Algebra over a Ring

An R-module equipped with an R-bilinear internal multiplication, with associativity, unity, and commutativity imposed only at the explicitly declared convention tier.

Version

v2 · 2026-09-19 · History

Domain-specific #

1259

Domain group

Formal Sciences

Origin domain

Mathematics

Aliases

R-algebra, Algebra over R

Core Idea

Let \(R\) be a commutative ring. An algebra over \(R\), in the broad bilinear sense used here, is an \(R\)-module \(A\) equipped with an internal multiplication

\[ \mu:A\times A\longrightarrow A, \]

that is \(R\)-bilinear. Thus multiplication distributes over addition in each input and is compatible with scalar multiplication:

\[ (a+a')b=ab+a'b,\quad a(b+b')=ab+ab',\quad (ra)b=r(ab)=a(rb). \]

The multiplication may then be required to be associative, unital, commutative, Lie, Jordan, alternative, or subject to another identity. These properties are convention tiers, not consequences of bilinearity. Nonassociative-algebra literature uses the vector-space/module-plus-bilinear-product core; commutative algebra and the Stacks Project conventionally use an associative unital ring \(A\) with a unital ring map \(R\to Z(A)\).^[1][2]

For associative unital \(A\), the two presentations agree: a central unital homomorphism \(\eta:R\to Z(A)\) defines \(r\cdot a=\eta(r)a\), while an \(R\)-module algebra with compatible unit defines \(\eta(r)=r1_A\). The recognition invariant is one carrier simultaneously supporting linear combination over a declared base and an internal product compatible with that scalar action. The explicit convention tier is part of a reference-grade identification.

How would you explain it like I'm…

The Add, Stretch, and Multiply Box

An algebra over a ring is a box of things you can do three jobs with: add them together, stretch them by numbers from a chosen number system, and multiply them together. The special part is that multiplying plays fair with adding and stretching, so it doesn't matter whether you stretch before or after you multiply.

Add, Scale, and Multiply

In math, a 'ring' is a system of numbers where you can add, subtract, and multiply, like the whole numbers. An algebra over a ring is a collection of objects that does three things at once. You can add the objects together. You can scale them by numbers from the ring, like 'three times this object'. And you can multiply two objects to get another object. The multiplication has to get along with the other two: it spreads over addition, and scaling before or after multiplying gives the same result. Extra rules, like whether the order of multiplying matters, are chosen separately.

Module With a Bilinear Product

Start with a commutative ring R, a number system where multiplication commutes. An algebra over R is a set A that is an R-module, meaning you can add elements and multiply them by scalars from R, together with a multiplication A × A → A that is bilinear. Bilinear means the product distributes over addition in each slot and is compatible with scaling: (ra)b = r(ab) = a(rb). Nothing in that basic definition forces the product to be associative, commutative, or to have a 1; those are additional requirements that particular fields of math add on. For example, in commutative algebra an 'R-algebra' usually means an associative ring with 1 plus a ring map from R into its center, while Lie algebras use a product that is not associative at all. For associative unital algebras the two ways of defining it match up. What stays constant is one set that supports both scaling by R and an internal product that respects the scaling.

 

For a commutative ring R, an algebra over R in the broad bilinear sense is an R-module A with an R-bilinear multiplication μ: A × A → A, so (a + a′)b = ab + a′b, a(b + b′) = ab + ab′, and (ra)b = r(ab) = a(rb). Associativity, unitality, commutativity, or identities defining Lie, Jordan, or alternative algebras are additional convention tiers rather than consequences of bilinearity. Nonassociative-algebra literature works with this module-plus-bilinear-product core, whereas commutative algebra and the Stacks Project take an R-algebra to be an associative unital ring A with a unital ring homomorphism R → Z(A). For associative unital A these presentations coincide: a central unital homomorphism η: R → Z(A) defines the scalar action r·a = η(r)a, and conversely a unital module algebra gives η(r) = r·1_A. The recognition invariant is a single carrier supporting linear combination over a declared base together with an internal product compatible with that scalar action, and a reference-grade identification must state which convention tier is in force.

Structural Signature

Recognition roles:

• The base ring \(R\) — normally commutative and usually unital under the selected convention.
• The carrier \(R\)-module \(A\) — addition and scalar multiplication satisfy module axioms.
• The internal product \(\mu\) — maps two elements of \(A\) back into \(A\).
• Separate \(R\)-linearity — fixing either factor makes multiplication an \(R\)-linear map in the other.
• The convention tier — associative/nonassociative, unital/nonunital, commutative/noncommutative, and any polynomial identities are stated.
• The scalar-structure map — in the associative unital convention, \(R\to Z(A)\) makes the base action central.
• Structure-preserving morphisms — \(R\)-linear maps preserving multiplication and, where stipulated, unit.
• Construction operations — subalgebras, quotients by appropriate ideals, tensor products, scalar extension, or free algebras under the selected tier.

Recognition test: identify the base ring, module action, internal product, and verify all bilinearity equations. Then ask which additional axioms are actually declared. A ring with no chosen central \(R\)-action, or an \(R\)-module with no internal product, is not yet an \(R\)-algebra.^[2][1]

What It Is Not

An algebra over a ring is not merely an algebraic expression or a school subject called algebra. It is a specific structured object. It is not just a ring: the base ring and compatible scalar action are additional data. The same underlying ring can carry different \(R\)-algebra structures when different structure maps are selected.

It is not merely an \(R\)-module. Modules support linear combination but no required product of two module elements. It is not a multilinear form: a bilinear form maps \(A\times A\) to the scalar ring, while algebra multiplication maps to \(A\) itself. Nor is every bilinear map an algebra structure unless its domain factors and codomain are the same carrier in the intended way.

It is not automatically associative, commutative, or unital. Matrix algebras are associative and unital but generally noncommutative. Lie algebras are generally nonassociative and anticommutative in characteristic not two. The real three-dimensional cross product gives a nonassociative algebra without a multiplicative identity. Conversely, many commutative-algebra texts use “\(R\)-algebra” only for an associative unital object; the convention must not be silently changed mid-proof.^[1][2]

Scope of Application

Algebras over rings organize linear algebra with multiplication. They occur as polynomial, matrix, group, tensor, exterior, Clifford, universal-enveloping, Lie, Jordan, and operator algebras, as coordinate rings in algebraic geometry, and as deformation or representation objects. Changing \(R\) changes what counts as scalar-linear and which scalar extensions are available.

The associative unital tier dominates commutative and noncommutative ring theory. The broader bilinear tier is necessary to state nonassociative families uniformly. The entry therefore does not pretend one convention is universal; it identifies the shared object and requires each use to lock its tier. For a noncommutative base, left/right/bimodule compatibility needs additional care and falls outside the unqualified commutative-base definition.

Clarity

The concept separates three operations often denoted similarly: addition in \(A\), multiplication by a scalar \(r\in R\), and internal multiplication of \(a,b\in A\). For a polynomial algebra, \(r\cdot f\) uses the structure map while \(fg\) is the internal product. For matrices over \(R\), scalar multiplication scales entries, while matrix multiplication combines rows and columns.

It also clarifies relative language. Saying “\(A\) is finitely generated” can mean finitely generated as an \(R\)-module or as an \(R\)-algebra; these are different. \(R[x]\) is generated by one element as an \(R\)-algebra but generally has infinite rank as an \(R\)-module. The abstraction makes the generating operations explicit.

Manages Complexity

The \(R\)-algebra package lets linear and multiplicative reasoning interact. Bilinearity reduces products of sums to products of basis or generating elements. A finite-rank algebra can be encoded by structure constants \(c_{ij}^k\) satisfying

\[ e_i e_j=\sum_k c_{ij}^k e_k, \]

after a basis is chosen. Polynomial identities then become equations among structure constants. In the associative unital tier, ideals, quotients, modules over \(A\), and scalar extension can be developed uniformly.^[3][2]

The compression does not solve classification. Structure constants depend on basis, and associative/nonassociative identities may be difficult to decide. Over a general ring, modules need not be free, tensor products need not preserve exactness, and scalar extension can alter properties. The abstraction makes these difficulties expressible without erasing them.

Abstract Reasoning

Bilinearity is equivalently an \(R\)-linear multiplication map

\[ \widetilde\mu:A\otimes_R A\to A, \]

by the tensor product's universal property.^[4] Associativity becomes equality of the composites \(\widetilde\mu\circ(\widetilde\mu\otimes1)\) and \(\widetilde\mu\circ(1\otimes\widetilde\mu)\) on \(A\otimes_R A\otimes_R A\). A unit is a map \(R\to A\) satisfying the left and right unit diagrams. This formulation separates the common bilinear core from optional axioms.

For an associative unital algebra, centrality of \(\eta(R)\) proves the compatibility equation: \((ra)b=(\eta(r)a)b=\eta(r)(ab)=a\eta(r)b=a(rb)\). Conversely, setting \(\eta(r)=r1_A\) and using bilinearity gives a central structure map. This equivalence is why module-product and central-ring-map definitions can coexist without being identical in the nonunital or nonassociative tiers.^[2]

Knowledge Transfer

Literal transfer occurs whenever new products are built on modules: matrix multiplication, polynomial multiplication, Lie brackets, tensor-algebra concatenation, and cross products can all be tested against the same base/module/bilinearity roles. Once the tier is fixed, homomorphism, subalgebra, quotient, and scalar-extension reasoning can transfer.

The strongest parent-prime residue is Linearity: multiplication respects superposition in each input. Closure, Composition, and Associativity may be related, but associativity is optional at the broad tier. Calling a business organization an “algebra” because activities combine is metaphor unless a genuine module and bilinear product are present.

Examples

Matrix algebra. \(M_n(R)\) is an associative unital \(R\)-algebra for commutative unital \(R\). Scalars act entrywise, the structure map sends \(r\) to \(rI_n\), and the image is central. Matrix multiplication is \(R\)-bilinear. For \(n>1\) and nonzero suitable \(R\), it is noncommutative, demonstrating that commutativity is not part of the base identity.

Polynomial algebra. \(R[x]\) is a commutative associative unital \(R\)-algebra. It is generated by \(x\) as an algebra, because coefficients come from \(R\) and multiplication produces powers. It is not generally finitely generated as an \(R\)-module, since \(1,x,x^2,\ldots\) are required. This maps the base, module, internal product, unit, and convention tier.

Cross-product algebra. \(\mathbb R^3\) with usual vector addition/scalars and \(u\cdot v=u\times v\) is a real algebra in the broad nonassociative sense. The cross product is bilinear and closed in \(\mathbb R^3\), but it is anticommutative and generally \((u\times v)\times w\ne u\times(v\times w)\). It has no unit. This proves why the broad core cannot silently import associative-unital axioms.^[1]

Central-map boundary. A ring homomorphism \(R\to A\) whose image is not central does not produce the ordinary symmetric \(R\)-module algebra structure on \(A\) without a more careful bimodule convention. The centrality clause is substantive.

Structural Tensions

Broad definition versus local convention. A shared bilinear core unifies associative and nonassociative objects, but many subfields reserve “algebra” for associative unital objects. Diagnostic: are associativity, unit, and commutativity explicitly stated before any theorem uses them?

Relative structure versus underlying ring. Forgetting the base makes comparison easier but loses scalar-map data; retaining it distinguishes objects that may be the same ring. Diagnostic: would changing \(R\to A\) change legal morphisms or scalar extension?

Autonomy versus reduction. Ring, Module, Bilinear Map, and Closure are ingredients. Their mere list does not enforce that one carrier's product is bilinear over one chosen base or fix convention-tier morphisms. Diagnostic: can the central structure map or tensor multiplication be recovered? If not, composite closure fails and the residual is autonomous.

Structural–Framed Character

This node is purely structural within abstract algebra. Its roles and equations are formal and carry no empirical or evaluative content. Convention choice introduces community framing, but the selected tier is explicit and mathematical once chosen.

It remains domain-specific because “ring,” “module,” “bilinear,” and “internal multiplication” are algebraic structures. The generic pattern of combining linearly is carried by primes, while literal algebra-over-\(R\) identity does not survive replacement by unrelated substrates.

Structural Core vs. Domain Accent

The portable core is an internal binary operation compatible with external linear combination. This skeleton supports decomposition into generators and recombination through distributivity.

The indispensable accent is a commutative base ring, an \(R\)-module carrier, an \(R\)-bilinear product, a stated axiom tier, and appropriate morphisms. Removing these yields generic composition or linearity, not an \(R\)-algebra. The abstraction is therefore domain-specific and independently recurrent.

Instantiates / Related Primes

This entry presupposes Linearity.

An algebra over a ring presupposes Linearity, because its multiplication is linear separately in both arguments. It relates to Closure because products return to the carrier, and to Associativity only in an important subtype. Linearity is the broader abstraction this entry instantiates; Associativity would incorrectly exclude Lie, Jordan, cross-product, and other nonassociative algebras from the broad identity.

Ring is a close neighbor rather than a broader form under the broad convention, since nonassociative algebras do not have associative ring multiplication.

Relationships to Other Abstractions

Current abstraction Algebra over a Ring Domain-specific

Parents (1) — outgoing typed relations

• Algebra over a Ring presupposes Linearity Prime

  Algebra over a Ring presupposes Linearity because its multiplication is linear separately in both arguments.

Children (10) — abstractions linked here by typed relations

• Associative algebra Domain-specific is a kind of Algebra over a Ring

  Associative algebra is a domain-specific kind of Algebra over a Ring: An algebra over a commutative ring whose internal multiplication satisfies associativity.

• Banach Algebra Domain-specific is a kind of Algebra over a Ring

  A Banach algebra is an algebra over the real or complex scalars with a compatible complete norm.

• Cohomology Ring Domain-specific is a kind of Algebra over a Ring

  A Cohomology Ring is an Algebra over a Ring whose graded module is cohomology and whose internal multiplication is the cup product.

▸ Show 7 more

Hierarchy path (1) — routes to 1 parentless root

• Algebra over a Ring → presupposes Linearity

Neighborhood in Abstraction Space

Algebra over a Ring sits in a moderately populated region (57^th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Unclustered & Miscellaneous (2769 abstractions)

Nearest neighbors

• Associative algebra — 0.90
• Flexible Algebra — 0.86
• Commutative ring — 0.85
• Module (Algebra) — 0.85
• Ring — 0.84

Computed from structural-signature embeddings · 2026-10-25

Not to Be Confused With

• Ring: has additive and multiplicative laws but no necessarily chosen compatible base-ring action.
• Module/vector space: has scalar linearity but no internal multiplication.
• Associative algebra: a major subtype adding associativity, often unity.
• Commutative algebra: an associative unital subtype/field of study; not the broad object.
• Multilinear form: scalar-valued; algebra multiplication is carrier-valued.
• Étale algebra: a tightly constrained commutative finite algebra, not a synonym.
• Algebraic structure: generic umbrella for sets with operations.
• Algebraic stack: a geometric/categorical object, lexically related but structurally distinct.

References

[1] Richard D. Schafer, An Introduction to Nonassociative Algebras, Academic Press, 1966; Dover reprint. Project Gutenberg edition. registry ↩a ↩b ↩c ↩d

[2] The Stacks Project Authors, “Noncommutative algebras,” Tag 073Y. The introductory paragraph states the associative-unital algebra convention, and Definition 11.2.1 defines a finite algebra. Stacks Project. registry ↩a ↩b ↩c ↩d ↩e

[3] Michael Artin, Noncommutative Rings, graduate lecture notes, MIT-hosted edition. MIT OpenCourseWare resource. registry ↩

[4] Encyclopedia of Mathematics, “Tensor product,” definition and universal property for modules. Reference entry. registry ↩ Show verification details