The Column Space, the Null Space and the Rank
An invertible transformation preserves the full dimension of space. But a non-invertible transformation squashes space flat — and then you need to ask two questions:
- Where do all the output vectors “land”? → the column space
- Which vectors get squashed to the origin? → the null space
Definition
The column space of a matrix is the set of all possible outputs:
Geometric meaning: what the transformed space looks like.
Why it is called the “column space”
Recall that
The output is always a linear combination of the columns! Hence:
Example: M = [[1, 2], [2, 4]] (column 2 = column 1 × 2)
column space = span{(1,2)} = a straight line
⟹ every output of this transformation lies on the line y = 2x
Deciding whether lies in the column space
The equation has a solution ⟺ .
Geometrically: has to lie inside the transformed space for some to map onto it.
Definition
The null space of a matrix is the set of all vectors mapped to the origin:
Geometric meaning: the vectors that get “squashed flat to the origin” in the course of the transformation.
Why the Null Space Matters
Because it is “the direction in which information is lost”. If has a nonzero solution , then and map to the same output:
The larger the null space, the more information is lost.
The definition of the rank
The rank = the dimension of the column space = the dimension of the transformed space.
- Full rank (rank = 2 for a 2×2 matrix): the transformation squashes nothing, and it is invertible
- Rank deficient (rank = 1): squashed into a line
- rank = 0: squashed into a single point
The Rank-Nullity Theorem
where is the dimension of the input space (the number of columns of the matrix).
Geometric intuition: the input space is split into two parts — “the part that becomes the output space” (the column space) and “the part that gets squashed to the origin” (the null space). The two together make up exactly the whole dimension of the input space.
Example: M = [[1, 2], [2, 4]] (2×2, squashed into a line)
column space: one line → dimension 1
null space: the solutions of Mx = 0 are x₂ = -x₁/2 (a line) → dimension 1
rank-nullity: 1 + 1 = 2 ✓
The unification with the determinant and invertibility
| Condition | Meaning |
|---|---|
| det ≠ 0 | full rank, the null space contains only the zero vector, invertible |
| det = 0 | rank deficient, the null space contains nonzero vectors, not invertible |
One Diagram for Four Concepts
Input space (n-dimensional)
├── the part mapped onto the column space (dimension = rank)
└── the part squashed onto the null space (dimension = n − rank)
↓
column space + null space = input space (the rank-nullity theorem)- Column space = where the transformation can “reach” (the set of outputs)
- Null space = what the transformation “squashes away” (the set of vectors sent to the origin)
- Rank = the dimension of the column space, measuring how much information the transformation retains
- Full rank ⟺ invertible ⟺ det ≠ 0 ⟺ the null space is 0
- Column space: the two columns are collinear, so (a line, dimension 1)
- Null space: solve
- This gives , that is, (a line, dimension 1)
- Rank: rank = 1
- Verify the rank-nullity theorem: 1 + 1 = 2 ✓
- Conclusion: is not invertible, and the equation either has no solution ( is not in the column space) or has infinitely many solutions
Symbols Used in This Article
| Symbol | Type | Pronunciation/Explanation | Meaning in This Article |
|---|---|---|---|
| Mathematical symbol | column space of M | The column space: the set of all possible outputs of the matrix | |
| Mathematical symbol | null space of M | The null space: the set of vectors mapped to the origin by | |
| Mathematical symbol | M x equals zero | The homogeneous system whose solutions form the null space | |
| Mathematical symbol | rank-nullity | The rank-nullity theorem: the dimension of the column space plus that of the null space equals the input dimension | |
| Mathematical symbol | n | The dimension of the input space (the number of columns of the matrix) | |
| Mathematical symbol | R n | The -dimensional real vector space, appearing in | |
| Mathematical symbol | vector z | A nonzero vector in the null space, so that |
