On the uniform dimension of subextensions in skew polynomial rings BertrandNguefack Abstract . This work investigates the invariance of the non-necessarily finite uniform dimension and related concepts for subextensions in skew polynomial rings 𝕊 = R ⁡ [ X ; α , δ ] \mathbb{S}=R[\mathbf{\mathrm{X}};\mathbf{\alpha},\mathbf{\delta}] of bijective type over a well-ordered set of variables. When the coefficient ring has enough uniform left ideals, in the commuting variables case we show that classical results on this topic for polynomial rings extend to subextensions of skew Laurent polynomial rings 𝕊 = R ⁡ [ X ± 1 ; α ] \mathbb{S}=R[\mathbf{\mathrm{X}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}};\mathbf{\alpha}] , generated over R R by any family of (standard) terms. The situation in the non-commuting variables context is more complex; easily formed polynomial-like subrings can behave very oddly from the ambient ring. We provide easy examples of a (semi)prime left Goldie skew polynomial ring of bijective type containing a monoid subring isomorphic to a free non-commutative polynomial ring. We then study the so-called subclass of essentially special subextensions and obtain for them the preservation of the uniform dimension and related concepts. 2020 Mathematics Subject Classification 16P60, 16P70, 16S36, 16S38 1. Introduction Throughout this paper, every ring is associative and has an identity; R denotes a ground ring (of coefficients); a ring extension of R is alternatively termed as an R -ring . An R -ring morphism is any ring map between R -rings that restricts to the identity map on the ground ring; a subextension of an R -ring S is generally any R -subring of S . The uniform (or Goldie) dimension \mathsf{udim}(M) of a module M (over some ring) is the supremum of those cardinal numbers \mathfrak{N} such that M contains an \mathfrak{N} -indexed direct sum of non-zero submodules [ GW04 , DGALP06 ] . Except otherwise specified, the uniform dimension of a ring S mean its left uniform dimension denoted and given by \mathsf{udim}(S)=\mathsf{udim}({{}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}S}}}S}) . Left Goldie conditions for a ring mean the finiteness of the uniform dimension together with the ascending chain condition on left annihilators. It is well-known that a subring needs not have the same uniform dimension as the ambient ring. Starting with a classical paper by [ Sho72 ] on polynomial rings with finite right uniform dimension, the invariance of the uniform dimension and the preservation of Goldie conditions under ring extensions has been an active research topic till date. Due to the importance of polynomial rings, and more generally, of skew polynomial rings and related polynomial-like rings, Shock’s result was immediately extended to group rings by [ Wil73 ] , and latter on, to (finitely) iterated Ore extensions of bijective type by [ Mat95 ] ; the case of iterated Ore extensions of injective type over semiprime left Goldie ring was dealt with by [ LM03 ] , and that of skew PBW extensions is started by [ Rey14 ] . It is a well-established fact that the uniform dimension and Goldie conditions have important applications in ring theory itself [ Jat72 , CC73 , Sho74 , CG86 , MR87 , Ker90 , Fai96 , VHJLP03 ] and references therein, as well as in module theory [ Cam77 , AHJLP92 , LM04 ] . Polynomial-like subrings of skew polynomial rings can behave very strangely from ordinary skew polynomial rings; however they arise naturally and in various specialized shapes in algebra and geometry. The present work considers an ambient R -ring \mathbb{S} together with its subextensions, where \mathbb{S} is ❍ either a multivariate skew Laurent polynomial ring R[\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}};\boldsymbol{\alpha}] of bijective type with commuting variables, ❍ or a multivariate skew polynomial ring R[\boldsymbol{X};\boldsymbol{\alpha},\boldsymbol{\delta}] of bijective type over an arbitrary well-ordered set \boldsymbol{X} of independent variables. Thus as left R -module, \mathbb{S} is free with basis given by the set \mathbb{T} of standard terms: for the skew polynomial ring R[\boldsymbol{X};\boldsymbol{\alpha},\boldsymbol{\delta}] , \mathbb{T}=\mathbb{T}(\boldsymbol{X}) consist of all ordered product x_{1}\dotsm x_{n} with n\in\mathbb{N} and x_{1}\leq x_{2}\leq\dotsm\leq x_{n} in \boldsymbol{X} ; and in the case of the skew Laurent polynomial ring in commuting variables, \mathbb{T}=\mathbb{T}(\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}}) simply consists of all commutative words on the alphabet \boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}}=\boldsymbol{X}\cup\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-1}}} with \boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-1}}}=\{x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-1}}}\hskip 3.0pt\colon\hskip 3.0ptx\in\boldsymbol{X}\} . We then address the following problem that arises naturally. Problem 1.1 . Let A be a polynomial-like subextension of \mathbb{S} ; essentially one may restrict attention to the case that A contains R and is generated as left R -module by a subset of standard terms. Which conditions on the shape of A will ensure that: \mathsf{udim}(A)=\mathsf{udim}(\mathbb{S})=\mathsf{udim}(R) , while A is left Goldie provided so is R ? A specialization of Problem 1.1 to the commuting variables context follows. Problem 1.2 . Let A be any subextension of the skew Laurent polynomial ring R[\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}};\boldsymbol{\alpha}] , generated over R by a family of standard terms from \mathbb{T}(\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}}) . Does it hold that \mathsf{udim}(A)=\mathsf{udim}(R) ? Is it true that A is ((semi)prime) left Goldie precisely when so is R ? The feeling about Problem 1.2 is that the answer should be positive, at least when the ambient ring is simply a polynomial ring R[\boldsymbol{X}] . However even in the latter case, it appears that one may not reach the expected positive answer by relying on the classical approach [ Sho72 ] . The issue becomes highly complex and very challenging in the non-commuting variables context of Problem 1.1 . Content and main results of the paper In section 2 , we review the invariance of the non-necessarily finite uniform dimension for arbitrary ring extensions. Indeed in the original paper by Shock and subsequent papers on the topic, it was tacitly or explicitly assumed that the ground ring has finite uniform dimension. For a non-zero left module M over some ring, we say that M has enough uniform submodules to mean that every non-zero submodule in M contains a uniform submodule. A key step towards the invariance of the uniform dimension is provided by Proposition 2.3 : M has enough uniform submodules if and only if M contains an essential submodule which is a direct sum \oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}} of uniform submodules for some indexing set I ; in this case, \mathsf{udim}(M) equals the cardinal number \lvert I\rvert . This key characterization allows to provide a general and quick strategy guaranteeing the invariance of the uniform dimension under general ring extension (Proposition 2.4 ). Specifically, we introduce and study a concept of nicely essential subring A of a given ring S (Definition 2.5 ); it results that A and S share many properties in common including the invariance of the uniform dimension, the preservation of Goldie conditions and (semi)primeness; see Propositions 2.6 and 2.7 . Section 3 focusses on the commuting variables setting since this case is enlightening for the more complex non-commuting setting. Relying only on the general strategy investigated in section 2 , we solve Problem 1.2 completely in Theorems 3.3 and 3.5 . In Section 4 , we address in full generality Problem 1.1 and formally introduce and investigate essentially special subextensions of \mathbb{S} . We then lift to this general setting the main classical results [ Sho72 , Mat95 ] , while allowing infinite uniform dimension. More precisely, Theorems 4.12 and 4.15 establish the invariance of the uniform dimension and the computation of left singular ideals for essentially special subextensions A . The main strategy is proving the lifting of uniform or essential left ideals from the ground ring R to A , generalizing classical techniques. The final theorem 4.17 is an application to (semi)prime left Goldie rings. Section 5 provides a family of essentially special subextensions in the two-variate case, highlighting the fact that familiar subextensions in \mathbb{S} generally fail to be skew polynomial rings in any meaningful way, while still having the same uniform dimension as \mathbb{S} . Though our results are sharp for an ambient skew polynomial ring of bijective type, an obvious perspective would be to further extend the class of ‘specialized’ subextensions of \mathbb{S} for which the invariance of the uniform dimension holds. 2. Invariance of the uniform dimension under general ring extensions We consider general ring extensions R\subset S and provide some strategy allowing, on one hand to quickly compare the uniform dimension of both rings R and S , and on the other hand to derive Goldie conditions for subrings of left Goldie rings. While \mathbb{N} denotes the set of natural numbers including 0 , for m,n\in\mathbb{Z} we write \llbracket m\hskip 1.00006pt,\hskip 1.00006ptn\rrbracket for the range of all integers k with m\leq k\leq n . For every subset U of in a left S -module M , the left annihilator of U in S is the left ideal \mathPzc{a}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}S}}}(U)=\{a\in S\hskip 3.0pt\colon\hskip 3.0ptaU=0\} ; it is an ideal of S whenever U is a submodule in M . We usually omit the subscript in \mathPzc{a}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}S}}}(U) when S is the base ring or when it is implied by the context. The singular submodule of M is defined as \boldsymbol{\zetaup}(M)=\{x\in M\hskip 3.0pt\colon\hskip 3.0pt\mathPzc{a}(x)\text{ contains an essential left ideal in }S\} . We write \boldsymbol{\zetaup}(S)=\boldsymbol{\zetaup}({{}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}S}}}S}) for the left singular ideal of the ring S . The notation for singular submodule is being borrowed from [ MR87 , §2.2.4] . Remark 2.1 . The ascending chain condition and the descending chain condition on left annihilators are inherited by subrings. Thus, a subring of a left Goldie ring is left Goldie if and only it has finite uniform dimension. 2.1. Module with enough uniform submodules Definition 2.2 . We say that a non-zero module M over some ring has enough uniform submodules provided every non-zero submodule in M contains a uniform submodule. Likewise, a ring S has enough uniform left ideals if every non-zero left ideal in S contains a uniform left ideal of S . Our first key result characterizes those modules whose uniform dimension is attained (in a strong sense), generalizing the finite dimensional case. Proposition 2.3 . Let M be any non-zero left module over some ring. Then M has enough uniform submodules if and only if M contains an essential submodule which is a direct sum \oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}} of uniform submodules for some indexing set I . In this case, \mathsf{udim}(M) coincides with the cardinal number \lvert I\rvert . Proof. Start with the assumption that M has enough uniform submodules. Thus the family of uniform submodules of M is non-empty and, by an obvious use of Zorn’s lemma, we may consider a maximal independent set \mathcal{U}=\{M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}\hskip 3.0pt\colon\hskip 3.0pti\in I\} of uniform submodules in M . Then the submodule \oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} must be essential in M . Indeed assuming the contrary, there will exist a non-zero submodule N\subset M with (\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}})\cap N=0 , while by assumption N contains some uniform submodule N^{\prime} . This yields a larger independent set \mathcal{U}\cup\{N^{\prime}\} , contradicting the maximality of \mathcal{U} . Hence as expected, M contains an essential submodule which is a direct sum of uniform submodules. Conversely, assume that M contains a direct sum M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}I}}}=\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}} of uniform submodules such that M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}I}}} is essential in M . Our gold is to prove that every non-zero submodule in M contains a uniform submodule and \mathsf{udim}(M)=\lvert I\rvert . In the case of a finite cardinal number n\geq 1 , it is well-known [ GW04 , Lemma 5.16, Proposition 5.20] that: a module has uniform dimension n precisely when it contains an essential submodule which is a direct sum of n uniform submodules. Thus we may continue with the case that \lvert I\rvert is infinite. Let N\subset M be any non-zero submodule; for every finite subset \Lambda\subset I we also let M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}=\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in\Lambda}}}M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} . Since M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}I}}} is essential in M , the module N\cap M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}I}}}=N\cap(\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}) is non-zero and consequently, for some finite non-empty subset \Lambda\subset I the module N\cap M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}} is non-zero as well. But then (by the final cardinal case), \mathsf{udim}(M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}})=\lvert\Lambda\rvert is finite and \mathsf{udim}(N\cap M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}})\leq\lvert\Lambda\rvert is also finite and non-zero. Thus, N\cap M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}} contains a uniform submodule. Now assuming that M contains a direct sum \oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s\in I^{\prime}}}}M_{s}^{\prime} of non-zero submodules with \lvert I^{\prime}\rvert\geq\lvert I\rvert , our next aim is to prove that \lvert I^{\prime}\rvert=\lvert I\rvert . Write \aleph_{{\cramped[\scriptstyle]{\mspace{-0.72mu}0}}}=\lvert\mathbb{N}\rvert for the countable cardinal; for each finite number n\in\mathbb{N} , set \mathcal{E}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}}=\{\Lambda\subset I\hskip 3.0pt\colon\hskip 3.0pt\lvert\Lambda\rvert=n\} . For every s\in I^{\prime} , since M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}}^{\prime} is a non-zero submodule of M while M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}I}}} is an essential submodule of M , the module M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}}^{\prime}\cap M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}I}}} must contain a non-zero cyclic submodule N_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}} and there is some finite subset \Lambda_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{s}}}}\subset I with N_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}}\subset M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{s}}}}}}} . Next fix 1\leq n\in\mathbb{N} . For every n -element subset \Lambda\subset I , form the subset: I_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}^{\prime}=\{s\in I^{\prime}\hskip 3.0pt\colon\hskip 3.0ptN_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}}\subset M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}\} , thus the finite direct sum M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}=\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in\Lambda}}}M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} of uniform submodules contains the direct sum N_{{\cramped[\scriptstyle]{\mspace{-0.72mu}{I_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}^{\prime}}}}}=\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s\in I_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}^{\prime}}}}N_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}} of non-zero cyclic submodules. But since M_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}} cannot contain a direct sum of more than n non-zero submodules, it holds that \lvert I_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}^{\prime}\rvert\leq n . Since moreover each element s\in I^{\prime} is by definition a member of I_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{s}}}}}}}^{\prime} , we also have that I^{\prime}=\cup_{{\cramped[\scriptstyle]{\mspace{-0.72mu}1\leq n\in\mathbb{N}}}}\cup_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda\in\mathcal{E}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}}}}}I_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}^{\prime} . But for each 1\leq n\in\mathbb{N} , the set I clearly embeds into each \mathcal{E}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}} and the latter embeds into the n -fold Cartesian product I^{n} , and since I is an infinite cardinal it follows that \lvert I\rvert\leq\lvert\mathcal{E}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}}\rvert\leq\lvert I^{n}\rvert=\lvert I\rvert . Hence for each 1\leq n\in\mathbb{N} and every \Lambda\in\mathcal{E}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}} , we have shown that \lvert I_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}^{\prime}\rvert\leq n and \lvert\mathcal{E}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}}\rvert=\lvert I\rvert , whence the following: \lvert I^{\prime}\rvert=\lvert\cup_{{\cramped[\scriptstyle]{\mspace{-0.72mu}1\leq n\in\mathbb{N}}}}\cup_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda\in\mathcal{E}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}}}}}I_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\Lambda}}}^{\prime}\rvert\leq\aleph_{{\cramped[\scriptstyle]{\mspace{-0.72mu}0}}}\lvert I\rvert=\lvert I\rvert. This shows that \lvert I^{\prime}\rvert=\lvert I\rvert , completing the proof of the proposition. ∎ 2.2. General strategies ensuring the invariance of the uniform dimension and related concepts The classical strategy successful for investigating the invariance of the finite uniform dimension for skew polynomial rings is summarized for arbitrary ring extensions by the following proposition, allowing infinite uniform dimension. Proposition 2.4 . Let A be any subring of an R -ring such that A\cap AL is essential as left A -submodule in AL for every non-zero left ideal L in R , and for all left ideals L,L^{\prime} with L\cap L^{\prime}=0 , AL\cap AL^{\prime}=0 . Then \mathsf{udim}(R)\leq\mathsf{udim}(A) . Assume additionally that if L is any uniform or essential left ideal in R , then so is the left ideal A\cap AL in A . Then, if R has enough uniform left ideals then so does A and \mathsf{udim}(A)=\mathsf{udim}(R) . Proof. To see that \mathsf{udim}(R)\leq\mathsf{udim}(A) , given any direct sum \oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}L_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} of non-zero left ideals in R , we quickly check that one gets a direct sum \oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}(A\cap AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}) of non-zero left ideals in A . By assumption, each A\cap AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} is already a non-zero left ideal in A for all i\in I . And if i_{1},\ldots,i_{n} are pairwise distinct indices from I for some integer n\geq 2 , then for every s\in\llbracket 1\hskip 1.00006pt,\hskip 1.00006ptn\rrbracket , since L_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}}}}}\cap\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}k\neq s}}}L_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i_{{\cramped[\scriptstyle]{\mspace{-0.72mu}k}}}}}}=0 , the first assumption from the proposition grants that AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i_{{\cramped[\scriptstyle]{\mspace{-0.72mu}s}}}}}}\cap\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}k\neq s}}}AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i_{{\cramped[\scriptstyle]{\mspace{-0.72mu}k}}}}}}=0 . Thus the sum \sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}(A\cap AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}) is still a direct sum of non-zero left ideals in A . Next continue by assuming that R has enough uniform left ideals and, if L is a uniform (or an essential) left ideal in R then so is the left ideal A\cap AL in A . Thus by virtue of Proposition 2.3 , the uniform dimension of R is attained in the stronger sense that R contains an essential left ideal arising as direct sum L=\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}L_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} of uniform left ideals, with \lvert I\rvert=\mathsf{udim}(R) . But then, the left ideal A\cap AL is essential in A and it contains the direct sum \widetilde{L}=\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in I}}}(A\cap AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}) of uniform left ideals of A . Let us show that \widetilde{L} is also essential as left ideal in A . Given any non-zero left ideal J in A , since J\cap AL=J\cap A\cap AL\neq 0 , there is a finite non-empty subset \Lambda\subset I such that J\cap\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in\Lambda}}}AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}\neq 0 . But since by assumption each A\cap AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} is an essential submodule of the left A -module AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} for all i\in\Lambda , it follows that \oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in\Lambda}}}(A\cap AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}) is still an essential submodule of the left A -module \oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in\Lambda}}}AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} , whence J\cap\oplus_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i\in\Lambda}}}(A\cap AL_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}})\neq 0 as well. Thus we have shown that \widetilde{L} is an essential left ideal in A , arising as I -indexed direct sum of uniform left ideals. Now another invocation of Proposition 2.3 yields that \mathsf{udim}(A)=\lvert I\rvert=\mathsf{udim}(R) and A has enough uniform left ideals. ∎ It is well-known that the uniform dimension is preserved under left localization with respect to a left Ore set of regular elements, [ MR87 , Lemma 2.2.12] . A slightly more general situation is captured by the following concept of nicely essential ring extension, as we will show. Definition 2.5 . Let L be a submodule of a left R -module M . We say that L is nicely essential in M or M is a nicely essential extension 1 1 1 An alternative terminology could have been ‘ strongly essential submodule ’, but the latter already appears in few places in the literature (for different purposes and) to mean a submodule L of an R -module M such that every I -indexed direct product of copies of L is essential in the corresponding I -indexed direct product of copies of M for all set I . For a submodule L of an R -module M , while the condition for L to be ‘strongly essential’ in M does not implies the property for L to be a nicely essential submodule in M , the latter is evidently stronger than requiring that every direct sum of copies of L is essential in the corresponding direct sum of copies of M . of L provided for every finite set E\subset M of non-zero elements, there exists a\in R with 0\neq ax\in L for all x\in E . If additionally it holds for every x\in M that cx\in L for some left regular element c of R , then we call L a strongly nicely essential submodule of M , (lacking a more shorter terminology). By a ( strongly ) nicely essential subring of a ring S we mean any subring A that is a (strongly) nicely essential left A -submodule in S . In this case, we also call S a ( strongly ) nicely essential ring extension of A . Proposition 2.6 . The following statements hold for a nicely essential subring A of a ring S . \mathsf{(a)} If L,L^{\prime} are left ideals in A with L\cap L^{\prime}=0 then SL\cap SL^{\prime}=0 ; \mathsf{udim}(S)=\mathsf{udim}(A) . The left singular ideal \boldsymbol{\zetaup}(A) is a nicely essential left A -submodule in \boldsymbol{\zetaup}(S) with \boldsymbol{\zetaup}(A)=A\cap\boldsymbol{\zetaup}(S) . \mathsf{(b)} Let L\subset S be any non-zero left A -submodule. Then the left ideal A\cap L is uniform (or resp., essential) in A if and only if so is the left ideal SL in S . Proof. Starting with the first claim, since A is at least essential as left A -submodule in S , it is already granted that \mathsf{udim}(S)\leq\mathsf{udim}(A) . Next assuming that L,L^{\prime} are two left ideals in A with L\cap L^{\prime}=0 , we will show that SL\cap SL^{\prime}=0 . Let v\in SL\cap SL^{\prime} ; then v=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n}}}}v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n^{\prime}}}}}v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}^{\prime}u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}^{\prime} for some n,n^{\prime}\in\mathbb{N} , and v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}},v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}j}}}^{\prime}\in S,\,u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}\in L,u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}j}}}^{\prime}\in L^{\prime} for 1\leq i\leq n and 1\leq j\leq n^{\prime} . The main assumption on S applied to the finite set E=\{v,v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}1}}},\ldots,v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}},v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}1}}}^{\prime},\ldots,v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n^{\prime}}}}^{\prime}\} yields some a\in A such that for every (i,j)\in\llbracket 1\hskip 1.00006pt,\hskip 1.00006ptn\rrbracket\times{\llbracket 1\hskip 1.00006pt,\hskip 1.00006ptn^{\prime}\rrbracket} we have: av_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}},av_{{\cramped[\scriptstyle]{\mspace{-0.72mu}j}}}^{\prime}\in A,\,av=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n}}}}(av_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}})u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n^{\prime}}}}}(av_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}^{\prime})u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}^{\prime}\in L\cap L^{\prime} and av is non-zero whenever so is v . But since L\cap L^{\prime}=0 , we must have av=0 and v=0 as well. Hence, SL\cap SL^{\prime}=0 . This also grants (just as in the first claim of Proposition 2.4 ) that \mathsf{udim}(A)\leq\mathsf{udim}(S) ; whence the equality, \mathsf{udim}(S)=\mathsf{udim}(A) . Turning to the claim about left singular ideals, the assumption that A is a nicely essential left A -submodule of S already grants for every left ideal J of S that A\cap J is still a nicely essential left A -submodule of J . Also, \boldsymbol{\zetaup}(A)\subset A\cap\boldsymbol{\zetaup}(S) since A is already an essential A -submodule in S . For the converse inclusion, let x\in A\cap\boldsymbol{\zetaup}(S) and 0\neq y\in A . The definition of the left singular ideal \boldsymbol{\zetaup}(S) yields some b\in S with by\neq 0 and by\cdot x=0 , so that by the assumption on S , there must also exist some a\in A with 0\neq ab,ab\cdot y\in A while (ab\cdot y)\cdot x=0 and x\in\boldsymbol{\zetaup}(A) . Whence, A\cap\boldsymbol{\zetaup}(S)\subset\boldsymbol{\zetaup}(A) . Next for statement \mathsf{(b)} , assuming that A\cap L is a uniform left ideal in A , let us show that SL is uniform as left ideal in S . So let f=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{m}}}}f_{i}u_{i} and g=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{j=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n}}}}g_{j}v_{j} be non-zero elements in SL with 0\neq m,n\in\mathbb{N} , f_{i},g_{j}\in S and u_{i},v_{j}\in L for 1\leq i\leq m , 1\leq j\leq n . Applying the main assumption to the finite subset \{f,g,f_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}},g_{{\cramped[\scriptstyle]{\mspace{-0.72mu}j}}}\,\hskip 3.0pt\colon\hskip 3.0pt1\leq i\leq m,\,1\leq j\leq n\} yields some v\in A with vf,vg,vf_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}},vg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}j}}}\in A (for all (i,j)\in\llbracket 1\hskip 1.00006pt,\hskip 1.00006ptm\rrbracket\times\llbracket 1\hskip 1.00006pt,\hskip 1.00006ptn\rrbracket ) and vf,vg\neq 0 . In particular, the non-zero elements vf and vg already live in the uniform left ideal A\cap L of A and there are two elements a,b\in A with 0\neq avf=bvg ; whence SL is uniform as left ideal of S . Conversely, assuming that SL is uniform as left ideal in S , we quickly check that A\cap L is uniform as left in A . Since A is already essential as left A -submodule in S , A\cap L must be non-zero. Given any non-zero elements x,y\in A\cap L , by assumption there exist c,d\in S with 0\neq cx=dy and another application of the main assumption yields some v\in A with vc,vd\in A and 0\neq(vc)x=(vd)y , which shows that A\cap L is a uniform left ideal in A . Next, to complete the proof of statement \mathsf{(b)} , it remains to address the claim about essential left ideals. Since A is already essential as left A -submodule in S , if A\cap L is essential as left ideal in A , then it is still essential as left A -submodule in S and so the left ideal SL is essential in S . Conversely, supposing that SL is essential as left ideal of S , let us check that A\cap L is essential as left ideal of A . Given any non-zero x\in A , it holds for some v\in S that 0\neq vx=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{m}}}}v_{i}u_{i}\in SL for some positive m\in\mathbb{N} and some (v_{i},u_{i})\in S\times L with v_{i}u_{i}\neq 0 for 1\leq i\leq m . The main assumption on S applied to the subset \{vx,v_{i}\hskip 3.0pt\colon\hskip 3.0pt1\leq i\leq m\} yields some a\in A with 0\neq avx,av_{i}\in A for all i\in\llbracket 1\hskip 1.00006pt,\hskip 1.00006ptm\rrbracket , so that, 0\neq avx=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{m}}}}(av_{i})u_{i}\in A\cap L and A\cap L is essential as left ideal in A as expected. This completes the proof of the proposition. ∎ We now characterize (semi)primeness under nicely essential extension. Here, ‘(semi)prime’ is a short form for ‘prime (or respectively, semiprime)’. Proposition 2.7 . Let A be a nicely essential subring of a ring S . Then, if A is (semi)prime then so is S ; if additionally A is left Goldie then so is S . Conversely, if S is left Goldie, then so is A . If S is (semi)prime left Goldie, then so is every strongly nicely essential subring in S . Proof. Starting with the first statement of the proposition and assuming that A is prime, we will show that so is S . Let I,J be ideals in S with IJ=0 and J\neq 0 . Since A is essential as left A -submodule in S , J\cap A is a non-zero ideal of A while (I\cap A)(J\cap A)=0 , which yields that I\cap A=0 and I=0 . Hence, S is prime. The corresponding statement for semiprimeness follows by letting I=J . Now, since a prime ring is semiprime while (by Goldie’s theorem [ MR87 , §2.3.6] ) a semiprime left Goldie ring is precisely a semiprime ring with finite uniform dimension and with zero left singular ideal, the second part of the first statement of the proposition follows thanks to Proposition 2.6 \mathsf{(a)} For the converse, suppose that S is left Goldie. Then since the ascending chain condition on left annihilators is inherited by subrings, A is left Goldie by virtue of Proposition 2.6 \mathsf{(a)} . Now assume the additional property that A is a strongly nicely essential subring in S . Assuming that S is semiprime and given any ideal N in A with N^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}=0 , let us show that N is the zero ideal. The right annihilator ideal I=\mathsf{r}\hskip-1.99997pt.\hskip-1.00006pt\mathPzc{a}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}A}}}(N) is known to be essential as left ideal in A . Indeed for each non-zero left ideal L in A , it holds for some k\in\{0,1\} that N^{{\cramped[\scriptstyle]{\mspace{-0.5mu}k}}}L\neq 0 and N^{{\cramped[\scriptstyle]{\mspace{-0.5mu}k+1}}}L=0 , hence 0\neq N^{{\cramped[\scriptstyle]{\mspace{-0.5mu}k}}}L\subset L\cap\mathsf{r}\hskip-1.99997pt.\hskip-1.00006pt\mathPzc{a}(N) . It follows by Proposition 2.6 \mathsf{(b)} that SI is also an essential left ideal of S . But then, because S is semiprime left non-singular with finite uniform dimension, we know (thanks to [ MR87 , Proposition 2.3.5] ) that SI contains some regular element c . Write c=u_{1}a_{1}+\dotsm+u_{n}a_{n} with u_{1},\ldots,u_{n}\in S and a_{1},\ldots,a_{n}\in I . By the additional assumption on A , there are left regular elements c_{1},\ldots,c_{n} of A with c_{n}u_{n},c_{n-1}c_{n}u_{n-1},\ldots,c_{1}\dotsm c_{n}u_{1}\in A . In particular, the product c^{\prime}=c_{1}c_{2}\dotsm c_{n}c is necessarily a left regular element of A that now lives in I ; thus Nc^{\prime}=0 which yields N=0 as expected. Finally, assuming that S is prime, we wish to show that A is prime as well. We know (by Goldie’s theorem [ MR87 , §2.3.6] ) that the set \mathscr{C}(S) of all regular elements of S is a left Ore set and the associated left Goldie quotient Q of S is a simple ring. The following arguments generalize those proving that S is prime from the assumption that Q is simple. Call an element c\in S , left A -regular provided the left annihilator \mathPzc{a}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}A}}}(c)=\{a\in A\hskip 3.0pt\colon\hskip 3.0ptac=0\} vanishes. Let I,J be any non-zero ideals in A . Since Q is simple, we have QJQ=Q and 1=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n}}}}d_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{-1}}}}u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}s_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{-1}}}}v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} for some positive n\in\mathbb{N} and some u_{i},v_{i}\in S , d_{i},s_{i}\in\mathscr{C}(S) and a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}\in J for 1\leq i\leq n . Letting d\in\mathscr{C}(S) be a common left denominator for the d_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} ’s, we get that d=d\times 1=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i=1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n}}}}u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}^{\prime}a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}s_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{i}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{-1}}}}v_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}} for some u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}i}}}^{\prime}\in S , 1\leq i\leq n . Now applying the additional assumption for A , there are left regular elements c_{1},\ldots,c_{n} of A with c_{n}u_{n}^{\prime},c_{n-1}c_{n}u_{n-1}^{\prime},\ldots,c_{1}\dotsm c_{n}u_{1}^{\prime}\in A . But then, since moreover d is a regular element of the larger ring S , the product c=c_{1}c_{2}\dotsm c_{n}d is necessarily a left A -regular element which now lives in JQ . It follows in particular that IJQ\supset Ic\neq 0 . Hence IJ\neq 0 and A is prime. This completes the proof of the proposition. ∎ 2.3. Easy examples illustrating the invariance of the uniform dimension or its failure Example 2.8 . Fix 1\leq q\in\mathbb{N} and two q -roots of unity \varepsilonup,\varepsilonup^{\prime} in the center of R . Consider the skew polynomial ring \mathbb{S}=R[x,y][z;\alpha] where \alpha is the R -ring automorphism with \alpha(x)=\varepsilonup y and \alpha(y)=\varepsilonup^{\prime}x ; clearly \alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{2q}}}} is the identity map. The subextension A\subset\mathbb{S} generated by x,z is free as left R -module and its left R -basis is the following subset of terms: \mathsf{T}=\{x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}}y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}n}}}z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}p}}}\hskip 3.0pt\colon\hskip 3.0ptm,n,p\in\mathbb{N}\text{ and, if }n\neq 0\text{ then }p\neq 0\} . The ring A cannot be presented as skew polynomial ring (even if one allows an infinite set of variables). But interestingly, the term z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2q}}}\in\mathsf{T} is a central element of \mathbb{S} with z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2q}}}\mathbb{S}\subset A . Thus A is a nicely essential subring in \mathbb{S} , in particular Proposition 2.6 grants that \mathsf{udim}(A)=\mathsf{udim}(\mathbb{S}) . Example 2.9 . A three-variate (semi)prime left Goldie graded skew polynomial ring of bijective type over R admitting a monoid R -subring that is isomorphic to a free non-commutative polynomial ring. Modifying the previous example, consider the graded Ore extension \mathbb{S}=R[x,y][z;\alpha] where this time \alpha is the R -ring automorphism with \alpha(x)=y and \alpha(y)=x+y . Thus we know that \mathbb{S} is (semi)prime left Goldie precisely when so is the coefficient R . Letting (a_{n})_{n} denote the Fibonacci series with a_{0}=0,a_{1}=1 and a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n+1}}}=a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n-1}}}+a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}} for all n\geq 1 , we have: \alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n}}}}(x)=a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n-1}}}x+a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}}y\text{ and }\alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n}}}}(y)=\alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n+1}}}}(x)=a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}}x+a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n+1}}}y\text{ for }1\leq n\in\mathbb{N}. It easily follows that \alpha will have finite order precisely when R has a non-zero characteristic. Assume that R has characteristic 0 , and let u=xz,\,v=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}z . Then the terms u and v lie in the monoid \mathscr{C}(\mathbb{S}) of all regular elements of \mathbb{S} since the variables x,y,z are regular in \mathbb{S} . We claim that the submonoid \Gamma=\cramped{\langle u,v\rangle}\subset\mathscr{C}(\mathbb{S}) generated by u and v is the free monoid on two non-commuting variables. Indeed given any k -length word w=u_{1}u_{2}\dotsm u_{k} with u_{i}\in\{u,v\} for 1\leq i\leq k , we have: \deg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}z}}}(u_{i})=1 and \deg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}z}}}(w)=k ; thus two words in the alphabet \{u,v\} , with different lengths, cannot be equal as elements of \Gamma . Next, there are exactly four distinct elements of \Gamma having length 2 : u^{2}=xyz^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}},\,uv=xy^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}},vu=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}yz^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}\text{ and }v^{2}=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}. The 8 three-length words in terms of u and v are also pairwise distinct in \Gamma : \displaystyle u^{3} \displaystyle=xy(x+y)z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}3}}}, \displaystyle v^{3} \displaystyle=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}(x+y)^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}3}}}; \displaystyle u^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}v \displaystyle=xy(x+y)^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}3}}}, \displaystyle v^{2}u \displaystyle=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}(x+y)z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}} \displaystyle uvu \displaystyle=xy^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}(x+y)z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}3}}}, \displaystyle vuv \displaystyle=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y(x+y)^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}3}}}, \displaystyle uv^{2} \displaystyle=xy^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}(x+y)^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}, \displaystyle vu^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}} \displaystyle=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y(x+y)z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}3}}}. Let \Gamma_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{\alpha}}}} be the submonoid of \mathscr{C}(\mathbb{S}) generated by the elements \alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n}}}}(x),\,1\leq n\in\mathbb{N} . For every k -length word w=u_{1}u_{2}\dotsm u_{k} with k\geq 3 and u_{1},\ldots,u_{k}\in\{u,v\} , there exists some \tau\in\Gamma_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{\alpha}}}} with w=x\tau z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}k}}} if u_{1}=u or w=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}\tau z^{{\cramped[\scriptstyle]{\mspace{-0.5mu}k}}} if u_{1}=v . Since moreover, \gcd(x,\tau)=1 for all \tau\in\Gamma_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{\alpha}}}} , these last observations together with the already computed 2 -length words in the alphabet \{u,v\} show that for all k,l\in\mathbb{N} and terms u_{1},\ldots,u_{k},v_{1},\ldots,v_{l}\in\{u,v\} , the products u_{1}\dotsm u_{k} and v_{1}\dotsm v_{l} are equal in \mathbb{S} precisely when k=l and u_{i}=v_{i} for all 1\leq i\leq k . Hence the submonoid \Gamma=\cramped{\langle u,v\rangle}\subset\mathscr{C}(\mathbb{S}) is the free monoid on two non-commuting variables. Consequently, the monoid subring R[u,v]=R\Gamma\subset\mathbb{S} is a two-variate free non-commutative polynomial ring, and it has infinite left or right uniform dimension (even when the ambient ring is (semi)prime left Goldie). The subring R[u,v] in the previous example is a very easy example of a monoid R -subextension of \mathbb{S} generated by a subset of standard terms that obviously fails to be essentially special in \mathbb{S} (as introduced by Definition 4.7 ). This shows that the main results of section 4.3 are also sharp. For (semi)prime left Goldie skew polynomial rings of injective type over R , it is even more easy to form monoid R -subextensions with infinite uniform dimension. Example 2.10 . A family of two-variate (semi)prime left Goldie skew polynomial rings of injective type over R containing a two-variate free non-commutative polynomial ring. Fix an integer e\geq 2 and consider the monoid ring \mathbb{S}=R\Gamma_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{e}}}} where \Gamma_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{e}}}}=\cramped{\langle x,y\hskip 3.0pt\colon\hskip 3.0ptyx=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}e}}}y\rangle} is the quotient of the free monoid on two generators x,y modulo the relation that yx=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}e}}}y . Alternatively, \mathbb{S}=R[x,y;\alpha]=R[x][y;\alpha] is the graded skew polynomial ring where \alpha:R[x]\mspace{1.8mu}\hbox to31pt{\vbox to5.44pt{\pgfpicture\makeatletter\hbox{\hskip 3.0pt\lower-2.44pt\hbox to0.0pt{\lxSVG@begingroup@{_scopebegin} \lxSVG@begingroup@{stroke} \lxSVG@begingroup@{fill} \lxSVG@setlinewidth{\the\pgflinewidth}\lxSVG@begingroup@{stroke-width} \lx@inpgf@ignorespaces\nullfont\lxSVG@begingroup@{_scopebegin} \lxSVG@closescope \hbox to0.0pt{\lxSVG@begingroup@{_scopebegin} { {{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{{}{}}}{{}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{-2.0pt}{0.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} { {}{}{}}{{{{}{}{}}}}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{27.00005pt}{0.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} {{}{}{}}{}{\lx@inpgf@ignorespaces}{}{{}}{}{ {}{}{}}{}{\lx@inpgf@ignorespaces}{}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {\lx@inpgf@ignorespaces}{{}}{} {\lx@inpgf@ignorespaces}{}{}{}{{{}{}}}{{}}{}{}{{{}{}}}{{}}{}{{ {\lxSVG@setdash{0.0pt}\lx@inpgf@ignorespaces\lxSVG@roundjoin\lx@inpgf@ignorespaces{} {}{}{} {}{}{} {}{}{} \lxSVG@fillstroke}} }{}{}{{}}\lxSVG@stroke\lxSVG@drawpath@unclipped{M 0 0 L 33.21 0}{fill:none} {{}{{}}{}{}{{}}{{{\lx@inpgf@ignorespaces}}{{{\lx@inpgf@ignorespaces}}{\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{24.00006pt}{0.0pt}\lxSVG@begingroup@{transform} \lxSVG@begingroup@{_scopebegin} \lxSVG@begingroup@{stroke-dasharray,stroke-dashoffset} \lxSVG@begingroup@{stroke-linejoin} \lxSVG@drawpath@unclipped{M 1.11 0 C -0.28 0.28 -1.66 0.83 -3.32 1.8 C -1.66 0.55 -1.66 -0.55 -3.32 -1.8 C -1.66 -0.83 -0.28 -0.28 1.11 0 Z}{} \lxSVG@closescope \lxSVG@closescope }}{{\lx@inpgf@ignorespaces}}}}\lx@inpgf@ignorespaces\hbox{\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}}\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{1.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} \lxSVG@closescope }}} {{}{}{}}{}{ {}{}{}}{}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} {{{(speciallabnode.north)}}{{\lx@inpgf@ignorespaces}}{{\lx@inpgf@ignorespaces}}{{}}{{}}{{}}{{}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{59}\lxSVG@closescope }; {{\lx@inpgf@ignorespaces}{}}{ {}{}{}}{}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{24.00005pt}{-1.72pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} } \lxSVG@closescope {\lx@inpgf@ignorespaces}{{{}{}{}}}{{ {}{}{}}}\hss}\lxSVG@discardpath\lxSVG@closescope \hss}}\lxSVG@closescope\endpgfpicture}}\mspace{1.8mu}R[x] is the R -ring injective map with \alpha(x)=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}e}}} . Thus when the coefficient ring R is semiprime left Goldie, we know that \mathbb{S} is semiprime left Goldie and \mathsf{udim}(\mathbb{S})=\mathsf{udim}(R) [ LM03 , Theorems 3.4, 3.8] . Next, we claim that the terms xy,x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y freely generate a non-commutative submonoid \cramped{\langle xy,x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y\rangle}\subset\Gamma_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{e}}}} . Given any k -length word w=u_{1}\dotsm u_{k} with u_{1}\ldots u_{k}\in\{xy,x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y\} , one observes that \deg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}y}}}(u_{i})=1 and \deg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}y}}}(w)=k , so that, two words in the alphabet \{xy,x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y\} with different lengths cannot be equal as elements of \Gamma_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{e}}}} . Next, \deg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}x}}}(w)\equiv\deg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}x}}}(u_{1})\mod e (while 1=\deg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}x}}}(xy)\neq\deg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}x}}}(x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y)=2 ). These two observations show that for all k,l\in\mathbb{N} and terms u_{1},\ldots,u_{k},\,v_{1},\ldots,v_{l}\in\{xy,x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y\} , the products u_{1}\dotsm u_{k} and v_{1}\dotsm v_{l} are equal in \Gamma_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{e}}}} precisely when k=l and u_{i}=v_{i} for all 1\leq i\leq k ; which yields our claim. Consequently, the monoid subring R[xy,x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y]\subset R\Gamma_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{e}}}} is a two-variate free non-commutative polynomial ring over R ; and R[xy,x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}2}}}y] has infinite left or right uniform dimension. 3. Subextensions in skew Laurent polynomial rings We fix an ambient skew Laurent polynomial ring R[\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}};\boldsymbol{\alpha}] , always assumed of bijective type; here \boldsymbol{X} is an arbitrary non-empty set of freely commuting variables over R , \boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}}=\boldsymbol{X}\cup\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-1}}} with \boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-1}}}=\{x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-1}}}\hskip 3.0pt\colon\hskip 3.0ptx\in\boldsymbol{X}\} . For each x\in\boldsymbol{X} , the associated conjugation map \alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{x}}}}:R\mspace{1.8mu}\hbox to31pt{\vbox to5.44pt{\pgfpicture\makeatletter\hbox{\hskip 3.0pt\lower-2.44pt\hbox to0.0pt{\lxSVG@begingroup@{_scopebegin} \lxSVG@begingroup@{stroke} \lxSVG@begingroup@{fill} \lxSVG@setlinewidth{\the\pgflinewidth}\lxSVG@begingroup@{stroke-width} \lx@inpgf@ignorespaces\nullfont\lxSVG@begingroup@{_scopebegin} \lxSVG@closescope \hbox to0.0pt{\lxSVG@begingroup@{_scopebegin} { {{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{{}{}}}{{}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{-2.0pt}{0.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} { {}{}{}}{{{{}{}{}}}}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{27.00005pt}{0.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} {{}{}{}}{}{\lx@inpgf@ignorespaces}{}{{}}{}{ {}{}{}}{}{\lx@inpgf@ignorespaces}{}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {\lx@inpgf@ignorespaces}{{}}{} {\lx@inpgf@ignorespaces}{}{}{}{{{}{}}}{{}}{}{}{{{}{}}}{{}}{}{}{}{}{{}}\lxSVG@stroke\lxSVG@drawpath@unclipped{M 0 0 L 33.21 0}{fill:none} {{}{{}}{}{}{{}}{{{\lx@inpgf@ignorespaces}}{{{\lx@inpgf@ignorespaces}}{\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{24.00006pt}{0.0pt}\lxSVG@begingroup@{transform} \lxSVG@begingroup@{_scopebegin} \lxSVG@begingroup@{stroke-dasharray,stroke-dashoffset} \lxSVG@begingroup@{stroke-linejoin} \lxSVG@drawpath@unclipped{M 1.11 0 C -0.28 0.28 -1.66 0.83 -3.32 1.8 C -1.66 0.55 -1.66 -0.55 -3.32 -1.8 C -1.66 -0.83 -0.28 -0.28 1.11 0 Z}{} \lxSVG@closescope \lxSVG@closescope }}{{\lx@inpgf@ignorespaces}}}}\lx@inpgf@ignorespaces\hbox{\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}}\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{1.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} \lxSVG@closescope }}} {{}{}{}}{}{ {}{}{}}{}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} {{{(speciallabnode.north)}}{{\lx@inpgf@ignorespaces}}{{\lx@inpgf@ignorespaces}}{{}}{{}}{{}}{{}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{59}\lxSVG@closescope }; {{\lx@inpgf@ignorespaces}{}}{ {}{}{}}{}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{24.00005pt}{-1.72pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} } \lxSVG@closescope {\lx@inpgf@ignorespaces}{{{}{}{}}}{{ {}{}{}}}\hss}\lxSVG@discardpath\lxSVG@closescope \hss}}\lxSVG@closescope\endpgfpicture}}\mspace{1.8mu}R is a ring automorphism with xa=\alpha(a)x and x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-1}}}a=\alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{x}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{-1}}}}(a)x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-1}}} for all a\in R . For every \lambda=x_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{k_{1}}}}}\dotsm x_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{n}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{k_{n}}}}} with n\in\mathbb{N} , k_{1},\ldots,k_{n}\in\mathbb{N} and x_{{\cramped[\scriptstyle]{\mspace{-0.72mu}1}}},\ldots x_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}}\in\boldsymbol{X} , the associated conjugation map is the composite automorphism \alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{\lambda}}}}=\alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{x_{1}}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{k_{1}}}}}\cdot\dotsm\alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{x_{n}}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{k_{n}}}}}:R\mspace{1.8mu}\hbox to31pt{\vbox to6.42pt{\pgfpicture\makeatletter\hbox{\hskip 3.0pt\lower-2.44pt\hbox to0.0pt{\lxSVG@begingroup@{_scopebegin} \lxSVG@begingroup@{stroke} \lxSVG@begingroup@{fill} \lxSVG@setlinewidth{\the\pgflinewidth}\lxSVG@begingroup@{stroke-width} \lx@inpgf@ignorespaces\nullfont\lxSVG@begingroup@{_scopebegin} \lxSVG@closescope \hbox to0.0pt{\lxSVG@begingroup@{_scopebegin} { {{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{{}{}}}{{}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{-2.0pt}{0.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} { {}{}{}}{{{{}{}{}}}}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{27.00005pt}{0.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} {{}{}{}}{}{\lx@inpgf@ignorespaces}{}{{}}{}{ {}{}{}}{}{\lx@inpgf@ignorespaces}{}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {\lx@inpgf@ignorespaces}{{}}{} {\lx@inpgf@ignorespaces}{}{}{}{{{}{}}}{{}}{}{}{{{}{}}}{{}}{}{}{}{}{{}}\lxSVG@stroke\lxSVG@drawpath@unclipped{M 0 0 L 33.21 0}{fill:none} {{}{{}}{}{}{{}}{{{\lx@inpgf@ignorespaces}}{{{\lx@inpgf@ignorespaces}}{\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{24.00006pt}{0.0pt}\lxSVG@begingroup@{transform} \lxSVG@begingroup@{_scopebegin} \lxSVG@begingroup@{stroke-dasharray,stroke-dashoffset} \lxSVG@begingroup@{stroke-linejoin} \lxSVG@drawpath@unclipped{M 1.11 0 C -0.28 0.28 -1.66 0.83 -3.32 1.8 C -1.66 0.55 -1.66 -0.55 -3.32 -1.8 C -1.66 -0.83 -0.28 -0.28 1.11 0 Z}{} \lxSVG@closescope \lxSVG@closescope }}{{\lx@inpgf@ignorespaces}}}}\lx@inpgf@ignorespaces\hbox{\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}}\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{9.79166pt}{1.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} \lxSVG@closescope }}} {{}{}{}}{}{ {}{}{}}{}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} {{{(speciallabnode.north)}}{{\lx@inpgf@ignorespaces}}{{\lx@inpgf@ignorespaces}}{{}}{{}}{{}}{{}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{59}\lxSVG@closescope }; {{\lx@inpgf@ignorespaces}{}}{ {}{}{}}{}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{24.00005pt}{-1.72pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} } \lxSVG@closescope {\lx@inpgf@ignorespaces}{{{}{}{}}}{{ {}{}{}}}\hss}\lxSVG@discardpath\lxSVG@closescope \hss}}\lxSVG@closescope\endpgfpicture}}\mspace{1.8mu}R. The skew Laurent polynomial ring R[\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}};\boldsymbol{\alpha}] may be alternatively termed as the skew monoid ring extension of R by the free abelian group \mathpzc{T}(\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}}) over \boldsymbol{X} : \mathpzc{T}(\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}})=\{x_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{1}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{k_{1}}}}}\dotsm x_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{n}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{k_{n}}}}}\hskip 3.0pt\colon\hskip 3.0ptn\in\mathbb{N},\,k_{1},\ldots,k_{n}\in\mathbb{Z}\text{ and }x_{{\cramped[\scriptstyle]{\mspace{-0.72mu}1}}},\ldots,x_{{\cramped[\scriptstyle]{\mspace{-0.72mu}n}}}\in\boldsymbol{X}\}. Thus, a subextension of R[\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}};\boldsymbol{\alpha}] (generated over R by a family of terms from \mathbb{T}(\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}}) ) is precisely a skew monoid ring extension of R by a submonoid \mathsf{T} in a free abelian group. Every skew polynomial ring of bijective type and in commuting variables embeds into a skew Laurent polynomial ring; see Proposition 3.2 below. For the investigation of left singular ideals and semiprimeness of subextensions in R[\boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}};\boldsymbol{\alpha}] , it shall be necessary to consider a well-ordering ‘ \leq ’ on \boldsymbol{X} , which is always possible (at least by axiom of choice). Then the latter is extended to a linear ordering on \boldsymbol{X}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\pm 1}}} by setting y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-1}}}\dotsm>\tau_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{m}}}} in \mathbb{T} , we have \boldsymbol{\mathrm{T}}(f)=\tau_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{1}}}} ; the leading coefficient of f is \boldsymbol{\mathrm{lc}}(f)=a_{{\cramped[\scriptstyle]{\mspace{-0.72mu}1}}} . For convenience, one sets \boldsymbol{\mathrm{T}}(0)=0=\boldsymbol{\mathrm{lc}}(0) and 0<\lambda for every \lambda\in\mathbb{T} . Lemma 4.3 . The set \mathbb{T} of standard terms is well-ordered by the reverse lexicographic ordering induced by the well-ordering on \boldsymbol{X} . Proof. Assume the non-trivial case, that is, \boldsymbol{X} is non-empty. We reason by (transfinite) induction on well-ordered sets (or if one prefers, on ordinals). The desired result is obvious if \boldsymbol{X} has only one variable, (that is, if \mathfrak{n} is the ordinal ‘ 1 ’). Suppose \boldsymbol{X} is not a singleton and the result holds for all proper well-ordered subset in \boldsymbol{X} . Let \Sigma be a non empty subset of \mathbb{T}=\mathbb{T}(\boldsymbol{X}) , whose set of ‘leading variables’ is \mathsf{Lvar}(\Sigma)=\{x\in\boldsymbol{X}\hskip 3.0pt\colon\hskip 3.0pt\Sigma\cap\mathbb{T}(\boldsymbol{X}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\leq x}}})x\neq\varnothing\} , where for each ordinal s\in\mathfrak{n} we have the truncated subsets \boldsymbol{X}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\dotsm>n_{k} and w_{1},\ldots,w_{k}\in\mathbb{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\kappa}}}L . Suppose that gf=0 . Then since gf=\alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{n_{1}}}}}(w_{1}p)x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}n_{1}+m}}}+g^{\prime} for some g^{\prime}\in\mathbb{S} with \deg_{{\cramped[\scriptstyle]{\mspace{-0.72mu}x}}}(g^{\prime})\mathpzc{v}_{x}(f-f(0)) . Thus one can inductively form elements u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}1}}},u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}2}}},\ldots in R[x] with: \displaystyle u_{0}=x,u_{1}=x^{e},\;\alpha(u_{m})=u_{m}+pu_{m+1},\,\mathpzc{v}_{x}(u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m+1}}})>\mathpzc{v}_{x}(u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m}}})\text{ and } \displaystyle\alpha^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}}(x)=\hskip-3.00003pt\tsum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{k=0}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{m}}}}\textstyle{\binom{m}{k}}p^{{\cramped[\scriptstyle]{\mspace{-0.5mu}k}}}u_{{\cramped[\scriptstyle]{\mspace{-0.72mu}k}}}\text{ for all }m\in\mathbb{N}. Thanks to Lemma 5.1 , the natural number \binom{p^{r}}{k}p^{k} is a multiple of the characteristic p^{r+1} of R for every k\geq 1 . And it follows from the above relation that p^{r} is the least positive integer m with \alpha^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}}(x)=x , proving that \alpha is an R -ring automorphism with finite order p^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}} . As an immediate consequence, \alpha(x^{p^{r}})=(x+px^{e})^{p^{r}}=x^{p^{r}}\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{k=0}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{p^{r}}}}}\binom{p^{r}}{k}p^{k}x^{k(e-1)}=x^{p^{r}} . Thus x^{p^{r}},y^{p^{r}} live in the center of \mathbb{S} . ∎ Next for every m\in\mathbb{N} , we define the following bounds \mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m}}}=\mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m}}}(e),\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}=\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}r,c}}}(e)\in\mathbb{N} and \overline{\mathrm{S}}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m}}}=\overline{\mathrm{S}}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m,r,c}}}(e) : (5.1) \displaystyle\mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m}}} \displaystyle=\tsum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{s=0}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{m-1}}}}e^{{\cramped[\scriptstyle]{\mspace{-0.5mu}s}}}=\frac{e^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}}-1}{e-1};\;\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}\text{ satisfies the relation: }c\mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}-1}}}\leq r\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}$}. Here, there are three subcases to consider. \begin{ItemList}\itemize@item@{The subcase: $m\geq\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}$}. Then, $m^{\prime}\geq\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}$ as well and $\mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m}}}\overline{\mathrm{S}}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m^{\prime}}}}\geq\overline{\mathrm{S}}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}}}}>r$. Thus setting $m=\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}+s$ and $m^{\prime}=\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}+s^{\prime}$ with $s,s^{\prime}\in\mathbb{N}$, we compute: \@@amsalign Since $e,\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}\geq 2$, the last expression above is clearly positive if $\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}\geq 3$. And if $\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}=2$, then by definition we must have $c=c\mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}1}}}\leq rr$, while this time, $\overline{\mathrm{S}}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m}}}=c\mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m}}} r ,sothatthebound S _ ap r ,whichisthehighest x -degreeoftermswith y -degree ap^ r in \mathsf{T} ,isgivenaccordingtothedefiningequation~\eqref{eq:Sn}by: S _ ap r =c⋅(ap^ r - m _ +S_ m ) +(ap^ r - m _ )r(e-1) .Sohavinginviewthat S_ m ≥m forall m∈ N while d ispositivedivisorof e-1 ,weget:\begin{aligned} \overline{\mathrm{S}}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}ap^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}}}}-(cap^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}+kd)&=c\cdot(-\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}+\mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}}}})+(ap^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}-\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}})r(e-1)-kd\\ &\geq[(ap^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}-\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}})r-k](e-1)\geq 0.\end{aligned}Hence,\mbox{$\upsilon x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}kd}}}\in\mathsf{T}$}asexpected.Next,letusshowthatonecanfind\mbox{$b,s\in\mathbb{N}$}suchthatbothterms\mbox{$\lambda=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}(cb+sd)p^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}}}}y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}bp^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}}}}$}and\mbox{$\lambda x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}-kd}}}=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}(cb+sd)p^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}-kd}}}y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}bp^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}}}}$}livein \mathsf{T} .Since\mbox{$(cb+sd)p^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}-cbp^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}\in d\mathbb{N}$}and\mbox{$[(cb+sd)p^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}-kd]-cbp^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}=(sp^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}-k)d$},weonlyneedtoensurethatonecanchoose b and s suchthat\begin{aligned} k\leq sp^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}\text{ and }(cb+sd)p^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}&\leq\overline{\mathrm{S}}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}bp^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}}}}=c\cdot(bp^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}-\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}+\mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}}}})+(bp^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}-\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}})r(e-1).\\ \@@ams@intertext{Since \mbox{$\mathrm{S}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}}}}-\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}\geq 0$} and \mbox{$1\leq d\leq e-1$}, it suffices to find the natural numbers $b,s$ such that}\cr\frac{k}{p^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}}\leq s&\leq(b-\frac{\boldsymbol{m}_{{\cramped[\scriptstyle]{\mspace{-0.72mu}}}}}{p^{{\cramped[\scriptstyle]{\mspace{-0.5mu}r}}}})r.\end{aligned}Butsince p^ r ≥2 ,wemaytake b=k + m _ and s=k .Thiscompletestheproofthatforall τ ∈ ~ \mathsf{T} thereexistssome υ _ τ ∈ \mathsf{T} ∩ T (x^ p r ,y^ p r ) with υ _ τ τ ∈ \mathsf{T} .\par Nowsince T (x^ p r ,y^ p r ) iscontainedinthecenterof S ,foreveryfinitesubset E⊂ ~ \mathsf{T} ofterms,theterm υ =∏_ τ ∈E υ _ τ liesin \mathsf{T} ∩ T (x^ p r ,y^ p r ) and υ E⊂ \mathsf{T} .Sincemoreovereverymemberof T (x,y) isalreadyaregularelementin S ,weconcludethat A isastronglynicelyessentialsubringof ~ A .\par\vskip 6.0pt plus 2.0pt minus 2.0pt\noindent{(\text{{\sc{iii}}}).~~{$\widetilde{A}$isaspecialsubextensionin$\mathbb{S}$}.\,}Regard S =R[x][y; α ] asunivariateOreextensionoverthepolynomialring R[x] .Recallingthedefinitions,\begin{aligned} \widetilde{A}&=R\widetilde{\mathsf{T}}\text{ with }\widetilde{\mathsf{T}}=\{x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}cm+dk}}}y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}}\hskip 3.0pt\colon\hskip 3.0ptm\in\mathbb{N},\,-cm/d\leq k\in\mathbb{Z}\};\text{ thus, }\\ \widetilde{A}&=\sum_{{\cramped[\scriptstyle]{\mspace{-0.72mu}m\in\mathbb{N}}}}(\widetilde{A}\cap R[x]y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}})\text{ and }\widetilde{A}\cap R[x]y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}}=R\cdot\{x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}cm+dk}}}\hskip 3.0pt\colon\hskip 3.0pt-cm/d\leq k\in\mathbb{Z}\}y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}}.\end{aligned}With m∈ N fixed,theterms x^ cm and λ_ m =x^ cm y^ m areregularelementsof S with λ_ m ∈ ~ A ∩R[x]y^ m ;andsetting ~ A _ (m) =R[x^ d ] ,itisclearthatforevery f∈ ~ A ∩R[x]y^ m thereisasuitable k∈ N suchthat x^ dk f liesin ~ A _ (m) λ_ m =R[x^ d ]x^ cm y^ m .Moreoversince d divides e-1 ,formula~\eqref{eq:power-conjmap}alreadygrantsthattheconjugation R -ringautomorphism\makebox{$\alpha:R[x]\mspace{1.8mu}\hbox to31pt{\vbox to5.44pt{\pgfpicture\makeatletter\hbox{\hskip 3.0pt\lower-2.44pt\hbox to0.0pt{\lxSVG@begingroup@{_scopebegin} \lxSVG@begingroup@{stroke} \lxSVG@begingroup@{fill} \lxSVG@setlinewidth{\the\pgflinewidth}\lxSVG@begingroup@{stroke-width} \lx@inpgf@ignorespaces\nullfont\lxSVG@begingroup@{_scopebegin} \lxSVG@closescope \hbox to0.0pt{\lxSVG@begingroup@{_scopebegin} { {{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{{}{}}}{{}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{-2.0pt}{0.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} { {}{}{}}{{{{}{}{}}}}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{27.00005pt}{0.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} {{}{}{}}{}{\lx@inpgf@ignorespaces}{}{{}}{}{ {}{}{}}{}{\lx@inpgf@ignorespaces}{}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {\lx@inpgf@ignorespaces}{{}}{} {\lx@inpgf@ignorespaces}{}{}{}{{{}{}}}{{}}{}{}{{{}{}}}{{}}{}{}{}{}{{}}\lxSVG@stroke\lxSVG@drawpath@unclipped{M 0 0 L 33.21 0}{fill:none} {{}{{}}{}{}{{}}{{{\lx@inpgf@ignorespaces}}{{{\lx@inpgf@ignorespaces}}{\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{24.00006pt}{0.0pt}\lxSVG@begingroup@{transform} \lxSVG@begingroup@{_scopebegin} \lxSVG@begingroup@{stroke-dasharray,stroke-dashoffset} \lxSVG@begingroup@{stroke-linejoin} \lxSVG@drawpath@unclipped{M 1.11 0 C -0.28 0.28 -1.66 0.83 -3.32 1.8 C -1.66 0.55 -1.66 -0.55 -3.32 -1.8 C -1.66 -0.83 -0.28 -0.28 1.11 0 Z}{} \lxSVG@closescope \lxSVG@closescope }}{{\lx@inpgf@ignorespaces}}}}\lx@inpgf@ignorespaces\hbox{\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}}\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{1.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} \lxSVG@closescope }}} {{}{}{}}{}{ {}{}{}}{}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} {{{(speciallabnode.north)}}{{\lx@inpgf@ignorespaces}}{{\lx@inpgf@ignorespaces}}{{}}{{}}{{}}{{}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{12.50002pt}{2.0pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{59}\lxSVG@closescope }; {{\lx@inpgf@ignorespaces}{}}{ {}{}{}}{}{{}}\lx@inpgf@ignorespaces\hbox{\hbox{{\lxSVG@begingroup@{_scopebegin} {{}{}{{ {}{}}}{ {}{}} {{}{{\lx@inpgf@ignorespaces}}}{{}{\lx@inpgf@ignorespaces}}{}{{}{\lx@inpgf@ignorespaces}} {\lx@inpgf@ignorespaces }{{{{\lx@inpgf@ignorespaces}}\lxSVG@begingroup@{_scopebegin} \lxSVG@transformcm{1.0}{0.0}{0.0}{1.0}{24.00005pt}{-1.72pt}\lxSVG@begingroup@{transform} \pgfsys@hbox{62}\lxSVG@closescope }}} \lxSVG@closescope }}} } \lxSVG@closescope {\lx@inpgf@ignorespaces}{{{}{}{}}}{{ {}{}{}}}\hss}\lxSVG@discardpath\lxSVG@closescope \hss}}\lxSVG@closescope\endpgfpicture}}\mspace{1.8mu}R[x]$}(with α (x)=x+px^ e )andits m thpower α _ ^ m restricttoautomorphismsof R[x^ d ] ,suchthat:\begin{aligned} \text{for all $f\in R[x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}d}}}]$},\;\,\lambdaup_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{m}}}}f&=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}cm}}}y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}}f=x^{{\cramped[\scriptstyle]{\mspace{-0.5mu}cm}}}\alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{m}}}}(f)y^{{\cramped[\scriptstyle]{\mspace{-0.5mu}m}}}=\alpha_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{}}}}^{{\cramped[\scriptstyle]{\mspace{-0.5mu}\scriptstyle{m}}}}(f)\lambdaup_{{\cramped[\scriptstyle]{\mspace{-0.72mu}\scriptstyle{m}}}}.\end{aligned}Hencewehaveprovedthat ~ A isaspecialsubextensionof S viewedasaunivariateOreextensionofbijectivetype(andwithvanishingderivationmap)overtheordinarypolynomialring R[x] .\penalty\qed\end@proof\par For n≥3 andleavingittothereader,oneseeshowtoeasilyform n -variatespecialsubextensions ~ A inanambient n -variateskewpolynomialrings S suchthat ~ A isnotanessentialsubringin S .Likewise,onecanalsoform n -variatestronglynicelyessentialsubextensionsin ~ A suchthat A isnotspecialin S ,thoughprovingthislastfactshouldbecomeverytedious(asillustratedbytheproofofProposition~\ref{prop:expl2-ncom+}).\par\par