In this lesson:
Why does the reciprocal require ? What goes wrong when ?
Three numbers do the heavy lifting:
Do matrices have analogs of all three?
The zero matrix has all entries zero:
Additive identity: for matching dimensions.
Also: .
: 1s on the main diagonal, 0s elsewhere:
Diagonal runs top-left to bottom-right.
For :
From VM.C.9: generally .
is the exception: for any square .
The diagonal structure of returns each entry unchanged — no cross-mixing. This is the one reliable commuting case.
We know works as the multiplicative identity.
Next question: is there a matrix such that ?
If so, (multiplicative inverse). Not every matrix has one — the condition is the determinant.
Mnemonic: main diagonal product minus anti-diagonal product.
The determinant is a scalar — it measures the area-scaling factor of the transformation represents.
has an inverse .
Just as requires , requires .
: row 1 = row 2
Proportional rows collapse the plane to a line — not reversible. Large entries don't guarantee nonzero determinant.
The inverse of is the unique matrix satisfying:
Note: both orders give — unlike general matrix products.
This is by definition: the inverse is defined to commute with in exactly this way. The VM.C.9 non-commutativity concern does not apply to .
For with :
Swap diagonal, negate off-diagonal, divide by .
, :
Verify: ✓
,
— undefined (division by zero)
No inverse exists. Collapsed information cannot be recovered.
A student's answer for where :
Find the error. What is the correct approach?
No scaffolding — all four steps on your own.
From VM.C.9: does not imply in general.
With inverses: if is invertible and :
Inverses restore cancellation — but only for invertible matrices.
With an invertible , the equation has the unique solution:
Next: VM.C.11 explores matrix-vector products in detail. VM.C.12 uses as the area scaling factor when defines a geometric transformation.
Click to begin the narrated lesson
Understand zero and identity matrices