Let $V$ be a vector space and let $v_1, v_2, \dots , v_n$ be a basis in $V$. For $x = \sum_{k=1}^{n}{\alpha_kv_k}, y = \sum_{k=1}^{n}{\beta_kv_k}$, define $\langle x,y \rangle := \sum_{k=1}^{n}{\alpha_k\bar{\beta_k}}$. Prove that $\langle x, y \rangle$ defines an inner product in V.

When I tried doing this question, I thought the best thing to do is to see if this definition satisfies the rules of an inner product. This wasn't entirely an issue when proving conjugate symmetry:

\begin{align*} \overline{\langle y,x \rangle} &= \overline{\sum_{k=1}^{n}{\beta_k\overline{\alpha_k}}} \\ &= \overline{\beta_1 \overline{\alpha_1}} + \dots + \overline{\beta_n \overline{\alpha_n}} \\ &= \alpha_1 \overline{\beta_1} + \dots + \alpha_n \overline{\beta_n} \\ &= \sum_{k=1}^{n}{\alpha_k \overline{\beta_k}} \\ &= \langle x,y \rangle \end{align*}

What I seem to be stuggling with is trying to prove linearity, non-negativity and non-degeneracy for this because for linearity the expression seems to become too big and complicated and for the other two properties I would need to have the inner product of a vector $x$ by itself.

Any suggestion of how I can go forward with this?

  • $\begingroup$ Linearity , non negativity are also as simple as you have done the conjugate symmetry part. Why do you think that those one will be lengthy? $\endgroup$
    – A learner
    Jul 20 at 16:20


Define $z = \sum_k \gamma_k v_k$, then $y + z = \sum_k (\beta_k + \gamma_k) v_k$ and

\begin{eqnarray} \langle x, y + z\rangle &=& \sum_k a_k\overline{(\beta_k + \gamma_k)} \\ &=& \sum_k (a_k \overline{\beta_k} + a_k \overline{\gamma}) \\ &=& \sum_k a_k \overline{\beta_k} + \sum_k a_k \overline{\gamma_k} \\ &=& \langle x, y\rangle + \langle x, z\rangle \end{eqnarray}


\begin{eqnarray} \langle x, x\rangle &=& \sum_k a_k\overline{a_k} \\ &=& \sum_k |a_k^2| \ge 0 \end{eqnarray}

which is only zero iff $a_k = 0$


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.