On the right-symmetric algebras with a unital matrix subalgebra (Q2221971)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | On the right-symmetric algebras with a unital matrix subalgebra |
scientific article |
Statements
On the right-symmetric algebras with a unital matrix subalgebra (English)
0 references
3 February 2021
0 references
An algebra \(\mathcal A\) is right-symmetric provided that the associator on \(\mathcal A\) is right-symmetric; i.e., it is symmetric in the last two elements: \((x, y, z)=(x, z, y)\) for all \(x, y, z \in \mathcal A,\) where \((x, y, z)=(xy)z - x(yz)\) is the associator of \(x, y,\) and \(z.\) In Section 1 the authors give the definition, some examples, and simplest properties of left-symmetric and right-symmetric algebras. In Section 2, granted an arbitrary algebra \(\mathcal A,\) the authors construct the algebra \(E(\mathcal A)\) that includes \(\mathrm{End}(\mathcal A)\) as a subalgebra, and provide some conditions for \(E(\mathcal A)\) to be right-symmetric and simple. Namely, let \((\mathcal A; \bullet)\) be an algebra over a field \(F,\) and let \(M := \mathrm{End}_F (\mathcal A).\) Consider \(\mathcal A\) as the natural unital right \(M\)-module. Fix a linear mapping \(\psi :\mathcal A \to M\) and a homomorphism \(\pi : \mathcal A \to M\) of right unital \(M\)-modules. Consider an algebra \(E(\mathcal A) := \mathcal A \oplus M,\) where \(\mathcal A\) is the natural unital right module over \(M\), while \(M\) is a unital matrix subalgebra of \(E(\mathcal A),\) and the remaining products are defined by the rules \[ A \cdot a = aA + [A, R^\bullet_a], \ a \cdot b = a \bullet b + \psi (b)\pi(a) \] for all \(a, b \in \mathcal A\) and \(A \in M.\) Then Theorem 1. Let \((\mathcal A; \bullet)\) be an algebra over a field \(F.\) Then \(E(\mathcal A)\) is right-symmetric if and only if \begin{align*} R^\bullet _{[a,b]} - [R^\bullet_ a, R^\bullet_ b ] + \psi(b) \pi(a) -\psi(a)\pi(b) = \alpha (a, b) \mathrm{id},\\ c \psi (b)\pi (a) - b \psi(c) \pi(a) = \alpha(b, c)a, \\((R^\bullet_b \psi(c) - R^\bullet_c \psi(b) - \psi([b, c])) \pi(a) = 0 \end{align*} for a skew-symmetric bilinear form \(\alpha\) on \(\mathcal A\) and for all \(a, b, c \in \mathcal A.\) Corollary 1. Let \(E(\mathcal A)\) be as above, and let \(\psi = 0.\) Then \(E(\mathcal A)\) is right-symmetric if and only if \(\mathcal A\) is right-symmetric. Let us now define two special types of algebras: (1) An algebra \(\mathcal A\) over a field \(F\) is an algebra of scalar multiplications if \(ab = \gamma (a)b\) (or \(ab = \gamma(b)a\)) for all \(a, b \in \mathcal A\) and some \(\gamma \in \mathcal A^*.\) Note that in this case, \(\mathcal A\) is associative. (2) Consider \(Q(\alpha, \beta) :=\mathcal A \oplus\mathrm{End}(\mathcal A)\) for a quadratic algebra \((A; \bullet)\) with the product \(a \bullet b := \alpha(b)a+\beta(a)b,\) where \(\alpha, \beta \in \mathcal A^*.\) Theorem 2. Let \(\mathcal A\) be an algebra over a field \(F,\) and let \(E(\mathcal A)\) be as in Theorem \(1.\) Assume that \(\mathcal A\) is not an algebra of scalar multiplications, and \(E(\mathcal A) \not\cong Q( \alpha, \beta).\) Then the right-symmetric algebra \(E(\mathcal A)\) is simple. In Section 3, the authors elaborate on the case when \(\mathcal A\) is a right-symmetric algebra with a unital matrix subalgebra \(M_2(F)\) (i.e., the unity \(E \in M_n(F)\) serves as a unity for \(\mathcal A\)), and \(\mathfrak{sl}_2\) acts irreducibly on \(\mathcal A/M_2(F).\) Lemma 3. Let \(\mathcal A\) be a right-symmetric algebra with the unital matrix subalgebra \(M_n(F).\) Then \(\dim \ \mathcal A = nk + n^2\) for some \(k \in \mathbb N.\) Let \(\mathcal A = V_2 \oplus M_2,\) where \(V_2\) is the standard two-dimensional module over \(M_2.\) Thus, we may choose a basis \(\{a, b\}\) for \(V_2\) such that \(V_2\) has the following right (non-zero) action of \(M_2:\) \[ a \cdot e_{21}=b, \ b\cdot e_1=b, \ b \cdot e_{12}=a. \] Fix the linear mapping \(\pi : V_2 \to M_2\) such that \(\pi(x,y)=\left( \begin{array}{cc} x& y\\ x& y\end{array} \right).\) It is easy to see that \(\pi\) is a homomorphism of right \(M_2\)-modules. Given a matrix \(A \in M_2,\) we define \(\psi_A : V_2 \to M_2\) by the rule \(\psi_A(x) = \pi(x)\odot A,\) where \(\odot\) is the Hadamard product. Theorem 3. Let \(\mathcal A = W \oplus M_2,\) where \(W\) is an irreducible module over \(\mathfrak{sl}_2.\) Then \(W \cong V_2\) is the standard module over \(M_2\) and there exist \( A \in M_2\) and a product \(\bullet\) on \(W\) such that the product on \(\mathcal A\) is given by the rules \[ B \cdot w = wB + [B,R^\bullet_w], \ w \cdot u = w \bullet u + \psi_A(u)\pi(w) \] for all \(w, u \in W\) and \(B \in M_2.\) Furthermore, the relations from Theorem 1 hold for \(\mathcal A.\)
0 references
right-symmetric algebra
0 references
left-symmetric algebra
0 references
simple algebra
0 references
Koszul-Vinberg algebra
0 references
pre-Lie algebra
0 references
0 references