Almost Inner Derivations: Definitions and Rigidity

• Almost inner derivations are maps that satisfy a weakened innerness condition by ensuring that D(x) lies within the commutator [x, g] for every element x.
• They emerge across various settings—from finite-dimensional Lie algebras to C*-algebras and von Neumann algebras—often revealing striking rigidity properties.
• Rigidity results indicate that under many structural conditions, almost inner derivations either collapse to true inner derivations or yield explicitly computable outer components.

Almost inner derivations are derivations that satisfy an innerness condition weaker than the existence of a single global commutator implementer. The precise meaning depends on context. For finite-dimensional Lie algebras, an almost inner derivation is a derivation $D$ such that $D(x)\in [x,\mathfrak g]$ for all $x$, equivalently $D(x)=[a_x,x]$ with $a_x$ depending on $x$ (Burde et al., 2017). In associative and operator-algebra settings, the same phrase can refer to local innerness on pairs, as in inner $2$-local derivations, or to norm-limits of inner derivations, as in separable unital $C^*$-algebras (Ayupov et al., 2013, Lupini, 29 Aug 2025). A recurring theme is rigidity: in many natural classes, the almost-inner condition collapses to genuine innerness, while in others the quotient by inner derivations is small, explicitly computable, or even infinite-dimensional (Ber et al., 2013, Burde et al., 2019, Shen et al., 13 Jul 2025).

1. Definitions and principal formulations

The foundational Lie-algebraic definition used by Gordon–Wilson and developed systematically in later work is pointwise innerness: for a finite-dimensional Lie algebra $\mathfrak g$ over a field of characteristic zero, a derivation $D\in \operatorname{Der}(\mathfrak g)$ is almost inner if

$D(x)\in [x,\mathfrak g]$0

The space of such derivations is denoted $D(x)\in [x,\mathfrak g]$1, and one has

$D(x)\in [x,\mathfrak g]$2

with the quotient $D(x)\in [x,\mathfrak g]$3 measuring outer almost inner derivations (Burde et al., 2019). A closely related exposition emphasizes the equivalent formulation that for every $D(x)\in [x,\mathfrak g]$4 there exists $D(x)\in [x,\mathfrak g]$5 such that $D(x)\in [x,\mathfrak g]$6 (Burde et al., 2017).

Several adjacent notions appear in the literature. For associative rings, an inner $D(x)\in [x,\mathfrak g]$7-local derivation on a ring $D(x)\in [x,\mathfrak g]$8 is a map $D(x)\in [x,\mathfrak g]$9 such that for every pair $x$0 there exists $x$1 with

$x$2

This is a local two-point version of innerness rather than a single global commutator identity (Ayupov et al., 2013). For derivations from a von Neumann algebra $x$3 into a quasi-normed bimodule $x$4, “almost inner” can mean that the derivation is implemented by some $x$5 before one proves that the implementer can in fact be chosen in $x$6 itself (Ber et al., 2013). In separable unital $x$7-algebras, the phrase can denote membership in the norm-closure of inner derivations inside the Banach space of derivations; in that setting, almost inner derivations exist precisely when the subspace of inner derivations is not norm-closed (Lupini, 29 Aug 2025).

A useful summary is that the phrase “almost inner” is not tied to a single universal definition. It ranges from pointwise commutator conditions, to local-pair innerness, to ambient-algebra implementability, to norm-approximation by inner derivations. What unifies these usages is that innerness is required only after weakening either the domain of comparison, the topology, or the location of the implementer.

Context	Almost-inner condition	Typical outcome
Lie algebras	$x$8 for all $x$9	Often $D(x)=[a_x,x]$0, but not always (Burde et al., 2017)
Matrix rings	Inner $D(x)=[a_x,x]$1-locality on each pair $D(x)=[a_x,x]$2	Collapse to a single inner derivation under commutativity hypotheses (Ayupov et al., 2013)
Von Neumann bimodules	Implementer first found in $D(x)=[a_x,x]$3	Refined to an implementer in the target module $D(x)=[a_x,x]$4 (Ber et al., 2013)
Unital $D(x)=[a_x,x]$5-algebras	Norm-limit of inner derivations	Equivalent to non-closedness of inner derivations (Lupini, 29 Aug 2025)

2. Associative, operator-algebraic, and $D(x)=[a_x,x]$6-algebraic rigidity

A strong rigidity phenomenon appears for matrix rings over commutative associative rings. If $D(x)=[a_x,x]$7 is a commutative associative ring with identity and $D(x)=[a_x,x]$8, every inner $D(x)=[a_x,x]$9-local derivation $a_x$0 is an inner derivation:

$a_x$1

for a single $a_x$2 (Ayupov et al., 2013). The proof uses matrix units $a_x$3, corner identities such as

$a_x$4

a decomposition of $a_x$5, and a normalization by the cyclic banded element $a_x$6 to pin down diagonal differences of local implementers (Ayupov et al., 2013). The same paper also proves that every derivation $a_x$7 extends entrywise to a derivation $a_x$8 by

$a_x$9

with $x$0 (Ayupov et al., 2013). This situates matrix rings as a setting where local innerness collapses to global innerness, but ordinary derivations may still contain both inner and entrywise components.

For von Neumann algebras, the rigidity is stronger. If $x$1 is a von Neumann algebra and $x$2 is a quasi-normed $x$3-bimodule of locally measurable operators, then every derivation $x$4 is inner: there exists $x$5 such that

$x$6

and moreover

$x$7

with the sharper bound $x$8 in the self-adjoint or skew-adjoint cases (Ber et al., 2013). The proof proceeds by automatic continuity, extension to $x$9, $2$0-continuity there, innerness on $2$1, and then a commutator inequality forcing the implementer into $2$2 rather than merely into $2$3 (Ber et al., 2013). This means that “inner in a larger ambient algebra” is not a genuinely weaker notion under the hypotheses of that theorem.

A related semifinite result treats derivations $2$4, where $2$5 is a symmetric operator space of $2$6-compact operators. Under mild assumptions—namely that $2$7 is the dual of a symmetric space $2$8 and either separability or reflexivity holds—every continuous derivation $2$9 is inner with implementer in $C^*$0 and

$C^*$1

(Ber et al., 2012). In particular, every derivation $C^*$2 is inner for $C^*$3 (Ber et al., 2012). Here again, almost-inner behavior in an ambient measurable-operator sense collapses to genuine innerness.

For separable unital $C^*$4-algebras, the perspective changes from algebraic localness to topological closure. A derivation is “almost inner” when it belongs to the norm-closure of the inner derivations $C^*$5 (Lupini, 29 Aug 2025). The main equivalence is that the subspace of inner derivations is norm-closed if and only if $C^*$6 is norm-closed in $C^*$7, and this is further equivalent to a descriptive-set-theoretic condition:

$C^*$8

(Lupini, 29 Aug 2025). Thus almost inner derivations, in this norm-closure sense, exist precisely when inner derivations fail to be norm-closed (Lupini, 29 Aug 2025). This suggests a conceptual bifurcation: in von Neumann and measurable-operator contexts almost innerness tends to collapse algebraically, whereas in $C^*$9-algebras it can survive as a genuine topological closure phenomenon.

3. Lie algebras: pointwise innerness, rigidity, and counterexamples

The Lie-algebraic theory begins with the observation that $\mathfrak g$0 is itself a Lie subalgebra of $\mathfrak g$1, that $\mathfrak g$2, and that for nilpotent $\mathfrak g$3, every almost inner derivation is a nilpotent endomorphism and $\mathfrak g$4 is a nilpotent Lie algebra (Burde et al., 2017). The same work introduced the method of fixed basis vectors: for $\mathfrak g$5, one chooses $\mathfrak g$6 with $\mathfrak g$7, and sufficient coordinate equalities imply that each basis vector is “fixed”; once every basis vector is fixed, $\mathfrak g$8 must be inner (Burde et al., 2017). This method yields $\mathfrak g$9 for broad classes, including $D\in \operatorname{Der}(\mathfrak g)$0-step nilpotent Lie algebras determined by graphs, free $D\in \operatorname{Der}(\mathfrak g)$1-step and $D\in \operatorname{Der}(\mathfrak g)$2-step nilpotent Lie algebras, free metabelian nilpotent Lie algebras on two generators over an infinite field, almost abelian Lie algebras over $D\in \operatorname{Der}(\mathfrak g)$3, and triangular Lie algebras (Burde et al., 2017).

A later paper extends this rigidity. Over a field of characteristic zero, one has

$D\in \operatorname{Der}(\mathfrak g)$4

for every free $D\in \operatorname{Der}(\mathfrak g)$5-step nilpotent Lie algebra $D\in \operatorname{Der}(\mathfrak g)$6 (Burde et al., 2019). If $D\in \operatorname{Der}(\mathfrak g)$7 has a codimension-one abelian ideal, then again $D\in \operatorname{Der}(\mathfrak g)$8 (Burde et al., 2019). More generally, if $D\in \operatorname{Der}(\mathfrak g)$9 over an algebraically closed field of characteristic zero, with $D(x)\in [x,\mathfrak g]$00 the abelian solvable radical and $D(x)\in [x,\mathfrak g]$01 semisimple, then all almost inner derivations are inner (Burde et al., 2019). These results identify several structurally rigid settings in which the pointwise condition $D(x)\in [x,\mathfrak g]$02 already forces a global commutator.

Filiform nilpotent Lie algebras show that rigidity is not universal. The standard filiform algebra $D(x)\in [x,\mathfrak g]$03 satisfies $D(x)\in [x,\mathfrak g]$04, but for the family $D(x)\in [x,\mathfrak g]$05 one has

$D(x)\in [x,\mathfrak g]$06

so there is exactly one outer almost inner derivation (Burde et al., 2019). For $D(x)\in [x,\mathfrak g]$07, the almost inner space is larger:

$D(x)\in [x,\mathfrak g]$08

with explicit witnesses $D(x)\in [x,\mathfrak g]$09 satisfying $D(x)\in [x,\mathfrak g]$10 (Burde et al., 2019). Most strikingly, the characteristically nilpotent filiform family $D(x)\in [x,\mathfrak g]$11 for $D(x)\in [x,\mathfrak g]$12 satisfies

$D(x)\in [x,\mathfrak g]$13

and $D(x)\in [x,\mathfrak g]$14 has dimension $D(x)\in [x,\mathfrak g]$15 (Burde et al., 2019). This is presented there as the first known family of nilpotent Lie algebras with all derivations almost inner (Burde et al., 2019).

Low-dimensional and family-based examples show that the quotient $D(x)\in [x,\mathfrak g]$16 can be arbitrarily large. In dimension $D(x)\in [x,\mathfrak g]$17, examples such as $D(x)\in [x,\mathfrak g]$18 and $D(x)\in [x,\mathfrak g]$19 have

$D(x)\in [x,\mathfrak g]$20

or

$D(x)\in [x,\mathfrak g]$21

(Burde et al., 2017). More dramatically, there exist $D(x)\in [x,\mathfrak g]$22-step nilpotent Lie algebras $D(x)\in [x,\mathfrak g]$23 of dimension $D(x)\in [x,\mathfrak g]$24 such that

$D(x)\in [x,\mathfrak g]$25

(Burde et al., 2017). This shows that even within nilpotent Lie algebras, the failure of innerness can be systematically large rather than sporadic.

4. Matrix pencils, algorithms, and explicit determination of $D(x)\in [x,\mathfrak g]$26

For $D(x)\in [x,\mathfrak g]$27-step nilpotent Lie algebras of genus $D(x)\in [x,\mathfrak g]$28, the theory becomes highly explicit through matrix pencils. Writing $D(x)\in [x,\mathfrak g]$29 with $D(x)\in [x,\mathfrak g]$30 and $D(x)\in [x,\mathfrak g]$31 abelian, the bracket is encoded by skew-symmetric matrices $D(x)\in [x,\mathfrak g]$32 via

$D(x)\in [x,\mathfrak g]$33

and the associated pencil is

$D(x)\in [x,\mathfrak g]$34

The canonical form of this skew pencil is described by minimal indices and elementary divisors, and these invariants determine the dimension of $D(x)\in [x,\mathfrak g]$35 (Burde et al., 2020). Over an algebraically closed field of characteristic not $D(x)\in [x,\mathfrak g]$36,

$D(x)\in [x,\mathfrak g]$37

while over $D(x)\in [x,\mathfrak g]$38 one adds the contribution

$D(x)\in [x,\mathfrak g]$39

from complex-conjugate quadratic blocks (Burde et al., 2020). In particular, over $D(x)\in [x,\mathfrak g]$40, a single real complex-conjugate block gives $D(x)\in [x,\mathfrak g]$41 blockwise (Burde et al., 2020).

A complementary computational approach starts directly from a structure-constant table. For a finite-dimensional Lie algebra $D(x)\in [x,\mathfrak g]$42 with basis $D(x)\in [x,\mathfrak g]$43 and structure constants $D(x)\in [x,\mathfrak g]$44, one first solves the linear derivation equations

$D(x)\in [x,\mathfrak g]$45

to obtain $D(x)\in [x,\mathfrak g]$46, then computes $D(x)\in [x,\mathfrak g]$47 from the matrices of $D(x)\in [x,\mathfrak g]$48 (Dietrich et al., 2024). The key test for almost innerness uses the matrix $D(x)\in [x,\mathfrak g]$49 defined by

$D(x)\in [x,\mathfrak g]$50

and the vector $D(x)\in [x,\mathfrak g]$51 corresponding to $D(x)\in [x,\mathfrak g]$52 for $D(x)\in [x,\mathfrak g]$53. Then

$D(x)\in [x,\mathfrak g]$54

(Dietrich et al., 2024). Over an algebraically closed field this becomes an ideal-membership condition in the polynomial ring generated by minors of $D(x)\in [x,\mathfrak g]$55 and of the augmented matrix (Dietrich et al., 2024).

This computational framework also answers a structural question about the quotient $D(x)\in [x,\mathfrak g]$56. For a specific $D(x)\in [x,\mathfrak g]$57-dimensional Lie algebra $D(x)\in [x,\mathfrak g]$58 in characteristic $D(x)\in [x,\mathfrak g]$59, one has

$D(x)\in [x,\mathfrak g]$60

and

$D(x)\in [x,\mathfrak g]$61

for a $D(x)\in [x,\mathfrak g]$62-dimensional subalgebra $D(x)\in [x,\mathfrak g]$63 (Dietrich et al., 2024). Two explicit derivations $D(x)\in [x,\mathfrak g]$64 have nonzero commutator modulo inner derivations, proving that

$D(x)\in [x,\mathfrak g]$65

can be non-abelian (Dietrich et al., 2024). This answers a question of Kunyavskii and Ostapenko in the affirmative.

5. Extensions to Leibniz algebras, Lie superalgebras, and generalized derivation theories

For Leibniz algebras, the right Leibniz identity changes the role of inner derivations: right multiplications

$D(x)\in [x,\mathfrak g]$66

are derivations and constitute the inner derivations in the convention of the cited work (Adashev et al., 2020). An almost inner derivation is then a derivation $D(x)\in [x,\mathfrak g]$67 such that for every $D(x)\in [x,\mathfrak g]$68 there exists $D(x)\in [x,\mathfrak g]$69 with

$D(x)\in [x,\mathfrak g]$70

In addition to $D(x)\in [x,\mathfrak g]$71, the paper introduces right central almost inner derivations and central almost inner derivations, denoted $D(x)\in [x,\mathfrak g]$72 and $D(x)\in [x,\mathfrak g]$73, with

$D(x)\in [x,\mathfrak g]$74

(Adashev et al., 2020). For null-filiform Leibniz algebras, $D(x)\in [x,\mathfrak g]$75, but in several filiform families the quotient acquires a one-dimensional extension generated by the map $D(x)\in [x,\mathfrak g]$76:

$D(x)\in [x,\mathfrak g]$77

under explicit parameter conditions (Adashev et al., 2020). A related paper on two-step nilpotent Leibniz algebras with one-dimensional commutator ideal proves that every almost inner derivation is inner, with exactly three exceptional families: the Heisenberg Leibniz algebras $D(x)\in [x,\mathfrak g]$78 and the Dieudonné algebras $D(x)\in [x,\mathfrak g]$79 (Chen et al., 2023).

For Lie superalgebras, the super-analogue of pointwise innerness is defined by requiring $D(x)\in [x,\mathfrak g]$80 for every homogeneous $D(x)\in [x,\mathfrak g]$81 (Serganova et al., 31 Aug 2025). Although finite-dimensional simple Lie superalgebras may admit outer derivations, the main theorem states that for every finite-dimensional simple Lie superalgebra $D(x)\in [x,\mathfrak g]$82 over $D(x)\in [x,\mathfrak g]$83,

$D(x)\in [x,\mathfrak g]$84

The proof shows that the known outer derivations—Euler derivations, odd derivations for $D(x)\in [x,\mathfrak g]$85, and the outer $D(x)\in [x,\mathfrak g]$86 for $D(x)\in [x,\mathfrak g]$87—are not almost inner (Serganova et al., 31 Aug 2025). At the same time, naturally occurring non-inner almost inner derivations do exist for quasireductive Lie superalgebras built from prehomogeneous vector spaces; there the Euler derivation can be almost inner whenever all orbits are conical, while perfectness of the even part prevents innerness (Serganova et al., 31 Aug 2025). This identifies a super-analogue of the gap between simple rigidity and non-semisimple flexibility.

A different generalization replaces the Leibniz rule by a $D(x)\in [x,\mathfrak g]$88-twisted one. A $D(x)\in [x,\mathfrak g]$89-derivation $D(x)\in [x,\mathfrak g]$90 satisfies

$D(x)\in [x,\mathfrak g]$91

For semisimple finite-dimensional Lie algebras over an algebraically closed field of characteristic $D(x)\in [x,\mathfrak g]$92, the classification is extremely rigid: $D(x)\in [x,\mathfrak g]$93 only for $D(x)\in [x,\mathfrak g]$94, $D(x)\in [x,\mathfrak g]$95, or $D(x)\in [x,\mathfrak g]$96, with the non-inner cases confined to scalar maps on adjoint modules for $D(x)\in [x,\mathfrak g]$97 and explicit $D(x)\in [x,\mathfrak g]$98 exceptions (Zohrabi et al., 2022). This is not the same notion as almost inner derivation, but it provides a parallel instance where derivations are “almost” ordinary only in a sharply delimited sense.

6. Cohomological and affine phenomena

A cohomological analogue of almost innerness appears for triple derivations on von Neumann algebras. Every triple derivation $D(x)\in [x,\mathfrak g]$99 is inner as a triple derivation, but for triple derivations $x$00 the picture changes (Pluta et al., 2013). For a factor $x$01, the quotient

$x$02

has dimension $x$03 if and only if $x$04 is finite, and dimension $x$05 if $x$06 is properly infinite (Pluta et al., 2013). In a properly infinite factor, every triple derivation decomposes as

$x$07

with $x$08 in the norm-closure of inner triple derivations and $x$09 representing the unique obstruction (Pluta et al., 2013). This is an “almost inner” phenomenon in the precise sense that innerness fails only by a one-dimensional quotient.

Loop algebras and affinizations furnish another sharply classified setting. Let $x$10 be a minimal $x$11-graded subalgebra of a semisimple Lie algebra over an algebraically closed field of characteristic $x$12. Its derived algebra

$x$13

is abelian, and every almost inner derivation of $x$14 is inner:

$x$15

(Shen et al., 13 Jul 2025). If $x$16, then under $x$17 the loop algebra satisfies

$x$18

and every almost inner derivation of $x$19 is inner (Shen et al., 13 Jul 2025). The affinization

$x$20

behaves differently: after modding out by inner derivations, almost inner derivations are exactly those vanishing on $x$21 and $x$22 and sending $x$23 into $x$24 (Shen et al., 13 Jul 2025). The quotient

$x$25

is spanned by explicit derivations $x$26, indexed by $x$27 and $x$28, and this infinite set is linearly independent (Shen et al., 13 Jul 2025). Thus a loop algebra can be rigid while its central extension acquires an infinite-dimensional almost-inner quotient.

From a broader perspective, these examples show that “almost inner” phenomena organize themselves around cohomological obstructions. Sometimes the obstruction vanishes completely, as in simple Lie superalgebras, matrix rings over commutative rings, and derivations into quasi-normed bimodules (Ayupov et al., 2013, Ber et al., 2013, Serganova et al., 31 Aug 2025). Sometimes it is finite-dimensional and explicit, as in triple derivations into the predual of a factor or filiform Lie algebras (Pluta et al., 2013, Burde et al., 2019). Sometimes it is genuinely infinite-dimensional, as for affinizations of minimal $x$29-graded subalgebras (Shen et al., 13 Jul 2025). This suggests that the topic is best understood not as a single theorem but as a family of rigidity and deformation problems centered on how far pointwise, local, ambient, or approximate innerness can deviate from actual innerness.