The characteristic polynomial det(lambda I-A) packages a square matrix’s eigenvalues into one algebraic object. Expanding a determinant symbolically becomes cumbersome as dimension grows. The Faddeev–LeVerrier algorithm instead builds the coefficients recursively from matrix products and traces. Mathematics matters because trace identities, the adjugate, Newton sums and the Cayley–Hamilton theorem connect local matrix operations to a global polynomial. The recursion is elegant and useful in exact or symbolic work, but it is not automatically the most stable floating-point route to eigenvalues. Choosing an algorithm requires knowing whether the real goal is coefficients, a determinant, an inverse or numerical eigenvalues.
The practical search question is not simply “what is this named method?” but “which mathematical structure makes the method work, what does it guarantee, and what evidence would show that it is a poor choice?” This article keeps those questions together so students can connect mathematics education with computing, engineering, data reasoning and careful everyday decisions without turning one technique into a career promise.
Quick navigation: Why it matters · Mechanism · Worked example · Deeper ideas · Investigations · Limits · Learning route · Parent guidance · FAQs · Sources
Why this mathematics matters
Mathematics is important here because it turns an apparently visual or computational trick into a chain of reasons. The representation tells us what information is kept. The equation or invariant explains each legal step. The complexity analysis describes the work. The boundary conditions say when a conclusion should not be trusted. That combination is a practical form of problem-solving literacy.
A strong learner should be able to move in both directions: from a real task to a mathematical model, and from a formal guarantee back to its real meaning. Naming an algorithm is not enough. The useful achievement is to explain why the next operation is valid, check a small instance independently, and notice which assumptions the implementation has added.
Source: SIAM Review proof of the LeVerrier–Faddeev characteristic-polynomial recursion.
Primary or implementation reference: Modern review of Faddeev–LeVerrier and faster characteristic-polynomial computation.
How the mechanism works
For an n by n matrix A, write p(lambda)=lambda^n+c_1lambda^(n-1)+…+c_n. Set B_0=I. For k=1 through n, calculate c_k=-(1/k)tr(A B_(k-1)), then set B_k=A B_(k-1)+c_k I. In exact arithmetic the final B_n is zero by Cayley–Hamilton. The coefficients appear in descending degree order; c_1=-tr(A), and c_n=(-1)^n det(A). Variants shift indices or update B before naming the coefficient, so an implementation must keep one convention throughout. Intermediate B matrices also connect to the adjugate and, when A is nonsingular, its inverse.
Trace sums diagonal entries and is invariant under similarity, matching the fact that characteristic-polynomial coefficients do not depend on the chosen basis. Newton identities connect power sums of eigenvalues, tr(A^k), to elementary symmetric functions, which are the polynomial coefficients. The matrix recurrence packages those identities without explicitly finding eigenvalues. Cayley–Hamilton states p(A)=0 and explains the terminating zero matrix. Dividing by k assumes the underlying arithmetic permits those divisions; over finite fields whose characteristic divides k, a different method may be required.
Keep a four-column notebook labelled object, rule, guarantee and caveat. In the first column name the inputs and stored state. In the second write one update precisely. In the third state what remains true after the update. In the fourth record a tie, approximation, domain restriction or hardware assumption. This routine prevents symbols from drifting away from their meaning.
Worked example and core calculations
Let A=[[2,1],[1,2]]. Set B_0=I. First c_1=-tr(A)=-4, so B_1=A-4I=[[-2,1],[1,-2]]. Next A B_1=[[-3,0],[0,-3]], whose trace is -6. Thus c_2=-(1/2)(-6)=3. Then B_2=A B_1+3I=0, providing the termination check. Therefore p(lambda)=lambda^2-4lambda+3=(lambda-1)(lambda-3). The eigenvalues are 1 and 3; their sum 4 matches the trace and their product 3 matches the determinant.
How to audit the calculation
- Recompute every intermediate value directly from the definition.
- Keep labels, indices, signs, units and normalisation visible.
- Test a case with a known exact answer.
- State tie rules, random seeds, tolerances and boundary conventions.
- Compare with a slower transparent baseline or trusted implementation.
- Separate agreement of numbers from suitability of the real-world model.
Small examples are not childish; they are diagnostic instruments. A five-line trace can reveal a reversed inequality, a missing scale factor or an inconsistent label convention before a large dataset hides the same error behind plausible output.
Four deeper ideas
Trace carries eigenvalue power sums
The trace of A^k equals the sum of kth powers of eigenvalues when counted algebraically, linking matrices to polynomial coefficients.
To transfer this idea, ask what would remain true if labels, scale or input order changed. Then identify the step whose justification would fail if the stated assumption were removed. This turns the paragraph into a reusable reasoning pattern rather than a fact to memorise.
The recurrence is basis-independent
Similarity changes matrix entries but preserves trace identities and the characteristic polynomial.
To transfer this idea, ask what would remain true if labels, scale or input order changed. Then identify the step whose justification would fail if the stated assumption were removed. This turns the paragraph into a reusable reasoning pattern rather than a fact to memorise.
Did you know? The final zero is a theorem check
B_n=0 is the Cayley–Hamilton relation appearing inside the computation.
To transfer this idea, ask what would remain true if labels, scale or input order changed. Then identify the step whose justification would fail if the stated assumption were removed. This turns the paragraph into a reusable reasoning pattern rather than a fact to memorise.
Outputs should match the real task
Computing all polynomial coefficients can be unnecessary and numerically risky when only eigenvalues are wanted.
To transfer this idea, ask what would remain true if labels, scale or input order changed. Then identify the step whose justification would fail if the stated assumption were removed. This turns the paragraph into a reusable reasoning pattern rather than a fact to memorise.
Twelve practical investigations
1. Symmetric 2×2 matrix
Use the worked example. Trace every matrix.
What the mathematics reveals: Polynomial factors transparently. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Check B2=0. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
2. Diagonal matrix
Choose known diagonal entries. Compare with product factors.
What the mathematics reveals: The recurrence reproduces symmetric sums. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Preserve signs. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
3. Triangular matrix
Use visible diagonal eigenvalues. Compare outputs.
What the mathematics reveals: Off-diagonal entries do not change eigenvalues. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Products still contain them. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
4. Repeated eigenvalue
Use a scalar multiple of I. Obtain a repeated factor.
What the mathematics reveals: Multiplicity is encoded. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Root finding may be sensitive. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
5. Zero matrix
Set A=0. Run every stage.
What the mathematics reveals: Polynomial is lambda^n. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Avoid dividing an uninitialised trace. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
6. Singular matrix
Use determinant zero. Inspect c_n.
What the mathematics reveals: Constant coefficient vanishes. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Inverse is unavailable. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
7. Similarity transform
Use P^-1AP. Compare coefficients.
What the mathematics reveals: The polynomial is invariant. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Numerical conditioning can change. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
8. Large scaling
Multiply A by a constant. Predict coefficient scaling.
What the mathematics reveals: Degree determines powers of scale. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Overflow can occur. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
9. Floating residual
Calculate norm of B_n. Compare precisions.
What the mathematics reveals: Termination becomes approximate. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Set a meaningful tolerance. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
10. QR comparison
Compute eigenvalues directly. Contrast goals.
What the mathematics reveals: Coefficient formation may be avoidable. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Choose the intended output. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
11. Sparse matrix
Track fill-in. Measure storage.
What the mathematics reveals: Matrix multiplication can destroy sparsity. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Exploit structure elsewhere. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
12. Finite-field case
Choose characteristic dividing k. Identify invalid division.
What the mathematics reveals: Algebraic setting matters. Evidence to collect: write the input, predict the next state, carry out the calculation, and compare the result with an independently produced reference. Explain the exact invariant, equation, inequality or counting rule that supports the conclusion. Then vary one assumption and record whether the result changes smoothly, discontinuously or not at all.
Limit to keep visible: Use a division-free method. A useful report places this boundary beside the result rather than hiding it in a final disclaimer.
Limits and misconceptions
Naive dense implementations require repeated matrix multiplication and are expensive for large matrices. Floating-point traces and products can accumulate error, while characteristic-polynomial coefficients may be extremely sensitive. Explicit coefficients are often a poor intermediate representation for numerical eigenvalue calculation; QR-based methods are usually preferred. Division by k prevents direct use over some rings or fields. Sparse structure can be destroyed by dense intermediate matrices. Repeated eigenvalues do not invalidate the polynomial but can make downstream root finding delicate. Symbolic expressions can swell rapidly.
Five misconceptions to challenge
- The algorithm finds eigenvectors directly: Its primary output is the characteristic polynomial coefficients.
- Trace alone determines every coefficient: Higher stages also need matrix products and previous coefficients.
- B_n close to zero always proves a correct floating result: Round-off and scaling affect the residual.
- Characteristic coefficients are the best path to numerical eigenvalues: They can be ill-conditioned.
- Every arithmetic system permits division by k: Finite characteristic can block the recurrence.
Another broad misconception is that advanced mathematics automatically creates intelligence, admission, income or employment. It does not. Deliberate study can strengthen modelling, abstraction, calculation and explanation, while real pathways also depend on interests, communication, domain knowledge, opportunities and circumstances outside one topic.
A learning route for students
Move from a visible trace to a symbolic rule, then to code and critique. For every step below, prepare three pieces of evidence: a correct hand example, a boundary or failure case, and a short explanation using the relevant invariant. Do not advance merely because software returned a number.
Step 1: Write a characteristic polynomial convention
Begin with the smallest non-trivial instance that lets you write a characteristic polynomial convention. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
Step 2: Calculate matrix traces
Begin with the smallest non-trivial instance that lets you calculate matrix traces. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
Step 3: Multiply small matrices
Begin with the smallest non-trivial instance that lets you multiply small matrices. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
Step 4: Apply the faddeev–leverrier recursion
Begin with the smallest non-trivial instance that lets you apply the Faddeev–LeVerrier recursion. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
Step 5: Check coefficient signs
Begin with the smallest non-trivial instance that lets you check coefficient signs. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
Step 6: Verify trace and determinant identities
Begin with the smallest non-trivial instance that lets you verify trace and determinant identities. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
Step 7: Use cayley–hamilton as a residual check
Begin with the smallest non-trivial instance that lets you use Cayley–Hamilton as a residual check. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
Step 8: Distinguish symbolic and numerical goals
Begin with the smallest non-trivial instance that lets you distinguish symbolic and numerical goals. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
Step 9: Recognise coefficient sensitivity
Begin with the smallest non-trivial instance that lets you recognise coefficient sensitivity. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
Step 10: Select an alternative eigenvalue method
Begin with the smallest non-trivial instance that lets you select an alternative eigenvalue method. Predict the result before calculating. Write each state change clearly enough that another student can reproduce it. Compare with a trusted implementation only after the hand trace is complete.
Next, change one feature—size, ordering, sign, tolerance, labelling or data distribution—and explain which part of the reasoning changes. Finish by teaching the step without code and by naming one input for which the method is unhelpful. This predict–calculate–vary–explain cycle is a strong test of transferable understanding.
A four-week practice plan
- Week 1 — vocabulary and diagrams: define each object, redraw the mechanism and reproduce the worked example.
- Week 2 — equations and manufactured tests: derive the key relation and test inputs with known answers.
- Week 3 — implementation and evidence: use a trusted library, record parameters and compare with a transparent baseline.
- Week 4 — transfer and critique: apply the method to a new context, report limits and identify when a different method is better.
Guidance for parents and educators
Ask “what stays true?” and “what could make this conclusion fail?” before asking for speed. Invite the student to draw the input and output, estimate the result and locate the first step that disagrees with a reference. These prompts reveal conceptual understanding more clearly than memorised terminology.
Advanced enrichment should sit beside secure school foundations in algebra, functions, geometry, statistics and computational thinking. It need not accelerate a student into a fixed career identity. A healthy goal is curiosity with discipline: the learner becomes comfortable reading unfamiliar notation, checking an example, revising a model and communicating uncertainty.
Evidence prompts for a useful conversation
- Which quantity is observed and which is inferred?
- Which step saves work, and why is it allowed?
- Is the claim exact, approximate, probabilistic or empirical?
- What units and conventions are in use?
- Which input makes the method struggle?
- What simple baseline could check the answer?
- What would count as evidence that the parameters are poor?
Mini-project assessment
A strong mini-project contains the original question, a small reproducible dataset, a hand-worked example, code or a spreadsheet trace, at least one boundary test, a comparison with a baseline, and a paragraph on limits. Assess the reasoning trail as well as the final output. Repairing a model after a surprising result is valuable mathematical progress.
Six evidence checks for a finished project
Check 1: Write a characteristic polynomial convention
Ask the student to demonstrate how to write a characteristic polynomial convention without relying on a hidden software step. The evidence should include a labelled input, the exact rule used, one intermediate state and a result that can be checked independently. If the answer changes after labels or order change, the student should say whether that reflects the mathematics or only a convention.
Then request one deliberately awkward example. The learner should predict the difficulty before running the method, explain what the output does and does not establish, and name a reasonable alternative or safeguard. This check joins correctness with judgement: a technically valid calculation is useful only when its assumptions and interpretation remain visible.
Check 2: Calculate matrix traces
Ask the student to demonstrate how to calculate matrix traces without relying on a hidden software step. The evidence should include a labelled input, the exact rule used, one intermediate state and a result that can be checked independently. If the answer changes after labels or order change, the student should say whether that reflects the mathematics or only a convention.
Then request one deliberately awkward example. The learner should predict the difficulty before running the method, explain what the output does and does not establish, and name a reasonable alternative or safeguard. This check joins correctness with judgement: a technically valid calculation is useful only when its assumptions and interpretation remain visible.
Check 3: Multiply small matrices
Ask the student to demonstrate how to multiply small matrices without relying on a hidden software step. The evidence should include a labelled input, the exact rule used, one intermediate state and a result that can be checked independently. If the answer changes after labels or order change, the student should say whether that reflects the mathematics or only a convention.
Then request one deliberately awkward example. The learner should predict the difficulty before running the method, explain what the output does and does not establish, and name a reasonable alternative or safeguard. This check joins correctness with judgement: a technically valid calculation is useful only when its assumptions and interpretation remain visible.
Check 4: Apply the faddeev–leverrier recursion
Ask the student to demonstrate how to apply the Faddeev–LeVerrier recursion without relying on a hidden software step. The evidence should include a labelled input, the exact rule used, one intermediate state and a result that can be checked independently. If the answer changes after labels or order change, the student should say whether that reflects the mathematics or only a convention.
Then request one deliberately awkward example. The learner should predict the difficulty before running the method, explain what the output does and does not establish, and name a reasonable alternative or safeguard. This check joins correctness with judgement: a technically valid calculation is useful only when its assumptions and interpretation remain visible.
Check 5: Check coefficient signs
Ask the student to demonstrate how to check coefficient signs without relying on a hidden software step. The evidence should include a labelled input, the exact rule used, one intermediate state and a result that can be checked independently. If the answer changes after labels or order change, the student should say whether that reflects the mathematics or only a convention.
Then request one deliberately awkward example. The learner should predict the difficulty before running the method, explain what the output does and does not establish, and name a reasonable alternative or safeguard. This check joins correctness with judgement: a technically valid calculation is useful only when its assumptions and interpretation remain visible.
Check 6: Verify trace and determinant identities
Ask the student to demonstrate how to verify trace and determinant identities without relying on a hidden software step. The evidence should include a labelled input, the exact rule used, one intermediate state and a result that can be checked independently. If the answer changes after labels or order change, the student should say whether that reflects the mathematics or only a convention.
Then request one deliberately awkward example. The learner should predict the difficulty before running the method, explain what the output does and does not establish, and name a reasonable alternative or safeguard. This check joins correctness with judgement: a technically valid calculation is useful only when its assumptions and interpretation remain visible.
Frequently asked questions
What does Faddeev–LeVerrier compute?
The coefficients of a square matrix’s characteristic polynomial.
What is the first coefficient?
The negative trace under the stated monic convention.
What is the constant coefficient?
(-1)^n times the determinant.
Why does the final matrix vanish?
By the Cayley–Hamilton theorem in exact arithmetic.
Does it calculate eigenvalues?
Only indirectly through the polynomial.
Is it good for large floating matrices?
Often not compared with dedicated eigenvalue algorithms.
Can it produce an inverse?
Related intermediate matrices yield one when A is nonsingular.
Does it work with repeated eigenvalues?
Yes, though root recovery can be sensitive.
Why can finite fields be problematic?
The recursion divides by stage number k.
What should students know first?
Matrix multiplication, trace, determinants, polynomials and eigenvalues.
Useful next reading and sources
Source: SIAM Review proof of the LeVerrier–Faddeev characteristic-polynomial recursion.
Source: Modern review of Faddeev–LeVerrier and faster characteristic-polynomial computation.
Related eduKateSG reading: QR iteration for eigenvalues.
Related eduKateSG reading: Bareiss exact determinants.
Related eduKateSG reading: Durand–Kerner polynomial roots.
Read sources with a question in mind: which definition, theorem, default or limitation is being supported? An authoritative link is useful evidence, but it does not replace a worked calculation or a clear statement of how the source applies to the example.
A cheerful final thought
Mathematics becomes friendlier when every symbol has a job and every claim has a boundary. Start small, make the state visible, test the awkward case, and celebrate the moment when a pattern becomes explainable. That is the practical benefit of learning mathematics: not certainty about every future, but a growing ability to ask better questions and build trustworthy answers.