Further Reading — Chapter 23: Eigenvalues and Eigenvectors

Eigenvalues are covered in every linear algebra text, but the treatments differ sharply in where they start. The ones below are chosen because, like this chapter, the best of them lead with meaning. Read at least one geometric source (3Blue1Brown) alongside one rigorous source (Strang or Axler).

Watch first — the geometric picture

  • 3Blue1Brown, Essence of Linear Algebra, Chapter 14: "Eigenvectors and eigenvalues." The single best visual introduction to this chapter's central idea. Grant Sanderson animates exactly what we drew by hand: most vectors get knocked off their span by a transformation, and eigenvectors are the rare ones that stay on their line, scaled by the eigenvalue. His treatment of why a rotation has no (real) eigenvectors, and of eigenvalues as the stretch factors, mirrors §23.1–§23.6 closely. Watch this before or right after reading the chapter; it makes the invariant-direction picture unforgettable. (Free on YouTube.)

Core textbooks

  • Gilbert Strang, Introduction to Linear Algebra (6th ed.), Chapter 6, "Eigenvalues and Eigenvectors," §6.1–6.2. Strang's introduction is the spiritual model for this book's approach: he opens §6.1 by applying a matrix repeatedly and noticing which directions survive, grounding eigenvalues in the dynamics of $A^k\mathbf{x}$ before the characteristic polynomial. His worked $2\times2$ examples, the trace-and-determinant relations, and the Markov-matrix steady state (our Case Study 1) are all here. The accompanying MIT OpenCourseWare lectures (18.06, Lectures 21–22) are free and superb. The primary companion reference for this chapter.

  • Sheldon Axler, Linear Algebra Done Right (4th ed.), Chapter 5, "Eigenvalues, Eigenvectors, and Invariant Subspaces." The rigorous, abstract counterpart. Axler defines eigenvalues through invariant subspaces of a linear operator — making precise the "invariant direction" language we used informally — and famously develops the theory with minimal reliance on determinants. His Theorem on the linear independence of eigenvectors for distinct eigenvalues (our Exercise 23.20) is a model of clean proof. Read this if you are a math major and want the operator-theoretic, determinant-free view; it is the natural home for §23.7's eigenspaces and the Math-Major Sidebars.

  • Stephen Boyd & Lieven Vandenberghe, Introduction to Applied Linear Algebra (VMLS). The applied/data-science perspective. While VMLS is light on eigen-theory itself, its chapters on linear dynamical systems and Markov chains show the steady-state and growth-rate applications (Case Study 1) in the language of data and engineering, and its companion numpy examples reinforce the computational track. Free PDF from the authors; best for CS/data-science readers who want to see eigenvalues at work in applications.

Deeper and complementary

  • David C. Lay, Steven Lay & Judi McDonald, Linear Algebra and Its Applications (5th ed.), Chapter 5. A gentle, example-rich treatment with an excellent section on eigenvectors and difference equations (the discrete dynamical systems behind our Fibonacci, Leslie-population, and Markov examples). Its "dynamical systems" application sections are particularly good for building intuition about $|\lambda|$ as a growth rate. A good middle ground in rigor between Strang and Axler.

  • Carl D. Meyer, Matrix Analysis and Applied Linear Algebra, Chapter 7. A thorough, careful reference that does not shy from the subtleties this chapter previews: defective matrices, the gap between algebraic and geometric multiplicity, and the spectral picture. Meyer's historical notes are reliable and well-sourced. Use as a reference when Chapter 24's multiplicity questions arise.

  • Trefethen & Bau, Numerical Linear Algebra, Lectures 24–27. For the computational reader who wants to know how eigenvalues are actually found at scale: power iteration (our toolkit function), the QR algorithm (Chapter 29), and why the naive "find roots of the characteristic polynomial" approach is numerically disastrous. The authoritative source on the computation track; read after Chapter 24.

On the history and the name

  • For the origin of the term eigen- and its attribution to Hilbert (≈1904), see the historical notes in Meyer (above) and the MacTutor History of Mathematics archive (St Andrews). Several of the specific dates and attributions in this chapter's Historical Note are marked [verify] precisely because secondary sources disagree on exactly who first used "eigenvalue" versus "characteristic value" versus "latent root" (Sylvester); consult a primary-source history before quoting them as settled.

Where to go next in this book

Chapter 24 (The Characteristic Polynomial) makes finding eigenvalues systematic and confronts multiplicity. Chapter 25 (Diagonalization) delivers $A = PDP^{-1}$ and easy matrix powers. Chapter 27 (The Spectral Theorem) explains why the symmetric stiffness matrix of Case Study 2 had orthogonal eigenvectors. Chapter 29 (PageRank) scales the dominant eigenvector to the whole web. If you read only one thing next, make it the 3Blue1Brown video, then Strang §6.1 — together they cement the geometry and the algebra of the most important idea in linear algebra.