Count dimensions in the domain
For a linear map with finite-dimensional domain,
For an real matrix , the domain is . Hence
Rank counts pivot columns in row-echelon form. Nullity counts free variables in . Every column corresponds to one unknown, which is either pivot or free; this explains the dimension count.
Row reduction serves two different jobs
Row operations preserve the homogeneous solution set, so a null-space basis can be obtained from the reduced equations. Assign one free variable 1 and the others 0, then solve for the pivot variables.
For a column-space basis, use row reduction to identify pivot indices, then select those columns from the original matrix. Row operations generally change the column space, although they preserve column-dependence relations.
A fresh 2 × 3 example
There are two pivots and one free variable. Set :
Therefore is a null-space basis, and the first two original columns form a column-space basis. The check is , the number of columns, not the number of rows.
To validate a proposed basis, check membership, independence and spanning. A correct dimension sum alone does not establish that the displayed vectors form the claimed bases.
Related question
Apply this knowledge
Use the concept guide to understand the reasoning, then return to the complete question and worked answer.
Rank Nullity Theorem: Find Null-Space and Column-Space BasesSources
These references support the core concepts and interpretation boundaries explained above.