The Frame, the Simple Nilpotents and the Gauge of $M_3(\mathbb{C})$
Introduction
A frame of $M_{3}(\mathbb{C})$ is a basis of the algebra by nine matrix units $E_{ij}$ with $E_{ij}E_{kl}=\delta_{jk}E_{il}$ and $\sum_{i}E_{ii}=I$. The diagonal units are three idempotents; the off-diagonal units are six nilpotents, of which two are the simple ones and the remaining four are their products and the products of the opposite pair. The frame is the matrix-unit form of the Peirce decomposition, and it is the object on which the gauge acts.
This article constructs the frame from a basis of $\mathbb{C}^{3}$, proves the multiplication table, isolates the two simple nilpotents $E_{12},E_{23}$ and the composite $E_{13}=E_{12}E_{23}$, and computes the gauge group. The gauge rescales the off-diagonal units by $E_{ij}\mapsto(t_{i}/t_{j})E_{ij}$; it fixes every idempotent, and its group is $(\mathbb{C}^{\times})^{2}$, the maximal torus of $PGL_{3}(\mathbb{C})$. The rescaling of the two simple nilpotents by $\lambda_{1}$ and $\lambda_{2}$ forces the composite to be rescaled by $\lambda_{1}\lambda_{2}$: there is no third independent scalar.
The idempotents do not depend on the gauge parameter. The symbols $\lambda_{1},\lambda_{2}$ below are gauge parameters of the nilpotents; they never multiply an idempotent. A statement that an idempotent depends on $\lambda$ is a confusion of the gauge with the frame.
The article assumes the algebra and its matrix units of The Matrix Algebra $M_{3}(\mathbb{C})$, the Peirce decomposition of The Peirce Decomposition of $M_{3}(\mathbb{C})$, and the root of Roots of Unity and the Lagrange Idempotents of $M_3(\mathbb{C})$. The two-parameter gauge is developed further in The Two-Parameter Gauge of $M_3(\mathbb{C})$; the modules built on the idempotents are The Modules and the Minimal Left Ideals of $M_3(\mathbb{C})$.
The Frame from a Basis
Definition. A frame of $M_{3}(\mathbb{C})$ is a family $(E_{ij})_{i,j=1}^{3}$ of elements with
$$ E_{ij}E_{kl}=\delta_{jk}E_{il},\qquad \sum_{i=1}^{3}E_{ii}=I,\qquad E_{ii}\neq0 . $$
The frame is normalized by the second condition; the first alone gives $E_{ii}^{2}=E_{ii}$, $E_{ii}E_{jj}=0$ for $i\neq j$, and $E_{ij}^{2}=0$ for $i\neq j$.
Theorem (the frame from a basis). Let $(v_{1},v_{2},v_{3})$ be a basis of $\mathbb{C}^{3}$ and let $(v_{1}^{\vee},v_{2}^{\vee},v_{3}^{\vee})$ be the dual basis, $v_{j}^{\vee}(v_{k})=\delta_{jk}$. Put
$$ E_{ij}=v_{i}v_{j}^{\vee}. $$
Then $(E_{ij})$ is a frame. Conversely every frame is of this form for a basis unique up to a common scalar.
Proof. The product is
$$ (v_{i}v_{j}^{\vee})(v_{k}v_{l}^{\vee})=v_{i}\,v_{j}^{\vee}(v_{k})\,v_{l}^{\vee}=\delta_{jk}\,v_{i}v_{l}^{\vee}=\delta_{jk}E_{il}, $$
which is the matrix-unit rule. The diagonal elements are idempotent, $E_{ii}^{2}=v_{i}v_{i}^{\vee}v_{i}v_{i}^{\vee}=v_{i}v_{i}^{\vee}=E_{ii}$, and pairwise orthogonal, $E_{ii}E_{jj}=v_{i}\delta_{ij}v_{j}^{\vee}=\delta_{ij}E_{ii}$. The sum is the resolution of the identity,
$$ \sum_{i}v_{i}v_{i}^{\vee}=I, $$
because both sides act as the identity on each basis vector $v_{k}$.
Conversely, let $(E_{ij})$ be a frame. The diagonal $E_{ii}$ is a rank-one idempotent, so its image is a line $L_{i}$; the three lines are independent because $\sum E_{ii}=I$, and they are the coordinate lines of a basis. Choose $v_{i}\in L_{i}$ with $E_{ii}v_{i}=v_{i}$. The off-diagonal $E_{ij}$ maps $L_{j}$ to $L_{i}$ and is nonzero; write $E_{ij}=v_{i}\varphi_{j}$ with $\varphi_{j}$ a linear functional. The relation $E_{ij}E_{jk}=E_{ik}$ gives $v_{i}\varphi_{j}(v_{j})\varphi_{k}=v_{i}\varphi_{k}$, hence $\varphi_{j}(v_{j})=1$, and $E_{ij}E_{kj}=0$ for $i\neq k$ gives $\varphi_{j}(v_{k})=0$ for $k\neq j$. So $\varphi_{j}=v_{j}^{\vee}$ and $E_{ij}=v_{i}v_{j}^{\vee}$. A second basis $v_{i}'$ giving the same frame satisfies $v_{i}'v_{j}'^{\vee}=v_{i}v_{j}^{\vee}$ for all $i,j$; taking $i=j$ gives $v_{i}'=s_{i}v_{i}$, and taking $i\neq j$ gives $s_{i}s_{j}^{-1}=1$, so all $s_{i}$ are equal. The basis is unique up to a common scalar.
Corollary. The frames of $M_{3}(\mathbb{C})$ form the homogeneous space
$$ \{\text{frames}\}\cong GL_{3}(\mathbb{C})/\mathbb{C}^{\times}I\cong PGL_{3}(\mathbb{C}), $$
of complex dimension $9-1=8$.
Proof. A basis is a point of $GL_{3}$; the frame depends on the basis only through its class modulo a common scalar, by the theorem. The action is free, so the quotient has dimension $9-1=8$.
The diagonal frame and the root. Given a frame, the three diagonal idempotents $P_{i}=E_{ii}$ are a complete orthogonal family, and
$$ H=\sum_{i=1}^{3}\omega_{i}E_{ii} $$
is a root with simple spectrum, by Roots of Unity and the Lagrange Idempotents of $M_3(\mathbb{C})$. Conversely, a root $H$ gives its three Lagrange idempotents $P_{i}$, and a frame with $E_{ii}=P_{i}$ is a choice of the six off-diagonal units. The root and the frame are therefore related by
$$ \{\text{frames}\}\longrightarrow\{\text{roots}\},\qquad (E_{ij})\longmapsto\sum_{i}\omega_{i}E_{ii}, $$
and the fibre is the gauge of §The Gauge below.
The standard frame. For the standard basis $v_{i}=e_{i}$ the frame is the standard matrix-unit frame $E_{ij}$, with $E_{ii}=\operatorname{diag}(0,\dots,1,\dots,0)$. Its root is $H_{0}=\operatorname{diag}(1,\omega,\omega^{2})$.
The Simple Nilpotents and the Composite
Definition. In a frame, the simple nilpotents are the two adjacent off-diagonal units
$$ N_{1}=E_{12},\qquad N_{2}=E_{23}. $$
They are square-zero,
$$ N_{1}^{2}=E_{12}E_{12}=0,\qquad N_{2}^{2}=E_{23}E_{23}=0, $$
and their only nonzero product is the composite
$$ N_{1}N_{2}=E_{12}E_{23}=E_{13}. $$
The opposite pair $E_{21},E_{32}$ has the composite $E_{31}=E_{32}E_{21}$.
Proposition. The algebra generated by $N_{1}$ and $N_{2}$ is the three-dimensional nilpotent algebra
$$ \mathfrak{n}_{+}=\operatorname{span}\{E_{12},E_{23},E_{13}\}, \qquad \mathfrak{n}_{+}^{2}=\mathbb{C}E_{13},\qquad \mathfrak{n}_{+}^{3}=0, $$
and $E_{13}$ is not an independent generator: it is the product $N_{1}N_{2}$.
Proof. The products of the two generators are $N_{1}^{2}=0$, $N_{2}^{2}=0$, $N_{1}N_{2}=E_{13}$ and $N_{2}N_{1}=E_{23}E_{12}=0$, because $E_{23}E_{12}=\delta_{31}E_{22}=0$. Every word in $N_{1},N_{2}$ of length at least two is therefore either $E_{13}$ or zero, and the span of $N_{1},N_{2},E_{13}$ is closed under multiplication. The product $E_{13}E_{13}=0$ and $E_{13}N_{1}=E_{13}E_{12}=0$, $N_{2}E_{13}=E_{23}E_{13}=0$, so the square of the span is $\mathbb{C}E_{13}$ and its cube is zero.
The composite is forced. The relation $E_{13}=E_{12}E_{23}$ is a consequence of the frame, not a choice. If the two simple nilpotents are rescaled by $\lambda_{1}$ and $\lambda_{2}$,
$$ E_{12}'=\lambda_{1}E_{12},\qquad E_{23}'=\lambda_{2}E_{23}, $$
then the composite is rescaled by the product,
$$ E_{13}'=E_{12}'E_{23}'=\lambda_{1}\lambda_{2}E_{13}. $$
There is no independent scalar on $E_{13}$: the third scalar is the product of the first two. This is the cocycle of §The Gauge.
The two simple nilpotents are the simple root vectors. In the basis of the root, $E_{12}$ and $E_{23}$ are the root vectors of the two simple roots of $\mathfrak{sl}_{3}$, and $E_{13}$ is the root vector of the highest root, whose weight is the sum of the two simple weights. The composite is therefore the highest-root vector, and its gauge weight is the product $\lambda_{1}\lambda_{2}$ of the two simple weights. The lower triangular pair $E_{21},E_{32}$ carries the negative roots, with the composite $E_{31}=E_{32}E_{21}$.
Explicit matrices. The standard simple nilpotents and their composite are
$$ N_{1}=\begin{pmatrix}0&1&0\\0&0&0\\0&0&0\end{pmatrix},\qquad N_{2}=\begin{pmatrix}0&0&0\\0&0&1\\0&0&0\end{pmatrix},\qquad N_{1}N_{2}=\begin{pmatrix}0&0&1\\0&0&0\\0&0&0\end{pmatrix}=E_{13}. $$
The products are checked directly: $N_{1}^{2}=N_{2}^{2}=0$, $N_{1}N_{2}=E_{13}$, $N_{2}N_{1}=0$. The algebra $\mathfrak{n}_{+}$ is the strictly upper triangular algebra, of dimension three; its cube is zero.
The Gauge
Definition. Let $t_{1},t_{2},t_{3}\in\mathbb{C}^{\times}$ and let $D=\operatorname{diag}(t_{1},t_{2},t_{3})$. The gauge transformation of the frame is
$$ E_{ij}\longmapsto E_{ij}'=D E_{ij}D^{-1}=\frac{t_{i}}{t_{j}}E_{ij}. $$
The diagonal units are fixed, $E_{ii}'=E_{ii}$; the off-diagonal units are rescaled by the ratios $t_{i}/t_{j}$.
Proposition. The gauge transformation preserves the frame. It fixes every idempotent $E_{ii}$, the root $H=\sum_{i}\omega_{i}E_{ii}$, the spectrum, the trace, the determinant and the characteristic polynomial of every element. It acts on the Peirce coordinates by
$$ \alpha_{ij}\longmapsto\frac{t_{j}}{t_{i}}\alpha_{ij}. $$
Proof. For the frame, $(D E_{ij}D^{-1})(D E_{kl}D^{-1})=D E_{ij}E_{kl}D^{-1}=\delta_{jk}D E_{il}D^{-1}$; the sum of the diagonal units is $D I D^{-1}=I$. The diagonal units are fixed because $D E_{ii}D^{-1}=(t_{i}/t_{i})E_{ii}=E_{ii}$. The root is a linear combination of the diagonal units, so it is fixed, and so are its spectrum and the polynomial invariants of every element, since conjugation by $D$ preserves the trace and the determinant. If $A=\sum\alpha_{ij}E_{ij}=\sum\alpha_{ij}'E_{ij}'$, then $\alpha_{ij}'E_{ij}'=\alpha_{ij}E_{ij}$ gives $\alpha_{ij}'=\alpha_{ij}(t_{j}/t_{i})$.
The gauge group is $(\mathbb{C}^{\times})^{2}$. The parameters $t_{1},t_{2},t_{3}$ and $ct_{1},ct_{2},ct_{3}$ give the same ratios, so the common scalar acts trivially and the effective group is
$$ (\mathbb{C}^{\times})^{3}/\mathbb{C}^{\times}\cong(\mathbb{C}^{\times})^{2}, $$
of complex dimension two. Its two coordinates are the ratios
$$ \lambda_{1}=\frac{t_{1}}{t_{2}},\qquad \lambda_{2}=\frac{t_{2}}{t_{3}}, \qquad \frac{t_{1}}{t_{3}}=\lambda_{1}\lambda_{2}. $$
In particular the simple nilpotents transform by
$$ E_{12}\longmapsto\lambda_{1}E_{12},\qquad E_{23}\longmapsto\lambda_{2}E_{23},\qquad E_{13}\longmapsto\lambda_{1}\lambda_{2}E_{13}. $$
Worked example. Take $\lambda_{1}=2$ and $\lambda_{2}=3$. The gauge parameters can be chosen as $t_{1}=6$, $t_{2}=3$, $t_{3}=1$, since $6/3=2$ and $3/1=3$. Then
$$ D=\begin{pmatrix}6&0&0\\0&3&0\\0&0&1\end{pmatrix}, \qquad D E_{12}D^{-1}=2E_{12},\quad D E_{23}D^{-1}=3E_{23},\quad D E_{13}D^{-1}=6E_{13}. $$
The diagonal idempotents are unchanged:
$$ D E_{11}D^{-1}=E_{11},\qquad D E_{22}D^{-1}=E_{22},\qquad D E_{33}D^{-1}=E_{33}, $$
and the root $H_{0}=\operatorname{diag}(1,\omega,\omega^{2})$ is unchanged. The composite is rescaled by $2\cdot3=6$, not by an independent number.
The cocycle that forbids a third scalar. Suppose that a rescaling $E_{ij}\mapsto\lambda_{ij}E_{ij}$ with $\lambda_{ij}\in\mathbb{C}^{\times}$ preserves the frame. Then for all $i,j,k$,
$$ \lambda_{ij}\lambda_{jk}=\lambda_{ik},\qquad \lambda_{ii}=1,\qquad \lambda_{ji}=\lambda_{ij}^{-1}. $$
Proposition. Every solution is of the form $\lambda_{ij}=t_{i}/t_{j}$ for some $t_{i}\in\mathbb{C}^{\times}$, and the solutions form a group isomorphic to $(\mathbb{C}^{\times})^{2}$.
Proof. Put $t_{i}=\lambda_{i1}$, with the index $1$ fixed. The cocycle gives $\lambda_{ij}=\lambda_{i1}\lambda_{1j}=t_{i}\lambda_{j1}^{-1}=t_{i}/t_{j}$, using $\lambda_{1j}=\lambda_{j1}^{-1}$. Conversely every assignment $t_{i}/t_{j}$ satisfies the three relations. Two assignments $t_{i}$ and $ct_{i}$ give the same $\lambda$, so the parameter space is $(\mathbb{C}^{\times})^{3}/\mathbb{C}^{\times}\cong(\mathbb{C}^{\times})^{2}$.
Corollary. If $\lambda_{1}$ is put on $E_{12}$ and $\lambda_{2}$ on $E_{23}$, then the scalar on $E_{13}$ is forced to be $\lambda_{1}\lambda_{2}$. A third independent scalar $\lambda_{3}$ on $E_{13}$ is incompatible with the frame unless $\lambda_{3}=\lambda_{1}\lambda_{2}$. The only freedom is the two-parameter torus, and the composite carries the product of the two simple weights.
The Count $8=6+2$
The frame has eight complex parameters, the root six, and the gauge two.
Theorem. The map
$$ \pi:\{\text{frames}\}\longrightarrow\{\text{roots with simple spectrum}\},\qquad (E_{ij})\longmapsto H=\sum_{i}\omega_{i}E_{ii}, $$
is surjective, and its fibre over every root is a torsor under the gauge group $(\mathbb{C}^{\times})^{2}$. Consequently
$$ \dim_{\mathbb{C}}\{\text{frames}\}=8=6+2=\dim_{\mathbb{C}}\{\text{roots}\}+\dim_{\mathbb{C}}(\mathbb{C}^{\times})^{2}. $$
Proof. The map is well defined by §The Frame from a Basis: the diagonal units of a frame are a complete orthogonal family, and the root built from them has simple spectrum. It is surjective because a root gives its three Lagrange idempotents $P_{i}$, and a frame with $E_{ii}=P_{i}$ exists: choose nonzero $E_{ij}\in A_{ij}$ normalized by $E_{ij}E_{ji}=P_{i}$; the Peirce pairing of The Peirce Decomposition of $M_{3}(\mathbb{C})$ makes the choice possible, and the matrix-unit table makes it a frame.
The fibre over $H$ consists of the frames with the same diagonal units. The Peirce lines $A_{ij}=P_{i}M_{3}P_{j}$ are one-dimensional, so a frame over $H$ is the choice of a nonzero $E_{ij}\in A_{ij}$ for each $i\neq j$, normalized by $E_{ij}E_{ji}=P_{i}$. The gauge $E_{ij}\mapsto(t_{i}/t_{j})E_{ij}$ acts on these choices, and it acts simply transitively: given two frames $E_{ij}$ and $E_{ij}'$ with the same diagonal, $E_{ij}'=\mu_{ij}E_{ij}$ with $\mu_{ij}\in\mathbb{C}^{\times}$; the matrix-unit table forces $\mu_{ij}\mu_{jk}=\mu_{ik}$ and $\mu_{ji}=\mu_{ij}^{-1}$, so by the cocycle proposition $\mu_{ij}=t_{i}/t_{j}$ for a unique class of $t$ modulo common scalar. The fibre is therefore a $(\mathbb{C}^{\times})^{2}$-torsor, of dimension two.
The dimensions are the orbit dimension $6$ of Roots of Unity and the Lagrange Idempotents of $M_3(\mathbb{C})$ and the torus dimension $2$; the total is the frame dimension $8$ of §The Frame from a Basis. The two counts agree, and the agreement is the content of the theorem.
The three counts of the same eight.
- Frames. $PGL_{3}$ has dimension $8$.
- Root plus gauge. The root orbit $PGL_{3}/T$ has dimension $6$; the torus $T$ has dimension $2$; $6+2=8$.
- Basis plus gauge. A basis of $\mathbb{C}^{3}$ has $9$ parameters; the common scalar of the frame removes $1$, leaving the $8$ parameters of the frame. Of these, the $2$ independent scalings of the gauge are the fibre over the root, leaving the $6$ parameters of the root orbit.
The gauge is not a redundancy of the root: it is the fibre of the frame over the root, and it is the freedom of the off-diagonal units at fixed idempotents.
The Maximal Torus of $PGL_{3}$
The gauge group is the maximal torus of the projective linear group.
The diagonal matrices of $GL_{3}$ form a maximal torus $T_{GL}\cong(\mathbb{C}^{\times})^{3}$. Its image in $PGL_{3}=GL_{3}/\mathbb{C}^{\times}I$ is
$$ T=T_{GL}/(T_{GL}\cap\mathbb{C}^{\times}I)\cong(\mathbb{C}^{\times})^{3}/\mathbb{C}^{\times}\cong(\mathbb{C}^{\times})^{2}, $$
which is a maximal torus of $PGL_{3}$, of dimension two. The gauge of §The Gauge is exactly this torus acting by conjugation on the frame; the root vectors $E_{ij}$ are the weight vectors, of weights $t_{i}/t_{j}$, and the simple nilpotents $E_{12},E_{23}$ carry the two simple roots. The composite $E_{13}$ carries the highest root, the sum of the two simple roots in the character lattice.
The root system. The weights $t_{i}/t_{j}$ for $i\neq j$ are the six roots of $A_{2}$, and the two simple ones are $\alpha_{1}=t_{1}/t_{2}$ and $\alpha_{2}=t_{2}/t_{3}$; the highest root is $\alpha_{1}\alpha_{2}=t_{1}/t_{3}$. The gauge parameters $\lambda_{1},\lambda_{2}$ are the values of the two simple roots, and the composite weight is their product. The root system is the classical $A_{2}$ of $PGL_{3}$; the frame is a point of the group, and the gauge is the torus that fixes the root.
The comparison with the biquaternion case. For $n=2$ the frame has two diagonal idempotents and two off-diagonal units $E_{12},E_{21}$; the gauge is
$$ E_{12}\longmapsto\mu E_{12},\qquad E_{21}\longmapsto\mu^{-1}E_{21}, $$
with one parameter $\mu\in\mathbb{C}^{\times}$. The torus is $(\mathbb{C}^{\times})^{2}/\mathbb{C}^{\times}\cong\mathbb{C}^{\times}$, of dimension one. The $M_{3}$ case has two simple nilpotents and two parameters; the composite is forced. An involution of the two-dimensional frame that anticommutes with $H$ is a point of this one-parameter gauge: it fixes the relative scale of $E_{12}$ and $E_{21}$, and it is not a second primitive idempotent. The pattern is $(\mathbb{C}^{\times})^{n-1}$ for $M_{n}(\mathbb{C})$: one parameter for the biquaternion algebra $M_{2}(\mathbb{C})$, two for $M_{3}(\mathbb{C})$, and $n-1$ in general.
Summary
A frame is a basis of $M_{3}(\mathbb{C})$ by matrix units. It is the same data as a basis of $\mathbb{C}^{3}$ up to a common scalar, so the frames form $PGL_{3}(\mathbb{C})$, of complex dimension $8$:
$$ \boxed{\;\{\text{frames}\}\cong PGL_{3}(\mathbb{C}),\qquad \dim_{\mathbb{C}}=8.\;} $$
The diagonal units are three idempotents; the simple nilpotents are
$$ N_{1}=E_{12},\qquad N_{2}=E_{23},\qquad N_{1}^{2}=N_{2}^{2}=0,\qquad N_{1}N_{2}=E_{13}. $$
The composite $E_{13}$ is the product of the two simple nilpotents, not an independent generator.
The gauge is
$$ E_{ij}\longmapsto\frac{t_{i}}{t_{j}}E_{ij},\qquad t_{i}\in\mathbb{C}^{\times}, $$
with group $(\mathbb{C}^{\times})^{3}/\mathbb{C}^{\times}\cong(\mathbb{C}^{\times})^{2}$. It fixes every idempotent, the root, the spectrum and the polynomial invariants. The two simple rescalings $\lambda_{1},\lambda_{2}$ force the composite rescaling $\lambda_{1}\lambda_{2}$; the cocycle $\lambda_{ij}\lambda_{jk}=\lambda_{ik}$ has only the solutions $\lambda_{ij}=t_{i}/t_{j}$, and there is no third independent scalar.
| Quantity | Value for $M_{3}(\mathbb{C})$ |
|---|---|
| Frame dimension | $8$ |
| Root orbit dimension | $6$ |
| Gauge torus dimension | $2$ |
| Simple nilpotents | $E_{12},E_{23}$ |
| Composite | $E_{13}=E_{12}E_{23}$ |
| Gauge parameters | $\lambda_{1},\lambda_{2}$ with $\lambda_{3}=\lambda_{1}\lambda_{2}$ |
| $\dim$ of the maximal torus of $PGL_{3}$ | $2$ |
The count $8=6+2$ is the orbit-stabiliser count for the action of $PGL_{3}$ on the root: the root is the orbit, the gauge is the stabiliser.
Summary of Notation
| Symbol | Meaning |
|---|---|
| $E_{ij}$ | the matrix units of a frame |
| $v_{i}^{\vee}$ | the dual basis functional, $v_{j}^{\vee}(v_{k})=\delta_{jk}$ |
| $N_{1}=E_{12}$, $N_{2}=E_{23}$ | the simple nilpotents |
| $\mathfrak{n}_{+}$ | the nilpotent algebra $\operatorname{span}\{E_{12},E_{23},E_{13}\}$ |
| $t_{i}$ | gauge parameters, $t_{i}\in\mathbb{C}^{\times}$ |
| $\lambda_{1}=t_{1}/t_{2}$, $\lambda_{2}=t_{2}/t_{3}$ | the two gauge parameters on the simple nilpotents |
| $T$ | the maximal torus $(\mathbb{C}^{\times})^{2}$ of $PGL_{3}$ |
| $\alpha_{1}=t_{1}/t_{2}$, $\alpha_{2}=t_{2}/t_{3}$ | the simple roots of $A_{2}$ |
Further Reading
- The Peirce Decomposition of $M_3(\mathbb{C})$, for the nine lines and the coarse split.
- Roots of Unity and the Lagrange Idempotents of $M_3(\mathbb{C})$, for the root $H$ and its idempotents.
- The Two-Parameter Gauge of $M_3(\mathbb{C})$, for the gauge read on the coherence entries and the comparison with the single $\lambda$ of $\mathbb{B}$.
- The Matrix Algebra $M_3(\mathbb{C})$, for the matrix units and the projective group.
- Matrix Algebras, for the matrix-unit table.
- Peirce Decompositions of the Biquaternion Algebra, for the two-dimensional frame and its single $\mu$.
- Root Systems and Classification, for the root system $A_{2}$ and the maximal torus.
- The Frame, the Simple Nilpotents and the Gauge of $M_n(\mathbb{C})$, for the general construction at order $n$.