Mathwords logoMathwords

Submatrix — Definition, Formula & Example

A submatrix is the smaller matrix you get by deleting some rows, some columns, or both from a larger matrix. The entries that remain stay in their original order.

Let A=(aij)A = (a_{ij}) be an m×nm \times n matrix, and choose nonempty index sets I⊆{1,…,m}I \subseteq \{1, \dots, m\} and J⊆{1,…,n}J \subseteq \{1, \dots, n\}. The submatrix A[I,J]A[I, J] is the ∣I∣×∣J∣|I| \times |J| matrix of entries aija_{ij} with i∈Ii \in I and j∈Jj \in J, with rows and columns kept in increasing index order. When I=JI = J, the submatrix is called principal. Many texts count AA itself as a submatrix and call any other submatrix proper.

Key Formula

A[I,J]=(aij)i∈I,  j∈J,size ∣I∣×∣J∣A[I,J] = \big(a_{ij}\big)_{i \in I,\; j \in J}, \qquad \text{size } |I| \times |J|
Where:
  • AA = The original m × n matrix
  • II = The set of row indices you keep
  • JJ = The set of column indices you keep
  • aija_{ij} = The entry of A in row i, column j
  • ∣I∣,∣J∣|I|, |J| = The number of rows and columns kept

Worked Example

Problem: Let A=[251430627189]A = \begin{bmatrix} 2 & 5 & 1 & 4 \\ 3 & 0 & 6 & 2 \\ 7 & 1 & 8 & 9 \end{bmatrix}. Form the submatrix obtained by deleting row 2 and columns 1 and 3. Then find the determinant of that submatrix.
Step 1: Deleting row 2 keeps rows I={1,3}I = \{1, 3\}. Deleting columns 1 and 3 keeps columns J={2,4}J = \{2, 4\}. Read off the entries where a kept row meets a kept column, preserving their order. The result is 2×22 \times 2, even though AA was 3×43 \times 4.
A[{1,3},{2,4}]=[a12a14a32a34]=[5419]A[\{1,3\},\{2,4\}] = \begin{bmatrix} a_{12} & a_{14} \\ a_{32} & a_{34} \end{bmatrix} = \begin{bmatrix} 5 & 4 \\ 1 & 9 \end{bmatrix}
Step 2: AA is not square, so it has no determinant. Its square submatrices do. The determinant of a square submatrix is called a minor of AA.
det⁡[5419]=5⋅9−4⋅1=45−4=41\det\begin{bmatrix} 5 & 4 \\ 1 & 9 \end{bmatrix} = 5 \cdot 9 - 4 \cdot 1 = 45 - 4 = 41
Answer: The submatrix is [5419]\begin{bmatrix} 5 & 4 \\ 1 & 9 \end{bmatrix}, and its determinant (a 2×22 \times 2 minor of AA) is 41.

Why It Matters

In a linear algebra course, you can find the rank of a matrix as the size of its largest square submatrix with a nonzero determinant. Sylvester's criterion tests whether a symmetric matrix is positive definite by checking the determinants of its leading principal submatrices. Positive definiteness matters for classifying critical points in multivariable calculus.

Common Mistakes

Mistake: Thinking a submatrix must be a solid block of adjacent entries, or rearranging the leftover entries.
Correction: The rows and columns you keep do not have to be next to each other. In the example, rows 1 and 3 and columns 2 and 4 were kept. You must keep the entries in their original relative order. Swapping rows or columns can change the sign of a determinant.

Related Terms

  • Cofactor — Signed determinant of the submatrix left after one deletion
  • Cofactor Matrix — Collects the cofactors from every one-row, one-column deletion
  • Adjugate — Transposed cofactor matrix built from these submatrix determinants
  • Coefficient Matrix — Submatrix of the augmented matrix without its last column
  • Augmented Matrix — Contains the coefficient matrix as a submatrix
  • Column of a Matrix — A single column is an m by 1 submatrix
  • Matrix Determinant — Determinants of square submatrices give minors and rank tests