x2x^2, this is y=9+6(x−3)y=9+6(x-3). Taking h=0.01h=0.01, the line predicts 9+6(0.01)=9.069+6(0.01)=9.06, while the exact value is (3.01)2=9.0601(3.01)^2=9.0601. The error is h2=0.0001h^2=0.0001.

More generally, differentiability is equivalent to

f(a+h)=f(a)+f′(a)h+r(h),lim⁡h→0r(h)h=0.f(a+h)=f(a)+f'(a)h+r(h),\qquad \lim_{h\to0}\frac{r(h)}h=0.

The remainder is written r(h)=o(h)r(h)=o(h): its magnitude becomes negligible relative to ∣h∣|h|. Subtracting f′(a)f'(a) from the defining difference quotient gives this condition. For the square function at 33, r(h)=h2r(h)=h^2, but differentiability alone does not guarantee a quadratic error bound or accuracy far from aa.

When a local slope does not exist​

The same remainder identity implies f(a+h)→f(a)f(a+h)\to f(a), so differentiability implies continuity. The converse fails: f(x)=∣x∣f(x)=|x| is continuous at 00, but its difference quotient ∣h∣/h|h|/h is −1-1 for h<0h<0 and 11 for h>0h>0. The one-sided slopes disagree, so there is no two-sided derivative.

Return to the figure above and switch to ∣x∣|x| at 00. Move hh toward zero from both sides: unlike x2x^2 at 33, the secant slopes do not approach the same value. At h=0h=0, the difference quotient remains undefined for both functions.

Also, f(x)=x3f(x)=\sqrt[3]{x} has quotient h−2/3→+∞h^{-2/3}\to+\infty at 00; a vertical tangent is not a finite derivative.

Interpretation of the Derivative​

Rate of Change​

The tangent slope f′(a)f'(a) gives the instantaneous output change per unit input change near aa. The local approximation f(a+h)≈f(a)+f′(a)hf(a+h)\approx f(a)+f'(a)h expresses that rate as a prediction.

Physical Interpretation in Motion​

If f(t)f(t) is an object's position at time tt, then f′(a)f'(a) is its velocity at t=at=a. Position in metres and time in seconds give a derivative in metres per second. For several inputs, the analogous local linear term uses a gradient and a displacement vector.

Equivalent Notations for Derivatives​

For y=f(x)y=f(x), these notations all denote the derivative with respect to xx:

  • f′(x)f'(x)
  • y′y'
  • dydx\frac{dy}{dx}
  • ddxf(x)\frac{d}{dx} f(x)
  • Df(x)Df(x)

Notation for Derivatives at a Specific Point​

At x=ax=a, write

f′(a)=dydx∣x=a=ddxf(x)∣x=a=Df(a).f'(a) = \left. \frac{dy}{dx} \right|_{x=a} = \left. \frac{d}{dx} f(x) \right|_{x=a} = Df(a).

References​

Explore connectionsOpen network