Imagine you're standing on a curved road. The tangent line is the direction your feet would point if you just kept walking straight ahead — it's the "along the curve" direction. Now, if you wanted to walk directly away from the curve, stepping straight out from where you're standing, that's the normal line. It's the line that stands at a perfect right angle to the tangent at that exact point.
The whole secret: The normal line is just the tangent line's perpendicular partner at the same point. Everything else is details.
Why this works
If you have a curve y=f(x) and a point P=(x0,y0) on it, the tangent slope is mtan=f′(x0). For two lines to be perpendicular (and not vertical), their slopes must multiply to −1:
mtan⋅mnorm=−1
So the normal slope is simply:
mnorm=−f′(x0)1(as long as f′(x0)=0)
Once you have the slope, you write the line using the point-slope form:
y−y0=mnorm(x−x0)
Putting it together:
y−y0=−f′(x0)1(x−x0)
Step by step — a worked example
Let's find the normal line to y=x2 at x=1.
Step 1: Find the point on the curve.
y0=(1)2=1, so the point is (1,1).
Step 2: Find the derivative at that point.
f′(x)=2x, so f′(1)=2.
Step 3: Get the normal slope.
mnorm=−21.
Step 4: Write the equation using point-slope form.
The normal to x2=4y at a point (x1,y1) on the curve has slope −x12. Forcing this normal to pass through (−1,4) leads to the cubic x13−8x1+8=0, which has three real roots: x1=2,−1+5,−1−5. Each root gives a genuine normal line, so there are three normals through (−1,4): x+y=3, y=−21+5x+27−5, and y=25−1x+27+5.
Setting up the normal at a general point
The curve is x2=4y. Differentiating implicitly with respect to x:
2x=4dxdy⇒dxdy=2x
So at a point (x1,y1) on the parabola, the tangent slope is 2x1, and since the normal is perpendicular to the tangent, its slope is
mn=−x12(x1=0).
Since (x1,y1) lies on x2=4y, we also have y1=4x12.
The equation of the normal at (x1,y1) is therefore
y−y1=−x12(x−x1).
Imposing the condition that it passes through (−1,4)
First note (−1,4) does not itself lie on the curve, since (−1)2=1=4(4)=16 — so we are looking for the point(s) of contact (x1,y1) whose normal happens to pass through this external point.
Substituting x=−1,y=4 into the normal's equation:
4−4x12=−x12(−1−x1)=x12(1+x1).
Multiplying both sides by 4x1:
16x1−x13=8(1+x1)=8+8x1
8x1−x13−8=0⇒x13−8x1+8=0.
Solving the cubic
Testing x1=2: 23−8(2)+8=8−16+8=0✓. So (x1−2) is a factor:
x13−8x1+8=(x1−2)(x12+2x1−4).
Solving x12+2x1−4=0 by the quadratic formula:
x1=2−2±4+16=2−2±20=−1±5.
So all three roots are x1=2,−1+5,−1−5 — all real, so all three give genuine normal lines (this is why the question has three answers, not one).
Method: Finding a Normal Line to a Curve That Passes Through a Given External Point
Use this method whenever you're asked for the normal to a curve that passes through a point which is not stated to lie on the curve itself — you must first find which point(s) of contact on the curve produce a normal through the given point.
Steps
Step 1: Confirm the given point is not automatically the point of contact
Substitute the given point's coordinates into the curve's equation. If it doesn't satisfy the equation, the point of contact is unknown and must be found — this is the key difference from a "normal at a given point" problem.
Step 2: Set up the normal's equation at a general point of the curve
Let (x1,y1) be the (unknown) point of contact on the curve. Differentiate the curve implicitly to get the tangent slope in terms of x1, then take the negative reciprocal for the normal slope. Write the normal's equation in point-slope form using (x1,y1), and use the curve's own equation to express y1 in terms of x1 (so only one unknown, x1, remains).
Step 3: Force the normal through the given external point
Substitute the given point's coordinates into the normal equation from Step 2. This produces a single equation in x1 alone — typically a polynomial, since the slope term introduces a variable in the denominator that clears when you multiply through.
Mistake 1: Assuming the given point (−1,4) lies on the parabola
(−1,4) does not satisfy x2=4y (since (−1)2=1=4(4)=16), so it cannot be used directly as the point of contact (x1,y1). The point of contact must be an unknown (x1,x12/4) whose normal happens to pass through (−1,4) — treating (−1,4) itself as the point on the curve skips this step and gives a wrong slope.
Mistake 2: Stopping after finding one root of the cubic and not checking the others …