Papers
Topics
Authors
Recent
Search
2000 character limit reached

Quasi-Faithfully Flat Extensions

Updated 5 July 2026
  • Quasi-faithfully flat extensions are finite ring maps defined via an auxiliary bimodule, relaxing classical flatness to preserve k-torsionfreeness.
  • They enable a change-of-rings framework that compares classical and Gorenstein transposes, ensuring Ext-vanishing properties persist across rings.
  • They are pivotal in transferring extension-closedness, quasi-k-Gorenstein properties, and finite representation type, with applications to Frobenius extensions and skew group rings.

Searching arXiv for the specified paper and closely related work. Search 1: exact paper by title keywords. Searching arXiv for: "Base change of (Gorenstein) transpose, k-torsionfree modules, and quasi-faithfully flat extensions" Quasi-faithfully flat extensions are finite ring homomorphisms φ ⁣:RA\varphi\colon R\to A designed to retain much of the change-of-rings control usually associated with faithfully flat extensions, while permitting substantially weaker hypotheses. In the setting where RR is a two-sided Noetherian ring and modules are finitely generated, the notion is used to compare transpose constructions, kk-torsionfree modules, extension-closedness, quasi-kk-Gorensteiness, and finite representation type across RR and AA. The framework developed in "Base change of (Gorenstein) transpose, kk-torsionfree modules, and quasi-faithfully flat extensions" establishes a close relationship between the classical transpose of an AA-module and the Gorenstein transpose of a suitable syzygy over RR, extends a result of Zhao, and gives applications to Frobenius extensions and skew group rings (Liu, 16 Jul 2025).

1. Ambient setting and kk-torsionfree modules

Throughout, RR0 is a Noetherian ring on both sides, and

RR1

For RR2, choose a projective presentation

RR3

and apply RR4. This yields an exact sequence of right modules

RR5

where RR6 is the transpose of RR7 (Liu, 16 Jul 2025).

For each integer RR8, the module RR9 is called kk0-torsionfree when

kk1

The full subcategory of kk2-torsionfree modules is denoted by kk3. This definition packages Ext-vanishing of the transpose into a categorical condition that is stable enough to support descent and ascent arguments.

In the paper’s broader program, kk4-torsionfree modules are the main invariants transported across finite ring homomorphisms. A central feature is that their behavior can be controlled not merely by flatness of kk5 over kk6, but by the existence of an auxiliary bimodule implementing an appropriate base-change mechanism.

2. Definition of quasi-faithfully flat extensions

A finite ring homomorphism kk7 is called a quasi-faithfully flat extension if there exists an kk8-kk9-bimodule kk0 such that:

  1. kk1 is finitely generated and projective as a left kk2-module,
  2. kk3 is faithfully flat as a right kk4-module, and
  3. kk5 is flat over kk6 (Liu, 16 Jul 2025).

This definition weakens the classical requirement of faithful flatness on both sides for kk7 itself. When kk8 is flat and faithfully flat over both kk9 and RR0, one recovers the usual notion of faithfully flat extension. The key distinction is that quasi-faithful flatness is witnessed by a possibly different bimodule RR1, rather than by RR2 alone.

Several examples clarify the scope of the notion. Any faithfully flat extension is automatically quasi-faithfully flat. If RR3 is a field and RR4 is a Frobenius RR5-algebra in the sense that RR6 as RR7-RR8-bimodules, then for any RR9-algebra map AA0 one may take

AA1

In this case AA2 is free, hence projective, over AA3, faithfully flat over AA4, and AA5 is free over AA6. Consequently, every AA7-algebra homomorphism out of a Frobenius AA8-algebra is quasi-faithfully flat. There are also non-projective examples: the canonical surjection

AA9

is quasi-faithfully flat although kk0 is not projective over kk1 (Liu, 16 Jul 2025).

A common misconception is that a useful “faithful flatness” substitute must force kk2 itself to be projective or flat over kk3. The example above shows that quasi-faithful flatness is genuinely weaker: it is formulated so that the transfer of homological properties can still proceed through kk4.

3. Base change for transpose and detection of kk5-torsionfreeness

The mechanism behind the theory is a functorial comparison between transpose over kk6 and transpose after tensoring to kk7. If kk8 and kk9 is an AA0-AA1-bimodule that is projective and finitely generated over AA2, then there is a canonical isomorphism of right AA3-modules

AA4

This transfer-of-transpose statement is the basic change-of-rings computation from which the later results follow (Liu, 16 Jul 2025).

Under additional flatness hypotheses, transpose comparison yields preservation and detection of AA5-torsionfreeness. If AA6 is flat over AA7 and AA8 is flat over AA9, then for each RR0,

RR1

If, in addition, RR2 is faithfully flat on the right, then the converse also holds. Thus under a quasi-faithfully flat extension one obtains the equivalence

RR3

The paper’s conceptual summary identifies the heart of the argument as the transpose isomorphism together with flatness assumptions that allow one to compare Ext-groups and detect their vanishing across the two rings. This is the precise sense in which quasi-faithfully flat extensions serve as a substitute for classical faithfully flat base change: they preserve the homological criterion defining RR4-torsionfree modules (Liu, 16 Jul 2025).

4. Extension-closedness and change of rings

A full subcategory RR5 is extension closed if every short exact sequence

RR6

with RR7 also has RR8. For the subcategories RR9, quasi-faithfully flat extensions support both descent and ascent results (Liu, 16 Jul 2025).

The descent statement is direct: if kk0 is quasi-faithfully flat and kk1 is extension closed in kk2, then kk3 is extension closed in kk4. The ascent statement requires stronger homological input. Suppose kk5 is finite and satisfies

kk6

for some projective kk7 and kk8. If either kk9 or RR00 is commutative satisfying the usual Gorenstein-depth condition RR01, then extension-closedness of RR02 implies extension-closedness of RR03.

In the especially important case RR04, these two directions combine into a clean equivalence. If RR05 is a finite quasi-faithfully flat extension, with RR06 Noetherian, and

RR07

then for every RR08 the category RR09 is extension closed in RR10 if and only if RR11 is extension closed in RR12. The paper further notes that in this special case one has the stronger identification

RR13

so the theory collapses to a particularly transparent form of ascent and descent.

5. Consequences for quasi-RR14-Gorenstein algebras

For a Noetherian algebra RR15, being left quasi-RR16-Gorenstein means that in the minimal injective resolution

RR17

one has

RR18

Huang’s theorem identifies this condition with extension-closedness of each RR19 for RR20 (Liu, 16 Jul 2025).

Because quasi-faithfully flat extensions preserve extension-closedness under the hypotheses above, they also preserve quasi-RR21-Gorensteiness. If RR22 satisfies the hypotheses of the equivalence theorem for extension-closedness and both RR23 and RR24 are Noetherian algebras, then for every RR25,

RR26

This yields an affirmative answer to a question posed by Zhao in the case where both rings are Noetherian algebras.

A notable special case is that of Frobenius extensions. If RR27 is a Frobenius extension of Noetherian algebras and RR28 is also faithfully flat over RR29, then RR30 is left quasi-RR31-Gorenstein if and only if RR32 is. In the commutative case, the paper observes that this forces left-right symmetry of quasi-RR33-Gorensteinness for RR34 (Liu, 16 Jul 2025).

6. Frobenius extensions, finite representation type, and skew group rings

A full subcategory RR35 has finite representation type if, up to isomorphism, it contains only finitely many indecomposable objects. Under the Krull-Remak-Schmidt hypothesis, such as when RR36 is artinian or Henselian local, Frobenius extensions also transfer finiteness properties of RR37 (Liu, 16 Jul 2025).

If RR38 is a Frobenius extension and Krull-Remak-Schmidt holds over both RR39 and RR40, then two complementary statements hold. First, if RR41 is separable and RR42 has finite representation type, then RR43 has finite representation type. Second, if RR44 is split and RR45 has finite representation type, then RR46 has finite representation type. The separable case gives ascent, while the split case gives descent.

The standard application is to skew group rings. If RR47 is a finite group acting on an artinian ring RR48 and RR49 is invertible in RR50, then the skew group ring RR51 is a separable, split Frobenius extension of RR52. Hence, for each RR53,

RR54

The same argument, applied to Gorenstein-projective modules, gives an equivalence of Cohen-Macaulay finiteness: RR55 These consequences place quasi-faithfully flat and Frobenius-type change-of-rings phenomena in direct contact with representation-theoretic finiteness questions.

7. Conceptual role within transpose and syzygy theory

The topic is situated within a broader change-of-rings analysis of transpose and Gorenstein transpose. Under suitable homological conditions on RR56 over RR57, the paper establishes a close relationship between the classical transpose of a finitely generated left RR58-module and the Gorenstein transpose of a certain syzygy module of that module over RR59 (Liu, 16 Jul 2025). The quasi-faithfully flat framework then isolates the hypotheses needed for this comparison to control the more concrete invariant of RR60-torsionfreeness.

The resulting picture is structurally coherent. Transpose comparison yields equivalence of Ext-vanishing conditions; Ext-vanishing controls membership in RR61; preservation of RR62 feeds into extension-closedness; extension-closedness characterizes quasi-RR63-Gorensteiness in the Noetherian algebra setting; and in the Frobenius, separable, and split contexts, the same formalism reaches finite representation type and skew group rings. A plausible implication is that quasi-faithfully flatness is best understood not as a variant of flatness in isolation, but as a change-of-rings device tailored to transpose-based homological invariants.

Within this framework, the notion’s significance lies in its precision. It is weak enough to include examples where RR64 is not projective over RR65, yet strong enough to make RR66-torsionfreeness, extension-closedness, and several Gorenstein-flavored finiteness properties transport reliably across finite extensions.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (1)

Topic to Video (Beta)

No one has generated a video about this topic yet.

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Follow Topic

Get notified by email when new papers are published related to Quasi-Faithfully Flat Extensions.