Solvability and nilpotency of transposed Novikov-Poisson algebras

\eqref{i} \Rightarrow \eqref{ii}: Let B \neq P be a subalgebra of P, where P is nilpotent. Then there exists a minimal integer r \geq 2 such that all products (including Novikov product and associate product) of r elements of P belong to B. So there exists an element z, which is the product of (r-1) elements and does not belong to B. Then we have x\cdot z,z\cdot x ,x\circ z, z\circ x \in B for all x\in P.
\eqref{ii}\Rightarrow\eqref{iii}: Taking B = 0 in (b) we have that there is a non-zero element x_1 \in P such that x_1\cdot y = y\cdot x_1=x_1\circ y=y\circ x_1 = 0 for all y \in P. Now taking B = {\bf k}x_1, which is an ideal of P, we see that there is an element x_2 \in P, x_2 \notin {\bf k}x_1 such that x_2\cdot y,y\cdot x_2,x_2\circ y,y\circ x_2 \in {\bf k}x_1, for all y \in P. Then {\bf k}x_1 + {\bf k}x_2 is an ideal of P. Continuing in this way we construct n linearly independent elements x_1, \dots, x_n such that x_i \notin {\bf k}x_1 + \dots + {\bf k}x_{i-1}, but x_i\cdot y,y\cdot x_i, x_i\circ y, y\circ x_i \in {\bf k}x_1 + \dots + {\bf k}x_{i-1} for all y \in P, 2 \leq i \leq n. This is the required basis.
\eqref{iii}\Rightarrow \eqref{iv}: Let P_{(i)} = {\bf k}x_1 + \dots + {\bf k}x_i (1 \leq i \leq n).
\eqref{iv}\Rightarrow \eqref{i}: It is straightforward.

By [simple, Novikov], B\circ C is an ideal of P_N. By Eq. (?), we obtain

This completes the proof.

It is straightforward.

Since A \cdot A^{k}+A\circ A^{k}+A^{k}\circ A \subseteq A^{k+1} is obvious, we only need to prove that

We proceed by induction on k. If k=1, the statement is obvious. Assume that A^{n+1}\subseteq A^{n}\cdot A+A\circ A^{n}+A^{n}\circ A holds for all 1\leq n\leq k-1. Now consider A^{k+1}. It suffices to show

We prove this by a second induction on i. For i=1, the inclusion is clear. Suppose that A^{j}\cdot A^{k+1-j}+A\circ A^{k+1-j}+A^{k+1-j}\circ A\subseteq A^{k}\cdot A+A\circ A^{k}+A^{k}\circ A holds for all 1\leq j\leq i. Consider A^{i+1}\cdot A^{k-i}+A^{i+1}\circ A^{k-i}+A^{k-i}\circ A^{i+1}. Since i+1\leq k, applying the outer inductive hypothesis, and by Eqs. (?), (?) and (2.6), we obtain

and

This completes the inner induction, and consequently the outer induction. This completes the proof.

We claim that (P^2)^m\subseteq P_L^{m+1} for all m\geq 1. Obviously, it holds for m=1. Suppose that (P^2)^k\subseteq P_L^{k+1} holds for all 1\leq k\leq m. Consider (P^2)^{m+1}=P^2\cdot (P^2)^m+P^2\circ (P^2)^m+(P^2)^m\circ P^2. By the hypothesis and Proposition 2.5, we obtain

This completes the induction. Then this conclusion follows directly.

The proof is similar to that in [Nilpotent, Novikov]. We first claim that \big(P^{(m)}\big)_L^{q}\subseteq \big(P^{(m)}\big)_L^{q-3}\cdot P^{(m+1)}+\big(P^{(m)}\big)_L^{q-3}\circ P^{(m+1)} holds for all integer q\ge4. We prove it by induction on q. If q=4, it is clear. Note that \big(P^{(m)}\big)_L^{q+1}=\big(P^{(m)}\big)_L^{q}\cdot P^{(m)}+\big(P^{(m)}\big)_L^{q}\circ P^{(m)}. By the hypothesis, we obtain

Hence, we complete the induction, and the claim. Next, we prove \big(P^{(m)}\big)_L^{3n}\subseteq \big(P^{(m+1)}\big)_L^n for all n\ge 1, m\ge 0. We prove it by induction on n. When n=1, it is clear. By the hypothesis, we obtain

This completes the induction. Finally, we prove (P^{(m)})_L^{3^n} \subseteq P^{(m+n)} by induction on n. Then we obtain

This completes the conclusion.

It is straightforward by Lemmas 3.1 and 3.2.

By Proposition 2.4, B\circ N is an ideal of P. Since B\circ N\subseteq B, the minimality of B implies that either B\circ N=B or B\circ N=0. Because N is Novikov nilpotent, we must have B\circ N=0. Similarly, we obtain N\circ B=0. Moreover, by Eqs. (?) and (?), we have

and

It follows that B\cdot N is also an ideal of P contained in B. Thus, by the minimality of B, we have either B\cdot N=0 or B\cdot N=B. Since N is associative nilpotent, we conclude B\cdot N=0. Hence, we obtain B\subseteq \operatorname{Ann}_P(N).

If P is nilpotent, then P_A and P_N are nilpotent clearly. Conversely, assume that P_A and P_N are nilpotent. We prove that P is nilpotent by induction on \dim(P). If \operatorname{dim}(P)=1, then P\circ P=P\cdot P=0 and the conclusion holds trivially. Assume that the conclusion holds for all transposed Novikov-Poisson algebras of dimensions at most k. Suppose that \operatorname{dim}(P)=k+1. If P\circ P=0, then P=P_A. There is nothing to prove. Otherwise, by Proposition 2.4, there exists a minimal ideal B of P contained in P\circ P. Consider the quotient P/B. By the induction hypothesis, P/B is nilpotent. Hence there exists a non-negative integer m such that P^m\subseteq B. By the proof of Lemma 3.5, we have B\subseteq \operatorname{Ann}_P(P). Therefore, we obtain

This completes the proof.

It is straightforward by Theorem 3.6.

The proof is similar to that in Lemma 3.5.

The proof is similar to Theorem 3.6 by Lemma 3.5.

The necessity is clear. Now we prove the sufficiency. Since P_A and P_N are solvable, by Lemma 3.11 and Theorem 3.10, P is right nilpotent. Hence, P is solvable by Theorem 3.3.

Obviously, we have Nil(P)\subseteq Nil_N(P). By Theorem 3.3, we obtain that Nil_N(P)\cdot P\subseteq P^2\subseteq Nil(P)\subseteq Nil_N(P), which implies that Nil_N(P) is an ideal of P. Hence, we have Nil(P)=Nil_N(P).

Here we only prove the left nilpotent case. The right nilpotent case is similar. By Proposition 2.5, we obtain that P\cdot B+P\circ B+C is an ideal of P. Thus we have P\cdot B+P\circ B+C\subseteq C or P\cdot B+P\circ B+C=B. Suppose that the latter holds. Since P is left nilpotent, we have

which leads to a contradiction. Hence, P\cdot B+P\circ B\subseteq C. Similarly, if P is right nilpotent, then B\cdot P+B\circ P\subseteq C.

Let B/C be a chief factor of P. Then we have B\cdot P+B\circ P+P\circ B\subseteq C. It implies that \operatorname{dim}(B/C)=1. So there is a chain of ideals of P

where \dim P_{(i)} = i and P\cdot P_{(i)} + P_{(i)} \circ P+P\circ P_{(i)} \subseteq P_{(i-1)}. Hence, P is nilpotent by Theorem 2.2.

It is straightforward by Theorems 3.3 and  3.16.

\eqref{meta2}\Leftrightarrow \eqref{meta3} is obvious.
\eqref{meta1}\Rightarrow \eqref{meta2}: Assume that  (4.1) is the corresponding exact sequence of P. Then we have c_1*c_2 \in \operatorname{Ker}(\pi) = V_1 since \pi(c_1*c_2)=\pi(c_1)*\pi(c_2)\in V_2*V_2=(0), where c_1,c_2\in P and *\in \{\cdot,\circ \}. Hence, we obtain (c_1*_1c_2)*_2(c_3*_3c_4)\in V_1\cdot V_1+V_1\circ V_1= (0), where c_1,c_2,c_3,c_4 \in P and *_1,*_2,*_3\in \{\cdot,\circ\}.
\eqref{meta3}\Rightarrow \eqref{meta1}: If P^1 is an abelian subalgebra of P, then we have an exact sequence:

where i is the inclusion map and \pi the canonical projection.

By Proposition 4.2, we only need to show that (c_1*_1c_2)*_2(c_3*_3c_4)= 0, where c_1,c_2,c_3,c_4 \in P and *_1,*_2,*_3\in \{\cdot,\circ\}. For all c_1,c_2,c_3,c_4\in P, we have (c_1\cdot c_2)\cdot (c_3\cdot c_4)=0 directly. Let c_i=a_i+b_i, where i=1,2,3,4, a_i\in A and b_i\in B. Then we obtain

Similarly, we obtain (c_1\circ c_2)\cdot c_3\cdot c_4=(c_1\cdot c_2)\cdot (c_3\circ c_4)=0. By Eqs. (?) and (?), we obtain

and

for all a,a'\in A,b, b'\in B. Similarly, we have a\cdot (b'\circ a')=b\cdot (a'\circ b')=b\cdot (b'\circ a')=0. Thus, we can obtain (c_1\circ c_2)\cdot (c_3\circ c_4)=(c_1\cdot c_2)\circ (c_3\circ c_4)=(c_1\circ c_2)\circ (c_3\cdot c_4)=0 directly. And (c_1\circ c_2)\circ (c_3\circ c_4)=0 follows from Itô’s theorem for Novikov algebras [Nilpotent, Novikov].