- $$(1-k)\mathbf b + 3k\mathbf a = (1-r)\mathbf a + 3r\mathbf b\\\iff\\
1-k=3r\;\:\text{ and }\;\:3k=1-r$$ iff $\mathbf a$ and $\mathbf b$ are not collinear.
- $$p\mathbf a+q\mathbf b+r\mathbf c=\mathbf 0\\\iff\\p,q,r=0$$ iff $\mathbf a, \mathbf b,$ and $\mathbf c$ are not coplanar.
- In general: $$p_1\mathbf a_1+p_2\mathbf a_2+\ldots+p_n\mathbf a_n=\mathbf 0\\\iff\\p_1,p_2,\ldots,p_n=0$$ iff $\mathbf a_1, \mathbf a_2, \ldots, \mathbf a_n$ are linearly independent.
Is there a way to prove it? How do you know that there is no other combination of values that will allow both sides to equate?
The third point above is actually the definition of linear independence.
A more intuitive characterisation is this: $\{\mathbf a_1, \mathbf a_2, \ldots, \mathbf a_n\}$ is linearly dependent iff some vector in the set is a linear combination of the others.
Notice that in Euclidean space $\mathbb R^m,$
- when $n=2,$ the condition “some vector in the set is a linear
combination of the others” is equivalent to “the two vectors are
collinear” (i.e., one of them is a scalar multiple of the other); and
- when $n=3,$ the same condition is equivalent to “the three vectors
are coplanar”.
When the above boldfaced conditions are not met, “equating coefficients” is invalid. For example, $$p\begin{pmatrix} 1 \cr 0\end{pmatrix}+q\begin{pmatrix} 2 \cr 0\end{pmatrix}=-3\begin{pmatrix} 1 \cr 0\end{pmatrix}+4\begin{pmatrix} 2 \cr 0\end{pmatrix}\\\kern.6em\not\kern-.6em\implies p=-3,\;q=4;$$ the unknown tuple $(p,q)$ could well have been $(1,2),(3,1),$ etc.