Undoing a transformation, and when you cannot
What you’ll learn
Section titled “What you’ll learn”This is lesson 7 of Track 4 (Visual Math: Linear Algebra). The previous lesson said a zero determinant loses information; this one makes that precise and turns it into a working skill. By the end you will be able to look at a matrix and decide whether its transformation can be reversed, using three tests that turn out to be the same question: a nonzero determinant, full rank, and a null space of just the origin. You will see what the inverse actually is (the undo transformation, defined by M^-1·M = I), what the column space and rank measure (everything the transformation can output, and how many dimensions that uses), and what the null space names (the input directions crushed to zero, the lost information itself). The source curriculum is 3Blue1Brown’s Essence of Linear Algebra by Grant Sanderson, freely available at 3blue1brown.com.
Where this fits
Section titled “Where this fits”This is lesson 7 of 15, the second lesson of Phase 2 (geometry of operations). It is the payoff of The determinant: the det ≠ 0 flag becomes a full account of what invertibility means and what its failure destroys. It also reuses Spans and basis directly, since the column space is a span and the dependent-columns collapse appears here a third time. The next lesson, Nonsquare matrices, drops the assumption that input and output share a dimension, and shows that column space and rank still tell the story when they do not.
Before you start
Section titled “Before you start”Prerequisites: The determinant (the invertibility condition) and Spans and basis (span, linear dependence). You should be comfortable reading a matrix-vector product as a linear combination of the columns, because the column space is built directly on that. No new computation is introduced beyond solving a small system by hand. The practice is pen and paper.
About the math
Section titled “About the math”The lesson favors meaning over mechanics: it does not teach the recipe for computing inverse entries (software does that), but rather what the inverse is and when it exists. You will solve a couple of small 2x2 systems by hand, find a null space by solving M·v = 0, and check the rank-nullity sum. The arithmetic stays light; the work is interpreting determinant, rank, and null space as three views of one question.
By the end, you’ll be able to
Section titled “By the end, you’ll be able to”- Reframe a linear system M·v = b as finding the input a transformation sends to a target
- Define the inverse by M^-1·M = I and explain why it exists exactly when det(M) is nonzero
- Describe the column space as the span of the columns and the rank as its dimension
- Describe the null space as the inputs crushed to zero, and apply rank + nullity = input dimension
- Identify invertibility three equivalent ways (nonzero determinant, full rank, trivial null space) and predict how many solutions a system has
Time and difficulty
Section titled “Time and difficulty”- Read time: about 11 minutes
- Practice time: about 15 minutes (a test-invertibility-three-ways exercise, a solve-the-system drill, and flashcards)
- Difficulty: standard (a Phase 2 lesson; the arithmetic is small systems, the work is interpreting three views of invertibility)