Let g\mathfrak g be a . Recall that a is a linear map D:g→gD:\mathfrak g\to\mathfrak g satisfying the Leibniz rule

D([x,y])=[D(x),y]+[x,D(y)].D([x,y])=[D(x),y]+[x,D(y)].

Definition (Inner derivation). For each x∈gx\in\mathfrak g, the map

adx:g→g,adx(y)=[x,y]\mathrm{ad}_x:\mathfrak g\to\mathfrak g,\qquad \mathrm{ad}_x(y)=[x,y]

is a derivation. A derivation is called inner if it equals adx\mathrm{ad}_x for some x∈gx\in\mathfrak g.

The assignment x↦adxx\mapsto \mathrm{ad}_x is the ad:g→gl(g)\mathrm{ad}:\mathfrak g\to\mathfrak{gl}(\mathfrak g), and the space of inner derivations is ad(g)⊆Der(g)\mathrm{ad}(\mathfrak g)\subseteq \mathrm{Der}(\mathfrak g).

Remarks

Key relation to the center. The kernel of ad\mathrm{ad} is exactly the (see ), so inner derivations detect noncentral directions.

Context. Derivations modulo inner derivations measure “outer” symmetries of g\mathfrak g (compare ), and exponentiating adx\mathrm{ad}_x is the infinitesimal source of many .