Preprint
Article

This version is not peer-reviewed.

Stability of Functional Inequalities Associated with \( S(A) \)-Extended Jordan \( * \)-Derivations in \( JB^* \)-Algebras

Submitted:

27 October 2026

Posted:

28 October 2026

You are already at the latest version

Abstract
Let A be a unital JB*-algebra with unit 1, Jordan product ∘ and involution \( * \). Since the commutator formed from the Jordan product vanishes, we measure deviations by the Jordan associator {x, w} := (x ∘ w) ∘ w − x ∘ (w ∘ w). Let S(A) consist of those z ∈ A for which {z, a} = 0 for every a ∈ A. A mapping δ : A→ A is called an S(A)-extended Jordan \( * \)-derivation if additivity, the Jordan derivation identity and the involution condition hold modulo S(A). We assume that A ≠ C1, S(A) = C1, and the associator controls the distance to S(A); these conditions hold for Mn(C), n ≥ 2. Under dyadic summability and decay assumptions on the errors in a three-variable functional inequality and in the product and involution conditions, we obtain a unique S(A)-extended Jordan \( * \)-derivation L(x) = limn→∞2nδ(x/2n); an analogous result holds for limn→∞2−nδ(2nx). These limits exist in norm, whereas the quantitative approximation estimates control the associator error and the distance of L(x) − δ(x) to C1. The limit is additive and exactly \( * \)-preserving. An additional one-variable decay condition yields δ = L (superstability), and a unimodular version of the functional inequality yields C-linearity. Matrix examples demonstrate a nonzero limit and the distinction between associator control and full norm control.
Keywords: 
;  ;  ;  ;  ;  

MSC:  17C65; 39B72; 39B82; 46K70

1. Introduction

The stability problem for functional equations originates in a question posed by Ulam [1]: is every approximately additive map on a Banach space close to an exactly additive map? Hyers [2] answered affirmatively, initiating a vast literature; see [3,4,5] and the references therein. The related phenomenon of superstability, where every approximate solution is automatically an exact solution, was first observed by Baker, Lawrence and Zorzitto [6] for the exponential equation and later established for derivation-type maps by Šemrl [7] and Badora [8]; see also [9].
Centrally-extended derivations on rings were introduced by Bell and Daif [10], and their involutive analogues by El-Deken and Nabiel [11]. The Hyers–Ulam stability and superstability of centrally-extended derivations on Banach algebras and Banach *-algebras were studied in [12,13], where the functional inequality
∥ [ a 1 δ ( x 1 ) + a 2 δ ( x 2 ) + a 3 δ ( x 3 ) , w ] ∥ ≤ ∥ [ δ ( a 1 x 1 + a 2 x 2 + a 3 x 3 ) , w ] ∥ + Φ ( x 1 , x 2 , x 3 , w )
is analyzed by means of the commutator [ x , y ] = x y − y x .
The present paper extends this program to J B * -algebras. Since the Jordan product a ∘ b = b ∘ a is commutative, the commutator formed from that product cannot serve as a deviation measure. We replace it by the Jordan associator { x , w } : = ( x ∘ w ) ∘ w − x ∘ ( w ∘ w ) . For the Jordan product of a C * -algebra one has { z , a } = − 1 4 [ a , [ a , z ] ] (Lemma 3), so the associator is a second-order commutator, and its kernel S ( A ) contains the center; for the matrix algebras M n ( C ) it is exactly C 1 (Proposition 1).
Transferring the associative arguments to this setting requires three adjustments, which we now describe.
(a) The norm-control condition. A condition of the form ∥ x ∥ ≤ C sup ∥ a ∥ ≤ 1 ∥ { x , a } ∥ cannot hold on a unital algebra, since { 1 , a } = 0 for every a. In the spirit of Stampfli’s formula [15], the natural condition is to control the distance to S ( A ) :
dist ( x , S ( A ) ) ≤ C sup ∥ a ∥ ≤ 1 ∥ { x , a } ∥ .
This condition holds in every finite-dimensional J B * -algebra (Proposition 2).
(b) Recovering the scalar part. The associator does not see elements of S ( A ) , so the inequality above only gives convergence of the approximating sequence 2 n δ ( x / 2 n ) modulo S ( A ) . Under A ≠ C 1 and S ( A ) = C 1 , the space S ( A ) is not a Jordan ideal: otherwise x = x ∘ 1 would belong to S ( A ) for every x ∈ A . Thus the quotient A / S ( A ) is a Banach space but does not inherit the Jordan product, so convergence in the quotient alone is not enough to pass to the limit in the product condition. We show that the product inequality itself forces the scalar part of the sequence to converge (Lemma 5). This gives genuine norm convergence and makes the limit unique. The quantitative estimates nevertheless control the associator error and distance to C 1 , not the full norm of L ( x ) − δ ( x ) ; see Remark 1.
(c) The form of the derivation identity. In the convention of [14], an additive map on a ring with involution is called a *-derivation if it satisfies
d ( x y ) = d ( x ) y * + x d ( y ) .
Its literal Jordan analogue
δ ( a ∘ b ) = δ ( a ) ∘ b * + a ∘ δ ( b )
is not symmetric in a , b , whereas a ∘ b is; we show in Proposition 3 that it forces δ ( h ) ∈ S ( A ) for every self-adjoint h. We therefore use the Jordan derivation identity
δ ( a ∘ b ) = δ ( a ) ∘ b + a ∘ δ ( b )
together with the *-condition δ ( a * ) = δ ( a ) * , each modulo S ( A ) (Definition 3).
The paper is organized as follows. Section 2 collects the preliminaries, including the structural lemmas that drive the proofs. Section 3 contains the Hyers–Ulam stability theorems (Theorems 1 and 2), and Section 4 the superstability results (Theorems 3 and 4, Corollary 1). Section 5 gives matrix examples illustrating the hypotheses and the scope of the estimates, and Section 6 summarizes the results.

2. Preliminaries

2.1. J B * -Algebras and the Jordan Associator

A Jordan algebra over C is a commutative algebra ( A , ∘ ) satisfying the Jordan identity ( a ∘ a ) ∘ ( a ∘ b ) = a ∘ ( ( a ∘ a ) ∘ b ) . A J B * -algebra is a complex Jordan algebra A with a conjugate-linear involution * satisfying ( a ∘ b ) * = a * ∘ b * , which is a Banach space with ∥ a ∘ b ∥ ≤ ∥ a ∥ ∥ b ∥ and ∥ U a ( a * ) ∥ = ∥ a ∥ 3 , where U a ( b ) : = 2 a ∘ ( a ∘ b ) − ( a ∘ a ) ∘ b ; see [16,17,18]. Every C * -algebra is a J B * -algebra under a ∘ b : = 1 2 ( a b + b a ) . In a J B * -algebra one has ∥ a * ∥ = ∥ a ∥ , and if A is unital with unit 1 , then ∥ 1 ∥ = 1 . In this case we write C 1 : = { λ 1 : λ ∈ C } for the set of scalar multiples of the unit.
From now on, let A be a J B * -algebra with Jordan product ∘ and involution *. We write w 2 : = w ∘ w for all w ∈ A .
Definition 1.
For x , w ∈ A , the Jordan associator is { x , w } : = ( x ∘ w ) ∘ w − x ∘ w 2 , and
S ( A ) : = z ∈ A : { z , a } = 0 for all a ∈ A , ρ ( x ) : = sup ∥ a ∥ ≤ 1 ∥ { x , a } ∥ .
Lemma 1.
For all x , y , w ∈ A and α , β ∈ C :
(i)
{ α x + β y , w } = α { x , w } + β { y , w } ;
(ii)
∥ { x , w } ∥ ≤ 2 ∥ x ∥ ∥ w ∥ 2 and ρ ( x ) ≤ 2 ∥ x ∥ ; in particular
| ∥ { x , w } ∥ − ∥ { y , w } ∥ | ≤ ∥ { x − y , w } ∥ ≤ 2 ∥ x − y ∥ ∥ w ∥ 2 ;
(iii)
{ x , w } * = { x * , w * } ;
(iv)
{ x , α w } = α 2 { x , w } .
Proof. (i) follows from the C -bilinearity of ∘. (ii) follows from ∥ ( x ∘ w ) ∘ w ∥ ≤ ∥ x ∥ ∥ w ∥ 2 , ∥ x ∘ w 2 ∥ ≤ ∥ x ∥ ∥ w ∥ 2 and (i). (iii) follows from ( a ∘ b ) * = a * ∘ b * . (iv) follows from the bilinearity of ∘. □
Lemma 2.
S ( A ) is a closed, *-invariant, complex linear subspace of A , and ρ ( x + s ) = ρ ( x ) for all x ∈ A , s ∈ S ( A ) . If A is unital, then C 1 ⊆ S ( A ) .
Proof. 
Linearity and closedness follow from Lemma 1(i),(ii). If z ∈ S ( A ) , then by Lemma 1(iii)
{ z * , a } = { z , a * } * = 0 .
The identity ρ ( x + s ) = ρ ( x ) follows from Lemma 1(i). Finally, { 1 , a } = a ∘ a − a 2 = 0 . □
Thus ρ is a continuous seminorm on A whose kernel is exactly S ( A ) .
Lemma 3.
Let A be a C * -algebra with the Jordan product a ∘ b = 1 2 ( a b + b a ) . Then { z , a } = − 1 4 [ a , [ a , z ] ] for all a , z ∈ A . Consequently, the center of A is contained in S ( A ) .
Proof. 
We have
( z ∘ a ) ∘ a = 1 4 ( z a 2 + 2 a z a + a 2 z ) , z ∘ a 2 = 1 2 ( z a 2 + a 2 z ) ,
hence { z , a } = 1 4 ( 2 a z a − z a 2 − a 2 z ) = − 1 4 [ a , [ a , z ] ] for all a , z ∈ A . □
Proposition 1.
Let n ≥ 2 and let M n ( C ) carry the operator norm and the Jordan product a ∘ b = 1 2 ( a b + b a ) . Then S ( M n ( C ) ) = C 1 .
Proof. 
Let z ∈ S ( M n ( C ) ) and let p be an orthogonal projection. By Lemma 3, we have
0 = [ p , [ p , z ] ] = p z + z p − 2 p z p .
Multiplying by p on the left gives p z = p z p , and on the right gives z p = p z p ; hence p z = z p . Since every matrix is a linear combination of orthogonal projections (spectral theorem applied to the real and imaginary parts), z commutes with all of M n ( C ) , so z ∈ C 1 . The reverse inclusion is Lemma 2. □

2.2. The Norm-Control Condition

Definition 2.
A J B * -algebra A satisfies the norm-control condition if there exists C > 0 such that
dist x , S ( A ) ≤ C ρ ( x ) for all x ∈ A .
Proposition 2.
Every finite-dimensional J B * -algebra satisfies the norm-control condition (1).
Proof. 
By Lemma 2, ρ induces a seminorm ρ ˙ ( x + S ( A ) ) : = ρ ( x ) on the quotient space A / S ( A ) , and ρ ˙ ( x + S ( A ) ) = 0 if and only if x ∈ S ( A ) . Thus ρ ˙ is a norm on the finite-dimensional space A / S ( A ) , hence equivalent to the quotient norm dist ( x , S ( A ) ) . □
Assumption 1.
Throughout Section 3 and Section 4, A is a unital J B * -algebra such that
(H1)
A ≠ C 1 and S ( A ) = C 1 ;
(H2)
A satisfies the norm-control condition (1).
By Propositions 1 and 2, M n ( C ) with n ≥ 2 satisfies Assumption 1.

2.3. Structural Lemmas Under (H1)

Lemma 4.
Assume (H1). If e ∈ A satisfies x ∘ e ∈ S ( A ) for all x ∈ A , then e = 0 .
Proof. 
Taking x = 1 gives e ∈ S ( A ) = C 1 , say e = c 1 . Then c x = x ∘ e ∈ C 1 for every x ∈ A . Since A ≠ C 1 , we get c = 0 . □
By (H1) we may fix y 0 ∈ A ∖ C 1 and, since y 0 ∉ S ( A ) , some z 0 ∈ A with { y 0 , z 0 } ≠ 0 .
Lemma 5
(Scalar separation). Assume (H1).
(i)
If h : A → C satisfies h ( x ) { y , z } + h ( y ) { x , z } = 0 for all x , y , z ∈ A , then h ≡ 0 .
(ii)
Let c n : A → C ( n ∈ N ) be functions such that, for all x , y , z ∈ A , the sequence
{ c n ( x ) { y , z } + c n ( y ) { x , z } } n
converges in A . Then { c n ( x ) } n converges in C for every x ∈ A .
Proof. 
Put v 0 : = { y 0 , z 0 } ≠ 0 . (i) Taking x = y = y 0 , z = z 0 gives 2 h ( y 0 ) v 0 = 0 , so h ( y 0 ) = 0 . Taking y = y 0 , z = z 0 then gives h ( x ) v 0 = 0 , so h ( x ) = 0 for every x.
(ii) Taking x = y = y 0 , z = z 0 shows that { 2 c n ( y 0 ) v 0 } converges. Since
| c n ( y 0 ) − c m ( y 0 ) | ∥ v 0 ∥ = ∥ ( c n ( y 0 ) − c m ( y 0 ) ) v 0 ∥ ,
the sequence { c n ( y 0 ) } is Cauchy. Next, taking y = y 0 , z = z 0 , the sequence
c n ( x ) v 0 = c n ( x ) v 0 + c n ( y 0 ) { x , z 0 } − c n ( y 0 ) { x , z 0 }
converges, and the same argument shows that { c n ( x ) } converges. □
Lemma 6.
Assume (H1) and (H2). Let A ′ be the dual space of A . Fix τ ∈ A ′ with ∥ τ ∥ = 1 and τ ( 1 ) = 1 (Hahn–Banach), and put P ( x ) : = x − τ ( x ) 1 . Then P is linear, P ( S ( A ) ) = { 0 } and ∥ P ( x ) ∥ ≤ 2 C ρ ( x ) for all x ∈ A .
Proof. 
For every c ∈ C we have P ( x ) = ( x − c 1 ) − τ ( x − c 1 ) 1 , so ∥ P ( x ) ∥ ≤ 2 ∥ x − c 1 ∥ . Taking the infimum over c and using (1) gives ∥ P ( x ) ∥ ≤ 2 dist ( x , C 1 ) ≤ 2 C ρ ( x ) . □

2.4. S ( A ) -Extended Jordan *-Derivations

Definition 3.
A mapping δ : A → A is called an S ( A ) -extended Jordan *-derivation ( S ( A ) -E Jordan *-derivation) if, for all a , b ∈ A ,
(E1)
δ ( a + b ) − δ ( a ) − δ ( b ) ∈ S ( A ) ;
(E2)
D δ ( a , b ) : = δ ( a ∘ b ) − δ ( a ) ∘ b − a ∘ δ ( b ) ∈ S ( A ) ;
(E3)
δ ( a * ) − δ ( a ) * ∈ S ( A ) .
The next proposition explains why b * does not appear in (E2).
Proposition 3.
Let A be a unital J B * -algebra and suppose that δ : A → A satisfies
δ ( a ∘ b ) − δ ( a ) ∘ b * − a ∘ δ ( b ) ∈ S ( A ) for all a , b ∈ A .
Then δ ( h ) ∈ S ( A ) for every self-adjoint h ∈ A .
Proof. 
Since a ∘ b = b ∘ a , subtracting the two relations obtained by interchanging a and b gives δ ( a ) ∘ ( b * − b ) + δ ( b ) ∘ ( a − a * ) ∈ S ( A ) for all a , b ∈ A . For a = h = h * and b = i 1 this reads − 2 i δ ( h ) ∈ S ( A ) . □
Proposition 4
(Rigidity). Assume (H1). Every S ( A ) -E Jordan *-derivation L : A → A is additive and satisfies
L ( 0 ) = L ( 1 ) = 0 , L ( x * ) = L ( x ) * for all x ∈ A .
Proof. 
Put s ( x , x ′ ) : = L ( x + x ′ ) − L ( x ) − L ( x ′ ) ∈ S ( A ) . A direct computation gives
D L ( x + x ′ , y ) − D L ( x , y ) − D L ( x ′ , y ) = s ( x ∘ y , x ′ ∘ y ) − s ( x , x ′ ) ∘ y
for all x , x ′ , y ∈ A . The left-hand side and s ( x ∘ y , x ′ ∘ y ) lie in S ( A ) , hence
y ∘ s ( x , x ′ ) ∈ S ( A ) for every y ∈ A ,
and Lemma 4 gives s ( x , x ′ ) = 0 . Thus L is additive and L ( 0 ) = 0 . Finally,
D L ( 1 , y ) = − L ( 1 ) ∘ y ∈ S ( A ) for every y ∈ A ,
so L ( 1 ) = 0 by Lemma 4.
To prove exact preservation of the involution, put L ♯ ( x ) : = L ( x * ) * . By (E2) and the *-invariance of S ( A ) ,
D L ♯ ( x , y ) = D L ( x * , y * ) * ∈ S ( A ) .
By (E3), the map g : = L ♯ − L takes values in S ( A ) = C 1 ; write g ( x ) = h ( x ) 1 . Subtracting the product identities modulo S ( A ) for L ♯ and L gives
g ( x ∘ y ) − g ( x ) ∘ y − x ∘ g ( y ) ∈ S ( A ) .
The first term is scalar, so applying the associator yields
h ( x ) { y , z } + h ( y ) { x , z } = 0 for all x , y , z ∈ A .
Lemma 5(i) gives h = 0 . Hence L ♯ = L , which is equivalent to L ( x * ) = L ( x ) * . □

2.5. The Functional Inequalities

Fix real numbers a 1 , a 2 , a 3 with a 1 a 2 a 3 ≠ 0 . We consider mappings δ : A → A satisfying
∥ { a 1 δ ( x 1 ) + a 2 δ ( x 2 ) + a 3 δ ( x 3 ) , w } ∥ ≤ ∥ { δ ( a 1 x 1 + a 2 x 2 + a 3 x 3 ) , w } ∥ + Φ ( x 1 , x 2 , x 3 , w ) ,
∥ { δ ( a ∘ b ) − δ ( a ) ∘ b − a ∘ δ ( b ) , z } ∥ ≤ ϕ ( a , b , z ) ,
∥ { δ ( a * ) − δ ( a ) * , z } ∥ ≤ ψ ( a , z ) ,
{ δ ( 0 ) , a } = 0 for all a ∈ A , i . e . , δ ( 0 ) ∈ S ( A ) ,
for all x 1 , x 2 , x 3 , w , a , b , z ∈ A , where Φ : A 4 → [ 0 , ∞ ) , ϕ : A 3 → [ 0 , ∞ ) and ψ : A 2 → [ 0 , ∞ ) . For the C -linearity results we also use the unimodular version of (2):
∥ { λ a 1 δ ( x 1 ) + λ a 2 δ ( x 2 ) + a 3 δ ( λ x 3 ) , w } ∥ ≤ ∥ { δ ( a 1 x 1 + a 2 x 2 + a 3 x 3 ) , w } ∥ + Φ ( x 1 , x 2 , x 3 , w ) .
Here x 1 , x 2 , x 3 , w ∈ A and λ ∈ T : = { μ ∈ C : | μ | = 1 } . Taking λ = 1 recovers (2). This formulation is compatible with nonzero C -linear maps: if L is C -linear and s = a 1 x 1 + a 2 x 2 + a 3 x 3 , then
λ a 1 L ( x 1 ) + λ a 2 L ( x 2 ) + a 3 L ( λ x 3 ) = λ L ( s ) ,
and ∥ { λ L ( s ) , w } ∥ = ∥ { L ( s ) , w } ∥ . Thus (6) holds for such a map with Φ = 0 .
We write
Φ ^ ( x 1 , x 2 , x 3 ) : = sup ∥ w ∥ ≤ 1 Φ ( x 1 , x 2 , x 3 , w ) , Ψ ( x , w ) : = Φ − a 3 x a 1 , − a 3 x a 2 , 2 x , w + Φ − a 3 x a 1 , 0 , x , w + Φ 0 , − a 3 x a 2 , x , w
and
Ψ ^ ( x ) : = sup ∥ w ∥ ≤ 1 Ψ ( x , w )
for all x 1 , x 2 , x 3 , w , x ∈ A .
Lemma 7.
Suppose that δ : A → A satisfies (5) and
∥ { a 1 δ ( x 1 ) + a 2 δ ( x 2 ) + a 3 δ ( x 3 ) , w } ∥ ≤ ∥ { δ ( a 1 x 1 + a 2 x 2 + a 3 x 3 ) , w } ∥
for all x 1 , x 2 , x 3 , w ∈ A . Then δ ( x + y ) − δ ( x ) − δ ( y ) ∈ S ( A ) for all x , y ∈ A .
Proof. 
Setting ( x 1 , x 2 , x 3 ) = ( x , 0 , − a 1 x / a 3 ) , ( 0 , y , − a 2 y / a 3 ) and ( x , y , − ( a 1 x + a 2 y ) / a 3 ) in (7), the right-hand side is ∥ { δ ( 0 ) , w } ∥ = 0 in each case, so
{ a 1 δ ( x ) + a 2 δ ( 0 ) + a 3 δ ( − a 1 x / a 3 ) , w } = 0 , { a 1 δ ( 0 ) + a 2 δ ( y ) + a 3 δ ( − a 2 y / a 3 ) , w } = 0 , { a 1 δ ( x ) + a 2 δ ( y ) + a 3 δ − ( a 1 x + a 2 y ) / a 3 , w } = 0 .
Subtracting the first two identities from the third and using δ ( 0 ) ∈ S ( A ) and Lemma 1(i), we get { δ ( − ( a 1 x + a 2 y ) / a 3 ) − δ ( − a 1 x / a 3 ) − δ ( − a 2 y / a 3 ) , w } = 0 . Replacing x by − a 3 x / a 1 and y by − a 3 y / a 2 gives { δ ( x + y ) − δ ( x ) − δ ( y ) , w } = 0 for all x , y , w ∈ A . □
Lemma 7 also holds for normed Jordan algebras.
Lemma 8.
If δ : A → A satisfies (2) and (5), then
∥ { δ ( 2 x ) − 2 δ ( x ) , w } ∥ ≤ 1 | a 3 | Ψ ( x , w ) for all x , w ∈ A .
Proof. 
Setting ( x 1 , x 2 , x 3 ) equal to
( x , 0 , − a 1 x / a 3 ) , ( 0 , y , − a 2 y / a 3 ) , x , y , − ( a 1 x + a 2 y ) / a 3
in (2), the right-hand norm vanishes by (5) and we obtain
∥ { a 1 δ ( x ) + a 2 δ ( 0 ) + a 3 δ ( − a 1 x / a 3 ) , w } ∥ ≤ Φ ( x , 0 , − a 1 x / a 3 , w ) , ∥ { a 1 δ ( 0 ) + a 2 δ ( y ) + a 3 δ ( − a 2 y / a 3 ) , w } ∥ ≤ Φ ( 0 , y , − a 2 y / a 3 , w ) , ∥ { a 1 δ ( x ) + a 2 δ ( y ) + a 3 δ ( − ( a 1 x + a 2 y ) / a 3 ) , w } ∥ ≤ Φ x , y , − ( a 1 x + a 2 y ) / a 3 , w .
By the triangle inequality, Lemma 1(i) and δ ( 0 ) ∈ S ( A ) ,
| a 3 | ∥ { δ ( − a 1 x + a 2 y , a 3 } ) − δ ( − a 1 x a 3 ) − δ ( − a 2 y a 3 ) w ∥ ≤ Φ x , y , − a 1 x + a 2 y a 3 , w + Φ x , 0 , − a 1 x a 3 , w + Φ 0 , y , − a 2 y a 3 , w .
Replacing x by − a 3 x / a 1 and y by − a 3 x / a 2 gives (8). □

3. Hyers–Ulam Stability

Theorem 1.
Let A satisfy Assumption 1. Assume that Φ ^ is finite-valued and that, for all x 1 , x 2 , x 3 ∈ A , w ∈ A and x , y , z ∈ A ,
∑ j = 0 ∞ 2 j Φ x 1 2 j , x 2 2 j , x 3 2 j , w < ∞ , ∑ j = 0 ∞ 2 j Φ ^ x 1 2 j , x 2 2 j , x 3 2 j < ∞ ,
lim n → ∞ 4 n ϕ x 2 n , y 2 n , z = 0 , lim n → ∞ 2 n ψ x 2 n , z = 0 .
Suppose δ : A → A satisfies (2), (3), (4) and (5). Then the limit
L ( x ) : = lim n → ∞ 2 n δ x 2 n
exists in norm for every x ∈ A , L is an additive S ( A ) -E Jordan *-derivation with L ( 1 ) = 0 and L ( x * ) = L ( x ) * for all x ∈ A , and
∥ { L ( x ) − δ ( x ) , w } ∥ ≤ η ( x , w ) : = 1 | a 3 | ∑ j = 0 ∞ 2 j Ψ x 2 j + 1 , w for all x , w ∈ A .
Moreover, L is the unique S ( A ) -E Jordan *-derivation satisfying (12). Setting
η ^ ( x ) : = 1 | a 3 | ∑ j = 0 ∞ 2 j Ψ ^ x 2 j + 1 ,
we also have, for every x ∈ A ,
ρ ( L ( x ) − δ ( x ) ) ≤ η ^ ( x ) , dist ( L ( x ) − δ ( x ) , C 1 ) ≤ C η ^ ( x ) .
Proof. 
Each of the three terms of 2 j Ψ ( x / 2 j + 1 , w ) has the form 2 j Φ ( y 1 / 2 j , y 2 / 2 j , y 3 / 2 j , w ) for a fixed triple, e.g., ( y 1 , y 2 , y 3 ) = ( − a 3 x / ( 2 a 1 ) , − a 3 x / ( 2 a 2 ) , x ) for the first one. Hence, by (9), the series defining η ( x , w ) and η ^ ( x ) converge.
  • Step 1: Cauchy estimates. Put u n ( x ) : = 2 n δ ( x / 2 n ) . Applying (8) at x / 2 j + 1 and multiplying by 2 j , we get ∥ { u j ( x ) − u j + 1 ( x ) , w } ∥ ≤ | a 3 | − 1 2 j Ψ ( x / 2 j + 1 , w ) . Summing over j, for 0 ≤ m < n ,
∥ { u m ( x ) − u n ( x ) , w } ∥ ≤ 1 | a 3 | ∑ j = m n − 1 2 j Ψ x 2 j + 1 , w , ρ u m ( x ) − u n ( x ) ≤ 1 | a 3 | ∑ j = m n − 1 2 j Ψ ^ x 2 j + 1 .
Step 2: convergence of the non-scalar part. With P and τ as in Lemma 6, write
u n ( x ) = v n ( x ) + c n ( x ) 1 , v n ( x ) : = P ( u n ( x ) ) , c n ( x ) : = τ ( u n ( x ) ) .
By Lemma 6 and (13), ∥ v m ( x ) − v n ( x ) ∥ ≤ 2 C ρ ( u m ( x ) − u n ( x ) ) , which tends to 0 as m → ∞ uniformly in n > m , since it is bounded by a tail of the convergent series defining η ^ ( x ) . Hence the limit
V ( x ) : = lim n → ∞ v n ( x )
exists in norm for every x ∈ A .
  • Step 3: convergence of the scalar part. Fix x , y , z ∈ A . Applying (3) with a = x / 2 n , b = y / 2 n and multiplying by 4 n , we obtain
∥ { E n , z } ∥ ≤ 4 n ϕ x 2 n , y 2 n , z ,
where E n : = u 2 n ( x ∘ y ) − u n ( x ) ∘ y − x ∘ u n ( y ) . Since { 1 , z } = 0 and 1 ∘ y = y , Lemma 1(i) gives
{ E n , z } = { v 2 n ( x ∘ y ) − v n ( x ) ∘ y − x ∘ v n ( y ) , z } − c n ( x ) { y , z } + c n ( y ) { x , z } .
The left-hand side tends to 0 by (10), and the first term on the right converges to
{ V ( x ∘ y ) − V ( x ) ∘ y − x ∘ V ( y ) , z }
by Step 2 and Lemma 1(ii). Therefore c n ( x ) { y , z } + c n ( y ) { x , z } converges for all x , y , z , and Lemma 5(ii) shows that c ( x ) : = lim n → ∞ c n ( x ) exists. Consequently
u n ( x ) → L ( x ) : = V ( x ) + c ( x ) 1 in norm ,
which proves (11).
  • Step 4: the estimates. Take m = 0 in (13) and let n → ∞ , using u 0 = δ and Lemma 1(ii), to obtain (12). Since the terms in the series are nonnegative,
sup ∥ w ∥ ≤ 1 η ( x , w ) ≤ η ^ ( x ) .
Taking the supremum in (12) yields ρ ( L ( x ) − δ ( x ) ) ≤ η ^ ( x ) , and (H2) gives the asserted distance estimate.
  • Step 5: L satisfies (E1)–(E3). Replacing x i by x i / 2 n in (2) and multiplying by 2 n , we get
∥ { a 1 u n ( x 1 ) + a 2 u n ( x 2 ) + a 3 u n ( x 3 ) , w } ∥ ≤ ∥ { u n ( a 1 x 1 + a 2 x 2 + a 3 x 3 ) , w } ∥ + 2 n Φ x 1 2 n , x 2 2 n , x 3 2 n , w .
The last term tends to 0 by (9), so letting n → ∞ shows that L satisfies (7). Moreover { L ( 0 ) , a } = lim n → ∞ 2 n { δ ( 0 ) , a } = 0 , so Lemma 7 gives (E1) for L. Letting n → ∞ in Step 3 gives
{ L ( x ∘ y ) − L ( x ) ∘ y − x ∘ L ( y ) , z } = lim n → ∞ { E n , z } = 0 ,
which is (E2). Finally, by (4) with a = x / 2 n ,
∥ { u n ( x * ) − u n ( x ) * , z } ∥ = 2 n ∥ { δ ( x * / 2 n ) − δ ( x / 2 n ) * , z } ∥ ≤ 2 n ψ x 2 n , z → 0 ,
and since the involution is isometric, { L ( x * ) − L ( x ) * , z } = 0 , which is (E3). Hence L is an S ( A ) -E Jordan *-derivation, and by Proposition 4 it is additive, vanishes at 1 , and satisfies L ( x * ) = L ( x ) * for every x ∈ A .
  • Step 6: uniqueness. Let L ′ be another S ( A ) -E Jordan *-derivation satisfying (12). Both L and L ′ are additive (Proposition 4), hence L ( x ) = 2 n L ( x / 2 n ) and L ′ ( x ) = 2 n L ′ ( x / 2 n ) . Therefore
∥ { L ( x ) − L ′ ( x ) , w } ∥ = 2 n ∥ { L ( x 2 n ) − L ′ ( x 2 n ) , w } ∥ ≤ 2 · 2 n η x 2 n , w = 2 | a 3 | ∑ j = n ∞ 2 j Ψ x 2 j + 1 , w → 0 .
Thus g : = L − L ′ takes values in S ( A ) = C 1 ; write g = h 1 with h : A → C . Subtracting (E2) for L ′ from (E2) for L gives g ( x ∘ y ) − g ( x ) ∘ y − x ∘ g ( y ) ∈ S ( A ) , i.e., { h ( x ) y + h ( y ) x , z } = 0 for all x , y , z . By Lemma 5(i), h ≡ 0 , so L = L ′ . □
Theorem 2.
Let A satisfy Assumption 1. Assume that Φ ^ is finite-valued and that, for all x 1 , x 2 , x 3 ∈ A , w ∈ A and x , y , z ∈ A ,
∑ j = 0 ∞ 1 2 j Φ 2 j x 1 , 2 j x 2 , 2 j x 3 , w < ∞ , ∑ j = 0 ∞ 1 2 j Φ ^ 2 j x 1 , 2 j x 2 , 2 j x 3 < ∞ ,
lim n → ∞ 1 4 n ϕ 2 n x , 2 n y , z = 0 , lim n → ∞ 1 2 n ψ 2 n x , z = 0 .
Suppose δ : A → A satisfies (2), (3), (4) and (5). Then L ( x ) : = lim n → ∞ 2 − n δ ( 2 n x ) exists in norm for every x ∈ A , L is an additive S ( A ) -E Jordan *-derivation with L ( 1 ) = 0 and L ( x * ) = L ( x ) * for all x ∈ A , and
∥ { L ( x ) − δ ( x ) , w } ∥ ≤ η ˜ ( x , w ) : = 1 2 | a 3 | ∑ j = 0 ∞ 1 2 j Ψ 2 j x , w for all x , w ∈ A .
Moreover, L is the unique S ( A ) -E Jordan *-derivation satisfying (16). Setting
η ˜ ^ ( x ) : = 1 2 | a 3 | ∑ j = 0 ∞ 2 − j Ψ ^ ( 2 j x ) ,
we also have, for every x ∈ A ,
ρ ( L ( x ) − δ ( x ) ) ≤ η ˜ ^ ( x ) , dist ( L ( x ) − δ ( x ) , C 1 ) ≤ C η ˜ ^ ( x ) .
Proof. 
Put u n ( x ) : = 2 − n δ ( 2 n x ) . Then u j ( x ) − u j + 1 ( x ) = − 2 − ( j + 1 ) δ ( 2 j + 1 x ) − 2 δ ( 2 j x ) , so (8) at 2 j x gives ∥ { u j ( x ) − u j + 1 ( x ) , w } ∥ ≤ ( 2 | a 3 | ) − 1 2 − j Ψ ( 2 j x , w ) , and hence, for 0 ≤ m < n ,
∥ { u m ( x ) − u n ( x ) , w } ∥ ≤ 1 2 | a 3 | ∑ j = m n − 1 1 2 j Ψ ( 2 j x , w ) , ρ u m ( x ) − u n ( x ) ≤ 1 2 | a 3 | ∑ j = m n − 1 1 2 j Ψ ^ ( 2 j x ) .
The right-hand sides are tails of series that converge by (14). Applying (3) with a = 2 n x , b = 2 n y and dividing by 4 n gives ∥ { E n , z } ∥ ≤ 4 − n ϕ ( 2 n x , 2 n y , z ) with the same expression
E n = u 2 n ( x ∘ y ) − u n ( x ) ∘ y − x ∘ u n ( y )
as before. With these changes, Steps 2–5 of the proof of Theorem 1 apply verbatim, using (14) and (15) in place of (9) and (10). (In Step 5, (2) is applied at 2 n x i and divided by 2 n , and (4) is applied at 2 n x .) For uniqueness, additivity gives L ( x ) = 2 − n L ( 2 n x ) and L ′ ( x ) = 2 − n L ′ ( 2 n x ) , so
∥ { L ( x ) − L ′ ( x ) , w } ∥ ≤ 2 2 n η ˜ ( 2 n x , w ) = 1 | a 3 | ∑ j = n ∞ 1 2 j Ψ ( 2 j x , w ) → 0 ,
and we conclude as in Step 6. □
Remark 1
(Meaning of the stability estimates). Theorems 1 and 2 give convergence of their approximating sequences in the original norm of A , not merely in a quotient. Their quantitative error estimates, however, concern ρ ( L ( x ) − δ ( x ) ) and the distance of L ( x ) − δ ( x ) to C 1 . Since ρ ( s 1 ) = 0 for every s ∈ C , those estimates alone do not bound ∥ L ( x ) − δ ( x ) ∥ . Example 2 shows that the associator error can vanish identically while the full norm error is arbitrarily large.

4. Superstability

The following results show that, under an additional decay condition on the product control ϕ in which only the first variable is scaled, the mapping δ itself is an S ( A ) -E Jordan *-derivation. In fact it coincides with L.
Theorem 3.
Under the hypotheses of Theorem 1, assume in addition that
lim n → ∞ 2 n ϕ x 2 n , y , z = 0 for all x , y , z ∈ A .
Then δ = L . In particular, δ is an additive S ( A ) -E Jordan *-derivation satisfying δ ( 1 ) = 0 and δ ( x * ) = δ ( x ) * for all x ∈ A .
Proof. 
Applying (3) with a = x / 2 n , b = y and multiplying by 2 n , we get
∥ { u n ( x ∘ y ) − u n ( x ) ∘ y − x ∘ δ ( y ) , z } ∥ ≤ 2 n ϕ x 2 n , y , z ,
where u n ( x ) = 2 n δ ( x / 2 n ) . Letting n → ∞ and using (11), (17) and Lemma 1(ii), we obtain { L ( x ∘ y ) − L ( x ) ∘ y − x ∘ δ ( y ) , z } = 0 . Comparing this relation with (E2) for L yields
{ x ∘ L ( y ) − δ ( y ) , z } = 0 for all x , y , z ∈ A ,
that is, x ∘ ( L ( y ) − δ ( y ) ) ∈ S ( A ) for all x ∈ A . By Lemma 4, L ( y ) = δ ( y ) for every y ∈ A . □
Theorem 4.
Under the hypotheses of Theorem 2, assume in addition that
lim n → ∞ 1 2 n ϕ 2 n x , y , z = 0 for all x , y , z ∈ A .
Then δ = L . In particular, δ is an additive S ( A ) -E Jordan *-derivation satisfying δ ( 1 ) = 0 and δ ( x * ) = δ ( x ) * for all x ∈ A .
Proof. 
Apply (3) with a = 2 n x , b = y , divide by 2 n , let n → ∞ , and argue as in the proof of Theorem 3. □
Corollary 1.
Under the hypotheses of Theorem 1(resp. Theorem 2), suppose that δ satisfies (6) for all λ ∈ T . Then L is C -linear. If, moreover, (17)(resp. (18))holds, then δ = L is a C -linear S ( A ) -E Jordan *-derivation.
Proof. 
We treat the case of Theorem 1; the case of Theorem 2 follows by applying (6) at 2 n x i and dividing by 2 n . In the first case, replace x i by x i / 2 n in (6), multiply by 2 n , and let n → ∞ . Norm convergence, (9), and 2 n δ ( λ x 3 / 2 n ) → L ( λ x 3 ) give
∥ { λ a 1 L ( x 1 ) + λ a 2 L ( x 2 ) + a 3 L ( λ x 3 ) , w } ∥ ≤ ∥ { L ( a 1 x 1 + a 2 x 2 + a 3 x 3 ) , w } ∥ .
Fix λ ∈ T . Since L is additive, L ( 0 ) = 0 . Taking ( x 1 , x 2 , x 3 ) = ( x , 0 , − a 1 x / a 3 ) gives
{ λ a 1 L ( x ) + a 3 L ( − λ a 1 x / a 3 ) , w } = 0 .
The same identity with the unimodular parameter equal to 1 and x replaced by λ x reads
{ a 1 L ( λ x ) + a 3 L ( − a 1 λ x / a 3 ) , w } = 0 .
Subtracting these identities and using a 1 ≠ 0 , we obtain
r λ ( x ) : = L ( λ x ) − λ L ( x ) ∈ S ( A ) ( x ∈ A ) .
Since S ( A ) is a complex subspace, (E2) gives D L ( λ x , y ) − λ D L ( x , y ) ∈ S ( A ) . A direct computation yields
D L ( λ x , y ) − λ D L ( x , y ) = r λ ( x ∘ y ) − r λ ( x ) ∘ y .
The first term on the right belongs to S ( A ) . Consequently y ∘ r λ ( x ) ∈ S ( A ) for every y ∈ A , and Lemma 4 gives r λ ( x ) = 0 . Thus
L ( λ x ) = λ L ( x ) for all λ ∈ T , x ∈ A .
Now let t ∈ C ∖ { 0 } and choose an integer M > | t | . Since | t / M | < 1 , there exist λ 1 , λ 2 ∈ T such that t / M = 1 2 ( λ 1 + λ 2 ) . Additivity implies Q -linearity, hence
L ( t x ) = M 2 L ( λ 1 x + λ 2 x ) = M 2 λ 1 L ( x ) + λ 2 L ( x ) = t L ( x ) .
The case t = 0 follows from L ( 0 ) = 0 . This proves C -linearity without any continuity assumption on L. The final assertion follows from Theorem 3 (resp. Theorem 4). □

5. Examples

Example 1.
Let n ≥ 2 and let A = M n ( C ) carry the operator norm and the Jordan product a ∘ b = 1 2 ( a b + b a ) . By Propositions 1 and 2, A satisfies Assumption 1. Fix self-adjoint matrices h , u ∈ A ∖ C 1 and a real number r > 1 , and define
Δ ( x ) : = i [ h , x ] , δ ( x ) : = Δ ( x ) + ∥ x ∥ r u ( x ∈ A ) .
Δ is a Jordan *-derivation. Since Δ is an associative derivation and h = h * ,
Δ ( x ∘ y ) = Δ ( x ) ∘ y + x ∘ Δ ( y ) , Δ ( x ) * = − i [ h , x ] * = i [ h , x * ] = Δ ( x * ) .
Inequality (2). Put s : = a 1 x 1 + a 2 x 2 + a 3 x 3 and β : = ∑ k a k ∥ x k ∥ r . Then
a 1 δ ( x 1 ) + a 2 δ ( x 2 ) + a 3 δ ( x 3 ) = Δ ( s ) + β u = δ ( s ) + ( β − ∥ s ∥ r ) u .
By Lemma 1(ii), ∥ { u , w } ∥ ≤ 2 ∥ u ∥ ∥ w ∥ 2 , so (2) holds with
Φ ( x 1 , x 2 , x 3 , w ) = 2 ∥ u ∥ ∥ w ∥ 2 ∥ a 1 x 1 + a 2 x 2 + a 3 x 3 ∥ r + ∑ k = 1 3 | a k | ∥ x k ∥ r .
Here Φ ^ ( x 1 , x 2 , x 3 ) = 2 ∥ u ∥ ∥ s ∥ r + ∑ k | a k | ∥ x k ∥ r is positively homogeneous of degree r > 1 , so both series in (9) are dominated by a multiple of ∑ j = 0 ∞ 2 j ( 1 − r ) < ∞ .
The unimodular inequality (6). For λ ∈ T , put
β λ : = λ a 1 ∥ x 1 ∥ r + λ a 2 ∥ x 2 ∥ r + a 3 ∥ x 3 ∥ r .
Since Δ is C -linear and ∥ λ x 3 ∥ = ∥ x 3 ∥ ,
λ a 1 δ ( x 1 ) + λ a 2 δ ( x 2 ) + a 3 δ ( λ x 3 ) = λ δ ( s ) + ( β λ − λ ∥ s ∥ r ) u .
Also | β λ − λ ∥ s ∥ r | ≤ ∥ s ∥ r + ∑ k = 1 3 | a k | ∥ x k ∥ r . Applying the associator and using | λ | = 1 therefore proves (6) with the same control function Φ, uniformly for λ ∈ T .
Inequality (3). We have δ ( a ∘ b ) − δ ( a ) ∘ b − a ∘ δ ( b ) = ∥ a ∘ b ∥ r u − ∥ a ∥ r u ∘ b − ∥ b ∥ r a ∘ u . Using ∥ a ∘ b ∥ ≤ ∥ a ∥ ∥ b ∥ and Lemma 1(ii), (3) holds with
ϕ ( a , b , z ) = 2 ∥ u ∥ ∥ z ∥ 2 ∥ a ∥ r ∥ b ∥ r + ∥ a ∥ r ∥ b ∥ + ∥ a ∥ ∥ b ∥ r ,
and 4 n ϕ ( x / 2 n , y / 2 n , z ) = 2 ∥ u ∥ ∥ z ∥ 2 2 2 n ( 1 − r ) ∥ x ∥ r ∥ y ∥ r + 2 n ( 1 − r ) ( ∥ x ∥ r ∥ y ∥ + ∥ x ∥ ∥ y ∥ r ) → 0 .
Conditions (4) and (5). Since u = u * and ∥ x * ∥ = ∥ x ∥ , we have δ ( x * ) = δ ( x ) * , so (4) holds with ψ = 0 ; clearly δ ( 0 ) = 0 .
By Theorem 1,
L ( x ) = lim n → ∞ Δ ( x ) + 2 n ( 1 − r ) ∥ x ∥ r u = Δ ( x )
is the unique S ( A ) -E Jordan *-derivation satisfying (12). Indeed,
∥ { L ( x ) − δ ( x ) , w } ∥ = ∥ x ∥ r ∥ { u , w } ∥ ,
which is nonzero for suitable w when x ≠ 0 , because u ∉ S ( A ) . Hence δ ≠ L . Moreover, L = Δ is nonzero since h ∉ C 1 . Thus the unimodular hypothesis is compatible with a nonzero C -linear limit.
The additional decay condition. For the chosen product control,
2 n ϕ ( x / 2 n , y , z ) = 2 ∥ u ∥ ∥ z ∥ 2 2 n ( 1 − r ) ∥ x ∥ r ∥ y ∥ r + ∥ x ∥ r ∥ y ∥ + ∥ x ∥ ∥ y ∥ r ⟶ 2 ∥ u ∥ ∥ z ∥ 2 ∥ x ∥ ∥ y ∥ r .
For nonzero x , y , z this limit is nonzero. Thus the chosen control does not satisfy (17), consistently with δ ≠ L .
Example 2
(Scalar errors are invisible to the associator). Let A = M 2 ( C ) with its usual operator norm, Jordan product and involution. Fix M > 0 and r > 1 , and define
δ M ( x ) : = M ∥ x ∥ r 1 ( x ∈ A ) .
All values are scalar, so both (2) and (6) hold with Φ = 0 . Moreover, δ M ( 0 ) = 0 and δ M ( x * ) = δ M ( x ) * , so we may take ψ = 0 . The product defect satisfies
{ D δ M ( a , b ) , z } = − M ∥ a ∥ r { b , z } − M ∥ b ∥ r { a , z } .
Thus (3) holds with
ϕ M ( a , b , z ) : = 2 M ∥ z ∥ 2 ∥ a ∥ r ∥ b ∥ + ∥ a ∥ ∥ b ∥ r .
For every x , y , z ∈ A ,
4 n ϕ M ( x / 2 n , y / 2 n , z ) = 2 M ∥ z ∥ 2 2 n ( 1 − r ) ∥ x ∥ r ∥ y ∥ + ∥ x ∥ ∥ y ∥ r ⟶ 0 .
All hypotheses of Theorem 1 are therefore satisfied, and
L ( x ) = lim n → ∞ M 2 n ( 1 − r ) ∥ x ∥ r 1 = 0 .
Here η = 0 and η ^ = 0 , while
∥ L ( x ) − δ M ( x ) ∥ = M ∥ x ∥ r .
For fixed x ≠ 0 , this norm becomes arbitrarily large as M increases, while the associator error and quotient distance vanish. Thus (12) does not imply a full norm estimate based on η alone.

6. Conclusions

We have established Hyers–Ulam stability (Theorems 1 and 2) and superstability (Theorems 3 and 4) for the functional inequality (2), coupled with approximate product and involution conditions, in unital J B * -algebras with S ( A ) = C 1 satisfying the norm-control condition (1). This extends the program of [12,13] from the associative to the Jordan setting. The main features are the following.
(1)
The Jordan associator replaces the commutator. For a C * -algebra equipped with its Jordan product, it is the second-order commutator − 1 4 [ a , [ a , z ] ] , and its kernel S ( A ) contains the associative center.
(2)
The norm-control condition bounds dist ( x , S ( A ) ) rather than ∥ x ∥ . It is therefore compatible with the unit, and it holds in every finite-dimensional J B * -algebra.
(3)
The scalar part of the approximating sequence, which the associator cannot see, is recovered from the product inequality (Lemma 5). This yields norm convergence and the uniqueness of the limit.
(4)
Under (H1), every S ( A ) -E Jordan *-derivation is automatically additive, vanishes at the unit, and preserves the involution exactly (Proposition 4).
(5)
The quantitative estimates control the associator error and the distance to C 1 , not the full norm error. The unimodular inequality (6) gives C -linearity and admits nonzero limits, as demonstrated in Example 1.

Author Contributions

Conceptualization, J.R.; Writing—original draft, J.R.; Writing—review & editing, J.R.; Supervision, J.R.; Funding acquisition, J.R.; The author has read and agreed to the published version of the manuscript.

Funding

This paper was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2021R1A2C109489611).

Data Availability Statement

No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Conflicts of Interest

The author declares no conflicts of interest.

References

  1. Ulam, S.M. A Collection of Mathematical Problems; Interscience: New York, NY, USA, 1960.
  2. Hyers, D.H. On the stability of the linear functional equation. Proc. Natl. Acad. Sci. USA 1941, 27, 222–224. [CrossRef]
  3. Rassias, T.M. On the stability of the linear mapping in Banach spaces. Proc. Am. Math. Soc. 1978, 72, 297–300. [CrossRef]
  4. Găvruta, P. A generalization of the Hyers–Ulam–Rassias stability of approximately additive mappings. J. Math. Anal. Appl. 1994, 184, 431–436. [CrossRef]
  5. Gajda, Z. On stability of additive mappings. Int. J. Math. Math. Sci. 1991, 14, 431–434. [CrossRef]
  6. Baker, J.A.; Lawrence, J.; Zorzitto, F. The stability of the equation f(x+y)=f(x)f(y). Proc. Am. Math. Soc. 1979, 74, 242–246. [CrossRef]
  7. Šemrl, P. The functional equation of multiplicative derivation is superstable on standard operator algebras. Integr. Equ. Oper. Theory 1994, 18, 118–122. [CrossRef]
  8. Badora, R. On approximate derivations. Math. Inequal. Appl. 2006, 9, 167–173. [CrossRef]
  9. Chang, I.-S.; Kim, H.-M.; Roh, J. On superstability of derivations in Banach algebras. Open Math. 2025, 23, 1–9. [CrossRef]
  10. Bell, H.E.; Daif, M.N. On centrally-extended maps on rings. Beitr. Algebra Geom. 2016, 57, 129–136. [CrossRef]
  11. El-Deken, S.F.; Nabiel, H. Centrally-extended generalized *-derivations on rings with involution. Beitr. Algebra Geom. 2019, 60, 217–224. [CrossRef]
  12. Chang, I.-S.; Roh, J. Some centrally extended derivations on Banach algebras. Mathematics 2025, 13, 2271. [CrossRef]
  13. Chang, I.-S.; Roh, J. Stability of some inequalities in Banach *-algebras. Mathematics 2026, 14, 1407. [CrossRef]
  14. Brešar, M.; Vukman, J. On some additive mappings in rings with involution. Aeq. Math. 1989, 38, 178–185. [CrossRef]
  15. Stampfli, J.G. The norm of a derivation. Pac. J. Math. 1970, 33, 737–747. [CrossRef]
  16. Wright, J.D.M. Jordan C*-algebras. Michigan Math. J. 1977, 24, 291–302. [CrossRef]
  17. Hanche-Olsen, H.; Størmer, E. Jordan Operator Algebras; Pitman: Boston, MA, USA, 1984.
  18. McCrimmon, K. A Taste of Jordan Algebras; Springer: New York, NY, USA, 2004. [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.