Let be a set of linearly dependent vectors, and let .
Let be a set of linearly dependent vectors, and let and be non-zero scalars.
Let be a set containing the zero vector, say at . Let . Now, there exists a column of zeros at . There is no pivot in this column, so a free variable exists. The equation has a non-trivial solution. There exist weights , not all zero, specifically , such that: Thus, the set is linearly dependent.
Let be a set of only one vector.
{} is linearly dependent.
has a non-trivial solution.
Let , where each has entries. Then is an matrix. The equation corresponds to a system of equations and unknowns. If , then we have more unknowns than equations, and we will have a free variable in our system. By a previous theorem, has a non-trivial solution, so the set of the columns of is linearly dependent.