What Is a Homogeneous System of Linear Equations?
A homogeneous system of linear equations is a set of linear equations in which the constant term on the right side of every equation is zero. A system like $3x - 2y + z = 0$ and $x + y - z = 0$ is homogeneous; a system with any non-zero constant (say $x + y = 2$) is not — that one is called non-homogeneous.
In matrix form, a homogeneous system is written $AX = 0$, where $A$ is the coefficient matrix, $X$ is the column vector of unknowns, and $0$ is the zero column vector. Because the right side is the zero vector, substituting $X = 0$ (every unknown set to zero) always satisfies the system. That guaranteed answer is the trivial solution, and it is the feature that sets homogeneous systems apart from general linear equations.
What Are Trivial and Non-Trivial Solutions?
Every homogeneous system has the trivial solution — the one where all variables equal zero, $X = 0$. It is always there, so it is never interesting on its own.
A non-trivial solution is any solution in which at least one variable is non-zero. When a homogeneous system has a non-trivial solution, it actually has infinitely many, because scaling any non-trivial solution by a constant produces another valid solution. So a homogeneous system has exactly one of two outcomes: only the trivial solution, or the trivial solution plus infinitely many non-trivial ones. There is never a "finite handful of non-trivial answers" case.
When does a homogeneous system have a non-trivial solution?
For a square system written $AX = 0$, the test is the determinant of the coefficient matrix:
If $\det A \neq 0$, the matrix is invertible, and the only solution is the trivial one, $X = 0$.
If $\det A = 0$, the matrix is singular, and the system has infinitely many non-trivial solutions.
There is also a counting shortcut that needs no determinant: if a homogeneous system has more unknowns than equations, it always has a non-trivial solution. Three unknowns, two equations — non-trivial solutions are guaranteed, because there are not enough equations to pin every variable to zero. You can confirm the singular case with a determinant calculation or, for non-square systems, by row-reducing and counting free variables.
How Do You Solve a Homogeneous System?
Solving means describing all solutions, not just confirming the trivial one. The reliable method is row reduction.
Write the coefficient matrix $A$. The zero column on the right never changes under row operations, so you can drop it and work with $A$ alone.
Row-reduce $A$ to row echelon (or reduced row echelon) form.
Identify pivot variables and free variables. A free variable is any column without a pivot.
If there are no free variables, only the trivial solution exists. If there is at least one free variable, set it to a parameter (say $t$) and express the pivot variables in terms of it — that gives the infinite family of non-trivial solutions.
The presence of even one free variable is the signal that non-trivial solutions exist. This is the same row-reduction engine used to solve matrices generally, applied to the special case where the constants are all zero.
Examples of Homogeneous System of Linear Equations
The set runs from identifying a homogeneous system, through the trivial-only case, the most common misread, an infinite-solution case with a parameter, the more-unknowns-than-equations shortcut, and a determinant check.
Example 1
Which of these is a homogeneous system: (i) $2x + y = 0,\ x - 3y = 0$ or (ii) $2x + y = 5,\ x - 3y = 0$?
A system is homogeneous only when every constant term is zero. System (i) has zeros on both right sides. System (ii) has a $5$ in the first equation.
Final answer: System (i) is homogeneous; system (ii) is not.
Example 2
A common slip — does $x + y = 0,\ 2x - y = 0$ have non-trivial solutions?
Wrong attempt. A student notices both right sides are zero, recalls "homogeneous systems have infinitely many solutions," and answers "yes, infinitely many." But that rule only applies when $\det A = 0$. Skipping the determinant check and pattern-matching on the word "homogeneous" is the trap.
Correct. Compute the determinant of $A = \begin{bmatrix} 1 & 1 \ 2 & -1 \end{bmatrix}$:
$$\det A = (1)(-1) - (1)(2) = -1 - 2 = -3 \neq 0.$$
Since $\det A \neq 0$, the matrix is invertible and the only solution is the trivial one.
Final answer: Only the trivial solution $(x, y) = (0, 0)$. A homogeneous system does not automatically have non-trivial solutions — the determinant decides.
Example 3
Solve the system $x + 2y - z = 0,\ 2x + 4y - 2z = 0$.
The second equation is exactly twice the first, so it adds no new information — effectively one equation, three unknowns. With more unknowns than independent equations, non-trivial solutions exist. From $x + 2y - z = 0$, solve for $x$: $x = -2y + z$. Let $y = s$ and $z = t$ be free parameters:
$$(x, y, z) = (-2s + t,\ s,\ t).$$
Final answer: infinitely many solutions, $(x, y, z) = (-2s + t, s, t)$ for any real $s, t$. Setting $s = t = 0$ recovers the trivial solution.
Example 4
Solve $2x + 3y - z = 0,\ x - y + 2z = 0,\ x + 4y - 3z = 0$ given that $\det A = 0$.
Row-reduce the coefficient matrix. The third row turns out to be the difference of combinations of the first two, leaving one free variable. Carrying the reduction through gives the pivot variables in terms of $z = t$:
$$x = -t, \qquad y = t, \qquad z = t.$$
Final answer: $(x, y, z) = (-t, t, t)$ for any real $t$. For instance $t = 1$ gives the non-trivial solution $(-1, 1, 1)$; $t = 2$ gives $(-2, 2, 2)$.
Example 5
A homogeneous system has 4 unknowns and 2 equations. How many solutions does it have?
No computation needed for the qualitative answer. With more unknowns ($4$) than equations ($2$), there must be free variables — at least two of them.
Final answer: infinitely many solutions (non-trivial solutions guaranteed). The trivial solution is among them, but it is not the only one.
Example 6
For what value of $\lambda$ does $\lambda x + y = 0,\ x + \lambda y = 0$ have a non-trivial solution?
Non-trivial solutions require $\det A = 0$. The coefficient matrix is $\begin{bmatrix} \lambda & 1 \ 1 & \lambda \end{bmatrix}$:
$$\det A = \lambda^2 - 1 = 0 \implies \lambda = \pm 1.$$
Final answer: $\lambda = 1$ or $\lambda = -1$. For any other $\lambda$, only the trivial solution exists. (This is the seed of the eigenvalue idea — the values of $\lambda$ that force a singular matrix.)
Why Homogeneous Systems Sit at the Centre of Linear Algebra
"When does a square matrix collapse a non-zero vector to zero?"
That question — which is exactly the non-trivial-solution question — runs through far more than a textbook chapter.
Eigenvalues and eigenvectors. Finding an eigenvector means solving $(A - \lambda I)X = 0$, a homogeneous system. The eigenvalues are precisely the $\lambda$ values that make $\det(A - \lambda I) = 0$ — the same singular-matrix condition as Example 6. Quantum mechanics, vibration analysis, and Google's original PageRank all rest on this.
The null space. The set of all solutions to $AX = 0$ is the null space of $A$ — a foundational object that measures how much information the matrix loses. Engineers use it to find the "modes" of a structure that can move without stress, like the slack directions in a bridge truss.
Linear independence. A set of vectors is linearly independent exactly when the only solution to their homogeneous combination equalling zero is the trivial one. The determinant test from this article is the same test that decides independence.
Where Students Trip Up on Homogeneous Systems
Mistake 1: Assuming "homogeneous" automatically means infinitely many solutions
Where it slips in: The student sees all-zero constants and concludes non-trivial solutions must exist.
Don't do this: Skip the determinant (or free-variable) check and declare "infinitely many solutions" on sight.
The correct way: A homogeneous system has non-trivial solutions only when $\det A = 0$ (or there are more unknowns than equations). When $\det A \neq 0$, the only solution is trivial.
Mistake 2: Forgetting the trivial solution counts as a solution
Where it slips in: Asked "how many solutions," a student answers "zero" when only the trivial one exists.
Don't do this: Say a homogeneous system has "no solution." It is never empty — $X = 0$ always works.
The correct way: When $\det A \neq 0$, the answer is "exactly one solution — the trivial one," not "no solution." The memorizer who confuses non-homogeneous "no solution" cases with homogeneous ones is the usual culprit here.
Mistake 3: Misreading a dependent equation as a new constraint
Where it slips in: A system has three equations but one is a multiple or sum of the others, so it adds nothing — yet the student counts all three as independent.
Don't do this: Treat $x + 2y - z = 0$ and $2x + 4y - 2z = 0$ as two separate constraints (Example 3). The second is just the first doubled.
The correct way: Row-reduce first; the dependent row collapses to $0 = 0$ and reveals the true number of constraints. The second-guesser often spots this only after reduction — which is exactly why row reduction is the safe habit, not visual inspection.
A real-world version of the mistake. In control engineering, a singular matrix flags a system that has lost a degree of control — directions the operator can no longer steer. Treating a near-singular system as fully controllable has, in real aerospace and power-grid models, hidden exactly the failure modes the homogeneous null space was warning about. The arithmetic of $AX = 0$ runs cleanly; the danger is reading a singular system as if it behaved like an invertible one.
Key Takeaways
A homogeneous system of linear equations has all-zero constants, written $AX = 0$, and always has the trivial (all-zeros) solution.
Non-trivial solutions exist exactly when $\det A = 0$ (singular matrix) or when there are more unknowns than equations.
Solve by row-reducing the coefficient matrix and reading off free variables; one free variable means infinitely many solutions.
The most common mistake is assuming "homogeneous" automatically means infinitely many solutions — the determinant decides.
The homogeneous system $(A - \lambda I)X = 0$ is the gateway to eigenvalues, the null space, and linear independence.
Practice These Before Moving On
Determine whether $3x - y = 0,\ 6x - 2y = 0$ has non-trivial solutions, and describe all solutions.
For what value of $k$ does $kx + 2y = 0,\ 2x + ky = 0$ have a non-trivial solution?
A homogeneous system has 5 unknowns and 3 equations. State how many solutions it has and why.
Answer to Question 1: the second equation is twice the first, so $\det A = 0$; solutions are $(x, y) = (t, 3t)$ for any real $t$. Answer to Question 2: $\det A = k^2 - 4 = 0$, so $k = \pm 2$. Answer to Question 3: infinitely many, because there are more unknowns than equations. If Question 1 tripped you up, return to the determinant test in the non-trivial-solution section.
Want a live Bhanzu trainer to walk through more homogeneous system of linear equations problems? Book a free demo class — online globally.
Was this article helpful?
Your feedback helps us write better content