Home Towards a homological generalization of the direct summand theorem
Article Open Access

Towards a homological generalization of the direct summand theorem

  • Juan Diego Vélez-Caicedo and Danny Arlen de Jesús Gómez-Ramírez EMAIL logo
Published/Copyright: November 24, 2020

Abstract

We present a more general homological characterization of the direct summand theorem (DST). Specifically, we state two new conjectures: the socle-parameter conjecture (SPC) in its weak and strong forms. We give a proof for the weak form by showing that it is equivalent to the DST. Furthermore, we prove the SPC in its strong form for the case when the multiplicity of the parameters is smaller than or equal to two. Finally, we present a new proof of the DST in the equicharacteristic case, based on the techniques thus developed.

MSC 2010: 13D22; 13B02; 16E65

1 Introduction

Fundamental work of Hochster, Peskine, Szpiro, and Serre during the second half of the twentieth century allowed us to develop a theory of multiplicities, along with the introduction of powerful prime characteristic methods in commutative algebra [1,2,3]. As a by-product of this effort for understanding general properties of commutative rings and their prime spectra in terms of homological and algebraic invariants, a collection of homological conjectures emerged [1]. Further research of Hochster’s showed that most of these conjectures were essentially new ways of describing a quite central algebraic splitting phenomenon for a particular kind of ring extensions: most of these homological questions turned out to be equivalent to the direct summand conjecture (DSC) [1,4,5,6,7].

Specifically, the DSC states that if R S is a finite extension of rings and if R is a regular ring, then R is a direct summand of S as R-module. Or equivalently, this extension splits as a map of R-modules, i.e. there exists a retraction ρ : S R sending 1 S to 1 R .

After many decades, the DSC (or direct summand theorem [DST]) was finally proved in the former general form by Y. André, by first reducing to the case of unramified complete regular local rings, as it had been suggested by the seminal work of Hochster’s, and then by using Scholze’s theory of perfectoid spaces [8,9]. Even though this fundamental question was settled, the former results (regarded as a whole) suggest some directions for future research in local homological algebra.

In this work, we present an equivalent form of the DSC given in terms of an estimate for the difference of the lengths of the first two Koszul homology groups of quotients of Gorenstein rings by principal zero-divisor ideals. Dutta and Griffith (see [10, Theorem 1.5]) had obtained similar results in an independent manner, although in a rather different context (for the case of complete and almost complete intersections).

In Sections 3 and 4, we state an equivalent form of the DSC conjecture in terms of the existence of annihilators of zero divisors on Gorenstein local rings not belonging to parameter ideals [11,12]. Based on these results, we find a new conjecture equivalent, in its weak form, to the DSC (Section 6). In its strong form, this conjecture states that if ( T , η ) is a Gorenstein local ring of dimension d and { x 1 , , x d } T is any system of parameters, and if we denote by Q the ideal generated by these parameters, then for any zero divisor z T , and for any lifting u T of a socle element in T/Q (i.e. Ann T / Q ( η ¯ ) = ( u ¯ ) ) it must hold that u z Q ( z ) .

This rather technical condition allows for more flexibility when one tries to do computations in particular examples (see for instance the proof of Proposition 12). We call this conjecture the socle-parameter conjecture, strong form (SPCS). We obtain the weak form if we add the requirement that in the mixed-characteristic case char T / η = p > 0 and x 1 = p . The SPCS is at the same time equivalent to a very general and homological condition involving the lengths of the Koszul homology groups:

( H 0 ( x ̲ , T / ( z ) ) ) ( H 1 ( x ̲ , T / ( z ) ) ) > 0 .

Evidently, this last condition only involves homological estimates. Then, a theorem of Ikeda (see [13, Corollary 1.4]) helps one to show that the SPCS is true when the multiplicities of the parameters are smaller than or equal to two, suggesting induction as a way of solving (Section 8).This approach has its origins in the work of the first author (J. D. Vélez). Specifically, the idea of proving the DSC by means of annihilators is first introduced in his thesis (see [14, Lemma 3.1.2.] and [15]). The reduction to the case S = T / J , where T is a Gorenstein local ring, and J is a principal ideal, was stated in a private communication from Vélez to Hochster, in 1996, and appears more explicitly in [11] and in [12]. Similar results related to the SPCS were obtained independently by Strooker and Stückrad [16].

Finally, it is worth noting that due to their simplicity and conceptual clarity, the results exposed in this paper are able to be used, in the context of modern artificial intelligence and pure mathematics, for creating pseudo-pre-code regarding fulfilling artificial mathematical intelligence in the specific mathematical sub-discipline of commutative algebra [17]. More specifically, this can be done based on the new cognitive foundations of mathematics program [17, Chapters 1–6]; the new (initial) global taxonomy of cognitive(metamathematical) mechanisms [17, Chapters 7–10]; and the computational realization and artificial generation of the artificial mathematical intelligence program [17, Chapters 11 and 12].

2 Preliminary results

Let ( R , m ) S be a module finite extension. Let s 1 , , s n S be generators of S as an R-module. Then, there exist monic polynomials f i ( y i ) R [ y i ] such that f i ( y i ) = y i m i + a i 1 y i n i 1 + + a i m i , a i j R , and f i ( s i ) = 0 . One can define a homomorphism of R-algebras from

T = R [ y 1 , , y n ] / ( f 1 ( y 1 ) , , f n ( y n ) )

to S sending each y i to s i . If J denotes the kernel of this map, S T / J . Finally, since dim T = dim R = dim T / J , by the reason of the finiteness of the extension, J should be contained in a minimal prime ideal of T, that means, ht J = 0 . Later we develop all the necessary facts in order to prove that if the residue field is algebraically closed, then we can reduce the DST to the case where a i j m .

Remark 1

Let R S be a finite extension of Noetherian rings, where ( R , m ) is local. Then, it is an elementary fact to see that the maximal spectrum of S, Spec m S , is equal to V ( m S ) Spec S (see for example [18, Corollary 5.8]).

Lemma 1

Let ( R , m , k ) be a local complete ring and R S a finite extension. Assume that Spec S = V ( m S ) = { η 1 , , η n } . Then S is naturally isomorphic, as a ring, to S η 1 × × S η n .

Proof

This is a well-known consequence of the fact that a local complete ring is Henselian and one of the equivalence forms of being Henselian is essentially the statement of the lemma. See, for example [19, Lemma 10.152.3].[1]

Corollary 2

Let ( R , m , k ) be a complete local ring with algebraically closed residue field k and B = R [ x ] / ( F ( x ) ) , where F ( x ) is a monic polynomial of degree n. Then there exist monic polynomials G i ( x ) of degree n i such that i = 1 r n i = n , B = i = 1 r R [ x ] / ( G i ( x ) ) as rings, and

G i ( x ) = x n i + a i 1 x n i 1 + + a i n i ,

where all a i j m .

Proof

Since k is algebraically closed we can factor f ( x ) F ¯ ( x ) = i = 1 r ( x b i ) n i k [ x ] , for some b i k and n i , with i = 1 r n i = n . Besides, B / m B k [ x ] / ( f ( t ) ) is an Artinian ring with maximal ideals η i ¯ = ( x b i ) for i = 1 , , r ; therefore, B / m B i = 1 r k [ x ] / ( ( x b i ) n i ) (see [18, Theorem 8.7 and proof]). Now, by Hensel’s Lemma there exist monic polynomials F i ( x ) R [ x ] such that F [ x ] = n = 1 r F i ( x ) and F i ¯ ( x ) = ( x b i ) n i . By Remark 1, Spec m B = { η 1 , , η r } = V ( m B ) .

Now, let B i R such that B i ¯ = b i . Since η i ¯ = ( x b i ) in B/mB, we see by the correspondence between the ideals of B/mB and the ideals of B containing mB that η i = ( x B i ) + m B . Now, ( F ( x ) ) R [ x ] η j = ( F i ( x ) ) R [ x ] η i , because ( F j ( x ) ) + η i = R [ x ] for all i j , since mod m R [ x ] this ideal corresponds to ( ( x b j ) n j ) + ( x b i ) = rad ( x b i , x b j ) = k [ x ] for all i j . Therefore, F j ( x ) is a unit in R [ x ] η i . Besides, the ring R [ x ] / ( F i ( x ) ) is local with maximal ideal ( x B i ) + m e , where m e denotes the expansion of m in R [ x ] (which we denote again by η i ). This is because any maximal ideal should contain the expansion of m (the extension R R [ x ] / ( F i ( x ) ) is finite) and this ring module m e is the local ring ( k [ x ] / ( ( x b i ) n i ) , ( x b i ) e ) . So,

B η i = R [ x ] η i / ( F ( x ) ) R [ x ] η i R [ x ] η i / F i ( x ) R [ x ] η i ( R [ x ] / ( F i ( x ) ) ) η i R [ x ] / ( F i ( x ) ) .

Thus, by the previous lemma, B i = 1 r B η i i = 1 r R [ x ] / ( F i ( x ) ) .

Finally, the isomorphism of rings θ i : R [ x ] R [ t ] , sending x t + B i , induces an isomorphism of rings θ i ¯ : R [ x ] / ( F i ( x ) ) R [ t ] / ( F i ( t + B i ) ) . But the reduction of F i ( x ) mod m R [ x ] is exactly ( ( t + B i ) B i ) n i = t n i , because translation and reduction mod m commutes. But that means exactly that G i ( t ) = t n i + a i , 1 t n i 1 + + a i , n i R [ t ] with a i , j m , for any indices i, j. In conclusion, we get an isomorphism of rings between B and i = 1 r R [ x ] / ( G i ( t ) ) satisfying the conditions of our corollary.□

Remark 2

Let i : R S = S 1 × × S n be an extension of rings, where R is an integral domain. Then there exists an S i such that π i i : R S i is also an extension, where π i is the natural projection. Suppose by contradiction that for any i there exist a i 0 R such that π i ( i ( a i ) ) = 0 . Then, if a = i = 1 n a i 0 , for any j,

π j ( a ) = i = 1 n π j ( i ( a i ) ) = π j ( i ( a j ) ) i j π j ( i ( a i ) ) = 0 .

Therefore, i ( a ) = ( π 1 ( i ( a ) ) , , π n ( i ( a ) ) ) = 0 , contradicting the hypothesis that i is an extension.

Theorem 3

Let ( R , m , k ) be a regular complete local ring with algebraically closed residue field k. Then, to prove the DSC for R it is enough to consider finite extensions R S , where S = T / J ,

T = R [ y 1 , , y r ] / ( f 1 ( y 1 ) , , f r ( y r ) ) ,

f i ( y i ) = y i n i + a i , 1 y 1 n i 1 + + a i , n i with a i , j m and J T is an ideal of height zero.

Proof

Let us fix a finite extension R S . By the discussion above, we know that S = T / J , where T = R [ y 1 , , y r ] / ( f 1 ( y 1 ) , , f r ( y r ) ) (the coefficients of the monic polynomials f i ( y i ) are not necessarily in m), and ht ( J ) = 0 . It is elementary to see that

T i = 1 r R [ y i ] / ( f i ( y i ) ) .

Write B i = R [ y i ] / ( f i ( y i ) ) , then by Corollary 2 B i α = 1 m i R [ y i ] / ( f i α ( y i ) ) with f i α ( y i ) = y i n i α + a i α 1 y i n i α 1 + + a i α n i α and a i α j m . Furthermore, by the distributive law between tensor products and direct sums (see [20]) we get

T i = 1 r ( α = 1 m i R [ y i ] / ( f i α ( y i ) ) ) w ( i = 1 r R [ y i ] / ( f i α i ( y i ) ) ) w T w ,

where w = ( α 1 , , α r ) , 1 α i m i , and

T w i = 1 r R [ y i ] / ( f i α i ( y i ) ) R [ y i , , y r ] / ( f 1 α 1 ( y 1 ) , , f r α r ( y r ) ) ,

where each f i α i ( y i ) has lower coefficients in m, as desired. Besides, J = w J w and then S w T w / J w .

Finally, by Remark 2, there is an α such that R T w / J w is an extension. But if there is a retraction ρ w : T w / J w R , then ρ = π w ρ w : S R is also a retraction. Besides ht ( J w ) = 0 , because dim T w = dim R = dim T w / J w . In conclusion, S w = T w / J w has the desired form of our proposition, and then it is enough to prove the DSC in this case.□

Proposition 4

Let ( R , m , k ) be a regular local ring of dimension d, and

T = R [ y 1 , , y r ] / ( f 1 ( y 1 ) , , f r ( y r ) ) ,

f i ( y i ) = y i n i + a i , 1 y 1 n i 1 + + a i , n i , with a i , j m . Then T is a Gorenstein local ring with maximal ideal η = m + ( y 1 ¯ , , y r ¯ ) .

Proof

First, we see that T is a local Cohen-Macaulay (C-M) ring. In fact, let m 1 be any maximal ideal of T. Then m 1 R = m , because dim ( R / ( m 1 R ) ) = dim ( T / m 1 ) = 0 and m 1 R Spec ( R ) . Therefore, R / ( m 1 R ) is a field. Thus, y i n i m 1 and therefore y i m 1 . In conclusion, m 1 = η . Now, R [ y 1 , , y r ] is a C-M ring (see [21, Proposition 18.9]). Thus, B = R [ y 1 , , y r ] η is also C-M. Let m = ( x 1 , , x d ) , where d = dim R . Then { f 1 ( y 1 ) , , f r ( y r ) , x 1 , , x d } B is a system of parameters because dim B = ht ( η ) = dim R [ y 1 , , y r ] = dim R + r = d + r , and since R [ y 1 , , y r ] is C-M, then by previous comments is equidimensional, and rad ( f 1 , , f r , x 1 , , x d ) = η B . Then { f 1 , , f r , x 1 , , x d } is a regular sequence in B. Thus, B / ( f 1 , , f r ) T is C-M, due to the fact that dim B / ( f 1 , , f r ) = d and { x 1 ¯ , , x d ¯ } B / ( f 1 , , f r ) is a regular sequence.

Finally, let Q = ( y 1 , , y r , x 1 , , x d ) be the ideal generated by the system of parameters { y 1 , , y r , x 1 , , x d } . Then T / Q k [ y 1 , , y r ] / ( y 1 n 1 , , y r n r ) = k [ w 1 , , w r ] , where w i = y i ¯ . Let us see that

Ann T / Q ( ( w 1 , , w r ) ) = i = 1 r w i n i 1 .

In fact, if h ¯ Ann T / Q ( ( w 1 , , w r ) ) and c i = 1 r w i m i 0 is a monomial of h ¯ such that there exists m j with m j < n j 1 , then w j h ¯ 0 T / Q , because the monomial term w j c i = 1 r w i m i has the power m j + 1 < n j on w j , which is a contradiction, by the reason that h is on the socle. Therefore, m i n i 1 for all i and so h ¯ i = 1 r w i m i . The other contention is clear, and then the socle has dimension one. In conclusion, ( T , η ) is a Gorenstein local ring.

Another approach for proving this proposition can be summarized as follows: since f i ( y i ) involve different variables (for each i = 1 , , r ), one can use this fact explicitly to see that { f 1 ( y 1 ) , , f r ( y r ) } forms a regular sequence. Now, due to the fact that R is Gorenstein, we see that R [ y 1 , , y r ] so is. Moreover, one can also prove that when one mods out a Gorenstein ring by a regular sequence, the corresponding quotient ring remains Gorenstein. So, this proves that T = R [ y 1 , , y n ] / ( f 1 ( y 1 ) , , f r ( y r ) is Gorenstein. Finally, T is local, due to the fact that T/mT so is (because it is isomorphic to k [ y 1 , , y n ] / ( y 1 n 1 , , y r n r ) , with maximal ideal ( y 1 , , y r ) ) and due to the fact that all the maximal ideals of T contract to m by module-finiteness.□

3 DSC in terms of annihilators

Now we make preparations for the proof of the following fact: let h : ( R , m ) ( T , η ) be a finite homomorphism of local rings, i.e. h ( m ) η , where T is a local R - free ring, with T/mT Gorenstein, and let S = T / J , for some ideal J T . Then h : R S splits if and only if Ann T J m T (by abuse of notation we denote by h again its composition with the natural projection π : T S ).

Proposition 5

Let ( A , η , k ) be a local Gorenstein ring of dimension zero (i.e. dim k ( Ann A η ) = 1 ). Let u η such that ( u ) = Ann A η . Then u I for any ideal I ( 0 ) A .

Proof

Clearly, we can assume that I η . We know nil ( A ) = η ; therefore, there exists n such that η n = 0 . Let x 0 I . Then η n 1 x η n = ( 0 ) . Let r be such that η r x = ( 0 ) but η r 1 x ( 0 ) . Hence, η ( η r 1 x ) = ( 0 ) and so η r 1 x Ann A η = ( u ) . Then ( u ) contains a nonzero element of the form bx, where b η r 1 . This means that there exists c A \ η with c u = x b , so u = ( c 1 b ) x I .□

Remark 3

If h : R T is any homomorphism of rings, we can consider T = Hom R ( T , R ) as a T -module with the following action: fix t T and define ( t ϕ ) ( x ) ϕ ( t x ) , for ϕ Hom R ( T , R ) and x T .

Remark 4

Let ( R , m ) be a local ring, T a finitely generated R-free module, and θ : T T an R-homomorphism. Then θ is an isomorphism of R-modules if and only if θ ¯ : T / m T T / m T is an isomorphism of K-vector spaces. In fact, if A M n × n ( R ) is the matrix defining θ , then θ is an isomorphism if and only if det A is an unit, which means that det A m . But that is equivalent to saying that det A ¯ 0 k , where A ¯ is the reduction of A mod m. Finally, since A ¯ is the matrix defining θ ¯ , the last condition is equivalent to saying that θ ¯ is an isomorphism of k - vector spaces.

Theorem 6

Let ( R , m ) and ( T , η ) be local rings. Assume that T/mT is Gorenstein. Let h : R T be a finite homomorphism of local rings, such that T is R - free. Then there exists a T - isomorphism β : T T such that for any ideal J T , β 1 ( ( T / J ) ) = Ann T J .

Proof

We identify ( T / J ) with { f T : f ( J ) = 0 } . We know that dim T / m T = dim R / m = 0 , since T/mT is a finitely generated R / m - module. So, fix u 1 η such that ( u 1 ¯ ) = Ann T / m η ¯ , since T/mT is Gorenstein. Let { u 2 , , u d } T such that T / m T = ( u 1 ¯ , , u d ¯ ) (where T R d ). Then, by the Lemma of Nakayama T = ( u 1 , , u d ) , T is generated by u 1 , , u d as an R -free module. In fact, let { w 1 , , w d } T be an R - basis for T. Define θ : T T by w i u i , then the induced θ ¯ : T / m T T / m T is clearly an isomorphism of k - vector spaces. Since T is R - free, by Remark 4, θ is an isomorphism which means just that { u 1 , , u d } T is an R - basis for T. Let u 1 T be the dual element and define β : T T by t t u 1 , where ( t u 1 ) ( t 1 ) u 1 ( t t 1 ) , for all t 1 T . By definition, it is clear that β is a T - homomorphism. Now, we can make the natural identifications T = Hom R ( R d , R ) ( Hom R ( R , R ) ) d = R d . Therefore, T is an R - free module of dimension d and β is an R - isomorphism if and only if β ¯ : T / m T T / m T is so, due to Remark 4 ( T T as R - free modules). But β ¯ is an isomorphism of k - vector spaces if it is injective. Suppose by contradiction that ker β ¯ ( 0 ) . Then, by Proposition 5, u 1 ¯ ker β ¯ . This means that β ( u 1 ) m T , so there exist m 1 , , m d m such that β ( u 1 ) = u 1 u 1 = i = 1 d m i u i . But, evaluating at 1 we get: 1 = u 1 u 1 ( 1 ) = i = 1 d m i u i ( 1 ) m , a contradiction. In conclusion, β is a T - isomorphism.

For the last part, let a Ann T J . Then, for any j J , β ( a ) ( j ) ( a u 1 ) ( j ) = u 1 ( a j ) = u 1 ( 0 ) = 0 . Therefore, β ( a ) ( T / J ) and thus a = β 1 ( β ( a ) ) β 1 ( ( T / J ) ) .

Conversely, take ϕ 0 β 1 ( ( T / J ) ) . Then there exists a ϕ ( T / J ) such that ϕ 0 = β 1 ( ϕ ) , i.e. ϕ = ϕ 0 u 1 and ϕ ( J ) = 0 , which means that u 1 ( ϕ 0 j ) = 0 for all j J . Now, let us fix j 0 J . Then, for all t T , β ( ϕ 0 j 0 ) ( t ) = u 1 ( ϕ 0 j 0 t ) = u 1 ( ϕ 0 ( j 0 t ) ) = 0 , because j 0 t J . Therefore, β ( ϕ 0 j 0 ) 0 and then by the first part ϕ 0 j 0 = 0 , which means that ϕ 0 Ann T J .□

Theorem 7

Let h : R T / J be a finite homomorphism of local rings such that ( T , η ) is a local R-free ring, with T/mT Gorenstein, and J T an ideal. Assume that the structure of R-module of T/J inherited by the R-structure of T is the same as the one induced by h. Then R T / J splits if and only if Ann T J m T , where m is the maximal ideal of m.

Proof

Assume that ρ : T / J R is a splitting R-homomorphism. Then ρ ( T / J ) and ρ ( 1 ) = 1 . By the last theorem, there exists ρ 0 Ann T J such that ρ 0 = β 1 ( ρ ) , which means, in particular, that 1 = ρ ( 1 ) = ρ 0 u 1 ( 1 ) = u i ( ρ 0 ) m . Then ρ 0 m T , so Ann T J m T .

Conversely, take a Ann T J \ m T . Then, by Proposition 5, u 1 ¯ ( a ¯ ) T / m T , which means that there exists m i m , with u 1 a = i = 1 d m i u i , where { u 1 , , u d } T is an R - basis for T as in the last Theorem. Then,

a u i ( 1 ) = ( 1 m 1 ) u 1 u 1 ( 1 ) i = 2 d m i u i u i ( 1 ) = ( 1 m 1 ) i = 2 d m i u 1 ( u i ) = ( 1 m 1 ) + 0 = 1 m 1 m .

So ρ = β ( ( 1 m i ) 1 a ) satisfies that ρ ( T / J ) because β 1 ( ρ ) = ( 1 m 1 ) 1 a Ann T J , and by Theorem 6 β 1 ( ( T / J ) ) = Ann T J . Besides, ρ 1 = ( 1 m i ) 1 a u 1 ( 1 ) = ( 1 m 1 ) 1 ( 1 m 1 ) = 1 , which implies that ρ : T / J R is the desired splitting R - homomorphism.□

4 Reduction to the case where J is principal

In the next proposition, we prove that (a proof of) the DST can be reduced to (a proof for) the case where J is a principal ideal generated by an element in mT.

Proposition 8

Let ( R 0 , m 0 , k 0 ) be a regular local ring and R 0 T 0 be a finite extension, where

T 0 = R 0 [ y 1 , , y r ] / ( f 1 ( y 1 ) , , f r ( y r ) ) ,

and each f i ( y i ) = y i n i + a i 1 y i n i 1 + + a i n i , with a i j m , for all indices i, j. Let S = T 0 / J , with J = ( g 1 , . , g s ) T 0 such that J R 0 = ( 0 ) . Let x 1 , , x s be new variables, and let R be R 0 [ x 1 , , x s ] , and let m be m 0 + ( x 1 , , x s ) , a maximal ideal of R. Write

T = R R 0 T 0 R [ y 1 , , y r ] / ( f 1 ( y 1 ) , , f r ( y r ) ) .

Let g be the element x 1 g 1 + + x s g s . Then:

  1. ( R m , m R m , k 0 ) is a regular local ring.

  2. ( g ) T m R m = ( 0 ) .

  3. R m ( T / ( g ) ) m T m / ( g ) e is a finite extension, and g ( m R m ) T m .

  4. R m T m / ( g ) splits if and only if R 0 T 0 / J splits.

Proof

  1. In general, if R is regular, then so is the polynomial ring R [ T ] (see [20]). In particular, R m is a regular local ring and R m / m R m R / m R 0 / m 0 = k 0 .

  2. R 0 R is an R - free extension, then, in particular, it is flat. Therefore, by tensoring R 0 T 0 / J with R, we see that R R R 0 T 0 / J T / J e is also an extension, and since localization is flat too, we get an extension R m T m / J e . Because of g J e , we get g T m R m = 0 .

  3. Clearly, by definition g ( m R m ) T m . Now, by the previous paragraph R m T m / ( g ) e is an extension, and it is finite because R 0 T 0 is finite, in fact free. Then, after tensoring with R m we get a module finite extension R m T m , so T m / ( g ) e is also a finitely generated R m - module.

  4. Assume that ρ 0 : T 0 / J R 0 is an R 0 - homomorphism such that ρ 0 ( 1 ) = 1 . Then, by tensoring with the flat R 0 - module R m , we get an R m - homomorphism ρ : T m / J R m , with ρ ( 1 ) = 1 . Now, composing ρ with the natural map T m / ( g ) T m / J , we obtain a retraction from T m / ( g ) to R m .□

Conversely, it is clear that T m satisfies the hypothesis of Proposition 4. Therefore, by Theorem 7, Ann T m ( g ) ( m T m . So let us choose

w = α ( h α ( x ̲ ) / k α ( x ̲ ) ) y α ,

such that w g = 0 , where h α R and k α R \ m (which is equivalent to saying that k α ( 0 ̲ ) m 0 ). Here y α denotes y 1 α 1 y r α r , α = ( α 1 , , α r ) , 0 α i < deg f i and some h β m (that means exactly w m T m ). We have T = R R 0 T T 0 [ x 1 , , x s ] . Now, multiplying by the product of the k α ( x ̲ ) , we can assume that w = α p α ( x ̲ ) y α T , where some p α m and 0 = w g = i α p α ( x ̲ ) y α x i g i ( y ̲ ) in T. Now, the coefficient of x i , which is zero in T, is exactly α p α ( 0 ̲ ) y α g i ( y ̲ ) , because the terms y α g i ( y ̲ ) are constants in T = T 0 [ x 1 , , x s ] . Therefore, if w 0 = α p α ( 0 ̲ ) y α T 0 , we have w 0 g i = α p α ( 0 ̲ ) y α g i ( y ̲ ) = 0 , and thus w 0 Ann t 0 J . But p β ( 0 ̲ ) m 0 , because p β ( x ̲ ) m , so w 0 m 0 T 0 . In conclusion, Ann T 0 J m 0 T 0 , which is equivalent by Theorem 7 to the fact that R 0 T 0 / J splits.

5 Socle-parameter conjecture

In this section, and in the next section, we state two new conjectures the socle-parameter conjecture (SPC) in its strong (SPCS) and weak forms (SPCW), and we prove that the SPCW is equivalent to the DSC, and that the SPCS implies the SPCW. Besides, these two conjectures are equivalent in the equicharacteristic case and therefore both are equivalent to the DSC in the equicharacteristic case.

However, as far as we know, the mixed characteristic case remains open. The new approach shows that the DSC is, in essence, a problem concerning algebraic and homological properties of Gorenstein local rings.

Let us start by reviewing some elementary notions. Let R be an -graded ring such that R 0 is an Artinian ring, and such that R is finitely generated as an R 0 -algebra. Let M be a finitely generated graded R-module of dimension d. Then, it is elementary to see that each homogeneous part M n is a finitely generated R 0 -module, and therefore, it has finite length (see [18, Proposition 6.5.]). It is well known in this case that there exists a unique polynomial

p M ( t ) = a d 1 t d 1 + + a 0 [ t ]

of degree d 1 , the Hilbert polynomial, such that for n 0 , p M ( n ) = ( M n ) . The multiplicity of M, e ( M ) , is defined as ( M ) if d = 0 , and as ( d 1)! a d 1 , if d > 0 (see [22, Definition 4.1.5.]. In particular, if M has positive dimension, then e ( M ) > 0 , since a d 1 > 0 . For the local case, assume that ( R , m ) is a local ring, M a finitely generated R-module of dimension d , and I = ( x 1 , , x n ) is an ideal of definition of M. This last condition means that m r M I M for some r > 0 , which is equivalent to saying that x = x 1 , , x n is a multiplicity system on I (i.e. ( M / ( x 1 , , x n ) M ) < + ) (see [22, p. 185]). We define the filtered graded ring gr I R = i = 0 + I i / I i + 1 , and the filtered graded module gr I M = i = 0 + I i M / I i + 1 M , where I 0 = R . Then, gr I R is in a natural way a graded ring (where R/I is Artinian, because after reducing to the case Ann R M = 0 , it is easy to see that ( M / I M ) < + if and only if rad I = m ). Therefore, we can define the multiplicity of M on I, e ( I , M ) e ( gr I M ) (see [22, p. 180]) and the multiplicity of R, e ( R ) e ( m , R ) (see [18, p. 108]). In particular, we can define the multiplicity of M on I = ( x 1 , , x n ) , where x 1 , , x n R is a system of parameters of M, i.e. n = dim M and M is Artinian, i.e. satisfied the descending chain condition for submodules (see [18, p. 74]). Besides, under the former hypothesis and assuming that x = x 1 , , x n is a multiplicity system of M, we can define the Euler Characteristic as

χ ( x ̲ , M ) = i = 0 n ( 1 ) i ( H i ( x ̲ , M ) ) .

For a more technical reformulation of this notion due to Auslander and Buchsbaum, see [22]. Now, a theorem of Serre (see [22, Theorem 4.6.6.]) states that χ ( x ̲ , M ) = e ( I , M ) , if { x 1 , , x N } R is a system of parameters for M , and zero otherwise. In particular, if { x 1 , , x n } R is a system of parameters for M , and dim M > 0 , then χ ( x ̲ , M ) = e ( I , M ) = e ( gr I M ) > 0 , because dim M = dim ( gr I M ) (see [22, Theorem 4.4.6.]).

5.1 SPCS

Let ( T , η ) be a Gorenstein local ring of dimension d. Let { x 1 , , x d } T be a system of parameters and write Q = ( x 1 , , x d ) . Let u T be any lifting of a socle element in T/Q, i.e. Ann T / Q ( η ¯ ) = ( u ¯ ) . Let z T be a zero divisor. Then u z Q ( z ) .

Now, it holds that this strong form of the conjecture is equivalent to saying that ( H 0 ( x ̲ , T / ( z ) ) ) ( H 1 ( x ̲ , T / ( z ) ) ) > 0 .

We prove the last equivalence:

Proposition 9

In the situation of the SPCS the following are equivalent.

  1. u z Q ( z ) .

  2. Ann T ( z ) Q .

  3. ( H 0 ( x ̲ , T / ( z ) ) ) ( H 1 ( x ̲ , T / ( z ) ) ) > 0 .

Proof

( 1 ) ( 2 ) Consider the following natural short exact sequence

0 Ann T ( z ) T ( z ) 0 .

We know that T / Ann T ( z ) ( z ) , by the isomorphism sending t ¯ to tz. After tensoring with T/Q we get

( z ) / ( Q ( z ) ) T / Q ( z ) T / Q T / Ann T ( z ) T / ( Ann T ( z ) + Q ) ) .

Now, let us see that u z Q ( z ) implies Ann T ( z ) Q . Effectively, u z Q ( z ) is equivalent to u z ¯ = 0 ( z ) / ( Q ( z ) ) , and it is equivalent to u ¯ = 0 T / ( A n n T ( z ) + Q ) , under the last isomorphism. Therefore, there exists w Ann T ( z ) and q Q such that u = w + q , and so w = u q Q because u Q . Then Ann T ( z ) Q .

(2) (1) If Ann T ( z ) Q , by Proposition 5, u ¯ Ann T ( z ) ¯ T / Q . Then there exists w Ann T ( z ) such that u ¯ = w ¯ , which means that there is a q Q such that u = w + q . So,

u z = ( w + q ) z = w z + q z = q z Q ( z ) .

(2) (3) Ann T ( z ) Q if and only if Q Ann T ( z ) + Q if and only if ( T / ( Ann T ( z ) + Q ) ) < ( T / Q ) . Besides, we have the natural short exact sequence

0 ( z ) T T / ( z ) 0 .

Then, after considering the induced long exact sequence for Tor, and noting that Tor 1 T ( T , T / Q ) = 0 , because T is a T - free module, and therefore flat, we get the following exact sequence

0 T o r 1 T ( T / ( z ) , T / Q ) ( z ) / Q ( z ) T / Q T / ( Q + ( z ) ) 0 .

Now, since Q is generated by a system of parameters, the T - modules T / ( Q + ( z ) ) , T/Q and T / ( Ann T ( z ) + Q ) ( z ) / Q ( z ) are Noetherian rings of dimension zero and therefore Artinian. In particular, they have finite length as T - modules. Then the submodule Tor 1 T ( T / ( z ) , T / Q ) ) has finite length too. By the additivity of ( ) , we have

( T o r 1 T ( T / ( z ) , T / Q ) ) ( T / ( Ann T ( z ) + Q ) ) + ( T / Q ) ( T / ( Q + ( z ) ) = 0 .

Hence,

( T / Q ) ( T / ( Ann T ( z ) + Q ) ) = ( T / ( Q + ( z ) ) ( T o r 1 T ( T / ( z ) , T / Q ) ) .

Then, Ann T ( z ) Q if and only if ( T / ( Q + ( z ) ) ( Tor 1 T ( T / ( z ) , T / Q ) ) > 0 . But T is C-M and then { x 1 , , x n } is a regular sequence. Hence, the Koszul Complex is a free (and then projective) resolution of T/Q:

K 2 K 1 K 0 T / Q 0 .

Hence, after tensoring this resolution with T / ( z ) , and taking homology, we find that H 1 ( x ̲ , T / ( z ) ) Tor 1 T ( T / Q , T / ( z ) ) and H 0 ( x ̲ , T / ( z ) ) ( T / ( z ) ) / Q ¯ ) T / ( ( z ) + Q ) . In conclusion, Ann T ( z ) Q is equivalent to the condition

( H 0 ( x ̲ , T / ( z ) ) ( H 1 ( x ̲ , T / ( z ) ) > 0 .

Proposition 10

Assume the same hypothesis as in SPCS and, in addition, that depth ( η ¯ , T / ( z ) ) d 1 , then SPCS holds.

Proof

Write a = depth ( η ¯ , T / ( z ) ) . It is a well-known fact that if H r ( x ̲ , T / ( z ) ) denotes the Koszul homology and q = sup { r : H r ( x ̲ , T / ( z ) ) 0 } , then a = d q , therefore, q = d a 1 . In the case that d = 0 , Q = 0 , and dim T ( η ) = 1 , thus for any element u T holds u z = 0 Q ( z ) . Then assume d 1 . Besides, { x 1 , , x n } T is a system of parameters for the T - module T / ( z ) , because dim T = dim T / ( z ) and dim ( ( T / ( z ) ) / ( x 1 , , x d ) T / ( z ) ) = 0 . So, ( T / ( z ) ) / ( x 1 , , x d ) T / ( z ) is an Artinian ring. Hence, by previous results, we get

( H 0 ( x ̲ , T / ( z ) ) ( H 1 ( x ̲ , T / ( z ) ) = χ ( x ̲ , T / ( z ) ) = e ( Q ¯ , T / ( z ) ) .

Now, by previous comments and the fact that d 1 we see that e ( Q ¯ , T / ( z ) ) > 0 .□

5.2 SPCW

Let ( T , η ) be a Gorenstein local ring of dimension d. Let { x 1 , , x d } be any system of parameters (if T is mixed characteristic [ char T / η = p > 0 ] we assume that x 1 = p ). Let Q = ( x 1 , , x d ) , and let u T be any lifting of a socle element in T/Q, i.e. Ann T / Q ( η ¯ ) = ( u ¯ ) . Let z be a zero divisor. Then u z Q ( z ) , which is equivalent to the inequality ( H 0 ( x ̲ , T / ( z ) ) ) ( H 1 ( x ̲ , T / ( z ) ) ) > 0 .

Note that between the two forms of the SPC the only difference is the fact that in the mixed characteristic case one assume that x 1 = p . One needs this last condition in order to apply Cohen’s structure theorem in mixed characteristic.

Remark 5

For proving any of the two versions of the SPC, it is enough to assume that ( T , η ) is complete.

Proof

Let τ : T T ^ be the natural homomorphism to the completion. Then τ is a faithfully flat extension and I T ^ T = I for any ideal I of T (see [20, p. 63]). Besides, another elementary consequence of faithfully flatness is that for any ideals I , J of T, ( J : I ) T ˆ = ( J T ˆ : I T ˆ ) . Now, assume by contradiction that there exists a system of parameters { x 1 , , x d } (for the SPCW assume x 1 = p ), a zero divisor z T and u T a lifting of a socle element for T/Q such that u z Q ( z ) . Let us write τ ( y ) = y . Then { x 1 , , x d } T ˆ is a system of parameters in T ˆ , because η ˆ = η T ˆ = rad ( ( x 1 , , x d ) T ˆ ) rad ( x 1 , , x d ) η ˆ . Besides, ( Q ( z ) : z ) T ˆ = ( Q T ˆ ( z ) : z ) ; T ˆ / Q T ˆ ( T / Q ^ ) = T / Q , due to the fact that η ¯ n = ( 0 ) for some n > 0 ; therefore, ( u ) = Ann T ˆ / Q T ˆ ( η ˆ ) and so u is a socle element. Note that p = char ( T ˆ / Q T ˆ ) = char ( T / Q ) . Furthermore, T ˆ is also a Gorenstein ring (see [20, Theorem 18.3]).

Finally, ( ( Q T ˆ : z ) + ( u ) ) / ( Q T ˆ : z ) ( ( Q : z ) + u ) ) / ( Q : z ) ) T ˆ 0 , because ( Q : z ) + u ) / ( Q : z ) 0 , since u z Q ( z ) . Then u z Q T ˆ ( z ) which contradicts SPC in the complete case.□

6 Equivalence to the DSC

First, we review the notion of a coefficient ring: If ( R , m , k ) is equicharacteristic, a coefficient field is a field K 0 R such that the natural projection π : K 0 R R / m = k is an isomorphism. On the other hand, let ( R , m , k ) be a complete quasi-local (that means with a unique maximal ideal m but not necessarily Noetherian), mixed characteristic, and separated ring (i.e. n m n = ( 0 ) ). Then a coefficient ring for R is a sub-ring ( D , η ) R such that it is a local complete discrete valuation ring such that m D = η , and so that the inclusion induces an isomorphism D / η R / m . It is elementary to see that if char R = 0 and char k = p > 0 , then D is a domain, and therefore one dimensional, which means exactly that ( D , η ) is a discrete valuation domain (DVD). A theorem of I. S. Cohen states that for complete local rings there always exists a coefficient ring (see [23, Theorem 9, Theorem 11] and [24, p. 24]). In fact, in the mixed characteristic case ( char k = p > 0 ), there exist coefficient rings which are DVD-s ( D , p D , k ) , with local parameter p, or D / p m D , when char R = p m (see [24, Corollary p. 24]).

Theorem 11

The SPCW is equivalent to the DSC.

Proof

SPCW DSC . By previous comments, we may assume that R S is a finite extension and R is a unramified regular local mixed characteristic ring ( char ( R / m ) = p > 0 ) with algebraically closed residue field k. Besides, by Theorem 3 we can assume that S = T / ( J ) such that T = R [ y 1 , , y r ] / ( f 1 ( y 1 ) , , f r ( y r ) ) ; f i ( y i ) = y i n i + a i 1 y i n i 1 + + a i n i , where a i k m , for all indices i, k, and ht ( J ) = 0 . Now, T is a Gorenstein local ring by Proposition 4. Moreover, by Proposition 8 R S splits if and only if R 1 R [ x 1 , , x d ] m 1 S 1 = T 1 / ( z ) splits, where m 1 = m + ( x 1 , , x d ) , T 1 = R 1 [ y 1 , , y r ] / ( f 1 ( y 1 ) , , f r ( y r ) ) , and z T 1 .

Besides, since R is regular, R [ x 1 , , x d ] so is. In particular, R 1 is regular. Furthermore, R 1 is unramified, otherwise there exist elements a i , b i m 1 and s R 1 \ m 1 such that sp = i = 1 c a i b i , and then evaluating in ( 0 , , 0 ) we get p = s ( 0 ̲ ) 1 i = 1 c a i ( 0 ̲ ) B i ( 0 ̲ ) , where a i ( 0 ̲ ) , b i ( 0 ̲ ) m and s ( 0 ̲ ) R \ m , so p m 2 , which is a contradiction.

Hence, p m 2 and then p ¯ 0 m / m 2 is a part of a basis of m / m 2 as k-vector space, which is equivalent by the Lemma of Nakayama to the fact that p is a part of a minimal set of generators of m, say, { w 1 = p , , w n } m . Now, { w 1 = p , , w n } T 1 is a system of parameters in T 1 , because dim ( T 1 / m 1 T 1 ) = dim ( r 1 / m 1 ) = 0 and dim T 1 = dim R 1 , since R 1 S 1 is a finite extension.

On the other hand, z is a zero divisor in T 1 because dim T 1 / ( z ) = dim S 1 = dim R 1 = dim T 1 and therefore z is contained in a minimal prime of T 1 , since T is C-M.

Since T 1 is Gorenstein, choose u T 1 such that ( u ¯ ) = Ann T 1 / m 1 T 1 ( η ¯ ) . By SPCW, u z m 1 T 1 ( z ) , so there exists a m 1 T 1 such that u z = a z , hence ( u a ) z = 0 . But u a m 1 T 1 , because u m 1 T 1 . Therefore, Ann T 1 ( z ) m 1 T 1 , and then, by Theorem 7, R 1 S 1 splits.

DSC SPCW . Let ( T , η ) be a Gorenstein local ring and { x 1 , , x d } T a system of parameters ( x 1 = p in the mixed characteristic case). By Remark 5, we can assume that T is complete. Let D be a coefficient ring for T (which always exists for any complete local ring). Then, due to Cohen’s structure theorem (see [23, Lemma 16]), the ring generated (after completion) as D-algebra by the parameters R = D [ [ x 1 , , x d ] ] is a complete regular local ring with maximal ideal Q = ( x 1 , , x d ) such that the extension R T is finite. Since R is regular, by Serre’s theorem (see [20] Theorem 19.2) p d R ( T ) is finite. Hence, for the Auslander-Buchsbaum formula and the fact that depth ( Q , T ) = depth ( η , T ) (see [22, Exercise 1.2.26]), we know that

p d R ( T ) = depth ( Q , R ) depth ( Q , T ) = dim R depth ( η , T ) = d dim T = d d = 0 .

So T is a free R-module. Furthermore, z is contained in an associated prime of T because it is a zero divisor. Since T is C-M, any associated prime is, in fact, a minimal prime. Thus, z is contained in a minimal prime P Spec T . Moreover, since T is C-M, T is equidimensional, which means, in particular, that dim T / ( z ) dim T / P = dim T . In conclusion,

dim R / ( ( z ) R ) = dim T / ( z ) = dim T = dim R ,

so ( z ) R = ( 0 ) , because R is a domain (R is regular!). Now, to see that u z Q ( z ) , it is enough to see that Ann T ( z ) Q . In fact, if Ann T ( z ) Q , then Ann T ( z ) ¯ 0 in T/Q. Therefore, by Proposition 5, u ¯ A n n T ( z ) ¯ , there exists w Ann T ( z ) such that u w Q . Thus, u z = ( u w ) z Q ( z ) , because w z = 0 . By Theorem 7, R T / ( z ) splits if and only if Ann T ( z ) m T = Q , where m = ( x 1 , , x d ) R . Hence, the DSC for R T / ( z ) implies Ann T ( z ) Q T , and then u z Q ( z ) . Another way of proving this is using directly Proposition 9.□

7 SPCS for small multiplicities

A natural way one could attack the SPCS would be by induction on e ( T ) , the multiplicity of T. We note first that by Remark 5 one may assume that T is complete. In the case e ( T ) = 1 , then, since T is a complete C-M ring, it is equicharacteristic and therefore unmixed. Hence, by the Criterion for multiplicity one (see [25]) T must be a regular local ring; in particular, this ring is an integral domain, which implies z = 0 , from which the SPCS follows directly.

Suppose now that e ( T ) = 2 . Since T is C-M, it must satisfy the S 2 condition of Serre: for any P Spec T , depth ( P , T ) min (2, dim T P ) , due to the fact that depth ( P , T ) = dim ( T P ) . Hence, by a Theorem of Ikeda (see [13, Corollary 1.3.]) T is a hypersurface of the form B / ( f ) , where B is a complete regular local ring. Now, we prove a more general result, namely, that the SPCS holds for residue class ring of local Gorenstein rings which are UFD and C-M, which implies, in particular, the case of multiplicity two because regular local rings are UFD and C-M (see [21]).

Proposition 12

The SPCS holds for Gorenstein rings of the form T = B / ( f ) , where B is a local C-M ring which is an UFD and f 0 B .

Proof

Let z 0 T be a zero divisor and { y 1 ¯ , , y d ¯ } T a system of parameters. We will see that ( H 0 ( y ̲ , T / ( z ) ) ) ( H 1 ( y ̲ , T / ( z ) ) ) > 0 . The minimal prime ideals of T are just the principal ideals generated by the prime factors of f = f i c i , i.e. P i = ( f i ) , since B is an UFD. Besides, it is enough to prove SPCS for z = f i ¯ , because each zero divisor is a multiple of one of these, i.e. z = a f i ¯ for some a B , and thus if u f i ¯ Q ( f i ¯ ) , where Q = ( y 1 ¯ , , y d ¯ ) and ( u ¯ ) = Ann T / Q ( η ¯ ) , then u z = u a f i ¯ = u a f i ¯ Q ( a f i ¯ ) = Q ( z ) . Let us fix some f j , then T / ( f j ¯ ) = B / ( f j ) is a C-M ring, because it is a quotient of a C-M ring by an ideal generated by a regular element (B is an integral domain) (see [21, Proposition 18.13]). Since T is equidimensional, dim T = dim T / ( f j ) and dim ( T / ( f j ) / ( y 1 ¯ , , y d ¯ ) = dim B / ( f j , y 1 , , y d ) dim B / ( f , y 1 , , y d ) = dim T / ( y 1 ¯ , , y d ¯ ) = 0 , because ( f j , y 1 , , y d ) ( f , y 1 , , y d ) B . Hence, ( y ¯ ) = { y 1 ¯ , , y d ¯ } T / ( f j ) is a system of parameters and so it is a regular sequence. Thus, H 1 ( y ̲ , T / ( f j ¯ ) ) = 0 (see [20, Theorem 16.5]). In conclusion, ( H 0 ( ( y ¯ ) , T / ( f j ¯ ) ) ) ( H 1 ( ( y ¯ ) , T / ( f j ¯ ) ) ) = ( H 0 ( ( y ¯ ) , T / ( f j ¯ ) ) ) = ( ( T / ( f j ¯ ) / ( y 1 ¯ , , y d ¯ ) ) ) 1 > 0 , because T / ( f j ¯ ) / ( y 1 ¯ , , y d ¯ ) ( 0 ) . This proves the SPCS for T.□

Typical examples of local C-M rings which are UFD are localizations of polynomial rings in prime ideals, or rings of formal power series over DVD or fields. More generally, regular local rings fulfil these conditions (see [21, Corollary 18.7, Theorem 19.19]).

8 A new proof of DSC in the positive characteristic case

Now, we present a new proof of the DSC in the positive characteristic case by proving some particular case of the SPCW. The new key ingredient is the following Lemma. We refer to [26, Proposition 5.2.6] for a proof.

Lemma 13

Let R be a Noetherian ring, M an R-module, and x 1 , , x n be a sequence of elements in R such that M / ( x 1 , , x n ) M has finite length. Let i { 1 , , n } . Then, there exists a constant c such that the length of the Koszul homology module ( H i ( x 1 q , , x n q ; M ) ) c q n i , for all positive integers q .

Proposition 14

Let ( R , m , k ) be an equicharacteristic regular local ring, with char R = p > 0 , and R S a finite extension. Then R S splits.

Proof

After tensoring with the completion of R, which is faithfully flat, we can assume, by previous comments, that R is complete. By Cohen’s structure theorem (see [24, Theorem, p. 26]), R k [ [ x 1 , , x n ] ] , where char k = p > 0 . Now, we can assume that k is perfect (i.e. k p = k ), because each extension of the tower R k ¯ k R k ¯ [ [ x 1 , , x n ] ] is faithfully flat, where k ¯ denotes an algebraic closure of k. Effectively, R k ¯ k R is R - free and therefore faithfully flat. Besides, we can identify k ¯ k R with i I E i [ [ x 1 , , x n ] ] , where E i runs over all field extensions k E i k ¯ , such that [ E : k ] < + . From this, we see that the completion of the local ring k ¯ k R is exactly k ¯ [ [ x 1 , , x n ] ] and so k ¯ k R k ¯ [ [ x 1 , , x n ] ] is faithfully flat.

Again, by Theorem 3 and Proposition 4 we can assume that S T / J , where T is a Gorenstein local ring and ht J = 0 . Moreover, by Proposition 8 and after tensoring with the completion of R 1 = R [ w 1 , , w r ] ( m + ( w 1 , , w r ) ) , which is isomorphic to k ¯ [ [ X 1 , , x m ] ] (for some m n ), we see that R 1 ^ R 1 ( R 1 R T ) has exactly the same form as in Proposition 4. But, now we can assume that J is a principal ideal generated by a zero divisor. In conclusion, we can assume that R = k [ [ x 1 , , x m ] ] , where k is a perfect field and S = T / ( z ) , where T is a Gorenstein local ring and z is a zero divisor.

Now, we set P q = ( x 1 q , , x n q ) T , where q is a power of p. Note that R q = k [ [ x 1 q , , x n q ] ] R is finite and thus by the proof of Theorem 11 R q S splits if and only if

δ = δ R q ( x 1 q , ,   x n q ; T / ( z ) ) R q ( H 0 ( x 1 q , , x n q ; T / ( z ) ) R q ( H 1 ( x 1 q , , x n q ; T / ( z ) ) > 0 .

Besides, it is elementary to see that

δ = δ R ( x 1 q , , x n q ; T / ( z ) ) ,

because the degree of the residue field extension of R q R is one and then

R ( H i R ( x 1 q , , x n q ; T / ( z ) ) ) = R q ( H i R ( x 1 q , , x n q ; T / ( z ) ) ) = R q ( H i R q ( x 1 q , , x n q ; T / ( z ) ) ) .

The last equality holds because the last two Kozsul homology groups are isomorph as R q -modules.

We prove that lim q + δ R ( x 1 q , , x n q ; T / ( z ) ) = + . In fact, since { x 1 q , , x n q } T is a system of parameters for the R - module T / ( z ) , we know that

χ ( x 1 q , x n q ; T / ( z ) ) = q n χ ( x 1 , , x n ; T / ( z ) )

(see Corollary 5.2.4 [26]) and by previous comments,

χ ( x 1 , , x n ; T / ( z ) ) = e ( ( x 1 , , x n ) , T / ( z ) ) > 0 ,

because dim R > 0 (if dim R = 0 , then R is a field and the DSC is trivial). Besides, by the previous corollary we know that there is a constant c such that R ( H i ( x 1 q , , x n q ; T / ( z ) ) ) < c q n i for each i = 1 , , n and each q. Combining this we get the following estimate:

δ R ( x 1 q , , x n q ; T / ( z ) ) = q n e + i = 2 n ( 1 ) i + 1 R ( H i R ( x 1 q , , x n q ; T / ( z ) ) q n e i = 2 n c q n i ,

where e = e ( x 1 q , , x n q ; T / ( z ) ) > 0 . Let us write f ( q ) q n e i = 2 n c q n i . Then, the polynomial f ( q ) + , when q + , because it has positive leading coefficient. Therefore, δ R ( x 1 q , , x n q ; T / ( z ) ) = + . Let us fix b = p h > 0 . Then by the proof of Theorem 11 R b T / ( z ) splits. Denote by ρ 1 : T / ( z ) R b a splitting R b - homomorphism. Finally, the Frobenius homomorphism F b : R R b sending x x b is an isomorphism of rings. Hence, we can define ρ : T / ( z ) R , by ρ ( x ) F b 1 ( ρ 1 ( x b ) ) . Clearly, ρ ( 1 ) = 1 and if x T / ( z ) and r R , then

ρ ( r x ) = F b 1 ( ρ 1 ( ( r x ) b ) ) = F b 1 ( ρ 1 ( r b x b ) ) = F b 1 ( r b ρ 1 ( x b ) ) = F b 1 ( r b ) F b 1 ( ρ 1 ( x b ) ) = r ρ ( x ) .

So ρ is R-linear. In view of that R T / ( z ) splits.□

Acknowledgments

The authors would like to express their gratitude to the German Academic Exchange Service (DAAD) and to the Universidad Nacional de Colombia for all the support. Danny A. J. Gómez-Ramírez thanks all the members of the family of Gómez Torres and Jenn P. D. for all the support and motivation; and sincere thanks goes to the anonymous reviewers for all the useful comments. Finally, Danny A. J. Gómez-Ramírez was also supported by the Institución Universitaria Pascual Bravo, by the Consortium VisionReal (CVR), and previously by the Vienna Science and Technology Fund (WWTF) as part of the Vienna Research Group 12-004.

References

[1] M. Hochster, Homological conjectures, old and new, Illinois J. Math. 51 (2007), 151–159.10.1215/ijm/1258735330Search in Google Scholar

[2] C. Peskine and L. Szpiro, Dimension projective finie et cohomologie locale, I.H.E.S. Publ. Math. 42 (1973), 323–395.10.1007/BF02685877Search in Google Scholar

[3] J.-P. Serre, Local Algebra, Springer Science & Business Media, Berlin Heidelberg, 2012.Search in Google Scholar

[4] M. Hochster, The direct summand conjecture and canonical elements in local cohomology modules, J. Algebra 84 (1983), 503–553.10.1016/0021-8693(83)90092-3Search in Google Scholar

[5] T. Ohi, Direct summand conjecture and descent for flatness, Proc. Am. Math. Soc. 124 (1996), no. 7, 1967–1968.10.1090/S0002-9939-96-03270-4Search in Google Scholar

[6] P. Roberts, Heitmann’s proof of the direct summand conjecture in dimension three, arXiv:math/0212073, 2002.10.2307/3597204Search in Google Scholar

[7] J. McCullough, On the Strong Direct Summand Conjecture, PhD thesis, University of Illinois at Urbana-Champaign, 2009.10.1090/S0002-9939-09-09853-0Search in Google Scholar

[8] Y. André, La conjecture du facteur direct, Pub. Math. de l’IHÉS 127 (2018), no. 1, 71–93.10.1007/s10240-017-0097-9Search in Google Scholar

[9] B. Bhatt, On the direct summand conjecture and its derived variant, Invent. Math. 212 (2018), no. 2, 297–317.10.1007/s00222-017-0768-7Search in Google Scholar

[10] S. P. Dutta and P. Griffith, Intersection multiplicities, canonical element conjecture and the syzygy problem, Michigan Math. J. 57 (2008), no. 227–247.10.1307/mmj/1220879406Search in Google Scholar

[11] J. D. Vélez and R. Flórez, Failure of splitting from module-finite extension rings, Beitr. Algebra Geom. 41 (2000), no. 2, 345–57.Search in Google Scholar

[12] L. Junes, El Problema de Existencia de Retracciones en Extensiones Modulo Finitas de Anillos Conmutativos, PhD thesis, Universidad Nacional de Colombia (Sede Medellín), Medellín (Colombia), 2002.Search in Google Scholar

[13] C. Huneke, A remark concerning multiplicities, Proc. Amer. Math. Soc. 85 (1982), 331–332.10.1090/S0002-9939-1982-0656095-2Search in Google Scholar

[14] J. Vélez, Openness of F-rational locus, smooth base change and Koh’s conjecture. PhD thesis, Michigan University, Ann-Harbor, 1993.Search in Google Scholar

[15] J. D. Vélez, Splitting results in module-finite extension rings and Koh’s conjecture, J. Algebra 172 (1995), 454–469, 1995.10.1016/S0021-8693(05)80011-0Search in Google Scholar

[16] J. R. Strooker and J. Stückrad, Monomial conjecture and complete intersections, Manuscr. Math. 79 (1993) no. 2, 153–159.10.1007/BF02568334Search in Google Scholar

[17] Danny AJ Gómez Ramírez, Artificial Mathematical Intelligence: Cognitive, (Meta)mathematical, Physical and Philosophical Foundations, Springer International Publishing, Cham, ISBN: 978-3-030-50272-0, e-ISBN: 978-3-030-50273-7, 2020.10.1007/978-3-030-50273-7Search in Google Scholar

[18] M. F. Atiyah and I. Macdonald, Introduction to Commutative Algebra, Reading, MA: Addison-Wesley, 1969.Search in Google Scholar

[19] T. S. P. Authors, The Stacks Project, https://stacks.math.columbia.edu, 2020.Search in Google Scholar

[20] H. Matsumura, Commutative Ring Theory, Cambridge: Cambridge University Press, 1989.Search in Google Scholar

[21] D. Eisenbud, Commutative Algebra with a View Toward Algebraic Geometry, Springer-Verlag, New York, 1995.10.1007/978-1-4612-5350-1Search in Google Scholar

[22] W. Bruns and U. Vetter, Determinantal Rings, Lecture Notes in Mathematics, vol. 1327, Springer-Verlag Berlin Heidelberg, 1988.10.1007/BFb0080378Search in Google Scholar

[23] I. S. Cohen, On the structure and ideal theory of complete local rings, Trans. Amer. Math. Soc. 59 (1946), 54–106.10.1090/S0002-9947-1946-0016094-3Search in Google Scholar

[24] M. Hochster, The structure theory of complete local rings, http://www.math.lsa.umich.edu/hochster/615W10/supStructure.pdf.Search in Google Scholar

[25] K. J. Nowak, A proof of the criterion for multiplicity one, Univ. Iagellonicae Acta Math. 35 (1997), 247–249.Search in Google Scholar

[26] P. Roberts, Multiplicities and Chern Classes in Local Algebra, Cambridge University Press, Cambridge (England), 1998.10.1017/CBO9780511529986Search in Google Scholar

Received: 2019-10-01
Revised: 2020-06-24
Accepted: 2020-09-09
Published Online: 2020-11-24

© 2020 Juan Diego Vélez-Caicedo and Danny Arlen de Jesús Gómez-Ramírez, published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

  1. Regular Articles
  2. Non-occurrence of the Lavrentiev phenomenon for a class of convex nonautonomous Lagrangians
  3. Strong and weak convergence of Ishikawa iterations for best proximity pairs
  4. Curve and surface construction based on the generalized toric-Bernstein basis functions
  5. The non-negative spectrum of a digraph
  6. Bounds on F-index of tricyclic graphs with fixed pendant vertices
  7. Crank-Nicolson orthogonal spline collocation method combined with WSGI difference scheme for the two-dimensional time-fractional diffusion-wave equation
  8. Hardy’s inequalities and integral operators on Herz-Morrey spaces
  9. The 2-pebbling property of squares of paths and Graham’s conjecture
  10. Existence conditions for periodic solutions of second-order neutral delay differential equations with piecewise constant arguments
  11. Orthogonal polynomials for exponential weights x2α(1 – x2)2ρe–2Q(x) on [0, 1)
  12. Rough sets based on fuzzy ideals in distributive lattices
  13. On more general forms of proportional fractional operators
  14. The hyperbolic polygons of type (ϵ, n) and Möbius transformations
  15. Tripled best proximity point in complete metric spaces
  16. Metric completions, the Heine-Borel property, and approachability
  17. Functional identities on upper triangular matrix rings
  18. Uniqueness on entire functions and their nth order exact differences with two shared values
  19. The adaptive finite element method for the Steklov eigenvalue problem in inverse scattering
  20. Existence of a common solution to systems of integral equations via fixed point results
  21. Fixed point results for multivalued mappings of Ćirić type via F-contractions on quasi metric spaces
  22. Some inequalities on the spectral radius of nonnegative tensors
  23. Some results in cone metric spaces with applications in homotopy theory
  24. On the Malcev products of some classes of epigroups, I
  25. Self-injectivity of semigroup algebras
  26. Cauchy matrix and Liouville formula of quaternion impulsive dynamic equations on time scales
  27. On the symmetrized s-divergence
  28. On multivalued Suzuki-type θ-contractions and related applications
  29. Approximation operators based on preconcepts
  30. Two types of hypergeometric degenerate Cauchy numbers
  31. The molecular characterization of anisotropic Herz-type Hardy spaces with two variable exponents
  32. Discussions on the almost 𝒵-contraction
  33. On a predator-prey system interaction under fluctuating water level with nonselective harvesting
  34. On split involutive regular BiHom-Lie superalgebras
  35. Weighted CBMO estimates for commutators of matrix Hausdorff operator on the Heisenberg group
  36. Inverse Sturm-Liouville problem with analytical functions in the boundary condition
  37. The L-ordered L-semihypergroups
  38. Global structure of sign-changing solutions for discrete Dirichlet problems
  39. Analysis of F-contractions in function weighted metric spaces with an application
  40. On finite dual Cayley graphs
  41. Left and right inverse eigenpairs problem with a submatrix constraint for the generalized centrosymmetric matrix
  42. Controllability of fractional stochastic evolution equations with nonlocal conditions and noncompact semigroups
  43. Levinson-type inequalities via new Green functions and Montgomery identity
  44. The core inverse and constrained matrix approximation problem
  45. A pair of equations in unlike powers of primes and powers of 2
  46. Miscellaneous equalities for idempotent matrices with applications
  47. B-maximal commutators, commutators of B-singular integral operators and B-Riesz potentials on B-Morrey spaces
  48. Rate of convergence of uniform transport processes to a Brownian sheet
  49. Curves in the Lorentz-Minkowski plane with curvature depending on their position
  50. Sequential change-point detection in a multinomial logistic regression model
  51. Tiny zero-sum sequences over some special groups
  52. A boundedness result for Marcinkiewicz integral operator
  53. On a functional equation that has the quadratic-multiplicative property
  54. The spectrum generated by s-numbers and pre-quasi normed Orlicz-Cesáro mean sequence spaces
  55. Positive coincidence points for a class of nonlinear operators and their applications to matrix equations
  56. Asymptotic relations for the products of elements of some positive sequences
  57. Jordan {g,h}-derivations on triangular algebras
  58. A systolic inequality with remainder in the real projective plane
  59. A new characterization of L2(p2)
  60. Nonlinear boundary value problems for mixed-type fractional equations and Ulam-Hyers stability
  61. Asymptotic normality and mean consistency of LS estimators in the errors-in-variables model with dependent errors
  62. Some non-commuting solutions of the Yang-Baxter-like matrix equation
  63. General (p,q)-mixed projection bodies
  64. An extension of the method of brackets. Part 2
  65. A new approach in the context of ordered incomplete partial b-metric spaces
  66. Sharper existence and uniqueness results for solutions to fourth-order boundary value problems and elastic beam analysis
  67. Remark on subgroup intersection graph of finite abelian groups
  68. Detectable sensation of a stochastic smoking model
  69. Almost Kenmotsu 3-h-manifolds with transversely Killing-type Ricci operators
  70. Some inequalities for star duality of the radial Blaschke-Minkowski homomorphisms
  71. Results on nonlocal stochastic integro-differential equations driven by a fractional Brownian motion
  72. On surrounding quasi-contractions on non-triangular metric spaces
  73. SEMT valuation and strength of subdivided star of K 1,4
  74. Weak solutions and optimal controls of stochastic fractional reaction-diffusion systems
  75. Gradient estimates for a weighted nonlinear parabolic equation and applications
  76. On the equivalence of three-dimensional differential systems
  77. Free nonunitary Rota-Baxter family algebras and typed leaf-spaced decorated planar rooted forests
  78. The prime and maximal spectra and the reticulation of residuated lattices with applications to De Morgan residuated lattices
  79. Explicit determinantal formula for a class of banded matrices
  80. Dynamics of a diffusive delayed competition and cooperation system
  81. Error term of the mean value theorem for binary Egyptian fractions
  82. The integral part of a nonlinear form with a square, a cube and a biquadrate
  83. Meromorphic solutions of certain nonlinear difference equations
  84. Characterizations for the potential operators on Carleson curves in local generalized Morrey spaces
  85. Some integral curves with a new frame
  86. Meromorphic exact solutions of the (2 + 1)-dimensional generalized Calogero-Bogoyavlenskii-Schiff equation
  87. Towards a homological generalization of the direct summand theorem
  88. A standard form in (some) free fields: How to construct minimal linear representations
  89. On the determination of the number of positive and negative polynomial zeros and their isolation
  90. Perturbation of the one-dimensional time-independent Schrödinger equation with a rectangular potential barrier
  91. Simply connected topological spaces of weighted composition operators
  92. Generalized derivatives and optimization problems for n-dimensional fuzzy-number-valued functions
  93. A study of uniformities on the space of uniformly continuous mappings
  94. The strong nil-cleanness of semigroup rings
  95. On an equivalence between regular ordered Γ-semigroups and regular ordered semigroups
  96. Evolution of the first eigenvalue of the Laplace operator and the p-Laplace operator under a forced mean curvature flow
  97. Noetherian properties in composite generalized power series rings
  98. Inequalities for the generalized trigonometric and hyperbolic functions
  99. Blow-up analyses in nonlocal reaction diffusion equations with time-dependent coefficients under Neumann boundary conditions
  100. A new characterization of a proper type B semigroup
  101. Constructions of pseudorandom binary lattices using cyclotomic classes in finite fields
  102. Estimates of entropy numbers in probabilistic setting
  103. Ramsey numbers of partial order graphs (comparability graphs) and implications in ring theory
  104. S-shaped connected component of positive solutions for second-order discrete Neumann boundary value problems
  105. The logarithmic mean of two convex functionals
  106. A modified Tikhonov regularization method based on Hermite expansion for solving the Cauchy problem of the Laplace equation
  107. Approximation properties of tensor norms and operator ideals for Banach spaces
  108. A multi-power and multi-splitting inner-outer iteration for PageRank computation
  109. The edge-regular complete maps
  110. Ramanujan’s function k(τ)=r(τ)r2(2τ) and its modularity
  111. Finite groups with some weakly pronormal subgroups
  112. A new refinement of Jensen’s inequality with applications in information theory
  113. Skew-symmetric and essentially unitary operators via Berezin symbols
  114. The limit Riemann solutions to nonisentropic Chaplygin Euler equations
  115. On singularities of real algebraic sets and applications to kinematics
  116. Results on analytic functions defined by Laplace-Stieltjes transforms with perfect ϕ-type
  117. New (p, q)-estimates for different types of integral inequalities via (α, m)-convex mappings
  118. Boundary value problems of Hilfer-type fractional integro-differential equations and inclusions with nonlocal integro-multipoint boundary conditions
  119. Boundary layer analysis for a 2-D Keller-Segel model
  120. On some extensions of Gauss’ work and applications
  121. A study on strongly convex hyper S-subposets in hyper S-posets
  122. On the Gevrey ultradifferentiability of weak solutions of an abstract evolution equation with a scalar type spectral operator on the real axis
  123. Special Issue on Graph Theory (GWGT 2019), Part II
  124. On applications of bipartite graph associated with algebraic structures
  125. Further new results on strong resolving partitions for graphs
  126. The second out-neighborhood for local tournaments
  127. On the N-spectrum of oriented graphs
  128. The H-force sets of the graphs satisfying the condition of Ore’s theorem
  129. Bipartite graphs with close domination and k-domination numbers
  130. On the sandpile model of modified wheels II
  131. Connected even factors in k-tree
  132. On triangular matroids induced by n3-configurations
  133. The domination number of round digraphs
  134. Special Issue on Variational/Hemivariational Inequalities
  135. A new blow-up criterion for the Nabc family of Camassa-Holm type equation with both dissipation and dispersion
  136. On the finite approximate controllability for Hilfer fractional evolution systems with nonlocal conditions
  137. On the well-posedness of differential quasi-variational-hemivariational inequalities
  138. An efficient approach for the numerical solution of fifth-order KdV equations
  139. Generalized fractional integral inequalities of Hermite-Hadamard-type for a convex function
  140. Karush-Kuhn-Tucker optimality conditions for a class of robust optimization problems with an interval-valued objective function
  141. An equivalent quasinorm for the Lipschitz space of noncommutative martingales
  142. Optimal control of a viscous generalized θ-type dispersive equation with weak dissipation
  143. Special Issue on Problems, Methods and Applications of Nonlinear analysis
  144. Generalized Picone inequalities and their applications to (p,q)-Laplace equations
  145. Positive solutions for parametric (p(z),q(z))-equations
  146. Revisiting the sub- and super-solution method for the classical radial solutions of the mean curvature equation
  147. (p,Q) systems with critical singular exponential nonlinearities in the Heisenberg group
  148. Quasilinear Dirichlet problems with competing operators and convection
  149. Hyers-Ulam-Rassias stability of (m, n)-Jordan derivations
  150. Special Issue on Evolution Equations, Theory and Applications
  151. Instantaneous blow-up of solutions to the Cauchy problem for the fractional Khokhlov-Zabolotskaya equation
  152. Three classes of decomposable distributions
Downloaded on 14.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/math-2020-0087/html
Scroll to top button