Connecting You to the Latest News,
Study Tips, and Academic Resources
Linear Algebra Tutoring for Proof Based Courses
November 20,2026

Linear algebra is the first course in which many students meet proofs as the primary mode of assessment, and the difficulty is often attributed to the proofs when it is actually about the definitions. A proof is a sequence of deductions from definitions, and a student who does not hold the definitions precisely has nothing to deduce from.

The course is also unusual in having two distinct cultures. An applied linear algebra course emphasises computation, matrices, and algorithms. A proof-based course emphasises vector spaces, linear transformations, and theorems. A student who has prepared for one and enrolled in the other will find the first month disorienting.

The Two Courses Called Linear Algebra

The distinction matters for diagnosis and is worth stating plainly.

The computational version teaches Gaussian elimination, determinants, eigenvalues, and diagonalisation as procedures, with applications to systems and to data. It is a service course for engineering and science, and it rewards fluency.

The proof-based version begins with the axioms of a vector space and builds: subspaces, span, independence, basis, dimension, linear maps, kernels and images, determinants as multilinear functions, eigenvalues, and the spectral theorem. It rewards precision, and computation is a source of examples rather than the subject.

A student failing the proof-based course who came from a computational background usually has strong technical skills and no experience with the genre of argument. That is a different problem from not understanding the material, and it needs a different response.

Why Definitions Carry the Course

In earlier courses, a definition is a convenience: it names something the student already recognises. In linear algebra, a definition is the object, and everything is derived from it.

The definition of a vector space is eight axioms. Almost every result in the first month is proved by checking the axioms or by exhibiting a counterexample that fails one. A student who has not memorised the axioms cannot prove anything and cannot construct the counterexamples that the exercises require.

The same is true of linear independence, span, and basis. These are definitional, and the definitions are short. A student who can state them precisely to a friend, without notes, has most of what the first month requires.

The Vector Space Abstraction

The abstraction is the specific thing students resist, and it is worth addressing directly.

A vector is not an arrow and not a list of numbers. A vector is an element of a set that satisfies the axioms, and the arrows and lists are examples. The same theorems therefore apply to polynomials, to matrices, and to functions, which is the point of the abstraction and the reason the subject is powerful.

Students who insist on thinking of vectors as arrows can follow the early chapters and hit a wall when the examples become polynomials. The transition is easier if it is made early, with concrete non-arrow examples used from the first week.

Writing Proofs: The Genre

Proof writing is a genre with conventions, and it is rarely taught explicitly.

The conventions worth naming: begin by stating what is assumed and what is to be shown. Introduce every symbol before using it. Justify each step by reference to a definition, a hypothesis, or a previously established result. Use the words "suppose," "let," "since," and "therefore" in their technical senses. End by confirming that the original claim has been established.

Beyond conventions, three techniques cover most undergraduate exercises: direct proof, proof by contrapositive, and proof by contradiction. A fourth, induction, appears when the course reaches dimensions and bases.

A student who can name which technique a problem invites has a starting point. A student who cannot reads the statement and waits for inspiration.

Common Proof Errors

Four errors account for most lost marks, and each is diagnosable.

The first is assuming what is to be proved, which usually appears as starting from the conclusion and manipulating until something true emerges. Working backwards is legitimate as exploration and must be reversed for the write-up.

The second is proving an example rather than a general claim. Showing that a particular set is a subspace does not show that all such sets are.

The third is mishandling quantifiers. "For every" and "there exists" are not interchangeable, and a proof that swaps them is wrong in a way that reads as correct.

The fourth is an unjustified step: a claim that is true but not established by what has been written. Readers mark these, and students cannot see them.

Linear Transformations as the Real Subject

Many courses introduce linear transformations late and treat them as an application of matrices. They are better understood as the subject, with matrices as their representation.

A linear transformation is a function between vector spaces that preserves addition and scalar multiplication. A matrix is what a linear transformation looks like once a basis is chosen. Change the basis and the matrix changes while the transformation does not.

That framing explains why similar matrices represent the same transformation, which is otherwise a theorem to be memorised, and it explains what diagonalisation is for: finding a basis in which the transformation is as simple as possible.

Kernel, Image, and the Rank-Nullity Theorem

Rank-nullity is the central structural theorem of the course, and it rewards genuine understanding.

The kernel is the set of inputs mapped to zero. The image is the set of outputs actually attained. The theorem states that the dimensions of the two add to the dimension of the domain.

Why it is true is worth a session. The kernel measures collapse: how many dimensions are lost. The image measures what survives. Since every input either contributes to the image or is collapsed, the two account for the whole domain.

A student who understands this can answer questions about injectivity and surjectivity by inspecting dimensions rather than by solving systems, which is the fluency the course is trying to build.

Eigenvalues and What They Mean

Eigenvalues are usually introduced through the characteristic polynomial and computed mechanically, which leaves students with a procedure and no idea what they have found.

The meaning: an eigenvector is a direction that the transformation does not rotate, only stretches or compresses, and the eigenvalue is the factor. A transformation with a full set of eigenvectors is diagonal in the right basis, which means it acts independently along those directions.

That interpretation makes diagonalisation purposeful rather than procedural, and it makes the spectral theorem for symmetric matrices a statement about orthogonal directions rather than a formula.

How Tutoring Should Work

The productive structure for a proof-based course is different from a computational one.

Sessions should not be spent solving exercises for the student. They should be spent on three activities: reciting definitions until precise, planning proofs without writing them, and reading proofs critically.

Planning without writing is the highest-yield activity. Given a statement, the student states what is assumed, what is to be shown, which technique is likely, and the first two steps. Ten of these take forty minutes and build the skill that the exercises are assessing.

Reading proofs critically means taking a textbook proof and asking, of every sentence, what licenses this. Students who do this for a month stop writing unjustified steps.

Determinants and Why They Are Introduced Twice

Determinants appear in a computational course as a formula to be expanded, and in a proof-based course as a function with properties to be characterised. Students who have only the first version find the second baffling.

The characterisation worth knowing is that the determinant is the unique function on square matrices that is multilinear in the columns, alternating under column swaps, and equal to one on the identity. Everything else, including the product rule and the relationship to invertibility, follows from those three properties.

The geometric interpretation ties the two versions together: the determinant is the signed volume scaling factor of the transformation. A determinant of zero means the transformation collapses a dimension, which is exactly why the matrix is not invertible.

Reading Before Writing

One habit speeds up progress in a proof-based course more than any other: reading proofs actively rather than passively.

Passive reading means following the argument and agreeing with it. Active reading means, at every sentence, asking what licenses this claim, and marking the sentence where the answer is unclear. A textbook proof read this way takes three times as long and produces a student who can write proofs.

The specific thing to look for is the hinge sentence, the one where the hypothesis is used. Many proofs consist of a routine argument plus one place where the hypothesis does the work, and identifying that place is what makes the proof reconstructible.

The Pace Problem

Proof-based courses move slowly on the page and quickly in the term, and students misjudge this constantly.

A week may cover one theorem and its proof, which reads as little material. The following week's exercises, however, require the theorem to be usable rather than merely read, and usability comes from having worked examples, constructed counterexamples, and applied the statement in unfamiliar settings.

The practical implication is that reading the material is not preparation for the exercises; it is the first of three steps. The second is reproducing the proof from memory, and the third is finding a case where the hypotheses fail.

Proof-based linear algebra fails students for two different reasons: they lack the technical background, or they have never written proofs and do not know the genre.

Establish which course the student is in and which background he has. Then teach definitions to precision, since every proof is a deduction from them, and teach proof writing as a genre with conventions, three standard techniques, and four characteristic errors.

Present linear transformations as the subject and matrices as their representation, explain rank-nullity through collapse and survival, and interpret eigenvectors as directions the transformation preserves. Spend sessions on recitation, planning, and critical reading rather than on solving exercises for the student.

Let’s discuss your academic needs!
Join Our Community
Subscribe to our newsletter to get latest education news, study advice, academic resources, and online courses promotion info!
ALSO, GET $ 100 OFF
YOUR FIRST SINICA EXPERIENCE
By submitting this form, you acknowledge that Sinica Education may collect and use the information you provide to respond to your inquiry and provide requested services. Please review our Privacy Policy for more information.
Join Our Community
GET $100 OFF YOUR FIRST SINICA EXPERIENCE!
By submitting this form, you acknowledge that Sinica Education may collect and use the information you provide to respond to your inquiry and provide requested services. Please review our Privacy Policy for more information.
Schedule a Free Consultation
Parent / Guardian
Student
Agency Partner
Name *
Email *
Phone *
    Your Kid’s Grade Level *
    Tell us more about your situation, needs, and goals!  
    Verification
    Name *
    Email *
    Phone *
      Your Grade Level *
      Tell us more about your situation, needs, and goals!
      Verification
      Name *
      Email *
      Phone *
        Tell us more about your situation, needs, and goals!
        Verification
        By submitting this form, you acknowledge that Sinica Education may collect and use the information you provide to respond to your inquiry and provide requested services. Please review our Privacy Policy for more information.