Blocks of Semiperfect Rings | KEVOS® Mathematics

Q: Why does the block decomposition need semiperfectness rather than just a chain condition?

What is really needed is that 1 be a finite sum of orthogonal primitive idempotents, which is (22.5)'s hypothesis. Chain conditions on ideals also suffice, by (22.2). Semiperfectness delivers the idempotent decomposition without any chain condition, so it covers examples such as local rings that are far from artinian.

Q: Is the block decomposition the same as the Peirce decomposition?

No. The Peirce decomposition of R along a set of orthogonal idempotents splits R as a direct sum of right ideals or of corner bimodules; the block decomposition splits R as a direct sum of two-sided ideals, which happens only for central idempotents. Blocks are always coarser.

Q: Can a block be decomposable as a module?

Yes, and usually is. c_iR is indecomposable as a ring — equivalently as a two-sided ideal — but as a right module it decomposes into the principal indecomposables e_jR belonging to that block. Indecomposability of a ring and of the module underlying it are different statements.

Q: How do I actually detect a nonzero corner e_pRe_q in practice?

For a finite-dimensional algebra, compute the matrix of dimensions _k e_pAe_q; these are the entries of a Hom table. For a right artinian ring one can instead use the Cartan matrix: the (p,q) entry is nonzero exactly when e_pRe_q ≠ 0, since V_q is a composition factor of e_pR if and only if e_pRe_q ≠ 0.

Q: Do the blocks of R and of a basic ring of R correspond?

Yes. A basic ring B = eRe has A RAR as a lattice isomorphism between ideals of B and ideals of R, so indecomposable ideal direct summands correspond, and hence blocks correspond one-to-one. This is part of the basic-ring theorem (25.8).

Q: What does a block look like when R is semisimple?

The blocks are exactly the simple components M_n_j(D_j) of the Wedderburn decomposition, and s = t. All the interest in the theory comes from rings where rad R ≠ 0 and the two counts differ.

Executive Summary

A semiperfect ring $R$ always decomposes as a finite direct product of indecomposable rings. The summands, its blocks, are the ideals $c_{i} R$ for the centrally primitive idempotents $c_{1}, \dots, c_{s}$ , and the decomposition is unique up to order. This is Lam's $(25.4)$ , obtained by feeding the idempotent decomposition of a semiperfect ring into the general block machinery of §22.

The useful content is the combinatorial description of the partition: two primitive idempotents lie in the same block precisely when they are linked, and linkage is connectedness in the graph whose vertices are the principal indecomposable modules and whose edges record nonzero homomorphisms. Everything about $R$ — modules, Cartan invariants, homological algebra — decomposes accordingly.

$c_{1}, \dots, c_{s}$ Centrally primitive idempotents

UniqueUp to reordering

$Hom \neq 0$ Edge in the linkage graph

(25.4)Lam's theorem

Overview

Decomposing a ring into a direct product is the crudest possible simplification, and it is not always available: $\prod_{i = 1}^{\infty} ℚ$ has infinitely many central idempotents and no finest decomposition. What makes a block decomposition exist is the ability to write $1$ as a finite sum of orthogonal primitive idempotents, and that is exactly what semiperfectness delivers.

1 = e_{1} + \dots + e_{n} (orthogonal, local) ⟹ 1 = c_{1} + \dots + c_{s} (orthogonal, centrally primitive),

(23.6) → (22.5)

Each $c_{i}$ is the sum of one linkage class of the $e_{j}$ ; the passage from the left-hand decomposition to the right-hand one is the whole content of the block theorem.

Blocks are coarser than the simple components of $\bar{R} = R / rad R$ , sometimes strictly so, because a central idempotent of $\bar{R}$ need not lift to a central idempotent of $R$ . The ring of upper triangular matrices is the standard witness: it is indecomposable, yet its semisimple quotient is a product of $n$ copies of a division ring.

Right modules throughout, following §25. The block decomposition itself is two-sided — the $c_{i}$ are central — so nothing here depends on the side; only the description via right modules does.

Learning Objectives

Recall the definitions of central, centrally primitive and linked idempotents.
State $(25.4)$ with the hypothesis semiperfect in place and identify where it is used.
Prove that a sum of a linkage class of orthogonal primitive idempotents is central.
Deduce that ${Hom}_{R} (M, N) = 0$ for modules in distinct blocks.
Compare the number of blocks of $R$ with the number of simple components of $R / rad R$ .
Compute the blocks of $T_{n} (k)$ , of $ℤ / 12 ℤ$ and of a commutative semiperfect ring.

Definitions

Central idempotent: $c = c^{2}$ with $c r = r c$ for all $r \in R$ ; equivalently $c R$ and $(1 - c) R$ are two-sided ideals and $R ≅ c R \times (1 - c) R$ as rings.
Centrally primitive: $c \neq 0$ central idempotent that cannot be written as a sum of two nonzero orthogonal central idempotents; equivalently the ring $c R$ is indecomposable.
Block: The ring $c_{i} R$ for $c_{i}$ centrally primitive, with identity element $c_{i}$ . It is a two-sided ideal of $R$ but not a subring containing $1$ unless $c_{i} = 1$ .
$e \sim e'$: For $e, e' \in E$ : there exists $f \in E$ with $e R f \neq 0 \neq e' R f$ . Reflexive and symmetric, not transitive in general.
Linked: $e$ and $e'$ are related by the equivalence relation generated by $\sim$ , that is, by a finite chain $e \sim g_{1} \sim \dots \sim g_{m} \sim e'$ of primitive idempotents.

Remark(21.6)Linkage is a Hom condition

Since ${Hom}_{R} (f R, e R) ≅ e R f$ , the condition $e R f \neq 0$ says exactly that some nonzero homomorphism $f R \to e R$ of principal indecomposables exists. Linkage is therefore a statement about the category of modules, not about the chosen idempotents.

Core Concepts

From orthogonal idempotents to central ones

Fix a decomposition $1 = e_{1} + \dots + e_{n}$ into orthogonal primitive idempotents, available for any semiperfect ring. Partition ${e_{1}, \dots, e_{n}}$ into linkage classes and let $c_{1}, \dots, c_{s}$ be the class sums. Each $c_{i}$ is an idempotent because the $e_{j}$ are orthogonal, and the $c_{i}$ are orthogonal with $\sum_{i} c_{i} = 1$ . The point of the construction is that $c_{i} R c_{j} = 0$ whenever $i \neq j$ : an element of $e_{p} R e_{q}$ with $e_{p}, e_{q}$ in different classes must vanish, since $e_{p} R e_{q} \neq 0$ would link them.

orthogonal primitive

e_{j}

linkage classesclass sums

c_{i}

central, centrally primitive

The linkage graph

Build a graph $Γ$ with one vertex for each $e_{j}$ and an edge between $e_{p}$ and $e_{q}$ whenever $e_{p} R e_{q} \neq 0$ . Then $e_{p} \sim e_{q}$ implies $e_{p}$ and $e_{q}$ lie in the same component of $Γ$ , and conversely — for if $e R f \neq 0$ then also $f R f ∋ f \neq 0$ , so $e \sim f$ directly. A witness $f \in E$ outside the chosen list may always be replaced by one inside it, because $f R ≅ e_{k} R$ for some $k$ forces $e R f ≅ {Hom}_{R} (f R, e R) ≅ e R e_{k}$ .

So the blocks of $R$ are the connected components of $Γ$ , and the number of blocks is a purely graph-theoretic count once the products $e_{p} R e_{q}$ are known.

Why blocks are coarser than Wedderburn components

Reducing modulo $J = rad R$ sends the $c_{i}$ to nonzero central idempotents ${\bar{c}}_{i}$ of $\bar{R}$ — nonzero because $J$ contains no nonzero idempotent. Hence each block of $R$ accounts for at least one simple component of $\bar{R}$ , and the simple components are partitioned by the blocks. The partition can be non-trivial: nothing forces ${\bar{c}}_{i}$ to be centrally primitive in $\bar{R}$ , since the central idempotents of $\bar{R}$ need not lift to central idempotents of $R$ .

Key Results

Theorem(25.4)Block decomposition of a semiperfect ring

Let $R \neq 0$ be a semiperfect ring and let $E$ be its set of primitive idempotents. Then:

$1$ is a sum of finitely many orthogonal centrally primitive idempotents $c_{1}, \dots, c_{s}$ , so that $R = c_{1} R \oplus \dots \oplus c_{s} R$ is a direct sum of indecomposable two-sided ideals;
this decomposition is unique up to a permutation of the summands, and $c_{1}, \dots, c_{s}$ are the only centrally primitive idempotents of $R$ ;
two primitive idempotents $e, e' \in E$ lie in the same block $c_{i} R$ if and only if they are linked;
each block $c_{i} R$ , regarded as a ring with identity $c_{i}$ , is itself semiperfect and indecomposable.

Proof

Existence. Because $R$ is semiperfect, $(23.6)$ gives $1 = e_{1} + \dots + e_{n}$ with the $e_{j}$ mutually orthogonal local idempotents; local idempotents are primitive, so $e_{j} \in E$ . Let $c_{1}, \dots, c_{s}$ be the sums over the linkage classes of ${e_{1}, \dots, e_{n}}$ . These are orthogonal idempotents with $\sum_{i} c_{i} = 1$ , and $c_{i} R c_{j} = 0$ for $i \neq j$ as observed above.

Centrality. For $a \in R$ compute, using $\sum_{j} c_{j} = 1$ and $c_{i} R c_{j} = 0$ for $j \neq i$ : $c_{i} a = c_{i} a (\sum_{j} c_{j}) = c_{i} a c_{i} = (\sum_{j} c_{j}) a c_{i} = a c_{i}$ . Hence each $c_{i}$ is central.

Central primitivity and (3). If $c$ is a nonzero central idempotent, then by $(22.4)$ every $e \in E$ lies in $c R$ or in $(1 - c) R$ , and by $(22.3)$ linked idempotents make the same choice. So the set of $e_{j}$ lying in $c R$ is a union of linkage classes, whence $c$ is a sum of some of the $c_{i}$ . Applying this to a hypothetical splitting of a single $c_{i}$ shows $c_{i}$ is centrally primitive, and it shows simultaneously that $e, e' \in E$ belong to the same block exactly when they are linked, which is $(22.5)$ .

Uniqueness. By $(22.1)$ , once $1$ is a sum of orthogonal centrally primitive idempotents, those idempotents are the only centrally primitive ones and the decomposition is unique up to order.

(4). The ring isomorphism $R ≅ c_{1} R \times \dots \times c_{s} R$ identifies $rad R$ with $\prod_{i} c_{i} rad R$ , so $c_{i} R / c_{i} rad R$ is a direct factor of the semisimple ring $\bar{R}$ and is semisimple; idempotents lift componentwise. Hence each block is semiperfect, and it is indecomposable because $c_{i}$ is centrally primitive.

Corollary—Modules belong to blocks

Let $R$ be semiperfect with blocks $c_{1} R, \dots, c_{s} R$ . Every right $R$ -module $M$ decomposes as $M = M c_{1} \oplus \dots \oplus M c_{s}$ , each $M c_{i}$ being a module over the block $c_{i} R$ . Consequently an indecomposable module satisfies $M = M c_{i}$ for exactly one $i$ , and ${Hom}_{R} (M, N) = 0$ whenever $M$ and $N$ belong to distinct blocks.

Proof

The $c_{i}$ are central orthogonal idempotents summing to $1$ , so each $M c_{i}$ is a submodule and $M = ⨁_{i} M c_{i}$ . If $M$ is indecomposable exactly one summand is nonzero. For the last claim, let $f : M \to N$ with $M = M c_{i}$ and $N = N c_{j}$ , $i \neq j$ ; then $f (M) = f (M c_{i}) = f (M) c_{i} \subseteq N c_{i} = 0$ .

Corollary—Blocks refine to Wedderburn components

With $R$ semiperfect, $s$ its number of blocks, and $\bar{R} ≅ \prod_{j = 1}^{t} M_{n_{j}} (D_{j})$ the Wedderburn decomposition of $R / rad R$ , the images ${\bar{c}}_{1}, \dots, {\bar{c}}_{s}$ are nonzero orthogonal central idempotents of $\bar{R}$ summing to $1$ . Hence $s \leq t$ , and the $t$ simple components are partitioned into $s$ nonempty groups, one per block.

Remark(23.10)The extreme case

If $R$ is semiperfect and $\bar{R}$ is simple, then $t = 1$ , so $s = 1$ and $R$ is indecomposable; in fact $R ≅ M_{n} (k)$ for some local ring $k$ and some $n \geq 1$ , both uniquely determined. The converse fails: indecomposability of $R$ says nothing about the indecomposability of $\bar{R}$ .

Proof Techniques and Method

How these proofs work, and which move to reuse.

Move 1

Centrality from vanishing corners

To prove an idempotent $c$ is central it suffices to show $c R c' = c' R c = 0$ for the complementary idempotent $c' = 1 - c$ . Inserting $1 = c + c'$ on both sides of $c a$ and $a c$ then collapses everything.

Move 2

Equivalence classes as idempotent sums

A relation on a finite orthogonal decomposition of $1$ becomes an algebraic object by summing over classes. This is how a combinatorial partition produces genuine ring-theoretic direct factors.

Move 3

Test central idempotents on primitive ones

Any central idempotent $c$ splits each $e \in E$ : either $c e = e$ or $c e = 0$ . Comparing the two options across a linkage chain shows $c$ is a union of classes, which forces both primitivity and uniqueness.

Move 3 is the reusable one. It reduces every question about the lattice of central idempotents to a question about how the finite set $E$ is partitioned, and it is the reason the block decomposition is unique while decompositions into non-central idempotents are not.

Worked Example

Upper triangular matrices: one block, many components

Let $k$ be a division ring and $R = T_{n} (k)$ , with $e_{i} = E_{i i}$ and $1 = e_{1} + \dots + e_{n}$ . For $i < j$ we have $e_{i} R e_{j} = k E_{i j} \neq 0$ , so consecutive idempotents are linked and the linkage graph is connected.

e_{1} \sim e_{2} \sim \dots \sim e_{n} ⟹ s = 1, while \bar{R} ≅ {\underset{⏟}{k \times \dots \times k}}_{n}, t = n .

(E.1)

$T_{n} (k)$ is indecomposable as a ring for every $n$ , although its semisimple quotient has $n$ simple components. Central idempotents do not lift here.

Concretely for $n = 2$ : a central idempotent $c = (\begin{matrix} a & b \\ 0 & d \end{matrix})$ must commute with $E_{12}$ , and $c E_{12} = a E_{12}$ while $E_{12} c = d E_{12}$ , forcing $a = d$ ; idempotency then gives $a = d \in {0, 1}$ and $b = 0$ in either case. So the only central idempotents are $0$ and $1$ .

A commutative example: $ℤ / 12 ℤ$

By the Chinese Remainder Theorem $ℤ / 12 ℤ ≅ ℤ / 4 ℤ \times ℤ / 3 ℤ$ , and both factors are local, so the ring is semiperfect with exactly two blocks. The centrally primitive idempotents are the CRT idempotents $c_{1} = 9$ and $c_{2} = 4$ : indeed $9^{2} = 81 = 9$ and $4^{2} = 16 = 4$ in $ℤ / 12 ℤ$ , with $9 + 4 = 13 = 1$ and $9 \cdot 4 = 36 = 0$ .

ℤ / 12 ℤ = 9 (ℤ / 12 ℤ) \oplus 4 (ℤ / 12 ℤ), 9 R ≅ ℤ / 4 ℤ, 4 R ≅ ℤ / 3 ℤ .

(E.2)

Here $s = t = 2$ : for a commutative semiperfect ring the blocks are the local factors and each is its own Wedderburn component, so the inequality $s \leq t$ is an equality.

Process and Workflow

Decompose $1$ Write

1 = e_{1} + \dots + e_{n}

with the

e_{j}

orthogonal local idempotents; semiperfectness guarantees this exists.

Tabulate the cornersCompute which of the

n^{2}

corner spaces

e_{p} R e_{q}

are nonzero. For a finite-dimensional algebra this is a table of dimensions.

Take connected componentsThe components of the resulting graph are the linkage classes; there is one block per component.

Sum each classThe class sums

c_{i}

are the centrally primitive idempotents, and

R ≅ \prod_{i} c_{i} R

Is $R$ indecomposable?

Linkage graph connectedYes:

s = 1

, the only central idempotents are

0

and

1

. This is compatible with

\bar{R}

having many components.

$\bar{R}$ simpleYes, and more:

R ≅ M_{n} (k)

for a local ring

k

(23.10)

Graph disconnectedNo: each component contributes a proper direct factor, and the factors are again semiperfect.

Comparison and Classification

Blocks versus Wedderburn components
Ring $R$ (semiperfect)	Blocks $s$	Components $t$ of $\bar{R}$	Comment
Local ring $k$	1	1	Indecomposable and basic
$M_{n} (k)$ , $k$ local	1	1	$\bar{R}$ simple, so $s = t = 1$
$T_{n} (k)$ , $k$ a division ring	1	$n$	Strict inequality $s < t$
$k \times k \times \dots \times k$	$n$	$n$	Semisimple: blocks are the simple components
$ℤ / 12 ℤ$	2	2	Commutative: blocks are the local factors
$k G$ , $char k ∤ \| G \|$	number of simple components	same	Semisimple by Maschke
$k S_{3}$ , $char k = 3$	1	2	One block containing both simple modules

The final row is typical of modular representation theory: the interesting phenomena live in blocks that contain several simple modules, since a block with a single simple module and no radical is just a matrix ring.

Relationship Map

Semiperfect ring $R$ block decomposition exists and is unique

Block $c_{i} R$ indecomposable semiperfect ring with identity

c_{i}

Linkage class of primitive idempotentsthe

e_{j}

with

c_{i} e_{j} = e_{j}

Principal indecomposables

e_{j} R

for

e_{j}

in the class

Simple modules

e_{j} R / e_{j} J

, the tops

Reading the nesting outwards: simple modules group into principal indecomposables, those group into linkage classes, and the classes are the blocks. Each level of grouping is canonical.

e \sim e'

same linkage classsame block

Hom

and

Ext

can be nonzero

Applications and Industry Use

Applications here means where this structure is used — inside mathematics and in the engineering and computing disciplines that consume it.

Modular representation theory

Brauer blocks

For $R = k G$ with $char k = p$ dividing $| G |$ , the blocks of this theorem are exactly the $p$ -blocks of $G$ . Defect groups, Brauer correspondence and the major conjectures of the subject are all statements about individual blocks.

Computational algebra

Divide and conquer

Algorithms for modules over a finite-dimensional algebra split the problem along blocks first: each block is a smaller algebra, and no information crosses between blocks because the connecting Hom spaces vanish.

Homological algebra

Derived decompositions

Since $Ext$ groups between modules in different blocks vanish, the derived category of $R$ is the product of the derived categories of its blocks. Invariants such as global dimension are computed blockwise.

Order theory and integral representations

Blocks of orders

For an order over a complete discrete valuation ring the ring is semiperfect, so the block decomposition applies and organises the classification of lattices over the order.

Honestly stated, the block decomposition is a bookkeeping theorem — but it is the bookkeeping that makes modular representation theory finite and computable.

Failure Modes and Common Mistakes

A block $c_{i} R$ is a ring but not a subring of $R$ in the unital sense: its identity is $c_{i}$ , not $1$ .
Do not expect a block to be a matrix ring. Blocks are indecomposable semiperfect rings, and indecomposable semiperfect rings have not been classified.
The number of blocks is not a Morita invariant of anything smaller than $R$ itself — but it is preserved under passage to a basic ring, which is the content of the basic-ring theorem.
For rings that are not semiperfect a block decomposition may simply fail to exist; an infinite product of fields has no finest decomposition into indecomposable ideals.

Best Practices

Compute the corner spaces $e_{p} R e_{q}$ before doing anything else: they determine the block structure and the support of the Cartan matrix simultaneously.
State whether a claimed decomposition is into blocks (central, unique) or merely into indecomposable modules (non-central, unique only up to isomorphism).
When working with a specific block, replace $R$ by $c_{i} R$ and $1$ by $c_{i}$ explicitly; silent identification of the two identities is a frequent source of error.
Record the partition of simple modules into blocks alongside the list of simple modules; almost every later computation is blockwise.

Quick Reference

Hypothesis

R \neq 0

semiperfect;

J = rad R

;

E

the primitive idempotents

Decomposition

R = c_{1} R \oplus \dots \oplus c_{s} R

, the

c_{i}

centrally primitive

UniquenessUnique up to order; the

c_{i}

are the only centrally primitive idempotents

Linkage

e, e'

in the same block

\Leftrightarrow

linked

\Leftrightarrow

connected in the graph

Γ

Edge test

e_{p} R e_{q} \neq 0 \Leftrightarrow {Hom}_{R} (e_{q} R, e_{p} R) \neq 0

Modules

M = ⨁_{i} M c_{i}

; indecomposables live in one block

Vanishing

{Hom}_{R} (M, N) = 0

across distinct blocks

Counting

s \leq t

, where

t

is the number of simple components of

\bar{R}

Statement locator
Fact	Statement	Reference
Existence of blocks	$1$ a sum of orthogonal centrally primitive idempotents	(25.4), (22.5)
Uniqueness	Block decomposition unique up to order	(22.1)
Linkage criterion	same block $\Leftrightarrow$ linked	(25.4), (22.5)
Idempotent decomposition	$1 = e_{1} + \dots + e_{n}$ orthogonal local	(23.6)
Central idempotents and $E$	each $e \in E$ lies in $c R$ or $(1 - c) R$	(22.4)
Linked idempotents agree	$e$ linked to $e' \Rightarrow$ ( $e \in c R \Leftrightarrow e' \in c R$ )	(22.3)
Simple quotient case	$\bar{R}$ simple $\Rightarrow R ≅ M_{n} (k)$ , $k$ local	(23.10)

Frequently Asked Questions

Why does the block decomposition need semiperfectness rather than just a chain condition?

What is really needed is that $1$ be a finite sum of orthogonal primitive idempotents, which is $(22.5)$ 's hypothesis. Chain conditions on ideals also suffice, by $(22.2)$ . Semiperfectness delivers the idempotent decomposition without any chain condition, so it covers examples such as local rings that are far from artinian.

Is the block decomposition the same as the Peirce decomposition?

No. The Peirce decomposition of $R$ along a set of orthogonal idempotents splits $R$ as a direct sum of right ideals or of corner bimodules; the block decomposition splits $R$ as a direct sum of two-sided ideals, which happens only for central idempotents. Blocks are always coarser.

Can a block be decomposable as a module?

Yes, and usually is. $c_{i} R$ is indecomposable as a ring — equivalently as a two-sided ideal — but as a right module it decomposes into the principal indecomposables $e_{j} R$ belonging to that block. Indecomposability of a ring and of the module underlying it are different statements.

How do I actually detect a nonzero corner

e_{p} R e_{q}

in practice?

For a finite-dimensional algebra, compute the matrix of dimensions $\dim_{k} e_{p} A e_{q}$ ; these are the entries of a Hom table. For a right artinian ring one can instead use the Cartan matrix: the $(p, q)$ entry is nonzero exactly when $e_{p} R e_{q} \neq 0$ , since $V_{q}$ is a composition factor of $e_{p} R$ if and only if $e_{p} R e_{q} \neq 0$ .

Do the blocks of

R

and of a basic ring of

R

correspond?

Yes. A basic ring $B = e R e$ has $𝔄 \mapsto R 𝔄 R$ as a lattice isomorphism between ideals of $B$ and ideals of $R$ , so indecomposable ideal direct summands correspond, and hence blocks correspond one-to-one. This is part of the basic-ring theorem $(25.8)$ .

What does a block look like when

R

is semisimple?

The blocks are exactly the simple components $M_{n_{j}} (D_{j})$ of the Wedderburn decomposition, and $s = t$ . All the interest in the theory comes from rings where $rad R \neq 0$ and the two counts differ.

References

T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §25 (pp. 373–374), with §22 (pp. 336–344).
F. W. Anderson and K. R. Fuller, Rings and Categories of Modules, 2nd edition, Graduate Texts in Mathematics 13, Springer-Verlag, 1992, §7 and §27.
C. W. Curtis and I. Reiner, Methods of Representation Theory, Volume I, Wiley-Interscience, 1981, §55 (blocks and defect groups).
R. Brauer, “Investigations on group characters”, Annals of Mathematics 42 (1941), 936–958.
L. H. Rowen, Ring Theory, Volume I, Academic Press, 1988, Chapter 2.

AI Suggested Questions

Give an example of a semiperfect ring with three blocks whose semisimple quotient has five simple components.
Prove directly that the relation $e \sim e'$ on primitive idempotents fails to be transitive in $T_{3} (k)$ read naively, and identify the correct chain.
How are the defect groups of a $p$ -block of $k G$ defined, and what do they measure about the block?
Describe the block decomposition of the group algebra $k A_{5}$ in characteristic $2$ , $3$ and $5$ .
Show that the number of blocks of $R$ equals the number of connected components of the quiver of a basic algebra Morita equivalent to $R$ .
Under what conditions on $rad R$ do central idempotents of $R / rad R$ lift to central idempotents of $R$ ?
Explain how block decomposition interacts with tensor products of algebras over a field.

01Executive Summary

02Overview

03Learning Objectives

04Definitions

05Core Concepts

From orthogonal idempotents to central ones

The linkage graph

Why blocks are coarser than Wedderburn components

06Key Results

07Proof Techniques and Method

Centrality from vanishing corners

Equivalence classes as idempotent sums

Test central idempotents on primitive ones

08Worked Example

Upper triangular matrices: one block, many components

A commutative example: ℤ/12ℤ

09Process and Workflow

10Comparison and Classification

11Relationship Map

12Applications and Industry Use

Brauer blocks

Divide and conquer

Derived decompositions

Blocks of orders

13Failure Modes and Common Mistakes

14Best Practices

15Quick Reference

16Frequently Asked Questions

17Related KEVOS Topics

18References

19AI Suggested Questions

Continue learning