Linear Algebra: Common Mistakes and Misconceptions

Abstract

Linear algebra is foundational to mathematics, physics, computer science, and engineering, yet students frequently encounter conceptual pitfalls that impede deeper understanding. This article examines three prevalent misconceptions: the assumption that matrix multiplication is commutative, confusion about when a matrix can be inverted, and misunderstanding the conditions for diagonalizability. By clarifying these points with concrete examples, we aim to strengthen intuition and reduce errors in both coursework and applications.

Background

Linear algebra deals with vectors, matrices, and linear transformations. While the formal definitions are precise, students often develop intuitions based on analogies to scalar arithmetic that do not transfer directly. This gap between scalar and matrix operations creates systematic errors. Understanding where these intuitions fail is essential for mastery.

Key Results

Mistake 1: Assuming Matrix Multiplication is Commutative

A common error is treating matrix multiplication as commutative—that is, assuming $AB = BA$ for any matrices $A$ and $B$. This is false.

The definition of matrix multiplication ^{[matrix-multiplication]} specifies that for matrices $A$ (size $m \times n$) and $B$ (size $n \times p$), the product $AB$ is computed as:

$$(AB)_{ij} = \sum_{k=1}^{n} A_{ik} B_{kj}$$

This operation is fundamentally directional. The $i$-th row of $A$ is combined with the $j$-th column of $B$. Reversing the order changes which rows and columns interact, and may not even be defined if dimensions do not align.

Example: Let $A = \begin{pmatrix} 1 & 2 \\ 0 & 1 \end{pmatrix}$ and $B = \begin{pmatrix} 1 & 0 \\ 1 & 1 \end{pmatrix}$.

Then $AB = \begin{pmatrix} 3 & 2 \\ 1 & 1 \end{pmatrix}$ but $BA = \begin{pmatrix} 1 & 2 \\ 1 & 3 \end{pmatrix}$.

Clearly $AB \neq BA$. This non-commutativity is not an exception—it is the norm. Students must always be careful about the order of matrix multiplication, especially when solving equations or manipulating expressions.

Mistake 2: Confusing Invertibility with Other Matrix Properties

A second misconception is that a matrix is invertible if it "looks" invertible or has certain properties like being square or having non-zero entries. In reality, invertibility is determined by a single criterion: the determinant.

According to ^{[determinant-of-a-matrix]}, the determinant of a square matrix $A$ is a scalar that indicates invertibility. For a $2 \times 2$ matrix $A = \begin{pmatrix} a & b \\ c & d \end{pmatrix}$:

$$\det(A) = ad - bc$$

A matrix is invertible if and only if $\det(A) \neq 0$. If $\det(A) = 0$, the matrix is singular and has no inverse.

The determinant also encodes geometric meaning: it represents the scaling factor of the linear transformation ^{[determinant-properties]}. A zero determinant means the transformation collapses the space into a lower dimension, destroying information and making inversion impossible.

Common error: Students sometimes assume that because a matrix is square, it must be invertible. This is incorrect. A square matrix with determinant zero is singular.

Mistake 3: Misunderstanding Diagonalizability

A third misconception is that all square matrices can be diagonalized, or that diagonalizability depends only on having distinct eigenvalues.

According to ^{[diagonalizable-matrix]}, a matrix $A$ is diagonalizable if there exists an invertible matrix $P$ and a diagonal matrix $D$ such that:

$$A = PDP^{-1}$$

The columns of $P$ must be linearly independent eigenvectors of $A$. This is the key requirement: we need enough linearly independent eigenvectors to form a basis.

A matrix with repeated eigenvalues may or may not be diagonalizable. If an eigenvalue has algebraic multiplicity greater than its geometric multiplicity (the dimension of the corresponding eigenspace), the matrix cannot be diagonalized. Not all square matrices are diagonalizable.

Implication: Diagonalization is a powerful tool ^{[diagonalizable-matrix]} for simplifying matrix powers and solving differential equations, but it is not always available. Checking diagonalizability requires computing eigenvalues and verifying that enough linearly independent eigenvectors exist.

Worked Examples

Example 1: Non-Commutativity in Practice

Suppose we need to solve $B^{-1}(A - X) = AX$ for $X$, where $A$ and $B$ are invertible matrices.

Multiplying both sides by $B$ on the left: $A - X = B \cdot AX$

Rearranging: $A = X + BAX = X(I + BA)$

Thus: $X = A(I + BA)^{-1}$

Note that we cannot write $X = (BA + I)^{-1}A$ unless we are careful about order. In fact, ^{[matrix-equation-solution]} shows that the correct solution is $X = (BA + I)^{-1}A$, which requires understanding that $(I + BA)^{-1}$ and $(BA + I)^{-1}$ are the same (addition is commutative), but the position of $A$ relative to the inverse matters crucially.

Example 2: Determinant and Invertibility

Consider the matrix $A = \begin{pmatrix} 2 & 4 \\ 1 & 2 \end{pmatrix}$.

Computing the determinant: $\det(A) = (2)(2) - (4)(1) = 4 - 4 = 0$

Since $\det(A) = 0$, the matrix is singular and has no inverse. The second row is a scalar multiple of the first, so the columns are linearly dependent. The transformation collapses 2D space onto a line.

Example 3: Checking Diagonalizability

Consider $A = \begin{pmatrix} 2 & 1 \\ 0 & 2 \end{pmatrix}$.

The characteristic polynomial is $\det(A - \lambda I) = (2 - \lambda)^2 = 0$, giving $\lambda = 2$ with algebraic multiplicity 2.

To find eigenvectors, solve $(A - 2I)v = 0$: $\begin{pmatrix} 0 & 1 \\ 0 & 0 \end{pmatrix} \begin{pmatrix} v_1 \\ v_2 \end{pmatrix} = 0$

This gives $v_2 = 0$, so eigenvectors are of the form $\begin{pmatrix} v_1 \\ 0 \end{pmatrix}$. The eigenspace is 1-dimensional.

Since the geometric multiplicity (1) is less than the algebraic multiplicity (2), the matrix is not diagonalizable. We cannot find two linearly independent eigenvectors.

References

• ^{[matrix-multiplication]}
• ^{[determinant-of-a-matrix]}
• ^{[determinant-properties]}
• ^{[matrix-equation-solution]}
• ^{[diagonalizable-matrix]}
• ^{[determinant-properties]}
• ^{[determinant-of-a-matrix]}
• ^{[matrix-inversion-formula]}
• ^{[diagonalizable-matrix]}

AI Disclosure

This article was drafted with AI assistance. The structure, examples, and explanations were generated based on the provided class notes. All mathematical claims are grounded in the cited notes; no external sources were consulted. The worked examples were constructed to illustrate the misconceptions discussed and have been verified for correctness.