For a 2×2 matrix, the determinant ad−bc measures the signed area of the parallelogram formed by the matrix, but for larger matrices that formula grows to thousands of terms, so we need a simpler way to find determinants.
Why We Use Rules and Row Vectors
Instead of using columns to build our parallelogram, we can use the rows of the matrix so that row operations work directly on the edges of the shape. Turning the rows of A=[acbd] into columns is the transpose AT, which leaves the determinant unchanged:
det(AT)=det[abcd]=ad−cb=ad−bc=det(A)
Even though the row vectors and column vectors of a 2×2 matrix usually form two different parallelograms, their signed areas are always equal and flip orientation at the exact same moment.
The parallelogram made by the rows and the one made by the columns always have the exact same signed area.
1
The Three Defining Rules
The determinant of any matrix comes from three geometric rules that signed area and volume follow when we work with the rows.
The Three Defining Rules
Every property of determinants comes from these three rules:
Rule 1 (Identity):det(I)=det[1001]=1Rule 2 (Row swap):det[0110]=−det[1001]=−1Rule 3 (One row at a time):det[3ac3bd]det[a+a′cb+b′d]=3det[acbd]=det[acbd]+det[a′cb′d]
the original area det(A)
only the top row stretched area 3det(A)
3A: both rows stretched area 32det(A)=9det(A)
On a test, watch the difference between multiplying a single row by a number c and multiplying the entire matrix A by c. In an n×n matrix, cA multiplies every one of the n rows by c, so a factor of c pulls out n times:
det(cA)=cndet(A)
For the same reason, we cannot split addition across a whole matrix: det(A+B) usually does not equal det(A)+det(B). Adding two 2×2 matrices changes both rows at once, so splitting one row at a time produces two extra determinants that mix a row from A with a row from B, alongside det(A) and det(B):
When the bottom row ⟨c,d⟩ stays fixed as a shared base, the heights of the parallelograms for ⟨a,b⟩ and ⟨a′,b′⟩ add up, so their signed areas add up to the signed area for ⟨a+a′,b+b′⟩.
With the second row kept as a shared base, adding a vector to the first row adds its height and signed area to the parallelogram.
2
Equal Rows, Row Replacement, and Triangular Matrices
To find the determinant of a matrix of any size without a long formula, we use the three defining rules to track what happens when we simplify a matrix with row operations.
Equal Rows, Row Replacement, and Triangular Matrices
If a matrix has a row of all zeros, factoring 0 out of that row immediately gives a determinant of 0. If a matrix has two identical rows, the parallelogram is squished flat onto a single line, and swapping those two equal rows flips the sign of the determinant D while leaving the matrix unchanged:
D=−D⟹D=0
Because a matrix with two equal rows has determinant 0, adding a multiple of one row to another row never changes the determinant. Splitting the changed row gives the original determinant plus a multiple of a determinant with two equal rows, which is 0:
This lets us find the determinant of any triangular matrix, which is a matrix with all zeros below (or above) the main diagonal, in one step.
Text solution
Eliminate above the diagonalSubtract multiples of the bottom row and then the middle row to clear the entries above the diagonal without changing the determinant.
det200730954=det200030004
Factor out each row's diagonal entryPull the number out of each row one at a time to leave the identity matrix.
=2⋅3⋅4⋅det(I)=1det100010001=2×3×4×1=24
So the determinant of any triangular matrix is simply the product of its main diagonal entries, and if elimination produces a 0 on the diagonal, the determinant is 0 and the matrix is not invertible.
Adding a multiple of Row 1 to Row 2 (R2→R2+kR1) slides the second row vector ⟨c,d⟩ parallel to the fixed base ⟨a,b⟩. That keeps the height and signed area the same, even when the vector lands on the vertical axis and makes the matrix upper triangular.
Adding a multiple of the first row to the second slides the second vector parallel to the first without changing the height or signed area, even when it lands on the vertical axis to make the matrix upper triangular.
3
Building ad−bc from the Rules
We can also see how the 2×2 determinant formula ad−bc comes directly from splitting rows and applying the three defining rules.
Building ad − bc from the Rules
det[acbd]=?
using only Rules 1, 2 and 3
Text solution
Split the first rowWrite the top row ⟨a,b⟩ as ⟨a+0,0+b⟩ and split the addition while holding the bottom row fixed.
det[acbd]=det[ac0d]+det[0cbd]
Split the second row in each pieceBreak each bottom row ⟨c,d⟩ into ⟨c,0⟩+⟨0,d⟩ while holding the top rows fixed.
Factor out numbers and swap to match IDrop the two flat matrices whose rows lie on the same axis, multiply the diagonal of the second matrix, and factor b and c from the third matrix.
=0+ad+bc−1det[0110]+0=ad−bc
So the minus sign in ad−bc comes from the single row swap needed to put ⟨0,b⟩ and ⟨c,0⟩ in the same order as the rows of I.
4
Products, Transposes, Inverses, and Determinant Problems
When we know the determinants of A and B without knowing their entries, we can find the determinants of combined or modified matrices by tracking how each operation scales volume.
Products, Transposes, Inverses, and Determinant Problems
A,Bare 3×3
det(A)=5,det(B)=−2
det(2A)det(ATB)det(A−1)det(A+B)=?=?=?=?
Multiplying two matrices applies one volume stretch after the other, and transposing leaves the signed volume and main diagonal unchanged, so products and transposes split cleanly:
det(AB)=det(A)det(B),det(AT)=det(A)
Using these rules alongside our row rules, we can find determinants for 3×3 matrices A and B with det(A)=5 and det(B)=−2.
Text solution
Track row operations one step at a timeStarting from det(A)=5, swapping R1 and R3 flips the sign, multiplying R2 by 4 multiplies the determinant by 4, and subtracting 10R1 from R3 changes nothing.
5R1↔R3−54R24(−5)=−20R3−10R1−20
Factor 2 from all three rows of 2ABecause A is 3×3, multiplying the whole matrix by 2 scales all three rows by 2.
det(2A)=23det(A)=8×5=40
Split the product and transpose in det(ATB)Separate the matrix product and replace det(AT) with det(A).
Use AA−1=I to find det(A−1)Take the determinant of both sides of AA−1=I and divide by det(A).
det(A)⋅det(A−1)⟹det(A−1)=det(I)=1=det(A)1=51
For det(A+B), however, we have to stop: without the entries of A and B, we cannot find det(A+B) because addition never splits across a whole matrix.
Key concepts
When finding a determinant from a matrix's entries, look first for zeros or matching rows. If a matrix is already triangular (all zeros below or above the diagonal), multiply the diagonal entries directly, such as det200730954=2×3×4=24. If it isn't triangular yet, use row replacement (Ri→Ri+kRj, which never changes the determinant) to clear entries into a triangular shape, and flip the sign once for each row swap you make along the way. When working with whole-matrix expressions where entries aren't given, split products, transposes, and inverses directly using det(AB)=det(A)det(B),det(AT)=det(A), and det(A−1)=det(A)1. Before pulling out a number c, always check the matrix size n×n so you write det(cA)=cndet(A) (like det(2A)=23(5)=40 for a 3×3 matrix). Never split det(A+B) into det(A)+det(B); only split a sum in a single row when every other row stays fixed, or use row replacement to cancel added multiples of another row first.
5
Practice
1
Find the determinant of the matrix 3−4720−219005−60001.
Be careful not to add the diagonal entries (which gives 7) or drop the minus sign on −2 (which gives 30).Be careful not to add the diagonal entries (which gives 7) or drop the minus sign on −2 (which gives 30).
Look at all the entries above the main diagonal. What shape is this matrix already in?
All entries above the main diagonal are 0, so the matrix is lower triangular and we can multiply the diagonal entries directly without doing any row operations below the diagonal. Be careful not to add the diagonal entries (which gives 7) or drop the minus sign on −2 (which gives 30):
det=3×(−2)×5×1=−30
2
Suppose A is a 3×3 matrix with det(A)=−6. What is the determinant of the matrix obtained by swapping Row 1 and Row 2, multiplying Row 3 by 5, and then adding 4 times Row 1 to Row 3?
Don't multiply by 53=125 (which gives 750), because only Row 3 was scaled, not the whole 3×3 matrix.Finally, adding 4 times Row 1 to Row 3 is a row replacement and changes nothing (don't multiply by 4 to get 120), leaving the final answer at 30.
Trace the determinant one operation at a time: which step flips the sign, which multiplies by a number, and which leaves the determinant unchanged?
Swapping Row 1 and Row 2 flips the sign from −6 to 6. Multiplying Row 3 by 5 scales only a single row, so multiply by 5 to get 30. Don't multiply by 53=125 (which gives 750), because only Row 3 was scaled, not the whole 3×3 matrix. Finally, adding 4 times Row 1 to Row 3 is a row replacement and changes nothing (don't multiply by 4 to get 120), leaving the final answer at 30.
3
Use row operations and the triangular rule to find det03321154−2.
Don't forget that this row swap flips the sign of the determinant (forgetting the sign flip gives 3×2×(−6)=−36).
Notice how Row 2 and Row 3 start with the exact same two numbers. What happens if you subtract Row 2 from Row 3, and which row swap then makes the matrix upper triangular?
Subtracting Row 2 from Row 3 (R3→R3−R2) gives 03021054−6 without changing the determinant. Next, swap Row 1 and Row 2 to get the upper triangular matrix 30012045−6. Don't forget that this row swap flips the sign of the determinant (forgetting the sign flip gives 3×2×(−6)=−36):
det=−(3×2×(−6))=36
4
Let A and B be 4×4 matrices with det(A)=3 and det(B)=−2. Find det(2ATB−1).
Because the matrices are 4×4, factoring 2 out of all four rows gives 24=16. Don't pull out a single 2 (which gives −3).Watch out for using −2 instead of the reciprocal −21, which gives −96.
Before splitting the product, check how many rows get multiplied by 2 in a 4×4 matrix, and how det(B−1) relates to det(B).
Because the matrices are 4×4, factoring 2 out of all four rows gives 24=16. Don't pull out a single 2 (which gives −3). Next, split the product using det(AT)=det(A)=3 and det(B−1)=det(B)1=−21 (watch out for using −2 instead of the reciprocal −21, which gives −96):
det(2ATB−1)=24⋅det(AT)⋅det(B−1)=16⋅3⋅(−21)=−24
5
Suppose det[acbd]=7 and det[ecfd]=−4. Find det[2a−3ec2b−3fd].
Don't square the numbers to 4 and 9, which gives 64, and don't drop the minus sign on −4, which gives 2.Don't square the numbers to 4 and 9, which gives 64, and don't drop the minus sign on −4, which gives 2.
Since the bottom row [cd] stays fixed in both matrices, how can you split the top row and factor out the numbers 2 and −3?
Because the second row [cd] stays fixed, we can split the subtraction in the first row and factor 2 and −3 out of that single row (don't square the numbers to 4 and 9, which gives 64, and don't drop the minus sign on −4, which gives 2):
Factoring −7 out of the bottom row (be careful not to drop the negative sign, which gives 35) leaves [cd] in the bottom row.
Instead of splitting both rows into four separate determinants, what multiple of the top row can you subtract from the bottom row to cancel the 3a and 3b?
Subtracting 3 times the top row from the bottom row cancels the a and b terms in the bottom row without changing the determinant:
det[a+2c3a−cb+2d3b−d]=det[a+2c−7cb+2d−7d]
Factoring −7 out of the bottom row (be careful not to drop the negative sign, which gives 35) leaves [cd] in the bottom row, and subtracting 2 times that new bottom row from the top row clears the 2c and 2d: