No label defined (Q4700631)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1357424
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 1357424 |
Statements
4 November 1999
0 references
unitriangular group
0 references
quasi-unitriangular group
0 references
elementary theory
0 references
elementary equivalence
0 references
axiomatizability
0 references
decidability
0 references
model theory of the group of unitriangular matrices over some ring
0 references
spectrum functions
0 references
Turing degrees
0 references
0.9291433691978456
0 references
0.8519237637519836
0 references
This paper takes up roughly half of the book, and is a translation of the author's 1995 D.Sc. thesis from Kemerova State University. In it, Belegradek studies the model theory of the group of unitriangular matrices over some ring \(R\). He begins with an extended introduction stating the previously known facts and open questions (with particular reference to \textit{A. I. Mal'tsev}'s important paper ``On a correspondence between rings and groups'' [Mat. Sb., Nov. Ser. 50(92), 257-266 (1960; Zbl 0148.24601); Engl. transl. Am. Math. Soc., Transl., II. Ser. 45, 221-231 (1965)], and \textit{C. R. Videla}'s work on the group of niltriangular matrices [``On the model theory of the ring \(\text{NT}(n,R)\)'', J. Pure Appl. Algebra 55, No. 3, 289-302 (1988; Zbl 0658.03020)]), before explaining his own results and (partial) solutions. Let \(\text{UT}_n(R)\) (for \(n\geq 3\)) denote the group of upper triangular matrices over some ring \(R\) with unit (which is associative for \(n>3\)). Mal'tsev introduced a correspondence between \(R\) and an expansion of \(\text{UT}_3(R)\) by two parameters, the transvections \(t_{12}(1)\) and \(t_{23}(1)\). It follows that \(R\) and this expansion of \(\text{UT}_3(R)\) (or, more generally, \(\text{UT}_n(R)\)), have the same model-theoretic behaviour in terms of axiomatizability, decidability, stability, or the spectrum function. But what role do these parameters play? In studying models of Th\((\text{UT}_n(R))\), Belegradek is led to the notion of a quasi-unitriangular group \(\text{UT}_n(R,g_1,\ldots,g_n)\) by modifying the multiplication by means of symmetric cocycles \(g_1,\ldots,g_n\) from \(R^+\) to \(R^+\), setting \((a_{ij})\odot(b_{ij})=(a_{ij})(b_{ij})+ \sum_{k=1}^{n-1}g_k(a_{k,k+1},b_{k,k+1})e_{1n}\). This modification does not alter conjugation, commutation, or the lower central series, and \(\text{UT}_n(R,g_1,\ldots,g_n)\) remains nilpotent of class \(n-1\); moreover the class of all quasi-\(\text{UT}_n\)-groups is elementary. If \(\mathfrak t\) denotes the basis of the transvections \((t_{ij}:1\leq i<j\leq n)\), then \(R\) is interpretable in \((\text{UT}_n(R,g_1,\ldots,g_n),\mathfrak t)\). For \(0<i+1<j\leq n\) set \(U_{ij}=\{t_{ij}(r):r\in R\}\), let \(U_{i,i+1}=\{t_{i,i+1}(r)+r'e_{1n}:r,r'\in R\}\), and call \(\mathfrak t\) splitting (resp. (\(m\)-)pure) if the extension \(H_{1n}\leq H_{i,i+1}\) splits (resp.\ is (\(m\)-)pure). Then a quasi-unitriangular group is unitriangular iff it has a splitting basis. Th\((\text{UT}_n(R),\mathfrak t)\) is recursively axiomatizable and its models are exactly the expanded quasi-\(\text{UT}_n(S)\) groups with a pure basis for some \(S\equiv R\); similarly Th\((\text{UT}_n(R))\) is recursively axiomatizable and its models are the locally pure quasi-\(\text{UT}_n(S)\) groups for some \(S\) (i.e. for all finite \(I\subset\omega\) they have a basis which is \(m\)-pure for all \(m\in I\)), but \(S\) need not be elementarily equivalent to \(R\) or \(R^{\text{op}}\). However, if \(R\) is \(\aleph_0\)-categorical, any \(G\equiv \text{UT}_n(R)\) is \(\text{UT}_n(S)\) for some \(S\equiv R\). If \(R\) is commutative or integral and \(\text{UT}_n(R,g_1,\ldots,g_n)\cong \text{UT}_n(S,q_1,\ldots,q_n)\), then \(R\cong S\) or \(R\cong S^{\text{op}}\); the same result holds with \(\equiv\) instead of \(\cong\). In general, \(\text{UT}_n(R,g_1,\ldots,g_n)\cong \text{UT}_n(R^{\text{op}},g_1,\ldots,g_n)\), and Belegradek finds a suitable ring \(R\) such that \(R\not\cong R^{\text{op}}\), but \(\text{UT}_n(R\times R)\cong \text{UT}_n(R\times R^{\text{op}})\). If any two bases in a quasi-unitriangular group over \(R\) are conjugate under an automorphism, then \(R\cong R^{\text{op}}\) and \(\text{Ext}(R^+,R^+)=0\); the converse holds if \(R\) is commutative or integral, but fails in general. For a commutative associative ring \(R\) any two bases in \(\text{UT}_3(R)\) are conjugate under an automorphism iff the \(2\)-cocyle \(\text{pr} :(x,y)\mapsto xy\) is a coboundary. If \(g_1,\ldots,g_n\) are coboundaries, then \(\text{UT}_n(R,g_1,\ldots,g_n)\cong \text{UT}_n(R)\), but the converse need not be true. Belegradek conjectures that \(\text{UT}_n(R,g_1,\ldots,g_n)\cong \text{UT}_n(R)\) for all \(g_1,\ldots,g_n\) iff \(\text{Ext}(R^+,R^+)=0\), and proves partial results in this direction. In a mainly model-theoretic section, Belegradek then compares the spectrum functions for \(R\) and \(\text{UT}_n(R)\). In particular, \(I(\lambda,R)=I(\lambda, \text{UT}_n(R))\) for all infinite \(\lambda\) and commutative associative \(R\). The same holds for uncountable \(\lambda\) and associative integral \(R\), and \(I(\aleph_0,\text{UT}_n(R))\leq I(\aleph_0,R)\), with inequality iff \(R\) is a skew field and there is a countable \(S\equiv R\equiv S^{\text{op}}\) with \(S\not\cong S^{\text{op}}\); no such \(R\) is known. On general grounds, \(I(\lambda,R)\) is \(1\) (resp.\ finite) iff \(I(\lambda,\text{UT}_n(R))\) is \(1\) (resp.\ finite), for any uncountable \(\lambda\), iff either theory is (almost) uncountably categorical; an example is given where these numbers are finite and different. The last section studies decidability questions: for any Turing degrees \(d_1\leq d_2\) there is a ring \(R\) such that Th\((\text{UT}_n(R))\) has Turing degree \(d_1\) and Th\((R)\) has Turing degree \(d_2\). Moreover, for any \(m\)-degree \(d\) there is a group formula \(\varphi_d(x)\) such that the problem ``\(\exists? x \varphi_d(x)\)'' has \(m\)-degree \(d\) in \(\text{UT}_3(\)\textbf{Z}\()\). The paper is well written, complete, and proceeds comfortably. It is an excellent exposition of the theory of unitriangular (and quasi-unitriangular) groups and their relation to the underlying ring. For the entire collection see [Zbl 0924.00004].
0 references