Introduction to Reduced Biquaternion Linear Representations
Introduction
An element representation writes the reduced biquaternion $\breve Q=Q_{\blacktriangle}+jQ_{\blacktriangledown}$ as an object of another kind — a quadruple of real numbers, a pair of complex numbers, a pair of idempotent coordinates, a $2\times2$ matrix, a pair of polar data. This article names the representations of the block and fixes the two structural facts that make the block short. The first is that the algebra is commutative, associative and unital, so its multiplication is the general plain product. The second is that it is semisimple and a direct product of two copies of $\mathbb{C}$, so every representation is a direct sum of copies of the two one-dimensional characters and no irreducible representation has dimension greater than one. That is the exact opposite of the biquaternion case, where the algebra is simple and carries a single irreducible representation, of complex dimension two.
Representations and Modules
A representation of the algebra is a homomorphism of $\mathbb{C}$-algebras into an endomorphism algebra, and a module is the same data read as an action.
Definition. A representation of $\mathbb{B}_{\mathrm{red}}$ on a complex vector space $V$ is a $\mathbb{C}$-linear map $\rho:\mathbb{B}_{\mathrm{red}}\to\operatorname{End}_{\mathbb{C}}(V)$ with $$ \rho(\breve Q_{1}\breve Q_{2})=\rho(\breve Q_{1})\rho(\breve Q_{2}),\qquad \rho(1)=\operatorname{id}_{V}. $$ A (left) module is a complex vector space $V$ with a multiplication $\mathbb{B}_{\mathrm{red}}\times V\to V$ that is bilinear and satisfies $1\cdot q=q$ and $(\breve Q_{1}\breve Q_{2})\cdot q=\breve Q_{1}\cdot(\breve Q_{2}\cdot q)$; a representation and a module are the same object, the map $\breve Q\cdot q=\rho(\breve Q)q$ passing from one to the other. A subspace $U\subset V$ is a submodule when $\breve Q\cdot U\subset U$ for every $\breve Q$. A module is simple when it has exactly two submodules, $0$ and $V$; semisimple when it is a direct sum of simple ones. The words irreducible and completely reducible say the same of a representation.
Definition (the regular representation). The regular representation is $\rho_{\mathrm{reg}}(\breve Q)=L_{\breve Q}$, where $L_{\breve Q}(\breve Q_{2})=\breve Q \breve Q_{2}$, acting on $\mathbb{B}_{\mathrm{red}}$ itself; the space is complex of dimension two on the basis $1,j$ and real of dimension four on the basis $1,i,j,ij$. The operator $L_{\breve Q}$ is read in coordinates in Operators on the Reduced Biquaternions.
Theorem (the regular representation is faithful). $L_{\breve Q}=0$ if and only if $\breve Q=0$, so $\rho_{\mathrm{reg}}$ is injective and its image is an algebra isomorphic to $\mathbb{B}_{\mathrm{red}}$, $$ \boxed{\;\rho_{\mathrm{reg}}(\mathbb{B}_{\mathrm{red}})=\{L_{\breve Q}:\breve Q\in\mathbb{B}_{\mathrm{red}}\}\cong\mathbb{B}_{\mathrm{red}}\;}, $$ a commutative two-dimensional subalgebra of $\operatorname{End}_{\mathbb{C}}(\mathbb{B}_{\mathrm{red}})\cong M_{2}(\mathbb{C})$.
Proof. $L_{\breve Q}(1)=\breve Q$, so $L_{\breve Q}=0$ forces $\breve Q=0$; and $\rho_{\mathrm{reg}}(\breve Q_{1}\breve Q_{2})=L_{\breve Q_{1}\breve Q_{2}}=L_{\breve Q_{1}}L_{\breve Q_{2}}=\rho_{\mathrm{reg}}(\breve Q_{1})\rho_{\mathrm{reg}}(\breve Q_{2})$ is multiplicativity, because the multiplication is associative. The image is commutative because the multiplication is commutative, $L_{\breve Q_{1}}L_{\breve Q_{2}}=L_{\breve Q_{1}\breve Q_{2}}=L_{\breve Q_{2}\breve Q_{1}}=L_{\breve Q_{2}}L_{\breve Q_{1}}$.
The Algebra as a Module over Itself, over $\mathbb{C}$ and over $\mathbb{R}$
The algebra is a free module over itself and over the two base fields, of ranks two over $\mathbb{C}$ and four over $\mathbb{R}$, and the three readings carry the same module.
Theorem (the ground fields). The left regular module is $\mathbb{B}_{\mathrm{red}}$ acting on itself by left multiplication; over $\mathbb{C}$ it is free of rank two on the basis $1,j$, over $\mathbb{R}$ it is free of rank four on the basis $1,i,j,ij$, and as a module over the algebra itself it is free of rank one. Its composition length is two in every reading, with the same two composition factors, the two simple modules $\lambda_+$ and $\lambda_-$.
Proof. The ranks are the dimensions of the two bases of Reduced Biquaternions as a Vector Space over $\mathbb{R}$ and $\mathbb{C}$. The chain $$ 0\subset\mathbb{C}\Pi_+\subset\mathbb{B}_{\mathrm{red}} $$ is a composition series: the submodule $\mathbb{C}\Pi_+$ carries the character $\lambda_+$, and the quotient is isomorphic to $\mathbb{C}\Pi_-$ carrying $\lambda_-$, both simple by §The Simple Modules and Complete Reducibility.
Corollary (the module category). Every module is a direct sum of copies of the two simple modules; the category of modules is the product of two copies of the category of complex vector spaces, and a module is described by the pair of multiplicities $(m,n)$ of $\lambda_+$ and $\lambda_-$. The regular module is $(1,1)$.
Proof. Complete reducibility gives the decomposition; the multiplicity pair is the dimension of the $\lambda_+$-part and of the $\lambda_-$-part, and the module is determined by them because the two simples are non-isomorphic.
The Simple Modules and Complete Reducibility
The algebra has exactly two simple modules, both of complex dimension one; every module is a direct sum of copies of them, and the regular representation is their sum with multiplicity one.
Definition (the two characters). For $\breve Q=Q_{+}\Pi_++Q_{-}\Pi_-$ set $$ \lambda_+(\breve Q)=Q_{+},\qquad \lambda_-(\breve Q)=Q_{-}. $$ Each is a $\mathbb{C}$-algebra homomorphism into the one-dimensional algebra $\mathbb{C}$, $\lambda_{i}(\breve Q_{1}\breve Q_{2})=\lambda_{i}(\breve Q_{1})\lambda_{i}(\breve Q_{2})$ and $\lambda_{i}(1)=1$, because the idempotent coordinates multiply componentwise.
Theorem (the simple modules). Up to isomorphism the algebra has exactly two simple modules, the one-dimensional complex lines read through $\lambda_+$ and through $\lambda_-$; no simple module has dimension greater than one.
Proof. Let $V$ be a module and $\Pi_+,\Pi_-$ the two idempotents, $\Pi_++\Pi_-=1$ and $\Pi_+\Pi_-=0$. For $V$ simple, each of $\Pi_+$ and $\Pi_-$ acts as an idempotent operator, and each image $\Pi_+V$, $\Pi_-V$ is a submodule, so each is $0$ or $V$. Both zero is impossible, since $\operatorname{id}_{V}=(\Pi_++\Pi_-)$ would act as $0$. So exactly one acts as the identity and the other as $0$, and $V$ is a module of the one-dimensional algebra $\mathbb{C}$ through the surviving character; over the algebraically closed field $\mathbb{C}$ such a module is one-dimensional. The two choices of surviving character give $\lambda_+$ and $\lambda_-$, and the two are non-isomorphic because $\lambda_+(\Pi_+)=1\neq0=\lambda_-(\Pi_+)$.
Theorem (complete reducibility). Every module is a direct sum of copies of $\lambda_+$ and $\lambda_-$; the regular representation is $$ \boxed{\;\rho_{\mathrm{reg}}\cong\lambda_+\oplus\lambda_-\;}, $$ each character with multiplicity one.
Proof. The algebra is the product of the two fields $\mathbb{C}\Pi_+$ and $\mathbb{C}\Pi_-$, hence semisimple, and a module over a product of fields splits as the direct sum of its two components, each a complex vector space, that is a direct sum of copies of the corresponding character. On $\mathbb{B}_{\mathrm{red}}$ itself the two lines $\mathbb{C}\Pi_+$ and $\mathbb{C}\Pi_-$ are submodules carrying $\lambda_+$ and $\lambda_-$, and they span the two-dimensional space, so each multiplicity is one.
Corollary (the zero divisors carry the submodules). The non-zero submodules of the regular module are the two idempotent lines $\mathbb{C}\Pi_+$ and $\mathbb{C}\Pi_-$ and the whole space, and the two lines are exactly the minimal ideals of Ideals and Peirce Decomposition of the Algebra of the Reduced Biquaternions; every element of one line annihilates every element of the other, which is the module statement that the two simple components are mutually annihilating.
Proof. The submodules of a direct sum of two non-isomorphic simples are generated by subsets of the summands; here that gives only the four subspaces listed, and the annihilation is $\Pi_+\Pi_-=0$.
Schur's Lemma and the Intertwiners
Definition. For two modules $V$ and $W$ an intertwiner is a $\mathbb{C}$-linear map $f:V\to W$ with $f(\breve Q\cdot q)=\breve Q\cdot f(q)$ for every $\breve Q$ and $q$; the space of intertwiners is $\operatorname{Hom}_{\mathbb{B}_{\mathrm{red}}}(V,W)$.
Theorem (Schur). If $V$ and $W$ are simple then every non-zero intertwiner is an isomorphism; consequently $\operatorname{End}_{\mathbb{B}_{\mathrm{red}}}(V)=\mathbb{C}\operatorname{id}_{V}$ and $\operatorname{Hom}_{\mathbb{B}_{\mathrm{red}}}(\lambda_+,\lambda_-)=0$.
Proof. The kernel and the image of an intertwiner are submodules, so a non-zero one has zero kernel and full image, hence is an isomorphism. An endomorphism of the one-dimensional space $V$ is a scalar; and $\lambda_+\not\cong\lambda_-$, so there is no non-zero map between them.
Corollary (all endomorphisms are scalars). Over the algebraically closed field $\mathbb{C}$ the two simple modules are absolutely simple, $\operatorname{End}(\lambda_+)=\operatorname{End}(\lambda_-)=\mathbb{C}$, and the algebra is commutative but not central, its centre being the whole of $\mathbb{C}\times\mathbb{C}$. This is the whole of the intertwiner calculus, against the richer calculus of The Category of Modules in the Biquaternion Algebra, where the one simple module is two-dimensional and the intertwiners between its direct sums are matrices.
Proof. The centre of $\mathbb{C}\times\mathbb{C}$ is $\mathbb{C}\times\mathbb{C}$, which is larger than $\mathbb{C}$ — so the algebra is commutative but not central. The absolute simplicity of the two one-dimensional modules is immediate: an endomorphism of a one-dimensional space is a scalar, and scalar multiplication commutes with the action.
The List of Representations
The same element is written in six representations, and the table locates each in the block.
| representation | the element read as | article |
|---|---|---|
| idempotent coordinates | the pair $(Q_{+},Q_{-})$ of complex numbers, $\breve Q=Q_{+}\Pi_++Q_{-}\Pi_-$ | Ideals and Peirce Decomposition of the Algebra of the Reduced Biquaternions |
| four-vector | the real quadruple $(q_{\blacktriangle},q'_{\blacktriangle},q_{\blacktriangledown},q'_{\blacktriangledown})$ and the complex pair $(Q_{\blacktriangle},Q_{\blacktriangledown})$ | The Four-Vector Element Representation of the Reduced Biquaternions |
| matrix | the $2\times2$ matrix of $L_{\breve Q}$, complex or real | Introduction to the $2\times2$ Matrix Element Representation of Reduced Biquaternions |
| other algebraic | the Clifford and quotient presentations, the block array | Reduced Biquaternion Other Algebraic Element Representations |
| polar | the two polar complex forms of $Q_{+}$ and of $Q_{-}$ | Modulus–Unit Polar Decomposition of Reduced Biquaternions |
| operator | the multiplication operator $L_{\breve Q}$ and its companions | Operators on the Reduced Biquaternions |
The idempotent coordinates are the universal representation. Every other reading is the pair $(Q_{+},Q_{-})$ written in another alphabet: the four-vector splits each of $Q_{+}$ and $Q_{-}$ into two reals, the matrix puts them on the diagonal, the other algebraic representations formalise the isomorphism $\mathbb{B}_{\mathrm{red}}\cong\mathbb{C}\times\mathbb{C}$, the polar writes each in polar form, and the operator representation is the regular representation of the same pair. A statement is written in the representation that makes it shortest, and the two characters $\lambda_+,\lambda_-$ are the two coordinates of the universal one.
The Representations of the Group of Invertibles
The invertibles are $\mathbb{C}^{\times}\times\mathbb{C}^{\times}$, an abelian group all of whose irreducible complex representations are one-dimensional.
Theorem. $U=Q_{+}\Pi_++Q_{-}\Pi_-$ is invertible if and only if $\langle U\rangle_{\mathrm{ts}}=Q_{+}Q_{-}\neq0$, that is $Q_{+}\neq0$ and $Q_{-}\neq0$; the group of invertibles is $$ \boxed{\;{\mathbb{B}_{\mathrm{red}}}^{\times}\cong\mathbb{C}^{\times}\times\mathbb{C}^{\times}\cong (S^{1}\times S^{1})\times(\mathbb{R}_{>0}\times\mathbb{R}_{>0})\;}, $$ connected and abelian.
Proof. The criterion is that of Norm and Invertibility in the Algebra of the Reduced Biquaternions, and the decomposition $\mathbb{C}^{\times}\cong S^{1}\times\mathbb{R}_{>0}$ by modulus and argument gives the display; a product of connected sets is connected and abelian.
Corollary (the characters). The irreducible complex representations of the group of invertibles are the one-dimensional characters $$ \chi_{m,n}(U)=Q_{+}^{m}Q_{-}^{n},\qquad (m,n)\in\mathbb{Z}^{2}, $$ extending the two characters of the algebra; the full group of continuous characters is the larger family $$ U\mapsto Q_{+}^{m}|Q_{+}|^{it}Q_{-}^{n}|Q_{-}|^{is},\qquad (m,n)\in\mathbb{Z}^{2},\ (t,s)\in\mathbb{R}^{2}, $$ with character group $\mathbb{Z}^{2}\times\mathbb{R}^{2}$. The group is abelian, so every irreducible representation is one-dimensional; there is no representation of dimension greater than one, against the non-abelian group of invertibles of the biquaternion algebra and its higher-dimensional representations.
Proof. An abelian group has only one-dimensional irreducible complex representations, by Schur's lemma. The characters $\chi_{m,n}$ are the holomorphic ones, the two integers being the exponents on the two factors; adjoining the two real exponents of the moduli gives the continuous character group, because the continuous characters of $\mathbb{C}^{\times}$ are $z\mapsto z^{m}|z|^{it}$.
The Relation to the Split Biquaternions and to the Biquaternions
The reduced biquaternions are the commutative member of the semisimple family; the split biquaternions are its non-commutative member and the biquaternions the simple member.
The three algebras are the three ways of splitting a quaternionic algebra over $\mathbb{R}$: the biquaternions $\mathbb{B}=\mathbb{C}\otimes_{\mathbb{R}}\mathbb{H}$ are simple, the split biquaternions $\mathbb{B}_{\mathrm{spl}}=\mathbb{H}\oplus\mathbb{H}$ are a product of two quaternion division algebras, and the reduced biquaternions $\mathbb{B}_{\mathrm{red}}=\mathbb{C}\otimes_{\mathbb{R}}\mathbb{C}$ are a product of two complex fields, commutative.
| algebra | structure | simple modules | division ring of a simple | group of invertibles | regular representation |
|---|---|---|---|---|---|
| $\mathbb{B}_{\mathrm{red}}$ | $\mathbb{C}\times\mathbb{C}$, commutative | two, of complex dimension one | $\mathbb{C}$ | $\mathbb{C}^{\times}\times\mathbb{C}^{\times}$ | complex dimension two |
| $\mathbb{B}_{\mathrm{spl}}$ | $\mathbb{H}\oplus\mathbb{H}$, non-commutative | two, of real dimension four | $\mathbb{H}$ | $\mathbb{H}^{\times}\times\mathbb{H}^{\times}$ | real dimension eight |
| $\mathbb{B}$ | $M_{2}(\mathbb{C})$, simple | one, of complex dimension two | $\mathbb{C}$ | essentially $GL_{2}(\mathbb{C})$ | complex dimension four |
Theorem (the passage). The passage from the split biquaternions to the reduced biquaternions is the replacement of the division ring $\mathbb{H}$ by the field $\mathbb{C}$ in the two factors, together with the commutativity that this forces; the passage from the biquaternions to the reduced biquaternions is the splitting of the single simple module into the two characters.
Proof. The idempotent decomposition writes each of the two semisimple algebras as a product of two factors; the factors are $\mathbb{H}$ and $\mathbb{H}$ for the split biquaternions and $\mathbb{C}$ and $\mathbb{C}$ for the reduced ones, the latter thus commutative; and the biquaternion algebra, being simple, has a single simple module, whose splitting into two characters is the reduction. The structures of the companions are Split-Biquaternion Element Representation Theory and The Category of Modules in the Biquaternion Algebra.
Corollary (the smallest simples). The reduced algebra has the smallest simple modules of the three, of complex dimension one over the commutative base; its representation theory is therefore the product of two copies of the representation theory of $\mathbb{C}$, the simplest member of the family. All three algebras are semisimple, the reduced and the split biquaternion by being products, the biquaternion by being a full matrix algebra; none of the three needs the theory of modules over a non-semisimple ring.
Proof. The two product cases are semisimple as products of division algebras, and the matrix algebra is semisimple by Wedderburn; the simples of the reduced algebra are one-dimensional because each factor is a field, and a module over a field is a vector space.
Summary
The reduced biquaternion algebra is commutative, associative and unital, and it is the direct product of two copies of $\mathbb{C}$; its representation theory is therefore the simplest of the corpus. A representation is a module; the regular representation $\breve Q\mapsto L_{\breve Q}$ is faithful with image $\{L_{\breve Q}\}\cong\mathbb{B}_{\mathrm{red}}$; the simple modules are exactly the two characters $\lambda_+(Q_{+}\Pi_++Q_{-}\Pi_-)=Q_{+}$ and $\lambda_-(Q_{+}\Pi_++Q_{-}\Pi_-)=Q_{-}$, of complex dimension one, and every module is a direct sum of copies of them. The regular representation is $\lambda_+\oplus\lambda_-$, each with multiplicity one, its non-zero submodules are the two idempotent lines, and those two lines are the minimal ideals and the zero divisors. Schur's lemma gives $\operatorname{End}(\lambda_{i})=\mathbb{C}$ and no intertwiner between $\lambda_+$ and $\lambda_-$. The element is written in the six representations of the table, the idempotent coordinates being the universal one and the two characters its two coordinates. The group of invertibles is $\mathbb{C}^{\times}\times\mathbb{C}^{\times}$, abelian and connected, with the one-dimensional characters $\chi_{m,n}$ and no higher-dimensional irreducible representation; the multiplication has zero divisors and is not a division algebra. As a module over itself the algebra has composition length two, with the same two simple factors in the readings over $\mathbb{C}$ of rank two and over $\mathbb{R}$ of rank four, and the comparison of the three algebras places it as the commutative member of the semisimple family, against the non-commutative split biquaternions $\mathbb{H}\oplus\mathbb{H}$ and the simple biquaternions $M_{2}(\mathbb{C})$.
Summary of Notation
| Symbol | Meaning |
|---|---|
| $\rho$, $V$ | a representation and its module |
| $L_{\breve Q}$, $\rho_{\mathrm{reg}}$ | left multiplication by $\breve Q$; the regular representation |
| $\lambda_+,\lambda_-$ | the two characters, $\lambda_+(Q_{+}\Pi_++Q_{-}\Pi_-)=Q_{+}$, $\lambda_-=Q_{-}$ |
| $\operatorname{Hom}_{\mathbb{B}_{\mathrm{red}}}(V,W)$ | the intertwiners between two modules |
| ${\mathbb{B}_{\mathrm{red}}}^{\times}$ | the group of invertibles, $\mathbb{C}^{\times}\times\mathbb{C}^{\times}$ |
| $\chi_{m,n}(U)=Q_{+}^{m}Q_{-}^{n}$ | the holomorphic characters of the group of invertibles |
Further Reading
- Reduced Biquaternion Space, for the multiplication and the isomorphism $\mathbb{C}\oplus\mathbb{C}$.
- Introduction to the Algebra of the Reduced Biquaternions, for the algebra read as an algebra.
- Ideals and Peirce Decomposition of the Algebra of the Reduced Biquaternions, for the idempotent coordinates and the minimal ideals.
- The Group of Invertibles of the Algebra of the Reduced Biquaternions, for the group of invertibles and its complement.
- $M_2(\mathbb{C})$ Reduced Representation : Operators, for the multiplication operators of the four products.
- The Category of Modules, for the non-commutative companion and the one simple module of dimension two.
- Split-Biquaternion Element Representation Theory, for the non-commutative semisimple companion with its two four-dimensional simple modules.