The notion universal localization or Cohn localization of a ring is a variant of the notion of localization of a ring which forces not just elements of the ring to become invertible (which one may think of as $1 \times 1$-matrices) but forces more general matrices with coefficients in the ring to become invertible. One also considers the corresponding localization functor on the category of modules. It can be related to ($H^0$ of) some Bousfield localization (on chain complexes of modules).
Let $\Sigma$ be a set of finite square matrices (of different sizes) over a (typically noncommutative) ring $R$. Without loss of generality, one assumes that $\Sigma$ is left or right multiplicative. It is left multiplicative if for any matrices $A,B,C$ of right sizes such that $A,B\in\Sigma$ and $C$ fits into matrix $New = \left(\array{ A & 0\\ C & B}\right)$, matrix $New$ is also in $\Sigma$.
We say that a homomorphism of rings $f: R\to S$ is $\Sigma$-inverting if all matrices $f(A)$ over $S$ where $A\in \Sigma$ are invertible in $S$. The Cohn localization of a ring $R$, is a homomorphism of rings $R\to \Sigma^{-1} R$ which is initial in the category of all $\Sigma$-inverting maps (which is the subcategory of coslice category $R/Ring$). In the left hand version, the elements in the localized ring are thought of as solutions of linear equations $A x = b$ where $b$ is a column vector with elements in $R$ and $A\in\Sigma$.
Given a ring $R$ and a family $S$ of morphisms in the category $R$Mod of (say left) finitely generated projective $R$-modules, we say that a morphism of rings $f:R\to T$ is $S$-inverting if the extension of scalars from $R$ to $T$ along $f$
sends all morphisms of $S$ into isomorphism in the category of left $T$-modules.
P. M. Cohn has shown that there is a universal object $R\to Q_S R$ in the category of $S$-inverting morphisms. The ring $Q_S R$ (and more precisely the universal morphism itself) are called the universal localization or Cohn localization of the ring $R$ at $S$.
Cohn localization induces a hereditary torsion theory, i.e. a localization endofunctor on the category of all modules, but it lacks good flatness properties at the level of full module category. However when restricted to the subcategory of finite-dimensional projectives it has all good properties – it is not any worse than Ore localization.
Universal localization is much used in algebraic K-theory, algebraic L-theory and surgery theory – see Andrew Ranicki‘s slides in the references at Cohn localization and his papers, specially the series with Amnon Neeman.
The original articles:
Paul M. Cohn, Free rings and their relations, Academic Press (1971) [pdf]
Paul M. Cohn, Inversive localization in noetherian rings, Communications on Pure and Applied Mathematics 26 5-6 (1973 ) 679-691 [doi:10.1002/cpa.3160260510]
Reviews and lecture notes:
V. Retakh, R. Wilson, Advanced course on quasideterminants and universal localization (2007) [pdf]
Andrew Ranicki (ed.), Noncommutative localization in algebra and topology, (Proceedings of Conference at ICMS, Edinburgh, 29-30 April 2002) London Math. Soc. Lecture Notes Series 330 Cambridge University Press (2006) [pdf]
containing this article on applications to topology:
reviewed in:
Andrew Ranicki, Noncommutative localization in algebra and topology, talk at Knot theory meeting (2008) [slides pdf]
Andrew Ranicki, Noncommutative localization, Pierre Vogel 65th birthday conference, Paris, 27 October 2010 [slides pdf]
See also:
On localization with inverses just from one side:
Universal localization of group rings (and connections to certain noncommutative rational function rings and Fox derivatives) is discussed in
Last revised on August 22, 2024 at 10:33:56. See the history of this page for a list of all contributions to it.