3
$\begingroup$

Let $R$ be an associative ring with identity.

Recall that a pair of classes of modules $(\mathcal{A}, \mathcal{B})$ in $R\text{-Mod}$ is called a cotorsion pair if
$$ \mathcal{A} = {}^{\perp}\mathcal{B} \quad\text{and}\quad \mathcal{B} = \mathcal{A}^{\perp}, $$ where
$$ {}^{\perp}\mathcal{B} = \{ M \mid Ext^1_R(M, B) = 0 \text{ for all } B \in \mathcal{B} \}, \qquad \mathcal{A}^{\perp} = \{ N \mid Ext^1_R(A, N) = 0 \text{ for all } A \in \mathcal{A} \}. $$ The cotorsion pair is said to be complete if every module $M$ admits both a special $\mathcal{A}$-precover and a special $\mathcal{B}$-preenvelope.

It is well known that for a commutative domain $R$, the cotorsion pair $(\mathcal{P}_1, \mathcal{D})$ — where
$\mathcal{P}_1$ denotes the class of all modules of projective dimension at most $1$, and
$\mathcal{D}$ denotes the class of all divisible modules — is complete.

In this situation we have the following equivalences: $$ R \text{ is a field } \ \Longleftrightarrow\ \mathcal{D} = R\text{-Mod} \ \Longleftrightarrow\ \text{every module in } \mathcal{P}_1 \text{ is projective.} $$

In the general (not necessarily commutative) case, it is known that
$(\mathcal{P}_1, \mathcal{P}_1^{\perp})$ is a complete cotorsion pair.
Hence the following statements are equivalent:

  1. $\mathcal{P}_1^{\perp} = R\text{-Mod}$;
  2. Every module in $\mathcal{P}_1$ is projective;
  3. Every quotient of a projective module by a projective submodule is projective;
  4. $R$ has no (non-projective) module of projective dimension $1$.

Question.
Are rings satisfying these equivalent conditions already known or characterized in the literature?
Equivalently,

$$ \text{Which rings have the property that every module of projective dimension at most $1$ is projective?} $$

I could not find a reference or a specific name for such rings.
Are they related to hereditary, semihereditary, or perfect rings, or do they form a strictly smaller or larger class?

$\endgroup$
2
  • 3
    $\begingroup$ Your rings don't have much overlap with hereditary rings, since hereditary rings will often have modules of proj. dim. exactly 1. On the other hand, IIRC over a non-semisimple quasi-Frobenius ring you have some modules of projective dimension 0, and some of infinite proj. dim., but nothing in between, so I think quasi-Frob. rings (e.g. fin-dim'l Hopf algebras over a field) are a subclass of the class of rings you're asking about. Maybe more generally, 0-dim'l Gorenstein rings: see Buchweitz "Maximal Cohen-Macaulay modules and Tate cohomology over Gorenstein rings" section 8.1. $\endgroup$ Commented Oct 22 at 22:49
  • 3
    $\begingroup$ Here is a surprising example. Let $p$ be a prime number, and let $A$ be the mod $p$ Steenrod algebra. This is an infinite-dimensional graded connected Hopf algebra over $\mathbb{F}_p$. Margolis proved that the graded-projective dimension of any graded $A$-module is $0$, $1$, or $\infty$. But if you restrict to just the bounded below graded $A$-modules, then Margolis proved they all have graded-projective dimension $0$ or $\infty$. So the Steenrod algebra satisfies the graded version of your condition iff you restrict to just the bounded below graded $A$-modules. Weird! $\endgroup$ Commented Oct 22 at 23:15

1 Answer 1

5
$\begingroup$

These are the rings of finitistic dimension zero.

They are studied in detail in Bass's 1960 paper Finitistic dimension and a homological generalization of semi-primary rings (Trans. AMS, vol.95, no. 3, pp. 466-488), where he gives the following equivalent conditions:

  1. Every left $R$-module of finite projective dimension is projective.
  2. $R$ is left perfect and every finitely generated left $R$-module of finite projective dimension is projective.
  3. $R$ is left perfect and every finitely generated proper right ideal of $R$ has nonzero left annihilator.
  4. $R$ is left perfect and every simple left $R$-module is a quotient of an injective left $R$-module.
$\endgroup$

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.