2.6.7 · D5Matrices & Determinants — Introduction

Question bank — Determinant of 2×2 matrix

1,309 words6 min readBack to topic

True or false — justify

The square is scaled by , so its determinant must be .
False — scaling every side by scales the columns by , so area scales by ; for a matrix.
If , then must be the identity matrix.
False — only says area and orientation are preserved; a rotation or a shear like also has but is not the identity.
A matrix with all four entries non-zero always has non-zero determinant.
False — has no zeros yet ; what matters is whether one column is a multiple of the other, not whether entries are zero.
Swapping the two rows of leaves unchanged.
False — a single row swap multiplies the determinant by because it reverses the orientation of the two column vectors ().
for every matrix.
True — transposing swaps , and is the same number as ; area is unchanged.
If , then shrinks areas because the number is negative.
False — the magnitude gives the area factor (areas triple); the minus sign only reports an orientation flip, not shrinking.
is a valid rule.
False — the determinant is not additive. E.g. give but .
If both columns of point in the same direction, .
True — parallel columns span a line, not a parallelogram; the enclosed area is zero (see Linear independence).
A matrix and its inverse have reciprocal determinants.
True — since , we get , which is why can't exist when .
always holds.
True — both equal , and multiplication of numbers commutes, even though the matrix products and may differ.

Spot the error

A student writes . What went wrong?
The diagonals are subtracted in the wrong order; correct is main-diagonal-first, . Their answer is the exact negative — a determinant computed with reversed orientation.
Someone claims . Error?
They computed the trace (sum of the diagonal), not the determinant. The determinant is ; trace and determinant are different quantities.
", so is the zero matrix." Fix it.
means singular (columns linearly dependent), not empty. has yet is far from the zero matrix.
A student uses . What's wrong?
The diagonal entries must be swapped (), not left in place; the correct adjugate is (see Inverse of a 2×2 matrix).
"." Why is this false for a ?
Multiplying by scales both columns by , so area scales by : . The scalar comes out to the power equal to the matrix size.
Given and , a student says could be non-zero. Error?
; any product involving a singular matrix is singular. Composing with a flattening map still flattens.

Why questions

Why does a zero determinant guarantee that has no unique solution?
A zero determinant collapses the plane onto a line, so the map isn't reversible; different inputs land on the same output, leaving either no solution or infinitely many (this is the boundary case of Cramer's rule).
Why is the determinant of a rotation matrix always ?
Rotation rigidly turns the plane without stretching or flipping, so area is preserved () with factor — no scaling and no orientation reversal.
Why does the sign of the determinant encode orientation rather than size?
The area magnitude comes from ; the sign records whether going column-1-then-column-2 sweeps counterclockwise () or clockwise (), i.e. whether the plane was flipped like a pancake.
Why can two very different matrices share the same determinant?
Determinant captures only the area-scaling-plus-flip summary, discarding shape and rotation; a shear and a rotation can both preserve area, giving equal determinants despite acting differently.
Why does correspond exactly to linearly dependent columns?
Dependent columns are parallel (one is a scalar multiple of the other), so the parallelogram they span has zero width and hence zero area — precisely (see Linear independence).
Why is the determinant unchanged by transposing, but negated by a row swap?
Transposing reflects the parallelogram, preserving its area and its orientation relationship; a row swap re-labels which vector is "first," reversing the sweep direction and hence the sign.

Edge cases

If one entire column of is , what is and why?
: a zero column contributes a zero-length side, so the parallelogram degenerates to a segment (or point) with no area.
What is of a matrix whose two columns are equal, e.g. ?
Zero — identical columns are the most extreme parallel case; , so the "parallelogram" is just a line.
For a diagonal matrix , what does the determinant become and why is it easy to see?
It becomes (the product of diagonal entries) because the axes stretch independently by and , scaling area by with no shearing.
Is it possible for while is invertible and is not?
No — invertibility is equivalent to . If is singular then , forcing , which would make singular too.
What happens to if you multiply just one row by (not the whole matrix)?
It scales by once: becomes , since only one side of the parallelogram is stretched by , changing the area by that single factor.
Can a matrix with a negative determinant still be invertible?
Yes — invertibility only needs . A negative determinant (say ) means orientation flips, but the map is perfectly reversible.
At the exact value , what geometric object does the unit square map to?
A line segment or a single point — the collapse boundary between orientation-preserving () and orientation-reversing () transformations (see Transformations and scaling).