Suppose that $\underline{u},\underline{w}_1,...,\underline{w}_n\in\mathbb{R^n}$
$\text{span}\left\{\underline{w}_1,...,\underline{w}_n,\underline{u}\right\}=\text{span}\left\{\underline{w}_1,...,\underline{w}_n\right\}$ iff $\underline{u}$ is a linear combinations of others.
My suggested proof:
Let $\bf{v}\in$ $\underline{u},\underline{w}_1,...,\underline{w}_n$
So there are scalars $c_1,...c_n$ such that
$\bf{v}$$=c_1\underline{w}_1+...+c_n\underline{w}_n+0\underline{u}=c_1\underline{w}_1+...+c_n\underline{w}_n$
So $\text{span}\left\{\underline{w}_1,...,\underline{w}_n,\underline{u}\right\}=\text{span}\left\{\underline{w}_1,...,\underline{w}_n\right\}$?
I'm not sure if the proof is correct.
If $V=$ Span$\{u,w_1,\ldots,w_n\}=$ Span$\{w_1,\ldots,w_n\}$ then for any $v\in V$ there exists $s,c_i,d_i$ such that \begin{align} v &=su+c_1w_1+\cdots+c_nw_n\\ &=d_1w_1+\ldots+d_nw_n. \end{align} We can accept without loss of generality $s\neq 0$ so $u=\sum_{i=1}^n \frac{1}{s}(d_i-c_i)w_i$. Other implication is obvious so we are done.
EDIT $\frac{1}{s}$ added to the expression of $u$. Since Span$\{u,w_1,\ldots,w_n\}=$ Span$\{w_1,\ldots,w_n\}$, the set $\{u,w_1,\ldots,w_n\}$ is linearly dependent so the expansion of $v$ is not unique in terms of $\{u,w_1,\ldots,w_n\}$. When $s=0$, the $c_i$'s must be $d_i$ and when $s\neq 0$ there are different $c_i$ and $d_i$.
Actually I noticed shortest proof of the proposition while typing the edit. If $V=$ Span$\{u,w_1,\ldots,w_n\}=$ Span$\{w_1,\ldots,w_n\}$ then $u\in V=$ Span$\{w_1,\ldots,w_n\}$. Thus $u$ is a linear combination of $w_1,\ldots,w_n$