1. Introduction

This paper establishes the definitive notation for the D8 program. It contains no physical claims and derives no new results; rather, its goal is to provide a single, unambiguous reference for the algebra, basis, generators, and charge conventions. Laying this groundwork here ensures that subsequent papers do not need to restate these definitions or risk drifting into inconsistent notation.

The core setting is the Clifford algebra $\mathrm{Cl}(1,7)$. Crucially, we make one foundational architectural commitment that holds throughout this entire body of work: the algebra is treated as a three-complex-dimensional fiber over a four-dimensional spacetime base, rather than a single eight-dimensional manifold. This is a matter of interpretation rather than a theorem, but it is the interpretation that dictates the meaning of every symbol below. It specifies which generators belong to spacetime and which belong to the fiber, elevates the relationship between them from a trivial identity to an object in its own right, and ultimately allows a soldering between them to exist.

If one instead views $\mathrm{Cl}(1,7)$ as a single eight-dimensional space with one metric and one set of tetrads, the definitions from §3 onward will not function as intended. We define this soldering structurally in §3.3 and address its dynamics in Paper 2; in the present paper, it serves primarily as the conceptual reason to maintain the base/fiber distinction.

Two consequences of this interpretation directly shape how this material is presented:

• Split indices: We separate indices throughout the text, using greek letters (ex: $\mu$) for the base, and latin letters (ex: $m$) for the fiber, and capital letters (ex: $M$) when referring to both.
• Distinct actions: Left and right actions on an algebra element represent physically distinct operations and are never interchanged. Consequently, the generators introduced in §7 are stated twice—once acting from the right and once from the left—because the difference between them is a physical reality, not a convention.

To outline the organization, §2 details the algebra, its grade structure, and the conventions for reversion and conjugation. §3 introduces the base/fiber split alongside the fiber Hermitian metric. In §4, we provide the Wick-rotated oscillator basis and its explicit dictionary to the generators $\gamma_M$. §5 constructs the sixteen primitive idempotents, the resulting block decomposition of a general element $\psi$, and the conjugation relations that tie these blocks together. Following this, §6 introduces the four commuting ladder bivectors that generate the decomposition, while §7 defines the $u(3)$ basis in its right, left, and coset forms alongside its closure relations. Finally, §8 fixes the charge operators, §9 collects the necessary normalizations, and §10 aggregates the reference tables.

We conclude with three conventions applied silently throughout the text. First, reversion is always meant in the algebraic sense as a grade-dependent sign; when verifying claims via matrix realization, the intertwiner remains a property of that specific realization rather than the algebra itself. Second, the numbering of the $\lambda$ generators is defined strictly by §7.1 and does not follow external literature conventions. Third, the relative signs in the block table of §5.3 are determined by the ladder construction in §6 rather than chosen arbitrarily; §5.5 clarifies exactly which signs are structurally fixed and which remain free up to a global sign.

2. The Algebra

2.1 Definition and dimension

We work within the real Clifford algebra $\mathrm{Cl}(1,7)$, which is generated by the elements $\gamma_M$ for $M \in \{0, 1, \ldots, 7\}$. These generators satisfy the defining relation:

$\gamma_M \gamma_N + \gamma_N \gamma_M = 2 g_{MN}$

where the metric is $g = \mathrm{diag}(+1, -1, -1, -1, -1, -1, -1, -1)$, meaning $\gamma_0^2 = +1$ and $\gamma_i^2 = -1$ for $i = 1, \ldots, 7$. This relation spans the entire eight-dimensional range. Consequently, the base and fiber generators introduced in §3 are components of a single, unified algebra, rather than two separate algebras placed side by side.

An ordered basis is formed by the products $\gamma_{M_1} \cdots \gamma_{M_k}$ where $M_1 < \cdots < M_k$, corresponding to each subset of $\{0, \ldots, 7\}$. The algebra therefore has a dimension of $2^8 = 256$ over $\mathbb{R}$, which breaks down by grade as follows:

grade $k$	0	1	2	3	4	5	6	7	8
$\dim$	1	8	28	56	70	56	28	8	1

Two structural properties will be essential later. First, $\mathrm{Cl}(1,7) \cong \mathbb{R}(16)$, meaning it maps directly to the algebra of real $16 \times 16$ matrices. The fact that this representation space is strictly real explains why conjugate partners naturally emerge in the block decomposition of §5. Second, the even subalgebra $\mathrm{Cl}(1,7)^{+}$ is $128$-dimensional, with $\mathrm{Cl}(1,7)^{+} \cong \mathrm{Cl}(1,6) \cong \mathbb{C}(8)$. Its center is spanned by $\{1, \omega\}$, where $\omega$ is the pseudoscalar defined in §2.4.

Throughout this work, we treat any field $\psi$ as an actual element of the algebra, rather than as an external column spinor. As a result, gauge and frame transformations act on it from both sides via $\psi \to L \psi U$, and these two actions remain completely independent.

2.2 Grade projection and the scalar part

For $A \in \mathrm{Cl}(1,7)$ write $\langle A \rangle_k$ for the projection onto grade $k$, and $\langle A \rangle_0$ for the scalar part. The bilinear form used throughout is $\langle A \tilde{B} \rangle_0$, with reversion defined in §2.3.

Lemma (cyclicity). For all $A, B \in \mathrm{Cl}(1,7)$,

$\langle A B \rangle_0 = \langle B A \rangle_0 .$

Proof. In the faithful realization $\mathrm{Cl}(1,7) \cong \mathbb{R}(16)$ the scalar part is proportional to the matrix trace, $\langle A \rangle_0 = \tfrac{1}{16} \operatorname{tr} A$, since every non-scalar basis blade is traceless. Cyclicity of the trace gives the result, and extends to any number of factors. Equivalently, and without leaving the algebra: for basis blades $e_I e_J$ the scalar part vanishes unless $e_J \propto e_I^{-1}$, and on that support the coefficient is symmetric under exchange of the two factors.

This lemma is what permits the cyclic rearrangement of the fermion bilinear in Paper 5; it is stated here so that no later argument relies on it implicitly.

2.3 Reversion, grade involution, and conjugation

Reversion, written $\tilde{A}$, reverses the order of the generators in each blade and therefore acts on grade $k$ by the sign $(-1)^{k(k-1)/2}$. Grade involution $\hat{A}$ acts by $(-1)^{k}$, and Clifford conjugation is their composition, acting by $(-1)^{k(k+1)/2}$:

grade $k$	0	1	2	3	4	5	6	7	8
reversion $\tilde{\phantom{A}}$	$+$	$+$	$-$	$-$	$+$	$+$	$-$	$-$	$+$
grade involution $\hat{\phantom{A}}$	$+$	$-$	$+$	$-$	$+$	$-$	$+$	$-$	$+$
Clifford conjugation	$+$	$-$	$-$	$+$	$+$	$-$	$-$	$+$	$+$

The reversion row is used repeatedly and is worth reading off explicitly. The reversion-even grades are $\{0, 1, 4, 5, 8\}$, spanning $136$ dimensions, and the reversion-odd grades are $\{2, 3, 6, 7\}$, spanning $120$. In particular bivectors are reversion-odd, which is what makes $e^{B}$ a rotor satisfying $L \tilde{L} = 1$ for $B$ a bivector; and grade $8$ is reversion-even, which is why $e^{\theta \omega}$ is not a rotor in this sense.

Reversion is always meant algebraically, as the grade-dependent sign above. Where a matrix realization is used for verification, reversion is implemented there by an intertwiner whose form depends on the realization chosen and carries no algebraic content.

2.4 The pseudoscalar

The pseudoscalar is

$\omega \equiv \gamma_0 \gamma_1 \gamma_2 \gamma_3 \gamma_4 \gamma_5 \gamma_6 \gamma_7 .$

For $n$ generators of which $q$ square to $-1$, $\omega^2 = (-1)^{n(n-1)/2} (-1)^{q}$; here $n = 8$ and $q = 7$, giving

$\omega^2 = (-1)^{28} (-1)^{7} = -1 .$

This sign is not a convention and is used in Paper 3, where it forces the anti-de Sitter sign of the vacuum.

Since $n$ is even, $\omega$ commutes with elements of even grade and anticommutes with those of odd grade; it is central in $\mathrm{Cl}(1,7)^{+}$ but not in the full algebra. It is reversion-even, $\tilde{\omega} = \omega$, so $e^{\theta \omega}$ does not satisfy $L \tilde{L} = 1$. Multiplication by $\omega$ maps grade $k$ to grade $8 - k$ and implements Hodge duality on the algebra.

3. Fiber and base

3.1 The split

The eight generators partition into two distinct sets. The first four, $\gamma_\mu$ where $\mu \in \{0, 1, 2, 3\}$, span the four-dimensional spacetime base and inherit the $(1,3)$ signature from $g_{MN}$:

$$\gamma_{\mu} \gamma_{\nu} + \gamma_{\nu} \gamma_{\mu} = 2 g_{\mu\nu}, \qquad \mu, \nu \in \{ 0, 1, 2, 3 \}$$

The remaining four generators, $\gamma_m$ for $m \in \{4, 5, 6, 7\}$, span the fiber and are strictly spacelike. To keep notation clean across this program, we use $\mu, \nu$ for base directions, $m, n$ for fiber directions, and $M, N$ for the complete eight-dimensional range. Because base and fiber generators anticommute via the defining relation in §2.1, this split simply divides the generators of a single, unified algebra rather than breaking it apart into a product of two independent systems.

This assignment is permanent; no convention allows us to swap or relabel the two sets. The physical distinction between base and fiber is rigidly locked by the soldering mechanism discussed in §3.3, and the structural integrity of the entire framework relies on this layout.

3.2 The fiber Hermitian metric

The fiber carries a complex structure that organizes its six real directions into three complex directions. If we denote the generators adapted to this structure as $\alpha_m$ and $\alpha^{\dagger}_m$, the Clifford relations become:

$$\begin{equation} \begin{aligned} \alpha_m \alpha_n + \alpha_n \alpha_m &= 0 \\ \alpha_m \alpha^{\dagger}_n + \alpha^{\dagger}_n \alpha_m &= h_{m\bar{n}}, \qquad m, n \in \{ 1, 2, 3 \} \end{aligned} \end{equation}$$

Here, $h_{m\bar{n}}$ serves as a Hermitian metric on the fiber. In a canonical frame, this metric reduces to $h_{m\bar{n}} = \delta_{m\bar{n}}$, recovering the standard fermionic oscillator algebra used in §4.

We retain this metric explicitly because it does not have to remain flat. A non-trivial $h_{m\bar{n}}$ deforms both the oscillator relations and the resulting projector construction in §5 (which we initially define in the canonical frame). To handle a non-flat fiber metric smoothly, our overarching convention is to execute computations in a local frame where the generators obey canonical relations, allowing any geometric variations of $h_{m\bar{n}}$ to manifest elsewhere in the machinery rather than warping the underlying algebra.

Additionally, the rotations of the $\alpha_m$ operators that preserve this complex structure form a $U(3)$ group, whose generators are derived in §7.

3.3 Soldering

The base and fiber link together via a soldering mechanism - a structural identification that maps a specific direction in the fiber directly to the base. This relationship ensures the partitioning in §3.1 is physically meaningful rather than an arbitrary choice. Crucially, this mechanism behaves as a field rather than a rigid background, meaning its value can vary from point to point.

This spatial variation yields three immediate structural consequences:

• Inherent connection terms: Keeping the frame aligned as the soldering fluctuates requires a local rotor, generating a non-removable connection term.
• Symmetry breaking: Choosing a specific soldering direction breaks the fiber's $U(3)$ symmetry down to a subgroup, isolating the broken directions as a coset.
• Geometric anisotropy: Because the soldering singles out one fiber direction over the others, it naturally induces an anisotropic geometry.

We do not develop these dynamics here. Paper 2 provides the full dynamic framework for the soldering field, constructing the rotor, the coset, and the resulting field content. For the scope of this paper, the mechanism serves strictly as the structural justification for maintaining the base/fiber split and fixing the assignments in §3.1.

The oscillator base

Wick rotated generators

Generator dictionary

Projectors and block decomposition

The ladder

The $u(3)$ basis

Charges

Normalization registry

Summary tables