Covering Extended Affine Lie Algebras

• Covering extended affine Lie algebras are structures defined by extension phenomena from the centerless core to the full EALA through central and derivational augmentations.
• The framework employs Lie tori, universal central extensions, and multiloop algebra techniques to reconstruct and classify EALAs with precise invariant properties.
• Lifting automorphisms and integrating group- and module-theoretic coverings address the intrinsic hierarchy and invariance within the extended affine Lie algebra framework.

Covering extended affine Lie algebras refers, in the standard structural language of EALA theory, to several closely related extension phenomena: the passage from the centerless core to the core by central extension, the reconstruction of the full EALA by adjoining derivations and cocycle data, and, in more specialized settings, universal central extensions of Lie tori, multiloop and torsor realizations of centerless cores, local affine coverings by root subsystems, and Steinberg-type coverings on the group side (Neher, 2010). The subject is therefore not governed by a single universal notion of covering; rather, it is organized by the canonical hierarchy

$E_{cc}\;\longleftarrow\;E_c\;\longrightarrow\;E,$

together with the realization of $E_{cc}$ as a centerless Lie torus and the analysis of how automorphisms, integral structures, modules, and group constructions lift across these successive extensions (Azam et al., 2023).

1. Core, centerless core, and the canonical extension ladder

An extended affine Lie algebra is a triple $(E,(\cdot,\cdot),H)$ satisfying the usual axioms, with root system $R=R^\times\cup R^0$, where $R^\times$ is the set of nonisotropic roots and $R^0$ the isotropic roots. Its core is the subalgebra generated by the nonisotropic root spaces,

$$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$

and the centerless core is

$E_{cc}:=E_c/Z(E_c).$

A central structural fact is that $E_{cc}$ is precisely a centerless Lie torus; conversely, starting from a centerless Lie torus $\mathcal L$, one reconstructs an EALA by adjoining a central part and a derivation part (Azam et al., 2023).

In Neher’s realization, one starts with a centerless Lie torus $E_{cc}$0 carrying an invariant nondegenerate graded form $E_{cc}$1, chooses a permissible graded subalgebra

$E_{cc}$2

of skew centroidal derivations, and an affine cocycle

$E_{cc}$3

The resulting EALA is built on

$E_{cc}$4

with bracket

$E_{cc}$5

where

$E_{cc}$6

Its invariant form is

$E_{cc}$7

and its Cartan subalgebra is

$E_{cc}$8

Crucially,

$E_{cc}$9

so the core is literally a central extension of the centerless core $(E,(\cdot,\cdot),H)$0, while the full EALA is obtained from the core by adjoining $(E,(\cdot,\cdot),H)$1 (Azam et al., 2023).

Neher’s lectures formulate the same picture in the language of coverings and central extensions. An extension of a Lie algebra $(E,(\cdot,\cdot),H)$2 is a surjective homomorphism $(E,(\cdot,\cdot),H)$3; it is a central extension if $(E,(\cdot,\cdot),H)$4; it is a covering if it is a central extension with $(E,(\cdot,\cdot),H)$5 perfect; and a universal central extension exists precisely for perfect Lie algebras (Neher, 2010). In EALA theory, the most immediate covering object is therefore not the full algebra $(E,(\cdot,\cdot),H)$6, but the central extension $(E,(\cdot,\cdot),H)$7.

2. Lie tori, universal central extensions, and reconstruction

Lie tori are the structural base objects for EALAs. Neher’s lectures emphasize that the centerless core of every EALA is an invariant centreless Lie torus, and that every EALA is reconstructed from such a Lie torus by taking a suitable central extension and adjoining derivations with an affine cocycle (Neher, 2010). In this framework, covering theory is concentrated at the level of the Lie torus and its central extensions.

A basic fact is that Lie tori are perfect, hence admit universal central extensions. Moreover, if

$(E,(\cdot,\cdot),H)$8

is the universal central extension of a Lie torus $(E,(\cdot,\cdot),H)$9, then $R=R^\times\cup R^0$0 is again a Lie torus of the same type (Neher, 2010). This is the abstract form of the statement that the class of centerless-core models relevant to EALAs is stable under universal central covering.

The affine and toroidal examples provide the standard prototypes. For affine Kac–Moody theory, the loop algebra $R=R^\times\cup R^0$1 is the centerless core, its canonical central extension $R=R^\times\cup R^0$2 is the core, and the affine algebra $R=R^\times\cup R^0$3 is obtained by adjoining the degree derivation (Neher, 2010). For untwisted multiloop algebras

$R=R^\times\cup R^0$4

the universal central extension is larger than the one-variable affine analogue; Neher writes it using a graded center

$R=R^\times\cup R^0$5

and a universal cocycle

$R=R^\times\cup R^0$6

The higher-nullity EALA then has the form

$R=R^\times\cup R^0$7

with $R=R^\times\cup R^0$8 spanned by degree derivations (Neher, 2010).

This central-extension picture can also be formulated torsorially. Gille and Pianzola treat multiloop algebras as $R=R^\times\cup R^0$9-forms of split simple Lie algebras over Laurent polynomial rings,

$R^\times$0

split by finite étale extensions $R^\times$1. They are classified by torsors under $R^\times$2, and a torsor is a loop torsor if its cohomology class lies in the image of

$R^\times$3

For reductive groups over $R^\times$4, loop torsors coincide with toral torsors: $R^\times$5 and a reductive $R^\times$6-group is loop reductive iff it admits a maximal torus (Gille et al., 2011). This identifies multiloop centerless cores as the Lie algebras of loop reductive group schemes, thereby attaching a geometric classification to the covering layer beneath EALAs.

3. Multiloop algebras, nullity $R^\times$7, and the classification of centerless cores

The paper on multiloop algebras and nullity $R^\times$8 organizes the subject around three classes: $R^\times$9 where $R^0$0 denotes multiloop algebras of finite-dimensional simple Lie algebras, $R^0$1 iterated loop algebras, and $R^0$2 Lie algebras isomorphic to centerless cores of EALAs of nullity $R^0$3 (Allison et al., 2010).

Class	Meaning	Nullity $R^0$4 role
$R^0$5	multiloop algebras	first-kind affine loop algebras
$R^0$6	iterated loop algebras	larger than $R^0$7
$R^0$8	centerless cores of EALAs	centerless-core class

For $R^0$9 and $$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$0, the classes coincide: $$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$1 For $$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$2, they separate, and the paper gives a complete description. The main theorem states that for a Lie algebra $$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$3, the following are equivalent: $$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$4; $$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$5 with $$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$6 untwisted affine and $$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$7 a diagram automorphism; $$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$8 with centroid

$$E_c=\langle E_\alpha\mid \alpha\in R^\times\rangle,$$9

Thus, in nullity $E_{cc}:=E_c/Z(E_c).$0, multiloop algebras are exactly the iterated loop algebras of affine type with Laurent polynomial centroid (Allison et al., 2010).

The intersection with the EALA-core class is sharply characterized: $E_{cc}:=E_c/Z(E_c).$1 Equivalently, an algebra in $E_{cc}:=E_c/Z(E_c).$2 is the centerless core of a nullity-$E_{cc}:=E_c/Z(E_c).$3 EALA iff it is isotropic, or equivalently iff it comes from a nontransitive affine diagram automorphism (Allison et al., 2010). The anisotropic part

$E_{cc}:=E_c/Z(E_c).$4

consists of algebras such as

$E_{cc}:=E_c/Z(E_c).$5

while

$E_{cc}:=E_c/Z(E_c).$6

consists of loop algebras defined by automorphisms of second kind (Allison et al., 2010).

The same paper gives an explicit affinization that recovers the full EALA above its centerless core. For an affine Lie algebra $E_{cc}:=E_c/Z(E_c).$7 and a nontransitive diagram automorphism $E_{cc}:=E_c/Z(E_c).$8,

$E_{cc}:=E_c/Z(E_c).$9

with bracket

$E_{cc}$0

If $E_{cc}$1 is nontransitive, this is a nullity-$E_{cc}$2 EALA whose centerless core is

$E_{cc}$3

This is the most explicit nullity-$E_{cc}$4 realization of the extension

$E_{cc}$5

in the literature summarized here (Allison et al., 2010).

4. Lifting automorphisms and involutions through the extension hierarchy

A central structural problem is whether automorphisms or involutions of the centerless core can be lifted through the central extension $E_{cc}$6 and then through the derivation extension $E_{cc}$7. This is treated explicitly for Chevalley involutions by the paper on Lie tori and EALAs (Azam et al., 2023).

For a $E_{cc}$8-graded algebra $E_{cc}$9, a pre-Chevalley involution is an involution $\mathcal L$0 such that

$\mathcal L$1

For a centerless Lie torus $\mathcal L$2, it is a Chevalley involution if it also satisfies

$\mathcal L$3

The main existence theorem states that every centerless Lie torus of reduced type admits a Chevalley involution, with the additional assumption in type $\mathcal L$4, $\mathcal L$5, that its coordinate algebra is equipped with a $\mathcal L$6-grading anti-involution (Azam et al., 2023).

The lifting mechanism begins by conjugation on endomorphisms,

$\mathcal L$7

and dually on $\mathcal L$8. If $\mathcal L$9 is the degree derivation attached to $E_{cc}$00, then

$E_{cc}$01

More generally,

$E_{cc}$02

These formulas determine how the derivation and central pieces transform under the involution (Azam et al., 2023).

Given permissible $E_{cc}$03, the transformed derivation algebra and cocycle are

$E_{cc}$04

Then

$E_{cc}$05

is an isomorphism. If $E_{cc}$06 is $E_{cc}$07-invariant, then $E_{cc}$08 restricts to a pre-Chevalley involution of the core

$E_{cc}$09

If $E_{cc}$10 is $E_{cc}$11-invariant, then $E_{cc}$12 becomes a pre-Chevalley involution of the full EALA. If $E_{cc}$13 is already a Chevalley involution of $E_{cc}$14, the lifted involution on $E_{cc}$15 or $E_{cc}$16 is again a Chevalley involution (Azam et al., 2023).

The criterion for $E_{cc}$17-invariance is concrete. Writing

$E_{cc}$18

one has

$E_{cc}$19

In particular, the standard choices

$E_{cc}$20

are always $E_{cc}$21-invariant, so the core inherits a Chevalley involution in these common realizations (Azam et al., 2023).

This lifting viewpoint persists in recent work on integral structures. The paper on integral structures for reduced tame EALAs states that the essential structural features are encoded in the centerless core $E_{cc}$22, and one then promotes the results from $E_{cc}$23 to $E_{cc}$24 and then to $E_{cc}$25 (Azam et al., 2021). The rank-one elliptic paper makes the same hierarchy completely explicit for type $E_{cc}$26, nullity $E_{cc}$27: if $E_{cc}$28, then

$E_{cc}$29

so the centerless Lie torus, the core, and the full EALA are visibly separated (Azam, 14 Jul 2025).

5. Local, presentation-theoretic, group-theoretic, and module-theoretic variants

A different covering-like mechanism is localization by root subsystems. For a closed subsystem $E_{cc}$30, a subspace $E_{cc}$31 is called a cover if it contains $E_{cc}$32 for all $E_{cc}$33 and the form restricted to $E_{cc}$34 is nondegenerate. The associated Lie cover is

$E_{cc}$35

It is again an EALA with Cartan $E_{cc}$36 and root system $E_{cc}$37. If $E_{cc}$38 is affine, then the refined localization

$E_{cc}$39

is an affine Lie algebra. The same paper develops a filtration

$E_{cc}$40

by EALAs of increasing nullity, with $E_{cc}$41 affine, so that an EALA is approximated by finite and affine local pieces rather than by a single global cover (Azam, 9 Oct 2025).

Another covering-like construction is presentation-theoretic. For a connected non-negative unit form $E_{cc}$42, the paper on $E_{cc}$43 constructs an EALA with root system

$E_{cc}$44

and shows that $E_{cc}$45 is obtained from the generalized-Serre algebra $E_{cc}$46 by quotienting out the unique maximal ideal intersecting the Cartan trivially: $E_{cc}$47 This is not formulated as a universal central covering, but it is a canonical presentation-to-EALA quotient closely analogous to the Serre presentation of affine Kac–Moody algebras (Jasso, 2012).

On the group side, the paper on groups of extended affine Lie type introduces an $E_{cc}$48-covering of a reduced extended affine root system $E_{cc}$49: $E_{cc}$50 The Steinberg group associated to $E_{cc}$51 is called the universal Steinberg group of $E_{cc}$52, and there is a right-split exact sequence

$E_{cc}$53

hence

$E_{cc}$54

The same paper shows that the extended affine Weyl group is recovered as

$E_{cc}$55

These are covering constructions for groups attached to EALAs rather than for the Lie algebras themselves (Azam et al., 2019).

A fourth usage of covering appears in representation theory. In the paper on irreducible modules for EALAs, a thin covering of a module $E_{cc}$56 over a graded Lie algebra $E_{cc}$57 is a family $E_{cc}$58 such that

$E_{cc}$59

minimal under termwise inclusion. Thin coverings are then used to construct irreducible modules for twisted toroidal Lie algebras and the corresponding EALAs. This usage is module-theoretic, not Lie-algebraic: it concerns coverings of modules, not coverings of EALAs by central extensions (Billig et al., 2010).

6. Intrinsic invariants, conjugacy, and the limits of the covering viewpoint

A foundational issue for any covering theory of EALAs is whether the core, the centerless core, and the associated root data are intrinsic to the Lie algebra or depend on the chosen EALA structure. For fgc EALAs, the conjugacy theorem resolves this strongly. If $E_{cc}$60 and $E_{cc}$61 are two EALA structures on the same Lie algebra and the centerless core $E_{cc}$62 is fgc, then there exists

$E_{cc}$63

such that

$E_{cc}$64

The paper first proves that the core is independent of the EALA structure: if $E_{cc}$65 and $E_{cc}$66 are the cores attached to two EALA structures on $E_{cc}$67, then

$E_{cc}$68

Moreover, the core is characteristic: $E_{cc}$69 Thus, for fgc EALAs, the root system and quotient root system are intrinsic invariants, and the covering ladder

$E_{cc}$70

is attached to the Lie algebra itself, not merely to an auxiliary Cartan choice (Chernousov et al., 2016).

This intrinsic viewpoint also clarifies the limits of the terminology. Multiloop algebras are generally not EALAs themselves; in nullity $E_{cc}$71, isotropic algebras in $E_{cc}$72 are centerless cores of EALAs, while the full EALA sits above them as a central extension with derivations (Allison et al., 2010). Likewise, Neher’s lectures stress that one should not identify an EALA with the universal central cover of its centerless core: the general construction allows many suitable central extensions arising from graded subalgebras of skew-centroidal derivations, not only the universal one (Neher, 2010).

The literature also leaves clear open ends. The conjugacy theorem is proved only for fgc EALAs, and the non-fgc case remains unresolved (Chernousov et al., 2016). On the group side, integral-structure methods currently yield the adjoint form over arbitrary fields, while the universal-type group is explicitly left open (Azam et al., 2021). This suggests that covering questions in EALA theory are best understood as a family of related extension problems—central, derivational, local, and group-theoretic—organized around the intrinsic ideal $E_{cc}$73 and its quotient $E_{cc}$74, rather than as a single universal-covering theory valid uniformly across all extended affine Lie algebras.