Characteristic Polynomial - Linear Algebra

SECTION 1

Historical Context & Motivation

Imagine you have a machine that takes arrows (vectors) and stretches, flips, or rotates them. Mathematicians wanted to know: are there special arrows that the machine only stretches, without changing direction? Finding those special arrows — called eigenvectors — and the amount they get stretched — called eigenvalues — turned out to be one of the most important problems in all of mathematics. The characteristic polynomial is the key tool that unlocks those values.

1750s • Euler's Rotations

Leonhard Euler studied how rigid bodies rotate in 3D space. He found special axes that don't change direction during rotation — an early glimpse of eigenvectors.

1826 • Cauchy's Determinant Work

Augustin-Louis Cauchy formally studied determinants of matrices and laid the groundwork for the characteristic equation, connecting polynomials to matrix properties.

1855 • Cayley & Matrix Algebra

Arthur Cayley defined matrices as mathematical objects and showed that every square matrix satisfies its own characteristic polynomial, a result now called the Cayley–Hamilton theorem.

1904 • Hilbert's Spectral Theory

David Hilbert extended eigenvalue ideas to infinite-dimensional spaces. This opened the door to quantum mechanics, where eigenvalues represent measurable quantities like energy.

The central question these mathematicians were chasing was simple: given a matrix, how do you systematically find which values of λ make the equation Av = λv possible? The answer is the characteristic polynomial — a formula you build from the matrix that, when solved, hands you all the eigenvalues at once.

SECTION 2

Core Principles & Definitions

Before we build a characteristic polynomial, let's nail down a few key ideas. Think of a matrix as a rule that transforms every point in space — sliding, stretching, or rotating it. Some special directions only get scaled (stretched or squished) by the matrix. Those are the directions we want to find.

Matrix (A)

A square grid of numbers that represents a transformation. For example, a 2×2 matrix transforms points in a flat plane. A 3×3 matrix transforms points in 3D space.

Eigenvalue (λ)

A number (often written as the Greek letter lambda, λ) that tells you how much the matrix stretches a vector along a special direction. If λ = 2, the vector doubles in length.

Eigenvector (v)

A nonzero arrow (vector) that only gets scaled — not rotated — when the matrix acts on it. The equation Av = λv captures this: the matrix A just multiplies v by the scalar λ.

Determinant

A single number computed from a square matrix that tells you whether the matrix squishes space to zero (determinant = 0) or not. It's the key ingredient in building the characteristic polynomial.

Characteristic Polynomial

The polynomial p(λ) = det(A − λI). Its roots (the values of λ that make it equal zero) are exactly the eigenvalues of the matrix A.

✦

KEY TAKEAWAY

Think of a matrix like a funhouse mirror. Most objects look warped in weird ways. But an eigenvector is like standing at the one perfect angle where you only look taller or shorter — not twisted. The characteristic polynomial is the math formula that tells you exactly which "stretch factors" (eigenvalues) are possible for that particular mirror.

SECTION 3

Visual Explanation

Let's see what a matrix transformation actually looks like. The diagram below shows a 2×2 matrix acting on different vectors. Notice how most vectors change both their length and direction. But the eigenvectors (shown in bold colors) only change their length — they stay on the same line.

The cyan eigenvector v₁ = [1, 1] gets stretched by a factor of 3 (eigenvalue λ₁ = 3). The pink eigenvector v₂ = [1, −1] keeps its length (eigenvalue λ₂ = 1). The dashed gray vector w is not an eigenvector — the matrix changes its direction.

The characteristic polynomial is the tool that lets us calculate those eigenvalues (3 and 1 in the diagram above) without having to guess and check every possible vector. You build one polynomial from the matrix, set it equal to zero, and solve. The solutions are the eigenvalues.

SECTION 4

Mathematical Framework

Here's the big idea in math language. We want to find values of λ so that Av = λv for some nonzero vector v. We can rewrite this as (A − λI)v = 0. For a nonzero solution v to exist, the matrix (A − λI) must squish space down to zero — meaning its determinant must equal zero. That condition gives us the characteristic polynomial.

EIGENVALUE EQUATION

Av = λv

A = the matrix, v = an eigenvector (nonzero), λ = the eigenvalue (the stretch factor)

REARRANGED FORM

(A − λI)v = 0

I = the identity matrix (1s on the diagonal, 0s elsewhere). We subtract λ from each diagonal entry of A.

CHARACTERISTIC POLYNOMIAL

p(λ) = det(A − λI) = 0

det means determinant. The roots of p(λ) — the values that make p(λ) = 0 — are the eigenvalues of A.

Determinant of a 2×2 Matrix

For a 2×2 matrix, the determinant has a simple formula. If your matrix is [[a, b], [c, d]], then det = ad − bc. When we subtract λ from the diagonal, the matrix becomes [[(a − λ), b], [c, (d − λ)]], and the determinant becomes:

2×2 CHARACTERISTIC POLYNOMIAL

p(λ) = (a − λ)(d − λ) − bc = λ² − (a + d)λ + (ad − bc)

The sum (a + d) is called the trace of A. The product (ad − bc) is the determinant of A. So p(λ) = λ² − trace(A) × λ + det(A).

SECTION 5

Building the Characteristic Polynomial Step by Step

Let's walk through the construction process for a concrete 2×2 matrix. Then we'll look at what happens with a 3×3 matrix. The diagram below shows the assembly line: start with the matrix, subtract λ from the diagonal, compute the determinant, and you have your polynomial.

The four-step flow for a 2×2 matrix: (1) start with matrix A, (2) subtract λ from the diagonal to form A − λI, (3) compute the determinant to get the polynomial, (4) solve for λ. The "Key Observations" box shows handy checks using the trace and determinant.

Notice that for a 2×2 matrix, the characteristic polynomial is always a quadratic (degree 2). For a 3×3 matrix, you'd get a cubic (degree 3), and so on. The degree of the characteristic polynomial always matches the size of the matrix. This makes sense: a bigger matrix can have more independent "stretch directions," so there can be more eigenvalues.

💡

Quick Check Trick

After finding eigenvalues, verify them! The sum of all eigenvalues should equal the trace (sum of diagonal entries). The product of all eigenvalues should equal the determinant of the original matrix. If either check fails, go back and look for an algebra error.

SECTION 6

Worked Example

Let's find the characteristic polynomial and eigenvalues of the matrix A = [[4, 2], [1, 3]]. We'll follow the same four-step process from the diagram.

Finding the Characteristic Polynomial of a 2×2 Matrix

1
Step 1 — Write Down A and λI

Our matrix is A = [[4, 2], [1, 3]]. The identity matrix I for a 2×2 system is [[1, 0], [0, 1]]. So λI = [[λ, 0], [0, λ]]. We will subtract this from A.
2
Step 2 — Form A − λI

Subtract λ from each diagonal entry: A − λI = [[(4 − λ), 2], [1, (3 − λ)]]. Only the diagonal changes. The off-diagonal entries (2 and 1) stay the same.

A − λI = [[(4 − λ), 2], [1, (3 − λ)]]
3
Step 3 — Compute the Determinant

For a 2×2 matrix [[a, b], [c, d]], the determinant is ad − bc. Here: det(A − λI) = (4 − λ)(3 − λ) − (2)(1). First expand: (4 − λ)(3 − λ) = 12 − 4λ − 3λ + λ² = λ² − 7λ + 12. Then subtract bc: λ² − 7λ + 12 − 2 = λ² − 7λ + 10.

p(λ) = λ² − 7λ + 10
4
Step 4 — Solve p(λ) = 0

Set the polynomial equal to zero: λ² − 7λ + 10 = 0. Factor: (λ − 5)(λ − 2) = 0. So λ = 5 or λ = 2.

Eigenvalues: λ₁ = 5 and λ₂ = 2
5
Step 5 — Verify with Quick Checks

Trace check: the sum of diagonal entries of A is 4 + 3 = 7. The sum of eigenvalues is 5 + 2 = 7. ✓ Determinant check: det(A) = (4)(3) − (2)(1) = 10. The product of eigenvalues is 5 × 2 = 10. ✓ Both checks pass, so our answer is correct!

Both checks pass ✓

SECTION 7

Strengths & Limitations

The characteristic polynomial is a powerful and elegant tool, but like any tool, it works better in some situations than others. Let's compare its strengths and limitations.

Aspect	Strength	Limitation
Small matrices (2×2, 3×3)	Gives exact eigenvalues by factoring or using the quadratic formula. Fast by hand.	No real limitation at this size — this is where it shines!
Large matrices (100×100+)	Theoretically still defines all eigenvalues.	Computing the determinant of a huge matrix is extremely slow. Computers use faster iterative methods instead.
Complex eigenvalues	The polynomial automatically reveals complex roots, showing when the matrix involves rotation.	Students need familiarity with complex numbers (i = √(−1)) to interpret these results.
Repeated eigenvalues	The polynomial shows repeated roots clearly (e.g., (λ − 3)² = 0 means λ = 3 appears twice).	Repeated roots require extra work to determine if there are enough independent eigenvectors.
Understanding matrix behavior	Trace and determinant appear directly in the polynomial, connecting algebra to geometry.	The polynomial alone doesn't tell you the eigenvectors — you need additional computation after finding eigenvalues.

Strengths and limitations of the characteristic polynomial

✦

KEY TAKEAWAY

The characteristic polynomial is like a treasure map. It tells you where the treasure (eigenvalues) is buried, but you still need to dig (do more algebra) to actually find the eigenvectors. For small matrices (2×2 and 3×3), this approach is perfect and can be done by hand. For really large matrices, mathematicians use computer algorithms that skip the polynomial entirely.

SECTION 8

Connection to Advanced Topics

The characteristic polynomial is your gateway to a web of deeper ideas in linear algebra and beyond. Once you're comfortable finding eigenvalues, you can explore concepts like diagonalization (rewriting a matrix in its simplest form), the Cayley–Hamilton theorem (every matrix satisfies its own characteristic polynomial), and spectral decomposition (breaking a matrix into pieces based on its eigenvalues).

Concept	What It Means	How Characteristic Polynomial Connects
Diagonalization	Rewriting A = PDP⁻¹ where D is a diagonal matrix of eigenvalues.	The eigenvalues from the characteristic polynomial go on the diagonal of D. This only works when there are enough independent eigenvectors.
Cayley–Hamilton Theorem	If you plug the matrix A itself into its characteristic polynomial, you get the zero matrix.	p(A) = 0. This lets you express high powers of A (like A⁵) using lower powers, which is useful in applications.
Stability Analysis	In engineering and physics, eigenvalues determine whether a system is stable or will blow up over time.	If all eigenvalues have absolute value less than 1 (or negative real part in continuous systems), the system is stable. The characteristic polynomial directly reveals this.
Google's PageRank	Google originally ranked web pages using the dominant eigenvalue of a huge matrix of web links.	The concept is the same: find the eigenvalue that matters most. For huge matrices, numerical methods replace the characteristic polynomial, but the theory is identical.

As you advance in math and science, you'll find eigenvalues everywhere: quantum mechanics uses them to describe energy levels, data science uses them for dimensionality reduction (PCA), and engineers use them to predict vibration frequencies of bridges and buildings. The characteristic polynomial is where all of these applications begin.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

In your own words, explain what the characteristic polynomial of a matrix tells you. Why do we set it equal to zero?

Show answer

The characteristic polynomial p(λ) = det(A − λI) encodes all the information needed to find the eigenvalues of a matrix A. We set it equal to zero because eigenvalues are the values of λ for which the matrix (A − λI) has a nonzero solution to (A − λI)v = 0. A nonzero solution exists only when the determinant of (A − λI) equals zero, meaning the matrix is "singular" — it squishes space down so that some vectors map to the zero vector. The roots of the polynomial are exactly those special λ values.

Question 2: BASIC CALCULATION

Find the characteristic polynomial and eigenvalues of the matrix A = [[3, 0], [0, 5]].

Show answer

First, form A − λI = [[(3 − λ), 0], [0, (5 − λ)]]. Since this is a diagonal matrix, the determinant is simply the product of the diagonal entries: p(λ) = (3 − λ)(5 − λ) = λ² − 8λ + 15. Setting p(λ) = 0, we factor: (λ − 3)(λ − 5) = 0, giving eigenvalues λ₁ = 3 and λ₂ = 5. Quick check: trace = 3 + 5 = 8 and 3 + 5 = 8 ✓; determinant = 15 and 3 × 5 = 15 ✓. For diagonal matrices, the eigenvalues are always just the diagonal entries.

Question 3: INTERMEDIATE

Find the characteristic polynomial and eigenvalues of A = [[1, 4], [2, 3]]. Verify your answer using the trace and determinant checks.

Show answer

Form A − λI = [[(1 − λ), 4], [2, (3 − λ)]]. Compute the determinant: (1 − λ)(3 − λ) − (4)(2) = 3 − λ − 3λ + λ² − 8 = λ² − 4λ − 5. Set equal to zero: λ² − 4λ − 5 = 0. Factor: (λ − 5)(λ + 1) = 0, so λ₁ = 5 and λ₂ = −1. Verification: trace(A) = 1 + 3 = 4 and 5 + (−1) = 4 ✓. det(A) = (1)(3) − (4)(2) = −5 and (5)(−1) = −5 ✓. Notice that one eigenvalue is negative, which means the matrix flips vectors along that eigenvector's direction.

Question 4: APPLIED

A population model uses the matrix A = [[0.5, 0.3], [0.4, 0.8]] to predict next year's population distribution from this year's. Find the eigenvalues. Based on the eigenvalues, will the total population grow, shrink, or stabilize over many years? (Hint: look at the largest eigenvalue.)

Show answer

Form A − λI = [[(0.5 − λ), 0.3], [0.4, (0.8 − λ)]]. The determinant gives p(λ) = (0.5 − λ)(0.8 − λ) − (0.3)(0.4) = λ² − 1.3λ + 0.4 − 0.12 = λ² − 1.3λ + 0.28. Using the quadratic formula: λ = (1.3 ± √(1.69 − 1.12)) / 2 = (1.3 ± √0.57) / 2 = (1.3 ± 0.755) / 2. So λ₁ ≈ 1.028 and λ₂ ≈ 0.272. Since the largest eigenvalue (≈ 1.028) is slightly greater than 1, the population will slowly grow over time. If it were exactly 1, the population would stabilize; if less than 1, it would shrink.

Question 5: CRITICAL THINKING

Consider a 2×2 matrix A where the characteristic polynomial is p(λ) = λ² + 1. What are the eigenvalues? What does this tell you about what the matrix does geometrically? (Hint: think about what kind of numbers solve λ² = −1.)

Show answer

Setting p(λ) = λ² + 1 = 0, we get λ² = −1, so λ = i and λ = −i, where i = √(−1) is the imaginary unit. These are complex eigenvalues, not real numbers. Geometrically, this means the matrix has no direction that simply gets stretched — there are no real eigenvectors. Instead, the matrix performs a pure rotation. For example, the matrix [[0, −1], [1, 0]] rotates every vector by 90° and has exactly this characteristic polynomial. Complex eigenvalues always signal that rotation is involved in the transformation. This is a deep and beautiful connection between algebra (complex roots) and geometry (rotations).

SUMMARY

Characteristic Polynomial — Summary

Summary

The characteristic polynomial of a square matrix A is defined as p(λ) = det(A − λI). You build it by subtracting λ from each diagonal entry of A, then computing the determinant of the resulting matrix. The roots of this polynomial — the values of λ that make p(λ) = 0 — are the eigenvalues of the matrix. For a 2×2 matrix, the characteristic polynomial is always a quadratic: p(λ) = λ² − trace(A) × λ + det(A).

After finding eigenvalues, always verify using the trace check (sum of eigenvalues = sum of diagonal entries) and the determinant check (product of eigenvalues = determinant of A). Eigenvalues reveal whether a transformation stretches, compresses, flips, or rotates vectors. Complex eigenvalues signal rotation, while real eigenvalues correspond to pure stretching or flipping. The characteristic polynomial connects algebra to geometry and serves as the foundation for diagonalization, stability analysis, and countless applications across science and engineering.