2.6.9 · D5Matrices & Determinants — Introduction

Question bank — Properties of determinants

1,561 words7 min readBack to topic

True or false — justify

True or false: for all square .
False. Linearity (P5) works one row at a time with the other rows frozen; adding whole matrices changes every row at once, so it can't factor. Counterexample : .
True or false: .
False. Scaling by scales all rows, and each contributes one factor of (P4), so .
True or false: if then has a row of zeros.
False. A zero row forces (P7), but the converse fails: only means the rows are linearly dependent, e.g. with no zero row.
True or false: swapping two columns changes the sign of the determinant.
True. By P1 every row rule is a column rule, and swapping rows flips the sign (P2), so swapping columns does too.
True or false: even for non-square... trick question.
Only square matrices have determinants, so the premise is fine only for square ; for those P1 gives because reindexing the permutation sum by keeps every signed term.
True or false: always.
True. By the product rule , since the two numbers commute even when the matrices do not.
True or false: if all diagonal entries are zero, then .
False. The permutation sum uses off-diagonal entries too. .
True or false: adding a multiple of one row to itself leaves the determinant unchanged.
False. "Slide Saves" (P6) requires adding a multiple of a different row. scales the row, so by P4 the determinant is multiplied by .
True or false: for an matrix, .
False in general. so : unchanged for even , negated for odd .
True or false: a matrix and its inverse have reciprocal determinants.
True. From and the product rule, , so (which also shows is required for an inverse to exist). See Inverse of a Matrix via Adjoint.

Spot the error

" then , answer unchanged."
The swap was ignored. The slide (P6) is safe, but the swap (P2) multiplies the running value by ; forgetting that flips the sign of the final answer.
"I pulled a common factor out of a matrix, so I divide the answer by ."
If the came from one row, dividing by is right (P4). But if you factored from every row you pulled out , so you must divide by , not .
"Two rows are proportional, , but not equal, so ."
Proportional still kills it: factor the out of (P4) to get two equal rows, which force (P3). Dependence, not literal equality, is the trigger.
"."
Powers are products, not scalings. By P9, , not .
"To compute , I expanded along a row, then swapped two rows to make it easier, and kept expanding — value stays the same."
Each swap flips the sign (P2). If you swap during a computation you must track a factor of per swap; otherwise the "easier" route gives the wrong sign.
" because inverting flips things."
Inverting has nothing to do with sign flips. The correct relation is from the product rule.

Why questions

Why does one row of zeros force the determinant to be ?
Every term in the permutation sum picks exactly one entry from that row; with a zero factor available in each term, all terms vanish (this is exactly P7, i.e. P4 with ).
Why is the determinant of a triangular matrix just the diagonal product?
Any permutation that dips off the diagonal must grab a zero entry from the empty triangle, killing that term; only the identity permutation survives, giving (P8).
Why does "adding a multiple of one row to another" leave the value unchanged?
By additivity (P5) it splits into the original plus times a determinant with two equal rows; the second is (P3), so the value is "saved" — that is precisely how P6 is proved. See the sliding-card picture below.
Figure — Properties of determinants
Why must every row property also be a column property?
Because (P1): transposing turns rows into columns without changing the value, so any statement provable for rows transfers verbatim to columns.
Why does swapping two equal rows immediately prove ?
The swap changes nothing (rows were identical) yet must negate the value (P2), so , and the only number equal to its own negative is .
Why is called a signed scaling factor rather than just a scaling factor?
The magnitude tells how area/volume stretches, and the sign records whether orientation is preserved or mirror-flipped . The picture below shows a positive-area square and its flipped, negative-area mirror image — see Linear Transformations & Scaling of Area.
Figure — Properties of determinants

Edge cases

What is of the (empty) matrix?
By convention it is — the permutation sum has exactly one term, the empty product, which equals . This keeps rules like and cofactor expansion working when a submatrix runs out of rows.
What is of the matrix ?
Just itself — the permutation sum has a single identity term , so there is no cross-multiplication.
What is the determinant of the identity matrix?
, because it is triangular with all diagonal entries (P8); this is the "no stretching, no flip" baseline.
If is with , what does that say geometrically?
The two rows are dependent, so the unit square is squashed onto a line: zero area, no inverse. This is exactly the Area of a Triangle using Determinants going flat.
What happens to if you scale a single row by ?
It becomes a zero row, and (P4 with , which is P7); the map collapses one direction entirely.
Can a matrix with all positive entries have determinant ?
Yes — positivity of entries says nothing about dependence. has equal rows despite every entry being positive (P3).
For which does actually hold?
Only , where ; for all the correct factor is , so the naive rule is a genuine trap.

Connections