MATH0005 Chap.4 Bases, Dimension and Rank
Bases, Dimension and Rank
Bases, Dimension and Rank turns a spanning and independence argument into a numerical invariant. Every basis of a finite-dimensional subspace has the same number of vectors, called its dimension. For a matrix with n columns, rank-nullity divides the domain into directions visible in the image and directions lost in the null space: rank plus nullity equals n. RREF exposes the pivot count and free-variable count.
Fundamental solutions built from free variables form a null-space basis. Pivot positions identify a column-space basis, but those vectors must be selected from the original matrix because row operations change the actual column space. The nonzero RREF rows form a basis of the row space, and row rank equals column rank.
Dimension also converts containment into equality: if U is contained in V and the dimensions match, then U equals V. An independent sequence cannot contain more vectors than the ambient dimension, while a spanning sequence cannot contain fewer. The chapter uses one reduction to coordinate several spaces, then checks every count against rank-nullity and the ambient dimensions.
Sum and intersection dimensions add an overlap audit that prevents shared directions from being counted twice and constrains every result by the ambient space.
What this chapter covers
- 01
Basis size
- 02
Dimension
- 03
Fundamental solutions
- 04
Column rank
- 05
Row space
- 06
Rank-nullity
- 07
Dimension inclusion
Worked example · free
Find null and column-space bases
- 3Recognise that the third row is the sum of the first two and reduce to pivots in columns 1 and 2.
- 4Parameterise with x3=s and x4=t to obtain null-space basis vectors (2,-1,1,0) and (1,-1,0,1).
- 3Use columns 1 and 2 of the original matrix for a column-space basis: (1,0,1) and (2,1,3).
- 2State rank 2 plus nullity 2 equals the four columns.
Key terms
- Nullity
- The dimension of a matrix's null space, equal to the number of free variables.
- Pivot column
- A column position containing a leading entry in reduced echelon form.
- Row space
- The subspace spanned by the rows of a matrix.
Bases, Dimension and Rank FAQ
Why use original columns for the column space?
Row operations preserve linear relations needed to identify pivot positions, but they change the actual columns through left multiplication. Therefore the RREF tells you which positions to choose, while the basis vectors themselves must come from the original matrix.
How do I find a null-space basis?
Reduce Ax=0, choose free variables as parameters and write the general solution as a linear combination. Set one free parameter to one and the others to zero in turn. The resulting fundamental solutions form a basis of the null space.
When do equal dimensions imply equal subspaces?
Only when one subspace is contained in the other. Distinct lines can have the same dimension, so equal counts alone are insufficient. Containment plus equal finite dimension rules out any missing direction and forces the two subspaces to coincide.
Exam move
Use a single carefully checked RREF to organise rank, nullity and three related spaces. Mark pivot positions and free variables before extracting any vector. Build one fundamental null vector per free variable, then substitute each into Ax=0. Retrieve original pivot columns for the image and use nonzero RREF rows for the row space.
Reconcile the number of pivot columns with the number of nonzero rows, since row rank and column rank agree. Reconcile rank plus nullity with the number of columns. Keep ambient spaces visible: the column space lies in the codomain, while the row and null spaces lie in the domain coordinate space. Practise theorem-driven shortcuts only with their hypotheses.
Equal dimensions prove equality only alongside containment; n independent vectors form a basis only in an n-dimensional target; n spanning vectors likewise need the target dimension. Use the sum-dimension formula to avoid double-counting intersection directions and check the result against the ambient dimension.
Record errors under wrong source of column vectors, omitted free parameters, basis count used without a property proof and row-space coordinates put in the wrong ambient space. End each answer by naming the basis, dimension and numerical reconciliation that verifies it. Use a four-column audit table headed space, ambient space, basis source and dimension. Fill it for the null, column and row spaces after every full reduction.
Then create dimension-only exercises in which one of dim U, dim V, dim(U intersection V) or dim(U+V) is missing, and reject any result outside its legal bounds. For basis-extension problems, explain which independent directions are retained and which new directions are added. The purpose is to connect counting theorems to actual vectors, preventing a correct dimension statement from masquerading as a constructed basis.
Working through Bases, Dimension and Rank in MATH0005? Sia is AskSia’s AI Mathematics tutor — ask any MATH0005 Bases, Dimension and Rank question and get a clear, step-by-step explanation grounded in how MATH0005 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.