Definition

For an inclusion kg\mathfrak k\subseteq\mathfrak g of finite-dimensional Lie algebras, tk\mathfrak t\subseteq\mathfrak k and hg\mathfrak h\subseteq\mathfrak g are compatible with the inclusion if

t=kh.\mathfrak t=\mathfrak k\cap\mathfrak h.

In the common maximal-rank case, one can take t=hk\mathfrak t=\mathfrak h\subseteq\mathfrak k, so the two algebras use the same Cartan subalgebra.

What compatibility supplies

Compatible Cartans let the of the subalgebra be compared directly with that of the ambient algebra. If k\mathfrak k is a reductive in g\mathfrak g and normalized by h\mathfrak h, then k\mathfrak k is assembled from a subspace of h\mathfrak h and selected g\mathfrak g-root spaces. Its roots can therefore be recorded as a of the ambient , subject to the relevant span and closure conditions.

For inclusions that are not of maximal rank, roots of g\mathfrak g restrict from h\mathfrak h^* to t\mathfrak t^*. Several ambient roots may have the same nonzero restriction, and some may restrict to zero. Compatibility alone does not make the smaller root system a literal subset of the larger one.

Scope and terminology

“Compatible” describes a choice made for a particular inclusion; it is not an additional intrinsic structure on either algebra. Authors sometimes use the term more loosely for nested Cartans th\mathfrak t\subseteq\mathfrak h. Stating the intersection equation removes that ambiguity. Existence and conjugacy statements depend on hypotheses such as reductivity, regularity, and the ground field, and should not be inferred from the definition alone.

References
  1. Eugene B. Dynkin, “Semisimple subalgebras of semisimple Lie algebras,” American Mathematical Society Translations, Series 2, vol. 6, 1957, pp. 111–244.
  2. Anthony W. Knapp, Lie Groups Beyond an Introduction, 2nd ed., Birkhäuser, 2002, Chapters IV–VI. Publisher record.