The Pin Group of the Gamma Frame

Introduction

The Pin group of a Lorentzian four-space is generated by the unit vectors. The article keeps its earlier title; the set it uses is the frame of five elements, and its dictionary to the literature notation is in the Gamma section. It is not a subgroup of the biquaternion algebra: its odd elements include the vectors themselves, and the vectors are not in $\mathbb{B}$. The norm-one part of the even Pin is the spin group, and that part is a biquaternion object: $$ \mathrm{Pin}(1,3)\not\subseteq\mathbb{B},\qquad \mathrm{Spin}(1,3)=\{u\in\mathrm{Pin}(1,3)\cap\mathrm{Cl}^{+}:u u^{\natural}=1\}\cong\mathbb{B}^{\times}_{1}. $$ The frame gives five unit vectors and a finite subgroup of Pin, of order $32$; the full Pin group needs the whole unit sphere, not only the set. This article separates the two, computes the reflection sign, the determinant and the lift, and identifies the finite group of the set inside the continuous group.

The frame is that of The Gamma Frame and the Complexified Dirac Algebra of $M_4(\mathbb{C})$; the versor group and the double cover are Biquaternion Versors and the Orthogonal Group; the Clifford, Pin and Spin definitions are The Clifford, Pin and Spin Groups with Signed Inner Conjugation. The finite group of the two-sided matrices is The Extraspecial Group of the Two-Sided Matrices of $M_4(\mathbb{C})$; the five frame elements as two-sided operators are The $\mathbb{Z}$-Grading and the Odd Slot of $M_4(\mathbb{C})$.

Pin and Spin

The definitions. Let $V$ be a real four-dimensional space with a quadratic form $q$ of inertia $(1,3)$, and let $\mathrm{Cl}(V)$ be its Clifford algebra. A versor is a product $\zeta_1\cdots\zeta_k$ of non-isotropic vectors. The Clifford group is $$ \Gamma=\{u\in\mathrm{Cl}(V)^{\times}:\alpha(u)Vu^{-1}\subseteq V\}, $$ with $\alpha$ the grade involution. The Pin group is $$ \mathrm{Pin}=\{u\in\Gamma:u u^{\natural}=\pm1\}, $$ where ${}^{\natural}$ is the Clifford conjugation. The spin group is the norm-one part of the even Pin: $$ \mathrm{Spin}=\{u\in\mathrm{Pin}\cap\mathrm{Cl}^{+}:u u^{\natural}=1\}. $$ The convention here is the identity component, the group that acts on the spinor as $SL_2(\mathbb{C})$; the general article The Clifford, Pin and Spin Groups with Signed Inner Conjugation writes $\mathrm{Spin}$ for the full even Pin, which adds the norm $-1$ coset. The Clifford group maps onto $O(1,3)$ by $u\mapsto\rho_u$, with $\rho_u(\eta)=\alpha(u)\eta u^{-1}$, and kernel the scalars. The map $\mathrm{Pin}\to O(1,3)$ is surjective and two-to-one, with kernel $\{\pm1\}$; the spin group maps onto $SO^{+}(1,3)$ two-to-one. The full even part $\mathrm{Pin}\cap\mathrm{Cl}^{+}$ also contains the norm $-1$ coset, which covers the time-reversing component of $SO(1,3)$ and is not part of the identity component $\mathrm{Spin}(1,3)$.

The biquaternion part. The even part of the Clifford algebra is $\mathbb{B}$, and the norm-one part of the even Pin is the spin group, $$ \mathrm{Spin}(1,3)\cong\mathbb{B}^{\times}_{1} =\{b\in\mathbb{B}:b b^{\natural}=1\}. $$ The full even part of Pin is $\mathrm{Spin}\cup\varkappa_{\mathrm{crd}}\,\mathrm{Spin}$, where $\varkappa_{\mathrm{crd}}$ is any even unit of norm $-1$, for example the left multiplication by $-i\vec e_x$; the second coset covers the time-reversing component of $SO(1,3)$. The odd part is not in $\mathbb{B}$: every odd element is a sum of a vector and a trivector, and the vectors are the Lorentzian space itself. This is the exact statement of the corollary of Biquaternion Versors and the Orthogonal Group: $\mathrm{Pin}\not\subseteq\mathbb{B}$, and only the even part is a biquaternion object.

The decomposition. The Pin group is the disjoint union of its even and odd parts, and each part is the union of the two norm cosets: $$ \mathrm{Pin}(1,3)=\bigl(\mathrm{Spin}\cup\varkappa_{\mathrm{crd}}\,\mathrm{Spin}\bigr)\ \cup\ \zeta_0\bigl(\mathrm{Spin}\cup\varkappa_{\mathrm{crd}}\,\mathrm{Spin}\bigr), $$ for any odd unit vector $\zeta_0$ and any even unit $\varkappa_{\mathrm{crd}}$ of norm $-1$. In the biquaternion dictionary, with $\zeta_0=A_0=\bar{\cdot}$ and $\varkappa_{\mathrm{crd}}=L_{-i\vec e_x}$, $$ \boxed{\;\mathrm{Pin}(1,3) =L(\mathbb{B}^{\times}_{1})\ \cup\ L(\mathbb{B}^{\times}_{-1})\ \cup\ \bar{\cdot}\,L(\mathbb{B}^{\times}_{1})\ \cup\ \bar{\cdot}\,L(\mathbb{B}^{\times}_{-1})\;} $$ as operations on $\mathbb{B}$, where $\mathbb{B}^{\times}_{\pm1}=\{b\in\mathbb{B}:b b^{\natural}=\pm1\}$. The norm-one even coset is $\mathrm{Spin}(1,3)=\mathbb{B}^{\times}_1$; the norm $-1$ even coset is the other component of the even Pin. The odd half acts by the bar composed with a left multiplication. This is the continuous group generated by the even versors and the odd generator; it is the full Pin group, and it is not the finite frame subgroup.

Proof. The decomposition is the partition of Pin by parity and by the sign of the norm. The identification $\mathrm{Spin}\cong\mathbb{B}^{\times}_1$ is the versor theorem, and the left multiplication $L_b$ is the action of $b$ on $\mathbb{B}$. The odd generator $A_0$ is the bar in the real-linear dictionary.

The Unit Vectors of the Frame

The basis. The frame is written in the coefficient basis $\mathcal{E}=(e_t,\vec e_x,\vec e_y,\vec e_z)$ of the biquaternion algebra $\mathbb{B}=\mathbb{C}\otimes_{\mathbb{R}}\mathbb{H}$, with $\vec e_k^2=-e_t$ and $\vec e_x\vec e_y=\vec e_z$, $\vec e_y\vec e_z=\vec e_x$, $\vec e_z\vec e_x=\vec e_y$. The four are a $\mathbb{C}$-basis of $\mathbb{B}$, the bar is the coefficient conjugation and $L_b$ is the left multiplication $L_b\tilde X=b\tilde X$. The basis and the four left matrices are $M_4(\mathbb{C})$ Axis-Free Representation; the sixteen matrices $L_{e_\mu}R_{e_\nu}$ are $M_4(\mathbb{C})$ Axis-Free Representation: Operators.

The squares. The five frame elements, $A_0=\bar{\cdot}$, $A_k=\bar{\cdot}\circ L_{-i\vec e_k}$ for $k=1,2,3$, and $A_5=R_{-i\vec e_z}$, are unit vectors up to sign: $$ (A_0)^2=+I,\qquad (A_k)^2=-I\quad(k=1,2,3),\qquad (A_5)^2=+I . $$ The first four are the Lorentzian unit vectors; the fifth is the pseudoscalar, not a vector, but it is invertible in the Clifford group. The metric square is $$ q(A_\mu)=g_{\mu\mu}\in\{+1,-1\}. $$

The reflection. For a non-isotropic $\zeta\in V$, the reflection in the hyperplane orthogonal to $\zeta$ is $$ \rho_\zeta(\eta)=-\zeta\eta\zeta^{-1} =\eta-2\frac{q(\zeta,\eta)}{q(\zeta)}\,\zeta . $$ For a unit vector, $\zeta^{-1}=\zeta/q(\zeta)$, and with $\varepsilon(\zeta):=q(\zeta)=\zeta^2$, $$ \rho_\zeta(\eta)=-\varepsilon(\zeta)\,\zeta\eta\zeta . $$ The sign $\varepsilon(\zeta)$ is $+1$ for a timelike vector and $-1$ for a spacelike vector. The four frame reflections are $$ \rho_{A_0}(\eta)=-A_0\eta A_0,\qquad \rho_{A_k}(\eta)=+A_k\eta A_k . $$

The determinant. The reflection $\rho_\zeta$ has determinant $-1$ on $V$. A product of $k$ reflections has determinant $(-1)^k$. The even words give determinant $+1$ and form the even part of Pin; the norm-one even words form the spin group, and the norm $-1$ even words form its other component. The odd words give determinant $-1$ and form the odd part. The Lorentz determinant is the parity of the word length, not the determinant of the $4\times4$ matrix.

The matrix determinant. In the complex-linear description each frame element is an invertible multiple of the two-sided product $L_{a}R_{b}$, and every one has determinant $+1$ as a $4\times4$ matrix: $$ \det(iL_{e_t}R_{\vec e_x})=\det(iL_{\vec e_x}R_{\vec e_y})=\det(iL_{\vec e_y}R_{\vec e_y})=\det(iL_{\vec e_z}R_{\vec e_y})=\det(-iL_{e_t}R_{\vec e_z})=1 . $$ The $4\times4$ determinant is therefore blind to the Lorentz determinant: a reflection is an odd word with Lorentz determinant $-1$, but its spinor matrix has determinant $+1$. The two determinants are different invariants. The spinor determinant is the scalar $1$ of $SL_4(\mathbb{C})$; the Lorentz determinant is the parity of the word.

The Finite Group of the Frame

Theorem (the group of the frame). Let $U$ be the group of the sixteen signed biquaternion units $$ U=\{\pm1,\pm \vec e_x,\pm \vec e_y,\pm \vec e_z,\pm i,\pm i \vec e_x,\pm i \vec e_y,\pm i \vec e_z\}. $$ The four generators $A_0,A_1,A_2,A_3$ generate the group $$ G_A=L(U)\ \cup\ \bar{\cdot}\,L(U), $$ of order $32$. The even part $L(U)$ has order $16$ and lies in the even part of Pin; the odd part $\bar{\cdot}L(U)$ has order $16$ and lies in the odd part of Pin.

Proof. The even products of the frame elements are the left multiplications by the even Clifford basis elements: $I$, the six bivectors and $A_5$. The eight basis elements, with their signs, are the sixteen elements of $U$. The odd products are the bar times those left multiplications: the four vectors and four trivectors, the eight odd Clifford basis elements with their signs. The two sets are disjoint, and each has sixteen elements.

The elements. The even part is $$ L(U)=\{\pm L_b:b\in\{1,\vec e_x,\vec e_y,\vec e_z,i \vec e_x,i \vec e_y,i \vec e_z,i\}\}, $$ and the odd part is $$ \bar{\cdot}L(U)=\{\pm\bar{\cdot}L_b:b\in\{1,\vec e_x,\vec e_y,\vec e_z,i \vec e_x,i \vec e_y,i \vec e_z,i\}\}. $$ The group is a finite subgroup of $\mathrm{Pin}(1,3)$, of order $32$, generated by the four generators $A_0,A_1,A_2,A_3$ of the frame.

The odd generators. In the real-linear dictionary, $$ A_0=\bar{\cdot},\qquad A_k=\bar{\cdot}\circ L_{-i \vec e_k}, $$ so the odd generators are the bar and the bar times the three spacelike units. They satisfy the operator relation of The Star and the Conjugate-Linear Operators of $M_4(\mathbb{C})$. The finite group is the smallest subgroup of Pin that contains them.

The spin part. The even group $L(U)$ is a finite subgroup of the even part of Pin. It contains the quaternion group $$ Q_8=\{\pm1,\pm \vec e_x,\pm \vec e_y,\pm \vec e_z\} $$ as the subgroup of index two, and the eight elements $\pm i,\pm i \vec e_x,\pm i \vec e_y,\pm i \vec e_z$ in the other coset, of norm $-1$. The subgroup $Q_8$ lies in $\mathrm{Spin}(1,3)\cong\mathbb{B}^{\times}_1$; the other coset lies in the norm $-1$ even component. The group $L(U)$ is the Pauli group of order $16$; it is the finite even group of the frame, not the full even Pin.

The frame is not the whole Pin group. The unit sphere of $V$ is continuous, and the Pin group it generates is a Lie group of real dimension six (the spin group) plus the odd coset. The five frame elements generate only the finite subgroup $G_A$. The passage from the frame to the full group is the passage from the five vectors to all linear combinations $\zeta=\sum_\mu \zeta_\mu A_\mu$ with $q(\zeta)=\pm1$; the finite group is the skeleton, the continuous group is the envelope.

The Finite Group of the Two-Sided Matrices

The group. The thirty-two matrices $$ G=\{\pm L_{e_\mu}R_{e_\nu}:\mu,\nu\in\{t,x,y,z\}\} $$ form a group of order $32$, isomorphic to $Q_8\circ Q_8$, the central product of two quaternion groups. It is extraspecial: centre $\{\pm I\}$, quotient $(\mathbb{Z}/2)^4$, exponent four, nineteen involutions, seventeen conjugacy classes. It is the group of The Extraspecial Group of the Two-Sided Matrices of $M_4(\mathbb{C})$.

The action on the frame. Every $L_{e_\mu}R_{e_\nu}$ conjugates each frame element to plus or minus itself: $$ L_{e_\mu}R_{e_\nu}\,A_\rho\,(L_{e_\mu}R_{e_\nu})^{-1}=\pm A_\rho . $$ No permutation of the four generators occurs. The image of $G$ in $O(1,3)$ is the group of sign changes $(\mathbb{Z}/2)^4$, of order $16$; the kernel is $\{\pm I\}$. The group is a central extension of the sign group by $\{\pm I\}$.

The determinant. Every element of $G$ has determinant $+1$: $$ \det L_{e_\mu}R_{e_\nu}=1 . $$ Hence $G\subset SL_4(\mathbb{C})$, and the group is a finite subgroup of the Clifford group of the frame. The determinant is again blind to the Lorentz determinant: the sign changes include transformations of Lorentz determinant $-1$, while the matrix determinant stays $+1$.

The two finite groups. The group $G_A$ of the frame is generated by the five frame elements and acts on the spinor by left multiplication; the two-sided group $G$ is generated by the $L_{e_\mu}R_{e_\nu}$ and acts on the frame by conjugation. Both have order $32$, both are finite shadows of the same Clifford group, and they are different subgroups: $G_A$ contains the odd elements, $G$ contains the signed permutation matrices. The first is the versor skeleton of the frame; the second is the operator skeleton of the two-sided basis.

The Lift and the Double Cover

The map. The Clifford group maps onto $O(1,3)$ by $$ u\longmapsto\rho_u,\qquad \rho_u(\eta)=\alpha(u)\eta u^{-1}, $$ with kernel the scalars. The Pin group is the norm-$\pm1$ part of $\Gamma$; its map to $O(1,3)$ is surjective and two-to-one, with kernel $\{\pm1\}$. The spin group maps onto $SO^{+}(1,3)$ two-to-one, and the odd part maps into the determinant $-1$ components. The full even part, including the norm $-1$ coset, covers the time-reversing component of $SO(1,3)$.

The biquaternion form. On the norm-one even part, the map is the versor map $$ \mathbb{B}^{\times}_1\longrightarrow SO^{+}(1,3),\qquad b\longmapsto\bigl(\eta\mapsto b\,\eta\,b^{*}\bigr), $$ which is two-to-one with kernel $\{\pm1\}$. The norm $-1$ even coset maps to the time-reversing component of $SO(1,3)$, and the odd part maps into the determinant $-1$ components.

The lift of a reflection. A reflection $\rho_\zeta$ lifts to the two elements $\pm\zeta$ of Pin. The sign is the usual sign ambiguity of the spinor representation; it is not a sign of the Lorentz transformation, which is fixed by $\det\rho_\zeta=-1$. The two lifts differ by $-1$, the kernel of the double cover.

The sign $\varepsilon$. The reflection formula carries the metric sign $\varepsilon(\zeta)=q(\zeta)$. The timelike and spacelike generators of the frame therefore enter the two-sided operator with opposite signs: $$ \rho_{A_0}(\eta)=-A_0\eta A_0,\qquad \rho_{A_k}(\eta)=+A_k\eta A_k . $$ This is the only place where the inertia of the frame enters the reflection; the anticommutation relations themselves carry it in the squares $(A_0)^2=+I$, $(A_k)^2=-I$.

The spinor determinant. The spinor representation is a map $$ \mathrm{Pin}(1,3)\longrightarrow GL_4(\mathbb{C}), $$ and its image lies in $SL_4(\mathbb{C})$ because every frame element has determinant $1$ in the complex-linear description. The determinant of the spinor matrix is not the determinant of the Lorentz transformation; the latter is the parity of the word. The two-to-one cover is visible in the spinor matrices, not in their determinants.

The Legacy Pauli Gamma Formulation

The Pauli and gamma formulation is legacy: the reasons are explained in the article Why AXF Matrices instead of Pauli and Gamma Matrices. The generators carry no frame label: $\gamma_\mu$ is one set of matrices, $\gamma_0^2=I_4$ and $\gamma_k^2=-I_4$. The frame belongs to the coordinate, not to the array.

The gammas are a frame of $\mathrm{Cl}_{1,3}(\mathbb{C})=M_4(\mathbb{C})$, not a different algebra.

Here $L_a$ and $R_b$ are the left and right multiplications, $L_a\tilde{Ξ}=a\tilde{Ξ}$ and $R_b\tilde{Ξ}=\tilde{Ξ}b$, and $c(M)$ is the coefficient matrix of $M$ in the basis of the sixteen products.

$$ \gamma_0=iL_{e_t}R_{\vec e_x},\quad \gamma_1=iL_{\vec e_x}R_{\vec e_y},\quad \gamma_2=iL_{\vec e_y}R_{\vec e_y},\quad \gamma_3=iL_{\vec e_z}R_{\vec e_y},\quad \gamma_5=-iL_{e_t}R_{\vec e_z} . $$

In the real-linear frame, in which the body writes, the same five are $\gamma_0=\bar{\cdot}$, $\gamma_k=\bar{\cdot}\circ L_{-i\vec e_k}$ for $k=1,2,3$, and $\gamma_5=R_{-i\vec e_z}=-iL_{e_t}R_{\vec e_z}$: the four vectors are antilinear there, and the basis operators above are their complex-linear form.

A single product is a pure tensor; the products form the seven-dimensional cone $\operatorname{rank}c(M)\le1$, and a general element is a sum of at most four, with $\ell(M)=\operatorname{rank}c(M)$.

$\gamma_5$ is the reference-frame representative of the chirality operator; the central grading $\chi=i\tilde{Ζ}_{\mathrm{ad}}$ — the difference of the two central idempotents of the complexified algebra, so $\chi^2=e_t$ — is a different object, represented by $\gamma_5$ on the Dirac module.

The dictionary is complex-linear; over $\mathbb{R}$ the products span $\mathrm{Cl}_{1,3}(\mathbb{R})\cong M_2(\mathbb{H})$ and the basis operators span $\mathrm{Cl}_{3,1}(\mathbb{R})\cong M_4(\mathbb{R})$.

Summary

The Pin group of the Lorentzian four-space is not contained in $\mathbb{B}$; its even part contains the spin group $\mathrm{Spin}(1,3)\cong\mathbb{B}^{\times}_1$ as its norm-one component, together with the norm $-1$ coset. The group generated by the even versors and the odd generator is the union of the even and odd parts, each with its two norm cosets, $$ \mathrm{Pin}(1,3)=L(\mathbb{B}^{\times}_1)\cup L(\mathbb{B}^{\times}_{-1})\cup\bar{\cdot}L(\mathbb{B}^{\times}_1)\cup\bar{\cdot}L(\mathbb{B}^{\times}_{-1}), $$ in the real-linear dictionary. The five frame elements are unit vectors with squares $+1,-1,-1,-1$; the reflection in the hyperplane orthogonal to an invertible element $\zeta$ is $\rho_\zeta(\eta)=-\varepsilon(\zeta)\zeta\eta\zeta$ with $\varepsilon(\zeta)=q(\zeta)=\zeta^2$, and its determinant is $-1$. The frame generates a finite subgroup of order $32$, $$ G_A=L(U)\cup\bar{\cdot}L(U), $$ with $U$ the sixteen signed biquaternion units; its even part of order $16$ is a finite subgroup of the even Pin, and its subgroup $Q_8$ lies in $\mathrm{Spin}(1,3)\cong\mathbb{B}^{\times}_1$. The two-sided matrices form a second finite group $G=\{\pm L_{e_\mu}R_{e_\nu}\}$ of order $32$, extraspecial $Q_8\circ Q_8$, determinant $1$, acting on the frame by sign changes with image $(\mathbb{Z}/2)^4$ of order $16$. The spinor matrices all have determinant $1$, while the Lorentz determinant is the parity of the word; the two determinants are different invariants. The full Pin group is continuous; the frame gives only its finite skeleton.

object value
Pin versors of norm $\pm1$ in $\mathrm{Cl}(1,3)$
Spin norm-one even part of Pin, $\cong\mathbb{B}^{\times}_1$
Pin in $\mathbb{B}$ not contained; only the even part is
Pin decomposition $L(\mathbb{B}^{\times}_1)\cup L(\mathbb{B}^{\times}_{-1})\cup\bar{\cdot}L(\mathbb{B}^{\times}_1)\cup\bar{\cdot}L(\mathbb{B}^{\times}_{-1})$
frame squares $(A_0)^2=1$, $(A_k)^2=-1$, $(A_5)^2=1$
reflection $\rho_\zeta(\eta)=-\varepsilon(\zeta)\zeta\eta\zeta$, $\varepsilon(\zeta)=\zeta^2$
reflection determinant $-1$
group of the frame $G_A=L(U)\cup\bar{\cdot}L(U)$, order $32$
even part of that group $L(U)$, order $16$, finite subgroup of the even Pin; $Q_8\subset\mathrm{Spin}$
two-sided group $\{\pm L_{e_\mu}R_{e_\nu}\}\cong Q_8\circ Q_8$, order $32$
action on frame sign changes, image $(\mathbb{Z}/2)^4$, order $16$
spinor determinant $1$ for every frame element, in the complex-linear description
Lorentz determinant $(-1)^k$ for a word of $k$ reflections

Summary of Notation

Symbol Meaning
$\mathrm{Cl}(V)$ the Clifford algebra of the Lorentzian four-space
$\mathrm{Cl}^{+}$ its even part, isomorphic to $\mathbb{B}$
$\alpha$ the grade involution
$\mathrm{Pin},\mathrm{Spin}$ the pin and spin groups
$\mathbb{B}^{\times}_1$ the norm-one group of the biquaternion algebra
$A_\mu,A_5$ the frame elements $\bar{\cdot}$, $\bar{\cdot}\circ L_{-i\vec e_k}$, $R_{-i\vec e_z}$; the fifth is the chirality
$\varepsilon(\zeta)$ the metric square $q(\zeta)=\zeta^2$
$\rho_\zeta$ the reflection in the hyperplane orthogonal to $\zeta$
$U$ the sixteen signed biquaternion units
$G_A$ the group of the frame $L(U)\cup\bar{\cdot}L(U)$
$G$ the two-sided group $\{\pm L_{e_\mu}R_{e_\nu}\}$
$L_{e_\mu}R_{e_\nu}$ the two-sided operator $L_{e_\mu} R_{e_\nu}$

Further Reading