I am trying to deepen my understanding of linear independence and uniqueness of linear combination.

I want to prove this statement: Assume $\{v_1,v_2,...v_n\}$ span a vector space $S$. $x \in S$ can be written as a unique linear combination of vectors $x = c_1v_1 + c_2v_2+...c_nv_n$ iff $\{v_1,v_2,...v_n\}$ are linearly independent.

Proof that if any $x \in S$ can be written as a unique linear combination of vectors $x = c_1v_1 + c_2v_2+...c_nv_n$, then $\{v_1,v_2,...v_n\}$ are linearly independent by contradiction:

Assume the negation: that we can write $x = c_1v_1 + c_2v_2+...c_nv_n$, and $\{v_1,v_2,...v_n\}$ are linearly dependent. Thus, theres exist at least one non-zero constant $d_k$ such that $0 = d_1v_1 + d_2v_2+...d_kv_k$. Hence, we can also say that $x = c_1v_1 + c_2v_2+...c_nv_n + d_1v_1 + d_2v_2+...d_kv_k$. Since there exists at least one $d_k$ that is nonzero, we contradicted the assumption that we could write $x$ uniquely.

Proof the converse by contradiction:

Assume that $\{v_1,v_2,...v_n\}$ are linearly independent but there exists some $x \in S$ such that $x = c_1v_1 + c_2v_2+...c_nv_n$ and $x = d_1v_1 + d_2v_2+...d_nv_n$ such that for at least one $k$, $d_k \neq c_k$.

Then, we can rewrite as $$\vec{0} = c_1v_1 + c_2v_2+...c_nv_n - (d_1v_1 + d_2v_2+...d_nv_n) = (c_1-d_1)v_1 + (c_2-d_2)v_2 + (c_n-d_n)v_n$$

However, since for at least one $k$, $d_k \neq c_k$, one of these terms is non-zero, which by definition contradicts the assumption that $\{v_1,v_2,...v_n\}$ are linearly independent.

Is this a valid proof? If not, is are there any suggestions on how to improve it?


Seems fine.

Just a minor typo in the forward direction. I think you meant to say $$0=d_1v_1+d_2v_2+\ldots +d_nv_n$$

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.