Family of Lines

Suppose someone gives you two lines that cross at a point P, and asks you to find a third line that also passes through P and satisfies some additional condition — maybe it has a specific slope, or it passes through another given point, or it is perpendicular to some direction. The brute-force approach is: solve the two line equations simultaneously to find P, then write the equation of a line through P with the required property.

That works. But there is a far more elegant approach that skips the step of finding P entirely.

Consider the two lines x + y - 4 = 0 and 2x - y - 2 = 0. They meet somewhere — you could solve to find the intersection at (2, 2), but hold off. Instead, look at this expression:

Pick any number for \lambda. For \lambda = 0, you get the first line back. For \lambda = 1, you get 3x - 6 = 0, which is x = 2 — a vertical line. For \lambda = -1, you get -x + 2y - 2 = 0, which is x - 2y + 2 = 0. Every choice of \lambda gives a different line, and every single one of them passes through the intersection point (2, 2).

Why? Because the intersection point (2, 2) makes both x + y - 4 = 0 and 2x - y - 2 = 0 true. So it makes (x + y - 4) + \lambda(2x - y - 2) = 0 + \lambda \cdot 0 = 0 true, no matter what \lambda is. The intersection point is automatically on every member of the family.

This is the central idea, and the rest of the article unpacks it, makes it precise, and shows you how to use it.

The formal statement

Why "except L_2"? To get L_2 from the expression L_1 + \lambda L_2 = 0, you would need L_1 to vanish entirely, which would require "\lambda = \infty." Since \lambda is a finite real number, L_2 is the one member of the family you cannot reach. (If you want L_2 as well, just swap the roles: write L_2 + \mu L_1 = 0 and set \mu = 0.)

Why is this useful? Because you never need to compute the intersection point. If a problem says "find the line through the intersection of 3x + 2y - 1 = 0 and x - y + 3 = 0 that is perpendicular to 4x + 5y = 7," you write the family, impose the perpendicularity condition, solve for \lambda, and you have the answer — without ever finding the intersection.

Building intuition: what the parameter does

Think of \lambda as a dial. When \lambda = 0, the family equation is just L_1 = 0 — you are sitting on the first line. As you turn \lambda up from 0, the line rotates around P, sweeping through different directions. Large positive \lambda makes the L_2 terms dominate, so the line tilts toward L_2's direction. Large negative \lambda tilts the line the opposite way.

The line L_2 is the one direction the dial can never quite reach — it would require \lambda to be infinite. But every other direction through P corresponds to exactly one finite value of \lambda.

A proof that the family covers all other lines through P. Take any line M through P, other than L_2. Suppose M has equation \alpha x + \beta y + \gamma = 0, and P = (x_0, y_0) satisfies M. Since P is on L_1 and L_2, we have L_1(P) = 0 and L_2(P) = 0. Now consider L_1 + \lambda L_2 evaluated at P: it equals 0 + \lambda \cdot 0 = 0 for all \lambda, confirming that every member passes through P. To show that M is actually in the family, we need to find a specific \lambda such that L_1 + \lambda L_2 is proportional to M. Since L_1 and L_2 are not parallel, their coefficient vectors (a_1, b_1, c_1) and (a_2, b_2, c_2) are linearly independent. Any line through P has its coefficient vector in the two-dimensional span of these two vectors (because passing through P is one linear constraint on three coefficients, leaving a two-dimensional family). So (\alpha, \beta, \gamma) must be a linear combination of (a_1, b_1, c_1) and (a_2, b_2, c_2). As long as M \neq L_2, the coefficient of (a_1, b_1, c_1) is nonzero, and dividing through gives the form L_1 + \lambda L_2 for a finite \lambda.

Here is a table showing how different values of \lambda produce different lines from the family (x + y - 4) + \lambda(2x - y - 2) = 0:

Every line in the table passes through (2, 2).

Worked examples

Common confusions

• "The family L_1 + \lambda L_2 = 0 contains all lines in the plane." It does not. It contains only lines through the intersection of L_1 and L_2. Lines in other directions, through other points, are not in this family.

• "Setting \lambda = 0 gives L_2." No — \lambda = 0 gives L_1. To get L_2 you would need \lambda \to \infty. The family is built from L_1, with \lambda controlling how much of L_2 is mixed in.

• "If I get a contradiction when solving for \lambda, I made an error." Not necessarily. A contradiction means the required line is L_2 itself — the one line the family cannot produce. This is a legitimate outcome, not an error.

• "I can use this method even when L_1 and L_2 are parallel." No. If L_1 \parallel L_2, they don't intersect (or intersect "at infinity"), and the family L_1 + \lambda L_2 = 0 produces parallel lines, not a fan of concurrent lines. The method requires a proper intersection point.

• "The family equation is quadratic." It is not — L_1 + \lambda L_2 is always linear in x and y for every fixed \lambda. The parameter \lambda is a constant, not a variable. Each choice of \lambda gives one straight line.

The parametric view

The family L_1 + \lambda L_2 = 0 is a one-parameter family — one dial (\lambda) generates infinitely many lines. This idea shows up throughout mathematics. In the next chapter on circles, you will meet the family of circles through the intersection of two given circles, written S_1 + \lambda S_2 = 0, which works by exactly the same logic: the intersection points of S_1 and S_2 satisfy both equations, so they satisfy S_1 + \lambda S_2 = 0 for every \lambda.

The "parametric family" idea also connects to the concept of a pencil (an older term, still used in projective geometry). A pencil of lines is a one-parameter family of lines sharing a common point. The expression L_1 + \lambda L_2 = 0 is the algebraic way to write a pencil.

Applications

Finding a specific line through an intersection

This is the bread-and-butter application. Given two lines that intersect, you need the line through their intersection satisfying one extra condition (a slope, a point, a perpendicularity condition, a distance condition). Write the family, impose the condition, solve for \lambda. The conditions you can impose include:

• Given slope m: set -(a_1 + \lambda a_2)/(b_1 + \lambda b_2) = m and solve for \lambda.
• Passes through a given point (p, q): substitute into L_1(p, q) + \lambda L_2(p, q) = 0 and solve for \lambda.
• Perpendicular to a given line \ell x + my + n = 0: the perpendicularity condition (a_1 + \lambda a_2)\ell + (b_1 + \lambda b_2)m = 0 gives a linear equation in \lambda.
• At a given distance from a point: this condition is slightly more involved but still gives a solvable equation in \lambda.

Each of these imposes one equation on the one unknown \lambda, pinning down a unique member of the family (or, in the edge case, revealing that the answer is L_2).

Concurrence of three lines

Three lines are concurrent (all pass through one point) if and only if the third line can be written as a linear combination of the first two: L_3 = L_1 + \lambda L_2 for some \lambda. Equivalently, the determinant

This determinant condition is a direct consequence of the family-of-lines theory. If L_3 is a member of the family L_1 + \lambda L_2 = 0, then the three coefficient triples are linearly dependent, forcing the determinant to zero.

Simplifying locus problems

When a locus involves the condition "passes through a fixed point," you can sometimes express the varying equation as L_1 + \lambda L_2 = 0. The fixed point is the intersection of L_1 and L_2, and identifying those two lines from the equation immediately gives you the fixed point without solving.

For instance, consider the line y = mx + \frac{1}{m} as m varies. Rewrite it as mx - y + \frac{1}{m} = 0, or m^2 x - my + 1 = 0. This is a quadratic in m: m^2 x - my + 1 = 0. For the line to pass through a fixed point for all m, you need x = 0 and 1 = 0 (from the constant term), which is impossible — or you need a more careful analysis. Treating the equation as m(mx - y) + 1 = 0, you see this is of the form L_1 + mL_2 = 0 where L_1: 1 = 0 (the "line at infinity") and L_2: mx - y = 0. The parametric approach reveals that the family does not have a common finite point for all m — different values of m give lines through different points.

A cleaner example: the line (2 + \lambda)x + (1 - \lambda)y = 3 + \lambda can be rewritten as (2x + y - 3) + \lambda(x - y + 1) = 0. For all \lambda, this line passes through the intersection of 2x + y - 3 = 0 and x - y + 1 = 0. Solving: adding the two equations gives 3x - 2 = 0, so x = 2/3, and y = 2/3 + 1 = 5/3. The fixed point is (2/3, 5/3).

Going deeper

The linear algebra perspective

The set of all lines ax + by + c = 0 can be thought of as living in a space of triples (a, b, c). Two non-parallel lines give two linearly independent triples. The family L_1 + \lambda L_2 = 0 is the span of these two triples — the set of all linear combinations \alpha(a_1, b_1, c_1) + \beta(a_2, b_2, c_2), which is a two-dimensional subspace. (We only care about the direction of the triple, since 2x + 2y + 2 = 0 and x + y + 1 = 0 are the same line, so we really work in projective space, where triples that differ by a scalar factor are identified.)

In this view, \lambda = \beta/\alpha. The case \alpha = 0 gives L_2; every other ratio gives a distinct member of the family. This is why L_2 is "at \lambda = \infty" — it corresponds to \alpha = 0.

The dual picture

In projective geometry, points and lines are dual objects. A "pencil of lines" — a one-parameter family of lines through a common point — is dual to a "range of points" — the set of all points on a common line. The family L_1 + \lambda L_2 = 0 is the algebraic description of a pencil. The dual statement is: given two points P_1 and P_2, the parametric family P_1 + t(P_2 - P_1) traces out all points on the line P_1 P_2. The structures are the same; only the roles of "point" and "line" swap.

This duality is one of the most beautiful ideas in geometry, and the family-of-lines equation is your first encounter with it.

Connection to circle families

The same idea extends to conics. If S_1 = 0 and S_2 = 0 are two circles, the family S_1 + \lambda S_2 = 0 gives a one-parameter family of curves passing through the intersection points of the two circles. For most values of \lambda, the equation represents a circle. But for exactly one value — the \lambda that makes the x^2 and y^2 terms cancel — the equation drops to degree 1 and represents a straight line. That line is the radical axis of the two circles: the locus of points having equal power with respect to both circles.

The intersection points of S_1 and S_2 lie on every member of the family, just as the intersection point of two lines lies on every member of the line family. The logic is identical: if a point satisfies both S_1 = 0 and S_2 = 0, it satisfies S_1 + \lambda S_2 = 0 for every \lambda, because 0 + \lambda \cdot 0 = 0.

This is one of the most useful techniques in circle geometry, and you will meet it properly in the article on Circle - Standard Forms.

Where this leads next

The parametric family of lines connects naturally to several related topics:

• Angle Bisectors — the bisectors of two intersecting lines are specific members of the family L_1 + \lambda L_2 = 0, chosen by the equal-distance condition.
• Systems of Linear Equations — the intersection point of L_1 and L_2 is the solution of the corresponding 2 \times 2 system, which the family method lets you avoid computing directly.
• Locus — many locus problems reduce to identifying a fixed point that a one-parameter family of lines passes through.
• Circle - Standard Forms — the family S_1 + \lambda S_2 = 0 for circles is the direct analogue of the line family.
• Pair of Straight Lines — a pair of lines through the origin can be viewed as a degenerate conic, and the family idea extends to second-degree equations.