Canonical: https://knowlpedia.com/linear-algebra/eigenspace/

# Eigenspace

Set of vectors sent to scalar multiples of themselves for a fixed eigenvalue.

An **eigenspace** of a [linear operator](https://knowlpedia.com/linear-algebra/linear-operator/) \(T:V\to V\) associated to a scalar \(\lambda\in\mathbb{F}\) is the set
\[
E_\lambda(T)=\{v\in V:\ T(v)=\lambda v\}.
\]

## Equivalent characterizations

Equivalently,
\[
E_\lambda(T)=\{v\in V:\ (T-\lambda I)(v)=0\},
\]
where \(I\) is the identity operator on \(V\).

## Remarks

If \(\lambda\) is an [eigenvalue](https://knowlpedia.com/linear-algebra/eigenvalue/), then \(E_\lambda(T)\) contains nonzero [eigenvectors](https://knowlpedia.com/linear-algebra/eigenvector/); otherwise it is \(\{0\}\). In all cases, \(E_\lambda(T)\) is a [vector space](https://knowlpedia.com/linear-algebra/vector-space/) under the operations inherited from \(V\).

## Examples

- For the identity operator \(I\) on \(V\), the eigenspace for \(\lambda=1\) is all of \(V\).
- For the projection \(P(x,y)=(x,0)\) on \(\mathbb{R}^2\), the eigenspace for \(\lambda=1\) is \(\{(x,0):x\in\mathbb{R}\}\) and for \(\lambda=0\) is \(\{(0,y):y\in\mathbb{R}\}\).
- For a diagonal matrix \(\operatorname{diag}(d_1,\dots,d_n)\), the eigenspace for \(\lambda\) is spanned by those standard basis vectors whose diagonal entry equals \(\lambda\).
