Canonical elements in local cohomology modules and the direct summand conjecture (Q2266742)

The paper contains numerous important theorems and methods concerning some old and new homological conjectures. The author's introduction begins with: ''One of the objectives of this paper is to show that the usual homological consequences of the existence of big Cohen-Macaulay modules follow from the direct summand conjecture when the residual characteristic of the local ring is positive.'' At the same time he develops ''a theory of certain canonical elements in the local cohomology of special modules of syzygies'': Let (R,\({\mathfrak m},K)\) be a local ring of dimension d and consider an exact R-sequence \({\mathcal E}:0\to S\to P_{d- 1}\to...\to P_ 1\to R\to K\to 0\) with finitely generated free modules \(P_ i\). The sequence \({\mathcal E}\) represents an element of \(Ext^ d_ R(K,S)\) which is mapped to an element \(\eta_ R=\eta_ R({\mathcal E})\) of \(H^ d_{{\mathfrak m}}(S)\) by the natural transformation of functors \(Ext^*_ R(K,\quad)\to H^*_{{\mathfrak m}}(\quad).\) \(\eta_ R\) is called the ''canonical element'' in \(H^ d_{{\mathfrak m}}(S)\). One can identify the various \(\eta_ R's\) obtained from various \({\mathcal E}'s\). Furthermore \(\eta_ R\neq 0\) can be shown to be equivalent to the following property: (CE) For every projective resolution \(P_*:...\to P_ i\to...\to P_ 0\to K\to 0\) and for every system of parameters \(\underline x=x_ 1,...,x_ d\) for R, if \(\phi\) is any homomorphism of complexes from the usual Koszul complex \(K_*(\underline x;R)\) to \(P_*\) which lifts the natural projection R/\b{x}R\(\to K\), then \(\phi_ d: K_ d(\underline x;R)(=R)\to P_ d\) is nonzero. Some of the main results about the ''usual'' homological conjectures and property (CE): (1) If R satisfies (CE), then the new intersection conjecture of Peskine-Szpiro and Roberts is true for R. - (2) If R is a Cohen-Macaulay domain such that all local rings of homomorphic image domains of R satisfy (CE), then the syzygy problem of Evans-Griffith can be settled affirmatively for R. - (1) and (2) are in some sense improvements of results if \textit{E. G. Evans} jun. and \textit{P. Griffith} [Ann. Math., II. Ser. 114, 323-333 (1981; Zbl 0497.13013)] because: (3) If R has a big Cohen-Macaulay module, then R satisfies (CE). - With respect to these theorems the following result seems the most important one: (4) If the direct summand conjecture is true, then every local ring satisfies (CE). The converse of (4) also holds which gives property (CE) the maturity to become a conjecture. Canonical element conjecture: For every local ring R, \(\eta_ R\neq 0\). Of course \(\eta_ R\neq 0\) is true in case the residual characteristic of R is 0 since in that case R is known to have a big Cohen-Macaulay module. When R has positive prime characteristic the author gives two proofs of \(\eta_ R\) being nonzero, one based on the functorial behavior of \(\eta_ R\) and the other - more complicated - one by establishing property (CE). The author then studies the connection of property (CE) with canonical modules and, in a separate section, the behaviour of \(\eta\) passing from R to a quotient of R by a nonzero-divisor. - The last section of the paper is devoted to a detailed investigation of the direct summand conjecture. The main theorem contains a reduction of the general case to the case of a formal power series ring over a complete unramified discrete valuation ring. Part of the ideas used to prove the theorem gives a new proof of the conjecture in the case of positive characteristic.