Check dimensions before using Rank-nullity theorem
Write the size and role of every vector, matrix, space, and transformation. An expression can look algebraically familiar while being dimensionally undefined. Rebuild missing background in the linear algebra subject guide, then use linear algebra practice to retrieve the relationship without being told its name.
Follow one object through Rank-nullity theorem
Track both the coordinates and the geometric or structural meaning:
Carry out the transformation in a valid order
- Name the unknown, and the unit the answer has to come out in.
- Match the symbols to your values. $n$ is number of columns of $A$.
- Check the conditions before substituting. $A$ represents a linear map from an $n$-dimensional domain.
- Substitute, keep exact values to the last line, then test the sign, size, and unit against a rough estimate - The check that catches most linear algebra slips.
An order or dimension error to avoid
What goes wrong: Students plug in the number of rows for n when the matrix is not square.
Why it fails: The theorem describes what happens to the inputs, and the inputs are counted by columns. Using rows gives the wrong total whenever the matrix is taller or wider than it is square.
Repair: Count the columns. For [[1, 2, 3], [2, 4, 6]] there are 3 columns, the rank is 1 because the second row is double the first, so the nullity is 2 even though the matrix has only 2 rows.
Multiply back or test a simple vector
After using Rank-nullity theorem, repeat the calculation with units attached and test the result against a boundary case or a quick numerical estimate.
When the transformation statement holds
The statement is about a linear map, so it says nothing about a nonlinear transformation, and it needs the input space to be finite-dimensional for n to be a number at all. Within those conditions it never breaks.
Questions about Rank-nullity theorem
Is n the number of rows or columns?
Columns, always. A useful memory hook is that each column corresponds to one input variable, and the theorem splits the input variables into pivot ones and free ones.
What does nullity 0 mean?
That the only vector sent to zero is the zero vector, so no two different inputs collide. For a square matrix that also means the determinant is nonzero and the system has one unique solution.
Can the rank be larger than the number of rows?
No. The rank cannot exceed either dimension, so it is capped by the smaller of the row and column counts. A 2 by 5 matrix can have rank at most 2, forcing a nullity of at least 3.
How do I find the rank by hand?
Row reduce until the matrix is in echelon form and count the pivots, one per nonzero row. Then read the nullity off as the number of columns without a pivot, with no extra work.
Change one input, predict whether the output should rise, fall, or stay fixed, and only then recalculate.