Projective Covers

Q: Is the projective cover functorial in M?

Not naively. A map M → M' lifts to a map of covers when both exist, by projectivity, but the lift is not unique and the assignment does not preserve composition on the nose. What is canonical is the isomorphism class of the cover, guaranteed by (24.10).

Q: If radM ≠ M, does M have a projective cover?

No. That condition is necessary by (24.11)(4) but not sufficient. Z/2Z over Z has zero radical, hence proper radical, yet has no projective cover because Z is J-semisimple and Z/2Z is not projective.

Q: Can a projective cover have zero kernel?

Yes, exactly when M is projective, in which case the identity map is the cover. So *projective* and *equal to its own projective cover* are the same condition, and this is what makes (24.11)(5) work over J-semisimple rings.

Q: How do projective covers give minimal resolutions?

Take a cover P_0 → M, then a cover P_1 → (P_0 → M), and iterate. Each syzygy is covered with no redundancy, so the resulting complex has rad-small differentials: the induced maps in P_ ⊗_R R/J are all zero, and the ranks of the P_i are exactly the Betti numbers of M.

Executive Summary

Every module is a quotient of a free module, so projective approximations always exist. A projective cover demands the approximation be efficient: the epimorphism $θ : P ↠ M$ must have $\ker θ$ small in $P$ , so that no proper submodule of $P$ already maps onto $M$ .

Two things follow. Covers are unique up to isomorphism when they exist — the proof is a three-line splitting argument — and they frequently do not exist. Over $ℤ$ the only modules with projective covers are the free abelian groups. The classes of rings for which they always exist are exactly the semiperfect and right perfect rings, which is the content of the two characterisation theorems later in §24.

$\ker θ \subseteq_{s} P$ Defining condition

UniqueUp to isomorphism over

M

$e R \to e R / e J$ The standard example

Free groupsThe

ℤ

-modules with covers

Overview

Modules are right $R$ -modules and $J = rad R$ . Projective covers are the dual of injective hulls: an injective hull is a monomorphism $M ↪ E$ with $E$ injective and essential image, while a projective cover is an epimorphism $P ↠ M$ with $P$ projective and small kernel. The duality is formal, the behaviour is not — injective hulls exist for every module over every ring, projective covers usually do not exist at all.

θ : P ↠ M is a projective cover ⟺ P projective and (θ (P') = M \Rightarrow P' = P) \forall P' \subseteq P

(24.9)

The right-hand condition is the practical form: no proper submodule of $P$ suffices.

The equivalence is immediate. If $\ker θ \subseteq_{s} P$ and $θ (P') = M$ , then $P' + \ker θ = P$ since any $p \in P$ differs from some $p' \in P'$ by an element of the kernel; smallness gives $P' = P$ . Conversely, applying the condition to $P' = \ker θ + N$ recovers smallness.

This page treats the definition, uniqueness and the elementary supply of examples. Existence for all finitely generated modules is the subject of Projective Covers over Semiperfect Rings; existence for all modules characterises right perfect rings.

Learning Objectives

State $(24.9)$ and prove the equivalence with the no-proper-submodule-suffices formulation.
Prove $(24.10)$ : any epimorphism from a projective module factors through a cover by a split epimorphism, and two covers are isomorphic.
Verify that $P \to P / P J$ is a projective cover when $P$ is finitely generated projective, or when $J$ is right T-nilpotent.
Verify that $e R \to e R / e J$ is a projective cover for every idempotent $e \in R$ and every right ideal $J \subseteq rad R$ .
Prove that a projective cover induces a bijection between the maximal submodules of $P$ and of $M$ .
Decide which $ℤ$ -modules and which $ℤ / 12 ℤ$ -modules have projective covers.

Definitions

Definition(24.9)Projective cover

Let $M$ be a right $R$ -module. A projective cover of $M$ is an epimorphism $θ : P \to M$ where $P$ is a projective right $R$ -module and $\ker θ \subseteq_{s} P$ . One often refers to $P$ itself as the projective cover of $M$ , leaving $θ$ implicit — legitimate because of the uniqueness theorem $(24.10)$ .

$θ : P ↠ M$: An epimorphism of right $R$ -modules. Projectivity of $P$ means every such map lifts along any epimorphism onto $M$ .
Small kernel: $\ker θ \subseteq_{s} P$ : no proper submodule $N ⊊ P$ satisfies $N + \ker θ = P$ .
Injective hull: The dual notion: a monomorphism $M ↪ E$ with $E$ injective and $M$ essential in $E$ . Always exists; the asymmetry with projective covers is the striking feature.
Local idempotent: An idempotent $e$ with $e R e$ local; then $e R$ is indecomposable projective and $e R / e J$ is simple.
Minimal projective resolution: A resolution in which each map onto the next syzygy is a projective cover; it exists exactly when all the syzygies have covers.

Throughout, $J = rad R$ , modules are unital right modules, and projective means projective as a right $R$ -module.

Core Concepts

Why smallness is the right minimality condition

Any module admits a free presentation $F ↠ M$ , but $F$ typically carries redundancy: a proper direct summand of $F$ may already surject onto $M$ . Requiring the kernel to be small removes exactly this slack, since $\ker θ \subseteq_{s} P$ says no submodule of $P$ other than $P$ itself maps onto $M$ .

F ↠ M

freediscard redundant summands

P ↠ M

with

\ker \subseteq_{s} P

unique up to isomorphism

The middle step is exactly what can fail. Discarding redundancy is an infinite process in general, and there is no reason for it to terminate — over $ℤ$ , the surjections $ℤ \to ℤ / 2$ can be trimmed no further, yet the kernel $2 ℤ$ is a direct summand of nothing and is certainly not small in $ℤ$ .

Covers detect maximal submodules

If $θ : P \to M$ is a projective cover, then every maximal submodule of $P$ contains $\ker θ$ , because maximal submodules absorb small ones. Hence $N \mapsto θ^{- 1} (N)$ is a bijection from the maximal submodules of $M$ onto those of $P$ , and consequently

rad P = θ^{- 1} (rad M), θ (rad P) = rad M .

(24.11)(4)

Combined with $rad P ⊊ P$ for nonzero projective $P$ , this yields the first and cheapest non-existence criterion: **a nonzero module with $rad M = M$ has no projective cover.**

Where the supply of covers comes from

All elementary examples come from one construction: quotient a projective module by $P J$ . Smallness of $P J$ is exactly $(24.2) (2)$ , so the construction works whenever $P$ is finitely generated, and for arbitrary $P$ when $J$ is right T-nilpotent. Applying it to $P = e R$ for an idempotent $e$ gives the building blocks used to assemble covers over semiperfect rings.

Key Results

Proposition(24.10)Uniqueness of projective covers

Let $θ : P \to M$ be a projective cover of the right $R$ -module $M$ , and let $θ' : P' \to M$ be any epimorphism with $P'$ projective. Then there is a split epimorphism $α : P' \to P$ with $θ α = θ'$ . If in addition $θ'$ is itself a projective cover, then $α$ is an isomorphism.

Proof

Since $P'$ is projective and $θ$ is onto, the map $θ'$ lifts: there is $α : P' \to P$ with $θ α = θ'$ . Then $θ (α (P')) = θ' (P') = M$ , so by the reformulation of $(24.9)$ applied to the submodule $α (P') \subseteq P$ we get $α (P') = P$ ; that is, $α$ is onto. As $P$ is projective, the epimorphism $α$ splits, so $P' = \ker α \oplus P''$ with $α |_{P''}$ an isomorphism onto $P$ .

Now suppose $θ'$ is also a projective cover. From $θ' = θ α$ we get $\ker α \subseteq \ker θ'$ , and $\ker θ' \subseteq_{s} P'$ , so $\ker α \subseteq_{s} P'$ by $(24.2) (3)$ . But $\ker α$ is a direct summand of $P'$ , and a nonzero direct summand is never small by $(24.2) (1)$ . Hence $\ker α = 0$ and $α$ is an isomorphism.

Proposition(24.11)(1)The standard construction

Let $P$ be a projective right $R$ -module and $J \subseteq rad R$ a right ideal of $R$ . If either $P$ is finitely generated, or $J$ is right T-nilpotent, then the canonical surjection $P \to P / P J$ is a projective cover of $P / P J$ .

Proof

The map is onto with kernel $P J$ , and $P$ is projective by hypothesis. Under either hypothesis $P J \subseteq_{s} P$ by $(24.2) (2)$ — the finitely generated case is Nakayama's Lemma, the T-nilpotent case is the criterion $(23.16)$ . A surjection from a projective module with small kernel is a projective cover.

Corollary(24.11)(2)Covers of the modules

e R / e J

Let $e = e^{2} \in R$ and let $J \subseteq rad R$ be a right ideal. Then $e R \to e R / e J$ is a projective cover. In particular, when $e$ is a local idempotent and $J = rad R$ , the simple module $e R / e rad R$ has projective cover $e R$ .

Proof

$R = e R \oplus (1 - e) R$ exhibits $e R$ as a direct summand of $R_{R}$ , so $e R$ is projective; it is generated by the single element $e$ , hence finitely generated. Now apply $(24.11) (1)$ , noting $(e R) J = e J$ .

Proposition(24.11)(3)Finite direct sums of covers

If $θ_{i} : P_{i} \to M_{i}$ is a projective cover for $1 \leq i \leq n$ , then $⨁_{i = 1}^{n} θ_{i} : ⨁_{i = 1}^{n} P_{i} \to ⨁_{i = 1}^{n} M_{i}$ is a projective cover. The statement fails for infinite direct sums in general.

Proof

A finite direct sum of projective modules is projective, and the kernel of $⨁ θ_{i}$ is $⨁ \ker θ_{i}$ , which is small in $⨁ P_{i}$ by $(24.2) (5)$ .

Proposition(24.11)(4)Covers and maximal submodules

Let $M \neq 0$ and let $θ : P \to M$ be a projective cover. Then $N \mapsto θ^{- 1} (N)$ is a bijection from the maximal submodules of $M$ to those of $P$ , $θ (rad P) = rad M$ , and $rad M ⊊ M$ . Consequently a nonzero module with $rad M = M$ admits no projective cover.

Proof

Every maximal submodule of $P$ contains the small submodule $\ker θ$ by $(24.2) (6)$ , so the usual correspondence for the surjection $θ$ restricts to a bijection between maximal submodules of $P$ and of $M$ . Intersecting, $rad P = θ^{- 1} (rad M)$ , and applying $θ$ gives $θ (rad P) = rad M$ .

If $rad M = M$ then $rad P = θ^{- 1} (M) = P$ . But $M \neq 0$ forces $P \neq 0$ , and $(24.7)$ says $rad P = P rad R ⊊ P$ for a nonzero projective module — a contradiction. Hence $rad M ⊊ M$ whenever a cover exists.

Proposition(24.11)(5)Covers over a J-semisimple ring

Let $R$ be a $J$ -semisimple ring, i.e. $rad R = 0$ . Then a right $R$ -module $M$ has a projective cover if and only if $M$ is already projective. In particular, over $ℤ$ the modules admitting projective covers are precisely the free abelian groups.

Proof

If $M$ is projective then the identity $M \to M$ is a projective cover with zero kernel. Conversely let $θ : P \to M$ be a projective cover. Then $\ker θ \subseteq_{s} P$ , so $\ker θ \subseteq rad P$ by $(24.4) (1)$ together with $(24.2) (6)$ , and $rad P = P rad R = 0$ by $(24.7)$ (the case $P = 0$ being trivial). Hence $\ker θ = 0$ and $θ$ is an isomorphism, so $M ≅ P$ is projective.

For $R = ℤ$ we have $rad ℤ = 0$ , and projective $ℤ$ -modules are exactly the free abelian groups, since every subgroup of a free abelian group is free.

Counterexample—A module with no projective cover

$ℤ / 2 ℤ$ has no projective cover as a $ℤ$ -module: it is not free, and $ℤ$ is $J$ -semisimple. Note that $rad (ℤ / 2 ℤ) = 0 \neq ℤ / 2 ℤ$ , so the criterion of $(24.11) (4)$ does not detect this failure — properness of the radical is necessary but far from sufficient.

Proof Techniques and Method

How these proofs work, and which move to reuse.

Move 1

Lift, then use minimality to get surjectivity

Projectivity of the source gives a map $α$ ; smallness at the target upgrades *the image maps onto $M$ * to the image is everything. This converts a lifting into an epimorphism for free.

Move 2

Split, then kill the kernel

An epimorphism onto a projective module splits, so its kernel is a direct summand. If that kernel is also small, $(24.2) (1)$ forces it to vanish. Summand plus small equals zero is the workhorse identity of the whole section.

Move 3

Build covers out of $P J$

Every elementary example is $P \to P / P J$ . Choose $P$ finitely generated, or make $J$ right T-nilpotent, and $(24.2) (2)$ supplies the smallness.

Move 2 deserves the emphasis. It is the reason uniqueness is so cheap here, and it is used again in $(24.16)$ to lift idempotents: an isomorphism $P \oplus Q ≅ R$ produced by comparing two covers of the same module is precisely a decomposition of $1$ into orthogonal idempotents.

Worked Example

All projective covers over $R = ℤ / 12 ℤ$

$R = ℤ / 12 ℤ$ is a finite commutative ring, hence artinian and semiperfect. Its Jacobson radical is $J = (6) = {0, 6}$ , and $1 = e_{1} + e_{2}$ with

e_{1} = 4, e_{2} = 9, e_{1}^{2} = 16 = 4, e_{2}^{2} = 81 = 9, e_{1} e_{2} = 36 = 0 in ℤ / 12 ℤ .

(E.1)

Orthogonal idempotents; $4 + 9 = 13 = 1$ .

This gives $R = e_{1} R \oplus e_{2} R$ with $e_{1} R = {0, 4, 8} ≅ ℤ / 3 ℤ$ and $e_{2} R = {0, 3, 6, 9} ≅ ℤ / 4 ℤ$ . Computing the radicals of the two blocks: $e_{1} J = 4 \cdot {0, 6} = {0}$ and $e_{2} J = 9 \cdot {0, 6} = {0, 6}$ , consistent with $J = e_{1} J \oplus e_{2} J = {0, 6}$ .

Projective covers of the simple $ℤ / 12 ℤ$ -modules
Simple module	Realised as	Projective cover $P$	$\ker θ$
$S_{1} ≅ ℤ / 3 ℤ$	$e_{1} R / e_{1} J = e_{1} R$	$e_{1} R ≅ ℤ / 3 ℤ$	$0$
$S_{2} ≅ ℤ / 2 ℤ$	$e_{2} R / e_{2} J$	$e_{2} R ≅ ℤ / 4 ℤ$	${0, 6}$

So $S_{1}$ is projective and is its own cover, while $S_{2}$ has cover $e_{2} R$ of length $2$ with small kernel $e_{2} J = {0, 6}$ — small because $e_{2} R$ is a finitely generated module and ${0, 6} = e_{2} R \cdot J$ .

A composite: the module $ℤ / 6 ℤ$

As an $R$ -module, $M = ℤ / 6 ℤ = R / (6) = R / J$ . By $(24.11) (1)$ applied to the finitely generated projective module $P = R$ , the quotient map $R \to R / J$ is a projective cover, with kernel $J = {0, 6}$ .

Cross-check via $(24.11) (3)$ : $M ≅ S_{1} \oplus S_{2}$ , whose covers are $e_{1} R$ and $e_{2} R$ ; their direct sum is $e_{1} R \oplus e_{2} R = R$ . The two computations agree, as uniqueness $(24.10)$ requires.

The contrast over $ℤ$

The same module $ℤ / 6 ℤ$ , viewed over $ℤ$ , has no projective cover at all: $ℤ$ is $J$ -semisimple, so by $(24.11) (5)$ only free abelian groups have covers, and $ℤ / 6 ℤ$ is torsion. Existence of covers is therefore a property of the ring, not of the module.

Process and Workflow

Check projectivity of the sourceExhibit

P

as a direct summand of a free module, or as

e R

for an idempotent

e

. Without this, nothing else matters.

Check surjectivityUsually automatic by construction; over a semilocal ring, surjectivity onto

M

can be tested modulo

M J

and lifted by Nakayama when

M

is finitely generated.

Check the kernel is smallIdentify

\ker θ

and try to place it inside

P J

. Then invoke

(24.2) (2)

: finite generation of

P

, or right T-nilpotence of

J

Conclude uniquenessOnce one cover is found,

(24.10)

says every other projective surjection onto

M

splits onto it, so the cover is determined up to isomorphism and may be named without ambiguity.

If step 3 fails, the fastest disproof of existence is $(24.11) (4)$ : show $rad M = M$ . If that is not the obstruction, look for a $J$ -semisimple quotient and apply $(24.11) (5)$ .

Comparison and Classification

Projective covers versus injective hulls
Feature	Projective cover	Injective hull
Shape	$P ↠ M$ , $P$ projective	$M ↪ E$ , $E$ injective
Minimality condition	$\ker$ is small in $P$	image is essential in $E$
Existence	Only for special rings and modules	Always, over every ring
Uniqueness	Up to isomorphism over $M$	Up to isomorphism over $M$
Characterising class of rings	Semiperfect (finitely generated $M$ ), right perfect (all $M$ )	No condition needed
Typical use	Minimal projective resolutions, Betti numbers	Minimal injective resolutions, Bass numbers

When does $M$ have a projective cover?
Situation	Verdict	Reason
$M$ projective	Yes, $θ = id$	Zero kernel is small
$M = P / P J$ , $P$ finitely generated projective	Yes	(24.11)(1)
$M = e R / e J$ , $e$ idempotent	Yes	(24.11)(2)
$M$ finitely generated, $R$ semiperfect	Yes	(24.12)
$M$ arbitrary, $R$ right perfect	Yes	(24.12), (24.18)
$M \neq 0$ with $rad M = M$	No	(24.11)(4)
$M$ non-projective, $R$ $J$ -semisimple	No	(24.11)(5)

Relationship Map

\ker θ \subseteq_{s} P

\ker θ \subseteq rad P = P J

maximal submodules correspond

rad M ⊊ M

Each arrow is a one-line consequence of the previous page's results, and the composite is the standard non-existence test. The converse chain is false at the last step: $rad M ⊊ M$ does not produce a cover, as $ℤ / 2 ℤ$ over $ℤ$ shows.

covers for all

M

R

right perfectcovers for finitely generated

M

R

semiperfect

Read as two separate equivalences, these are the theorems $(24.18)$ and $(24.16)$ that close out the section. Both are proved by turning a supply of covers into idempotent-lifting data.

Failure Modes and Common Mistakes

Do not conclude from $rad M ⊊ M$ that a cover exists; the condition is necessary only.
Do not confuse the projective cover of $M$ with a minimal free presentation: over a non-local ring the cover need not be free, only projective.
Do not drop the hypothesis $P$ projective in $(24.10)$ ; the splitting step uses projectivity of the target $P$ , not of the source alone.
Do not assume the cover of a submodule embeds in the cover of the module; covers are not functorial in any naive sense, and only the lifting statement of $(24.10)$ is available.

Historical Notes and Lessons Learned

1956Homological algebra arrivesCartan and Eilenberg codify projective resolutions and homological dimension, making the efficiency of a projective approximation a natural question.
1953Eckmann–Schöpf: injective hullsInjective hulls are shown to exist over any ring, with essential extensions supplying the minimality. The dual question — minimal projective approximations — is immediately visible and immediately harder.
1960Bass introduces perfect ringsIn Finitistic dimension and a homological generalization of semi-primary rings, Bass defines projective covers, proves their uniqueness, and characterises the rings over which every module has one.
1960sSemiperfect rings and block theoryThe finitely generated version becomes the standard tool for finite-dimensional algebras: projective covers of simple modules are the principal indecomposable modules, and Cartan invariants count their composition factors.
1970s–presentCovers beyond projectivesEnochs and others generalise to flat covers and to covers relative to arbitrary classes of modules; the flat cover conjecture, proved in 2001, shows flat covers exist over every ring — a striking contrast with the projective case.

The lesson is that minimality conditions dualise formally but not practically. Essential extensions are cheap because injective modules are abundant; small kernels are expensive because projective modules are rigid. Recognising which side of a duality carries the existence theory is a recurring judgement in homological algebra.

Quick Reference

Definition

θ : P ↠ M

P

projective,

\ker θ \subseteq_{s} P

Working form

θ (P') = M \Rightarrow P' = P

for all

P' \subseteq P

UniquenessTwo covers of

M

are isomorphic over

M

ComparisonAny projective surjection onto

M

splits onto the cover

Standard example

e R \to e R / e J

for

e = e^{2}

J \subseteq rad R

Finite sums

⨁_{i = 1}^{n} θ_{i}

is a cover of

⨁_{i = 1}^{n} M_{i}

Necessary condition

M \neq 0 \Rightarrow rad M ⊊ M

$J$ -semisimple $R$ Cover exists

iff M

is projective

Over $ℤ$ Only free abelian groups have covers

The statements of (24.9)–(24.11)
Item	Statement	Hypotheses
(24.9)	Definition of projective cover	$P$ projective, $\ker θ$ small
(24.10)	Comparison and uniqueness	$θ$ a cover, $θ'$ any projective epimorphism
(24.11)(1)	$P \to P / P J$ is a cover	$J \subseteq rad R$ ; $P$ finitely generated or $J$ right T-nilpotent
(24.11)(2)	$e R \to e R / e J$ is a cover	$e = e^{2}$ , $J \subseteq rad R$
(24.11)(3)	Finite direct sums of covers are covers	finitely many summands
(24.11)(4)	Maximal submodules correspond; $rad M ⊊ M$	$M \neq 0$ has a cover
(24.11)(5)	Cover exists iff $M$ projective	$rad R = 0$

Frequently Asked Questions

Why do injective hulls always exist while projective covers usually do not?

Injectives are abundant — every module embeds in an injective, and one can shrink to a maximal essential extension by Zorn's Lemma. Projectives are scarce and rigid: trimming a free presentation means splitting off summands, and the process need not terminate. The asymmetry is genuine, not an artefact of the proofs.

Is the projective cover functorial in

M

Not naively. A map $M \to M'$ lifts to a map of covers when both exist, by projectivity, but the lift is not unique and the assignment does not preserve composition on the nose. What is canonical is the isomorphism class of the cover, guaranteed by $(24.10)$ .

What is the projective cover of a simple module?

Over a semiperfect ring, the simple right modules are exactly the $e_{i} R / e_{i} J$ for a complete set of orthogonal local idempotents $e_{i}$ , and the cover of $e_{i} R / e_{i} J$ is $e_{i} R$ . These are the principal indecomposable modules; their composition factors are recorded in the Cartan matrix of the algebra.

rad M \neq M

, does

M

have a projective cover?

No. That condition is necessary by $(24.11) (4)$ but not sufficient. $ℤ / 2 ℤ$ over $ℤ$ has zero radical, hence proper radical, yet has no projective cover because $ℤ$ is $J$ -semisimple and $ℤ / 2 ℤ$ is not projective.

Can a projective cover have zero kernel?

Yes, exactly when $M$ is projective, in which case the identity map is the cover. So projective and equal to its own projective cover are the same condition, and this is what makes $(24.11) (5)$ work over $J$ -semisimple rings.

How do projective covers give minimal resolutions?

Take a cover $P_{0} \to M$ , then a cover $P_{1} \to \ker (P_{0} \to M)$ , and iterate. Each syzygy is covered with no redundancy, so the resulting complex has $rad$ -small differentials: the induced maps in $P_{∙} \otimes_{R} R / J$ are all zero, and the ranks of the $P_{i}$ are exactly the Betti numbers of $M$ .

References

T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §24, (24.9)–(24.11) (pp. 361–363).
H. Bass, “Finitistic dimension and a homological generalization of semi-primary rings”, Transactions of the American Mathematical Society 95 (1960), 466–488.
F. W. Anderson and K. R. Fuller, Rings and Categories of Modules, 2nd edition, Graduate Texts in Mathematics 13, Springer-Verlag, 1992, §17 (semiperfect rings and projective covers).
C. W. Curtis and I. Reiner, Methods of Representation Theory, Volume I, Wiley-Interscience, 1981, §6 (principal indecomposable modules and Cartan invariants).
B. Eckmann and A. Schöpf, “Über injektive Moduln”, Archiv der Mathematik 4 (1953), 75–78.

AI Suggested Questions

Prove that a minimal projective resolution over a semiperfect ring computes $Ext$ with zero differentials after applying $- \otimes_{R} R / rad R$ .
Give a module over a commutative Noetherian ring that has no projective cover but does have a flat cover.
State and prove the dual of $(24.10)$ for injective hulls, and identify precisely where the two arguments diverge.
Which non-semiperfect rings still admit projective covers for all simple modules, and what does that class look like?
Show that over a local ring the projective cover of a finitely generated module is free, and compute its rank.
Explain how projective covers of simple modules produce the Cartan matrix of a finite-dimensional algebra.
Compare projective covers with minimal free resolutions in commutative algebra over a local Noetherian ring.

01Executive Summary

02Overview

03Learning Objectives

04Definitions

05Core Concepts

Why smallness is the right minimality condition

Covers detect maximal submodules

Where the supply of covers comes from

06Key Results

07Proof Techniques and Method

Lift, then use minimality to get surjectivity

Split, then kill the kernel

Build covers out of PJ

08Worked Example

All projective covers over R=ℤ/12ℤ

A composite: the module ℤ/6ℤ

The contrast over ℤ

09Process and Workflow

10Comparison and Classification

11Relationship Map

12Failure Modes and Common Mistakes

13Historical Notes and Lessons Learned

14Quick Reference

15Frequently Asked Questions

16Related KEVOS Topics

17References

18AI Suggested Questions

Continue learning

Executive Summary

Overview

Learning Objectives

Definitions

Core Concepts

Key Results

Proof Techniques and Method

Build covers out of $P J$

Worked Example

All projective covers over $R = ℤ / 12 ℤ$

A composite: the module $ℤ / 6 ℤ$

The contrast over $ℤ$

Process and Workflow

Comparison and Classification

Relationship Map

Failure Modes and Common Mistakes

Historical Notes and Lessons Learned

Quick Reference

Frequently Asked Questions

Related KEVOS Topics

References

AI Suggested Questions