Linearly Dependent Sequence of Vector Space
Contents |
Theorem
Let $\left({G, +}\right)$ be a group whose identity is $\mathbf 0$.
Let $\left({G, +, \circ}\right)_K$ be a $K$-vector space.
Let $\left \langle {a_n} \right \rangle$ be a sequence of distinct non-zero vectors of $G$.
Then $\left \langle {a_n} \right \rangle$ is linearly dependent iff:
- $\exists p \in \left[{2 .. n}\right]: a_p$ is a linear combination of $\left \langle {a_{p-1}} \right \rangle$
Proof
Necessary Condition
Suppose $\left \langle {a_n} \right \rangle$ is linearly dependent.
By hypothesis, the set of all integers $r \in \left[{1 .. n}\right]$ such that $\left \langle {a_r} \right \rangle$ is linearly dependent is not empty.
Let $p$ be its smallest element.
Then from Singleton is Linearly Independent‎, $p \ge 2$, as $a_1 \ne \mathbf 0$ and hence $\left\{{a_1}\right\}$ is linearly independent.
Also, there exist scalars $\lambda_1, \ldots, \lambda_p$, not all of which are zero, such that $\displaystyle \sum_{k=1}^p \lambda_k \circ a_k = \mathbf 0$.
Suppose $\lambda_p = 0$.
Then not all of $\lambda_1, \ldots, \lambda_{p-1}$ can be zero.
Then $\left \langle {a_{p-1}} \right \rangle$ is linearly dependent.
That contradicts the definition of $p$, so $\lambda_p \ne 0$.
So, because:
- $\displaystyle \lambda_p \circ a_p = - \sum_{k=1}^{p-1} \lambda_k \circ a_k$
we must have:
- $\displaystyle a_p = \sum_{k=1}^{p-1} \left({- \lambda_p^{-1} \lambda_k}\right) \circ a_k$
and thus $a_p$ is a linear combination of $\left \langle {a_{p-1}} \right \rangle$.
$\Box$
Sufficient Condition
Now suppose that $a_p$ is a linear combination of $\left \langle {a_{p-1}} \right \rangle$.
Then:
- $\displaystyle a_p = \sum_{k=1}^{p-1} \mu_k \circ a_k$
So we can assign values to $\lambda_k$ as follows:
- $ \forall k \in \left[{1 .. n}\right]: \lambda_k = \begin{cases} \mu_k & : k < p \\ -1 & : k = p \\ 0 & : k > p \\ \end{cases}$
Then $\displaystyle \sum_{k=1}^n \lambda_k \circ a_k = \mathbf 0$.
Hence the result.
$\blacksquare$
Sources
- Seth Warner: Modern Algebra (1965): $\S 27$: Theorem $27.6$
- For a video presentation of the contents of this page, visit the Khan Academy.