Definition: Let $v_0, v_1.. v_k$ be points in $\mathbb{R}^d$. These points are called affinely independent if there do not exist real numbers $\alpha_0, \alpha_1...\alpha_k$ that are not all zero such that $\sum_{i=0}^k \alpha_i v_i = 0$ and $\sum_{i=0}^k \alpha_i = 0$.

We need to prove the following: The points $v_0, v_1... v_k$ are affinely independent if and only if the vectors $v_1 - v_0, v_2 -v_0... v_k - v_0$ are linearly independent.

Thank you so much.

  • 4
    $\begingroup$ You can try expressing one of the coefficients, say the first one, with the remaining ones since $\sum_{i=1}^{k-1} \alpha_i = -\alpha_0$. Then plug this into the definition of affine dependence and you will get a pattern. $\endgroup$ – user13838 Nov 15 '11 at 0:52

Assume affine independence. Assume some linear combination of the $v_i-v_0$ is 0. Write that linear combination down (with unknown coefficients) and manipulate it to a form where you can use the assumption of affine independence.

Conversely, assume linear independence. Let some linear combination of the $v_i$ be zero. Do some manipulation to relate it to a linear combination of the $v_i-v_0$. Use the linear independence hypothesis to draw a conclusion about the coefficients.

If all that is too abstract for you, try to do the case $k=1$ or $k=2$ where you can more easily see what's happening.


Statement A: $v_0,v_1,...,v_k$ are affinely independent.
Statement B: $v_1-v_0,v_2-v_0,...,v_k-v_0$ are linearly independent.

First let us prove that $A \implies B$.
Consider some $\lambda_1,...,\lambda_k$, such that:
$$\sum_{i=1}^{k} \lambda_i (v_i - v_0) = 0 \tag{1}$$ We have to show that, if affine independence holds, all the cofficients ($\lambda$'s) must be zero.
Now, consider some $\lambda_0 \in \mathbb{R}$, such that: $$\sum_{i=0}^{k} \lambda_i = 0\tag{2}$$ Also, we have: $$\sum_{i=0}^{k} \lambda_i v_i = \sum_{i=1}^{k} \lambda_i (v_i - v_0) + (\sum_{i=0}^{k} \lambda_i)v_0 \tag{3}$$ Using equations 1 and 2, we observe that both terms on the RHS of (3) are zero. This means that: $$\sum_{i=0}^{k} \lambda_i v_i = 0 \tag{4}$$

From equations 2 and 4, we can say that $\lambda_i = 0$, $\forall i$ due to affine independence. Therefore, since all coefficients must be zero, this implies that B is true.

Now, let us prove the converse, that $B \implies A$.
Consider some $\lambda_0,\lambda_1,...,\lambda_k$, such that: $\sum_{i=0}^{k} \lambda_i v_i = 0$ and $\sum_{i=0}^{k} \lambda_k = 0$.
We have to show that all these coefficients must be zero under the condition of linear independence.
Using equation 3, and the above two conditions, we can conclude that $\sum_{i=1}^{k} \lambda_i (v_i - v_0) = 0$.
Therefore, due to linear independence of $(v_i - v_0)$, we conclude that: $$\lambda_1 = \lambda_2 = .... = \lambda_k = 0$$ Also, $\sum_{i=0}^{k} \lambda_k = 0 \implies \lambda_0 = 0$.
This proves that they are affinely independent (B is true).

We have shown $A \implies B$ and $B \implies A$.
$\therefore A \iff B$.


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.