Linear Algebra
1.1214 min read

Echelon Form and Basic/Free Variables

To solve a linear system systematically, we transform it into echelon form — a staircase shape where each equation's leading nonzero term is strictly to the right of the leading term of the equation above.

Once in echelon form, we can identify two types of variables: basic variables (also called pivot or leading variables) are those that appear as the leading term of some equation. Free variables are all others — they can take any value.

The dimension of the solution set equals the number of free variables. If there are ff free variables, the solution set is ff-dimensional. We express each basic variable in terms of the free variables.

Reading off the solution from echelon form is the job of back substitution — working from the last equation upward, substituting the already-solved variables into earlier equations.

Formal View

Definition 1.12 — Echelon Form
A matrix (or system of equations) is in echelon form (row echelon form) if: \begin{itemize} \item All zero rows are at the bottom. \item The leading nonzero entry (pivot) of each nonzero row is strictly to the right of the pivot of the row above. \end{itemize}
Definition 1.12b — Basic and Free Variables
In echelon form, a basic variable (pivot variable) corresponds to a column containing a pivot. A free variable corresponds to a non-pivot column. The solution set has dimension = (number of free variables).

Interactive Visualization

Back Substitution Animator

Why This Matters

Echelon form is the canonical simplified form for linear systems — all solution information is readable from it.

  • Every linear algebra software package reduces to echelon form internally
  • The number of free variables directly counts the degrees of freedom in engineering designs
  • Pivot positions identify which measurements constrain a system and which are redundant

Quiz

Question 1

A system in echelon form has 4 unknowns and 2 pivot columns. How many free variables are there?

Question 2

A system can have basic variables but no free variables, resulting in a unique solution.

Common Mistakes

  • Thinking free variables are "unknown" — they are free to be anything; we parameterize by them.
  • Confusing echelon form (staircase shape) with reduced echelon form (which also zeros entries above each pivot).
  • Forgetting to count zero rows as having no pivot — they reduce the effective number of constraints.