Mathematics
The Rank-Nullity Theorem
Quick fact
For any linear map from one finite-dimensional vector space to another, the dimension of its kernel plus the dimension of its image always equals the dimension of its domain—no matter what the map is.
Why this is interesting
Have you ever wondered how many independent directions a transformation can ‘reach’ versus how many it ‘collapses’? The rank-nullity theorem reveals a beautiful balance between these two numbers.
Read the full explanation
Understanding The Rank-Nullity Theorem
Imagine a machine that takes in objects and transforms them. Some objects might be crushed into the same output, while others are painted in a variety of colors. The rank-nullity theorem is about counting how many independent directions the machine preserves (the image) and how many independent directions it completely loses (the kernel). Formally, for a linear map T from a vector space V to a vector space W, the kernel is the set of all vectors in V that T sends to zero. The image is the set of all outputs in W. The rank is the dimension of the image, and the nullity is the dimension of the kernel. The theorem states: rank(T) + nullity(T) = dim(V). Think of a rectangular grid of points in the plane. If you project it onto a line, the image is just that line (dimension 1), while the kernel is the direction perpendicular to the line (dimension 1). The sum is 2, the dimension of the plane. If you rotate the plane, the image is the whole plane (dimension 2) and the kernel is {0} (dimension 0), again summing to 2.
A deeper explanation
The rank-nullity theorem works because of a fundamental relationship between the kernel and the image. When a linear map T acts on a vector space V, it partitions V into cosets of the kernel. Each coset corresponds to a distinct vector in the image. In fact, the image is isomorphic to the quotient space V / ker(T). Therefore, the dimension of V equals the sum of the dimension of the kernel and the dimension of this quotient, which is exactly the rank. This is a direct consequence of the first isomorphism theorem for vector spaces. The theorem is not just a neat relationship; it is profoundly useful. It tells us that if a linear map has a large kernel (many things collapse to zero), then its image must be small, and vice versa. This helps determine whether a linear system has solutions: if the rank of the matrix is less than the dimension of the target space, the system may be inconsistent. It also explains why a square matrix is invertible if and only if its kernel is trivial (nullity 0), because then the rank equals the dimension of the domain, meaning the map is onto and thus invertible.