Subspaces and Quotient Spaces: Linear Algebra Explained

A subspace is a smaller vector space sitting neatly inside a bigger one: a subset that is itself a vector space under the same operations. Lines and planes through the origin in R cubed are subspaces, while a line missing the origin is not, since it lacks the zero vector.

Subspaces matter because the interesting sets in linear algebra are usually subspaces: solution sets of homogeneous systems, outputs a matrix can produce, polynomials with given properties. A quotient space goes further and collapses a whole subspace to a single point, letting you study what remains, which is how new spaces are built from old ones and how working modulo something gets geometric meaning.

Key Concepts

• Subspace: a subset W of V that is a vector space under the same addition and scalar multiplication.
• Subspace test: a nonempty subset is a subspace exactly when closed under addition and scalar multiplication.
• Span: all linear combinations of given vectors, always a subspace, the smallest one containing them.
• Coset v + W: the set (v + w such that w in W), the subspace W shifted to pass through v.
• Quotient space V/W: the vector space of cosets, with (v1+W)+(v2+W) = (v1+v2)+W and c(v+W) = (cv)+W. The zero coset is W itself.

Formulas

• Coset equality: v1 + W = v2 + W exactly when v1 minus v2 belongs to W.
• Dimension formula: dim(V/W) = dim(V) minus dim(W) for finite-dimensional V.
• Canonical map: pi(v) = v + W is linear and onto, with kernel exactly W.

Worked Example

Let V = R squared and W be the x-axis, dimension 1. A coset (a, b) + W = (x + a, b) as x varies is the horizontal line at height b. The cosets (1, 1) + W and (2, 1) + W coincide because their difference (minus 1, 0) lies in W. But (0, 1) + W and (0, 2) + W differ, since (0, minus 1) is not on the x-axis. So each coset is identified by its height b alone, and R squared / W is 1-dimensional with basis the single coset (0, 1) + W, matching dim = 2 minus 1 = 1. Addition adds heights: ((0,1)+W) + ((0,2)+W) = (0,3)+W. Distinct cosets are disjoint parallel lines; sharing one vector forces equality.

Common Misconceptions

• Every subset is a subspace. The unit circle is not: (1,0)+(0,1) = (1,1) leaves it, and zero is missing.
• The union of two subspaces is a subspace. Usually false: the union of the axes contains (1,0) and (0,1) but not (1,1). Use the sum of subspaces instead.
• Cosets can partially overlap. Distinct cosets are fully disjoint; one shared vector forces them identical.

Real-World Applications

• Linear systems: solutions of Ax = b form a coset of the kernel: particular plus homogeneous solution.
• Error-correcting codes: decoding identifies which coset a received message lands in.
• Modular arithmetic: integers mod n are the quotient Z/nZ, the prototype of all quotients.

Subspaces are the well-behaved pieces of a vector space; quotient spaces flatten one piece to a point and study the rest. Keep the subspace test and the dimension formula dim(V/W) = dim(V) minus dim(W) close at hand.