Canonical: https://knowlpedia.com/linear-algebra/dual-basis/

# Dual basis

The coordinate linear functionals associated with a basis of a finite-dimensional vector space.

Let \((e_1,\ldots,e_n)\) be a [basis](https://knowlpedia.com/convex-analysis/basis-hamel-basis-and-dimension/) of a finite-dimensional vector space \(V\). Its **dual basis** is the family \((e^1,\ldots,e^n)\) of [linear functionals](https://knowlpedia.com/convex-analysis/dual-space-and-duality-pairing/) satisfying
\[
e^i(e_j)=\delta_{ij},
\]
where \(\delta_{ij}\) is one if \(i=j\) and zero otherwise. For \(v=\sum_j v_j e_j\), the functional \(e^i\) returns the coordinate \(v_i\), so
\[
v=\sum_i e^i(v)e_i.
\]

## Basis of the dual

Every linear functional \(\ell\) equals \(\sum_i\ell(e_i)e^i\). Evaluating a proposed linear relation among the \(e^i\) on each \(e_j\) shows independence. Hence the family is a basis of the [dual space](https://knowlpedia.com/convex-analysis/dual-space-and-duality-pairing/). No inner product or orthogonality is required.

## References

- [Sheldon Axler, Linear Algebra Done Right, 4th ed., Chapters 2–3, 6 and 9](https://linear.axler.net/LADR4e.pdf).
