Canonical Quantization of the Biquaternion Rarita–Schwinger Field

Introduction

The Rarita–Schwinger field is the field of spin $\tfrac32$: a vector-spinor $\psi_\mu$, carrying one Lorentz vector index and one Dirac spinor index, subject to the constraint that its $\gamma$-trace vanish. As a classical field equation the Rarita–Schwinger equation belongs to the companion subcategory on the relativistic quantum theory of higher spins; the present article takes that equation as given and addresses the canonical quantization of the field in the biquaternion framework: the constraint structure, the mode expansion, the anticommutators, the count of physical polarizations, and the gauge and ghost structure of the massless case.

The biquaternion framework is well suited to this field, because the two indices of $\psi_\mu$ correspond to the two kinds of object the framework distinguishes. The vector index carries the material vector representation, which the article on integer-spin quantization identifies with the adjoint action on the vector part $V\subset\mathbb{M}_-$; the spinor index carries the left-ideal representation that the Dirac article identifies as the spin-$\tfrac12$ carrier. The Rarita–Schwinger field is therefore the tensor product of the two, and its decomposition is the framework's own:

$$ \left(\tfrac12\right)\otimes 1 = \left(\tfrac12\right)\oplus\left(\tfrac32\right), \qquad 2\cdot 3 = 2+4 , $$

so that the tensor product of the spinor with the material vector contains a spin-$\tfrac12$ piece and a spin-$\tfrac32$ piece, and the $\gamma$-trace constraint removes the former. The same pattern appears for the graviton, where the symmetric trace-free square of the vector decomposes as $0\oplus2$; here the antisymmetric structure of the constraint selects the highest spin. The two articles together are the framework's statement of how higher-spin fields are built: tensor a lower-spin carrier with the material vector and project onto the top component.

The article is organized as follows. The field, its constraint, and the consequence that transversality follows from the constraint and the field equation are stated, and the projector onto the $\gamma$-traceless part is constructed and verified. The gauge structure of the massless case and the second-class structure of the massive case are contrasted. The canonical quantization is then carried out: momenta, Dirac brackets, anticommutators, the mode expansion, and the fermionic Fock space. The propagator and its gauge fixing are recorded, the biquaternion reading of the decomposition is developed, and the article closes with the accounting of what the algebra supplies and what is imported.

Companion article The Dirac Equation in Biquaternionic Form, for the spinor carrier as a left ideal and the Clifford structure.
Companion article Canonical Quantization of the Biquaternion Maxwell Field, for the massless gauge field whose gauge structure the massless Rarita–Schwinger case generalizes.
Companion article The Spin–Statistics Theorem in Biquaternionic Form, for the fermionic quantization of half-integer-spin fields.

Conventions. The algebra is $\mathbb{B}=\mathbb{C}\otimes_\mathbb{R}\mathbb{H}$ with basis $e_0=1,e_1,e_2,e_3$, $e_k^2=-e_0$, $e_je_k=-\delta_{jk}e_0+\varepsilon_{jkl}e_l$, central $i$, material sector $\mathbb{M}_-$ and informational sector $\mathbb{M}_+$. The vector part of the material sector is $V=\mathrm{span}_\mathbb{R}\{e_1,e_2,e_3\}$, on which the adjoint action of the rotation algebra acts as the spin-one representation. The spinor carrier is the left ideal of the Dirac article, with the Clifford algebra $\mathrm{Cl}_{1,3}\cong M_2(\mathbb{H})$ and the Clifford metric $g=\mathrm{diag}(+1,-1,-1,-1)$ defined by $\{\gamma^\mu,\gamma^\nu\}=2g^{\mu\nu}I_4$, so that $(\gamma^0)^2=+I_4$ and $(\gamma^k)^2=-I_4$; the $ict$ coordinate metric remains $\eta=\mathrm{diag}(-1,+1,+1,+1)$, and $g=-\eta$. The vector-spinor is $\psi_\mu$, with $\mu$ the material vector index and the spinor index suppressed; the $\gamma$-trace is $\gamma^\mu\psi_\mu$. The mass is $m$, natural units are used, and the metric-dependent contractions follow the Clifford convention $g$ while the coordinate contraction of the field with $dx^\mu$ follows the $ict$ convention $\eta$.

The Field, Its Constraint, and Transversality

The Equations

The Rarita–Schwinger field satisfies the field equation

$$ \left(i\partial\!\!\!/ - m\right)\psi_\mu = 0 , $$

together with the algebraic irreducibility constraint

$$ \gamma^\mu\psi_\mu = 0 . $$

The constraint is not an equation of motion; it removes the spin-$\tfrac12$ components that a bare vector-spinor would contain, and it is what makes the field irreducible. The fully antisymmetric form of the Rarita–Schwinger equation, $(i\gamma^{\mu\nu\rho}\partial_\rho-m\gamma^{\mu\nu})\psi_\nu=0$, whose variation gives the field equation and the constraint for $m\neq0$, is the standard Lagrangian origin of the two equations and is cited below; for the purposes of quantization the two equations above are the starting point.

Transversality from the Constraint

Contracting the field equation with $\gamma^\mu$ and using the Clifford identity

$$ \gamma^\mu \not p = 2p^\mu - \not p\gamma^\mu , $$

which is the antisymmetrized form of $\{\gamma^\mu,\gamma^\nu\}=2g^{\mu\nu}I_4$, gives, in momentum space,

$$ \gamma^\mu\left(\not p-m\right)\psi_\mu = 2p^\mu\psi_\mu - \not p\gamma^\mu\psi_\mu - m\gamma^\mu\psi_\mu = 2p^\mu\psi_\mu = 0 , $$

where the constraint has been used in the last two terms. Hence the constraint and the field equation together imply the transversality condition

$$ p^\mu\psi_\mu = 0 , $$

which is not independent but follows from the two. Both steps were verified with explicit gamma matrices in the Clifford convention: the Clifford algebra to machine precision, the identity $\gamma^\mu \not p=2p^\mu-\not p\gamma^\mu$ to machine precision at random momenta, and the resulting contraction to zero.

The Projector onto the Traceless Part

The constraint is imposed by the projector

$$ P^{\mu\nu} = g^{\mu\nu}I_4-\tfrac14\gamma^\mu\gamma^\nu , $$

which annihilates the $\gamma$-trace and is idempotent,

$$ P^{\mu\nu}\gamma_\nu = 0 , \qquad \gamma_\mu P^{\mu\nu} = 0 , \qquad P^{\mu}{}_\alpha P^{\alpha\nu} = P^{\mu\nu} , $$

so that $\psi_\mu$ is replaced by $P_\mu{}^\nu\psi_\nu$, whose $\gamma$-trace vanishes. This was verified with explicit $4\times4$ gamma matrices in the Clifford convention $g=\mathrm{diag}(+1,-1,-1,-1)$: the idempotency and both transversality conditions held to machine precision. The rank of the $16\times16$ matrix $P^{(\mu i)(\nu j)}$ — the projector acting on the $4\times4$ components of the vector-spinor — is

$$ \mathrm{rank}\,P = 12 = 16-4 , $$

verified by Gaussian elimination over the complex numbers, so that the $\gamma$-trace constraint removes exactly a four-component spinor from the sixteen components of the vector-spinor.

Gauge Structure: Massless and Massive

The Massless Case

When $m=0$ the Rarita–Schwinger field has a gauge symmetry,

$$ \psi_\mu \;\longmapsto\; \psi_\mu+\partial_\mu\epsilon , $$

with $\epsilon$ an arbitrary spinor parameter. The transformation changes $\psi_\mu$ by a gradient of a spinor, and it is the spin-$\tfrac32$ analogue of the gauge transformation of the vector field; the constraint $\gamma^\mu\psi_\mu=0$ is preserved by a gauge transformation whose parameter satisfies the Dirac equation, and the field equation is preserved exactly. The gauge symmetry is a first-class structure: the constraint $\gamma^\mu\psi_\mu=0$ and its conjugate are the generators, and the physical content is the quotient of the constraint surface by the gauge orbits. Taking the $\gamma$-trace gauge $\gamma^\mu\psi_\mu=0$ is the analogue of the Lorentz gauge, and it leaves a residual gauge freedom whose quantization requires Faddeev–Popov ghosts — Grassmann-odd spinor ghosts, in the Dirac representation of the gauge parameter. The physical states are the two helicities $\pm\tfrac32$; the helicities $\pm\tfrac12$ that the massless field would otherwise carry are removed by the gauge structure, exactly as the longitudinal and timelike states of the massless vector are removed.

The Count of Physical States

The component counting makes the role of the constraints explicit. A vector-spinor has sixteen complex components. The $\gamma$-trace constraint removes four, leaving twelve; the transversality condition $p^\mu\psi_\mu=0$ removes a further four on shell, leaving eight; and the field equation projects the remaining spatial spinors onto positive frequency, halving the count. The result is the four polarizations of the massive spin-$\tfrac32$ multiplet — the helicities $\pm\tfrac32$ and $\pm\tfrac12$ — for the particle, with the same four for the antiparticle of the complex field. The two helicities of largest magnitude are the genuinely new content of spin $\tfrac32$; the two helicities $\pm\tfrac12$ are the members that the two-index structure would also allow a spin-$\tfrac12$ field to have, and it is precisely those that the constraint is designed to remove. In the massless case the gauge symmetry removes them again, leaving the two states $\pm\tfrac32$, so that the massless field has just two physical polarizations. The count was confirmed componentwise: the projector's rank of twelve is the statement that the $\gamma$-trace removes four of the sixteen components, and the remaining steps are the standard on-shell reduction.

The Massive Case

When $m\neq0$ the gauge symmetry is absent: the field equation is not invariant under $\psi_\mu\mapsto\psi_\mu+\partial_\mu\epsilon$, because the mass term is not. The constraints are then second class: $\gamma^\mu\psi_\mu=0$ and its conjugate do not commute weakly, and the canonical quantization must use the Dirac bracket rather than the Poisson bracket. The count of physical polarizations is the standard $2s+1=4$ for $s=\tfrac32$ — the helicities $\pm\tfrac32$ and $\pm\tfrac12$ — and no ghosts appear, since there is no gauge redundancy to fix. This is the same contrast as in the vector case: the massless field has a gauge redundancy and two physical states, the massive field has no redundancy and three (or, for spin $\tfrac32$, four) states, and the extra states are the ones the gauge would have removed.

Canonical Quantization

Momenta, Constraints, and Dirac Brackets

The canonical momentum conjugate to $\psi_\mu$ is, from the standard Rarita–Schwinger Lagrangian,

$$ \pi^\mu = \frac{\partial\mathcal{L}}{\partial(\partial_0\psi_\mu)} = i\bar\psi_\nu\gamma^{\nu 0\mu} = i\psi_\nu^\dagger\gamma^0\gamma^{\nu 0\mu} , $$

with $\gamma^{\nu 0\mu}$ the antisymmetrized product of the three gamma matrices, and the canonical analysis gives the primary constraint $\pi^0\approx 0$ — automatic here, since $\gamma^{\nu 0 0}=0$ — together with the $\gamma$-trace constraint and the conjugate constraints. The constraint algebra is second class for $m\neq0$, so the equal-time bracket is the Dirac bracket

$$ \{A,B\}_D = \{A,B\}_P-\{A,\chi_a\}_P\,C^{ab}\,\{\chi_b,B\}_P , \qquad C_{ab}=\{\chi_a,\chi_b\}_P , $$

with $\chi_a$ the second-class constraints. The effect of the Dirac bracket is to replace the naive anticommutator by the projected one: the fields are restricted to the constraint surface, and the algebra closes on the projected field $P_\mu{}^\nu\psi_\nu$ rather than on $\psi_\mu$ itself. This is the standard treatment and is cited below; the point to record is that the projection is the projector of the preceding section, so the algebraic content of the constraints is the biquaternion projector and its rank.

Anticommutators and Modes

The quantized field is fermionic: the equal-time anticommutator of the field and its momentum is the projection of the naive anticommutator onto the constraint surface,

$$ \{\psi_\mu(t,\mathbf{x}),\,\pi^\nu(t,\mathbf{y})\} = i\,P_\mu{}^\nu\,\delta^{(3)}(\mathbf{x}-\mathbf{y}) , $$

and the mode expansion carries four polarizations,

$$ \psi_\mu(x) = \int\frac{d^3p}{(2\pi)^3}\frac{1}{\sqrt{2E_p}} \sum_{r=1}^{4}\left[ a_r(\mathbf{p})\,u_{\mu,r}(p)\,e^{-ip\cdot x} + b_r^\dagger(\mathbf{p})\,v_{\mu,r}(p)\,e^{+ip\cdot x} \right] , $$

with $E_p=\sqrt{\mathbf{p}^2+m^2}$, and with the Rarita–Schwinger polarization spinors satisfying

$$ \left(\not p-m\right)u_{\mu,r}=0 , \qquad \gamma^\mu u_{\mu,r}=0 , \qquad p^\mu u_{\mu,r}=0 , \qquad \sum_{r=1}^{4}u_{\mu,r}\bar{u}_{\nu,r} = -\left(\not p+m\right)\Pi_{\mu\nu} , $$

the last being the completeness relation of the four polarizations. The object $\Pi_{\mu\nu}$ is the transverse $\gamma$-traceless projector, the on-shell projector onto the intersection of the two constraints, $\{\gamma^\mu\psi_\mu=0,\ p^\mu\psi_\mu=0\}$, of rank eight on the vector-spinor space and four on the positive-energy solutions. It is not the constraint projector $P_{\mu\nu}$ of the preceding section: transversality requires the additional momentum-dependent terms, and $p^\mu P_{\mu\nu}\neq0$, so the completeness relation cannot be carried by $P_{\mu\nu}$ alone. It is the standard Rarita–Schwinger projector and is cited below. The creation and annihilation operators satisfy the fermionic algebra

$$ \{a_r(\mathbf{p}),\,a_s^\dagger(\mathbf{q})\} = \{b_r(\mathbf{p}),\,b_s^\dagger(\mathbf{q})\} = (2\pi)^3\,\delta_{rs}\,\delta^{(3)}(\mathbf{p}-\mathbf{q}) , $$

with all other anticommutators vanishing, and the Fock space is built by acting on the vacuum with $a_r^\dagger$ and $b_r^\dagger$. The four polarizations $r=1,\dots,4$ are the two helicities $\pm\tfrac32$ and the two helicities $\pm\tfrac12$.

The biquaternion reading of the mode expansion is the same as the Dirac article's, one vector index richer: each polarization spinor is an element of the left ideal, labelled by the internal direction $r$ that the adjoint action supplies; the four internal labels are the four vectors of the spin-$\tfrac32$ representation, that is, the (traceless) product of the spin-$\tfrac12$ label with the material triplet.

Fermionic Statistics

The field is quantized with anticommutators, not commutators, and the reason is the spin-statistics relation: the field carries half-integer spin, so its excitations are fermions. The framework's account of the relation, and the argument that only half-integer spins can be quantized with anticommutators while preserving positivity and microcausality, is the subject of the companion article on the spin–statistics theorem; here it is imported and used. The vector index does not change the statistics: a vector-spinor is still a spinor, and the material vector label is an internal label like any other, so the field is fermionic despite the integer index.

The Propagator and Its Gauge Fixing

The momentum-space propagator is the inverse of the quadratic form that the field equation and the constraint define on the $\gamma$-traceless subspace. It is of the general form

$$ G_{\mu\nu}(p) = -\frac{i}{p^2-m^2}\left[\left(\not p+m\right)\Pi_{\mu\nu} + \text{(lower-spin terms)}\right] , $$

where the transverse projector $\Pi_{\mu\nu}$ is the one of the completeness relation and the lower-spin terms are the parts that the constraint removes; in the standard covariant form the propagator carries the gauge-fixing dependence of the massless case and a $p_\mu p_\nu$ term whose coefficient involves $1/m^2$ for the massive case. The detailed form, including the dependence on the gauge-fixing parameter and the $D$-dimensional continuation, is the standard result and is cited below; the framework's contribution is the identification of the $\gamma$-trace part of the projector with the algebra's $\gamma$-traceless projector and the interpretation of the lower-spin terms as the removed $\left(\tfrac12\right)$ component.

The massless case requires the gauge fixing, and the gauge-fixing parameter enters the propagator in the same way as in the vector case, through the longitudinal-traceless decomposition of the field. The ghost fields are spinor and anticommuting, so they live in the Grassmann envelope of the algebra tensored with the spinor carrier; the BRST structure is the standard one, with the gauge parameter now a spinor.

The Biquaternion Reading

The algebraic content of this article is the tensor-product decomposition and the constraint that selects the top component, and it is worth stating in the framework's terms.

The carrier is the material four-vector index tensored with the left-ideal spinor carrier $S$ of the Dirac article, sixteen complex components before the constraint. Its spatial part is the material triplet $V$, the adjoint action's carrier, and the triplet tensored with the spinor decomposes under the rotation algebra as

$$ 1\otimes\tfrac12 = \tfrac12\oplus\tfrac32 , \qquad 3\cdot 2 = 4+2 , $$

the spin-$\tfrac32$ piece of dimension four and a spin-$\tfrac12$ piece of dimension two. The $\gamma$-trace constraint removes the spin-$\tfrac12$ piece, and the projector that does it is $P^{\mu\nu}=g^{\mu\nu}-\tfrac14\gamma^\mu\gamma^\nu$, whose rank on the sixteen components of the vector-spinor is twelve — the four removed components being exactly the spin-$\tfrac12$ trace. This was verified by the explicit rank computation. The massive multiplet is the spin-$\tfrac32$ piece, with its four polarizations; the massless field, when the gauge symmetry is present, keeps the two states of helicity $\pm\tfrac32$ and the gauge removes the $\pm\tfrac12$ members of the same multiplet.

The parallel with the integer-spin case is exact. There, the carrier is a tensor power of the material vector and the trace removal selects the top component: for the graviton, the symmetric trace-free square, $1\otimes1$ symmetric and traceless $=2$ with dimension five, verified by the transverse-traceless projection. Here, the carrier is the material vector tensored with the spinor, and the $\gamma$-trace removes the lower component: $1\otimes\tfrac12=\tfrac32\oplus\tfrac12$, with the $\tfrac32$ of dimension four retained. In both cases the framework's statement is the same: the higher-spin field is obtained by tensoring the material vector with a lower-spin carrier and projecting onto the top component with a projector built from the algebra's own invariant form — the metric pairing for the integer-spin case, the $\gamma$-trace for the half-integer one. The difference is that the spinor carrier is a left ideal rather than a tensor power of the vector, and that is why the half-integer case requires the Clifford structure while the integer case does not.

Consistency of the Coupled Theory

A spin-$\tfrac32$ field is not consistent as a fundamental field in an arbitrary background, and the reason is visible in the constraint structure developed above. Transversality followed from the field equation and the $\gamma$-trace constraint through the Clifford identity, and that derivation used the free field equation. Coupling the field to a background — an electromagnetic field or the gravitational field — changes the field equation, and the change modifies the contraction by terms involving the background field strength or curvature. The constraint then ceases to be preserved by the evolution unless the background satisfies its own equations of motion: for the gravitational coupling the consistency condition is that the background be a solution of the Einstein equations, which is the Deser–Kay–Stelle condition of supergravity, and for the electromagnetic coupling the analogous condition (the Velo–Zwanziger analysis) restricts the background field strength and, when it fails, admits modes propagating faster than light. In the framework's terms the statement is that the projector $P^{\mu\nu}$ commutes with the free dynamics but not with the coupled dynamics, so the removal of the $\left(\tfrac12\right)$ component is maintained only when the background is on shell. This is why the spin-$\tfrac32$ field is treated either as the gauge field of local supersymmetry, whose consistency is tied to the background geometry, or as an effective field with a cutoff; the framework does not change the conclusion, and it exhibits the mechanism through its projector.

The Pattern Across the Spins

The three quantized fields of this subcategory — the vector, the Rarita–Schwinger field, and the graviton — share one structure, and the sharing is the framework's main statement about higher spins.

In each case the carrier is a tensor product of a lower-spin object with the material vector index — the triplet $V$ for the integer cases, the material four-vector for the Rarita–Schwinger field, whose spatial part is the same triplet — on which the adjoint action generates the spin-one rotation: the vector field is $V$ itself, with dimension three; the Rarita–Schwinger field is the material four-vector tensored with the spinor $S$, of dimension sixteen, reduced to twelve by the $\gamma$-trace constraint; and the graviton is the symmetric trace-free square of $V$, with dimension five. In each case a projector built from the algebra's invariant form selects the top-spin component: the metric pairing for the integer cases, the $\gamma$-trace for the half-integer one, and the traceless-and-symmetric projection for the graviton. And in each case the physical count is the standard one: $2s+1$ states for the massive field, and two states of helicity $\pm s$ for the massless field.

$s$	$\text{carrier}$	$\dim$	$\text{massive} \to \text{massless}$
$1$	$V$	$3$	$3 \to 2$
$\tfrac32$	$(\text{material four-vector})\otimes S$	$12\ (4\ \text{on shell})$	$4 \to 2$
$2$	$(V\odot V)_{\text{traceless}}$	$5$	$5 \to 2$

The massless limit takes the count to two in every case, because the gauge symmetry removes the helicities below $s$; the framework sees this as the removal of the lower-spin components that the tensor product contains, and the projector that performs the removal is the same one that defines the field. The pattern is not a derivation of the fields — their existence and their couplings are inputs — but it is a uniform account of why the higher-spin fields have the carriers, the constraints, and the counts that they do.

What the Algebra Supplies and What It Imports

Supplied by the algebra, and recomputed here. The vector index as the material triplet on which the adjoint action acts; the spinor index as the left ideal of the Dirac article; the tensor-product decomposition $1\otimes\tfrac12=\tfrac32\oplus\tfrac12$ with $3\cdot2=4+2$; the $\gamma$-traceless projector $P^{\mu\nu}=g^{\mu\nu}-\tfrac14\gamma^\mu\gamma^\nu$, its idempotency, its transversality, and its rank twelve on the sixteen components, all verified with explicit gamma matrices in the Clifford convention; the transversality $p^\mu\psi_\mu=0$ as a consequence of the constraint and the field equation, verified through the identity $\gamma^\mu \not p=2p^\mu-\not p\gamma^\mu$; and the completeness relation of the four polarization spinors, whose transverse projector is built from the same Clifford structure.

Imported, and left visible. The Rarita–Schwinger equation and its antisymmetric Lagrangian origin, which belong to the classical theory of higher spins; the canonical momentum and the constraint algebra; the Dirac bracket and the quantization of second-class systems; the fermionic mode algebra and the Fock space; the spin–statistics relation, cited to the companion article on the theorem; the propagator, its gauge-fixing dependence, and its dimensional continuation; the Faddeev–Popov ghosts of the massless case and the Grassmann envelope; and the BRST structure, cited to the standard literature.

Not supplied. The value of the mass; the choice of gauge; the odd coordinates; the physical identification of the spin-$\tfrac32$ field with any observed particle; and any empirical content. The framework organizes the field and its constraints, and it does so by exhibiting them as a tensor product and a projection; it does not derive the field's existence.

Open Questions

The projector from the algebra alone. The $\gamma$-traceless projector is built from the Clifford metric. Is there a biquaternion expression of the projection — an operation on the tensor product $V\otimes S$ using only the algebra's structure — that does not pass through the gamma matrices, in the way the adjoint action expresses the spin-one projection without them?
The massless ghosts and the sector split. The ghosts of the massless Rarita–Schwinger field are spinor-valued and anticommuting. In which sector do they live, and does the sector split organize the ghost spectrum of higher-spin gauge fields in a way that differs from the vector case?
The supergravity connection. The spin-$\tfrac32$ field is the gravitino of supergravity, where it is the gauge field of local supersymmetry and the graviton is its partner. Does the biquaternion framework relate the gravitino's vector index to the material sector and its spinor index to the ideal in a way that organizes the supermultiplet, or is the pairing invisible to the framework?
Consistency and the supertrace. The consistency of a higher-spin field coupled to gravity is governed by the supertrace of the theory — the difference of the numbers of bosonic and fermionic degrees of freedom. Is there a biquaternionic accounting of the supertrace for the spin-$\tfrac32$ field?
The vZw and Fierz–Pauli structure. The massless Rarita–Schwinger field has the van Dam–Veltman–Zakharov discontinuity in its massive limit, as the graviton does. Is the discontinuity visible in the framework's decomposition, as a change in the dimension of the projected carrier?
Empirical content. As everywhere, whether any of this yields a prediction distinguishing the framework from standard higher-spin theory. The transcription given here does not.

Summary

The Rarita–Schwinger field is a vector-spinor $\psi_\mu$ satisfying $(i\partial\!\!\!/-m)\psi_\mu=0$ and the irreducibility constraint $\gamma^\mu\psi_\mu=0$. Contracting the field equation with $\gamma^\mu$ and using $\gamma^\mu \not p=2p^\mu-\not p\gamma^\mu$ gives the transversality condition $p^\mu\psi_\mu=0$, so the constraint and the field equation together imply transversality; this was verified with explicit gamma matrices. The constraint is imposed by the projector $P^{\mu\nu}=g^{\mu\nu}-\tfrac14\gamma^\mu\gamma^\nu$, which is idempotent and annihilates the $\gamma$-trace, and whose rank on the sixteen components of the vector-spinor is twelve, so that the constraint removes exactly a four-component spinor — all verified to machine precision.

The gauge structure differs between the two mass regimes. For $m=0$ the field has the gauge symmetry $\psi_\mu\mapsto\psi_\mu+\partial_\mu\epsilon$ with a spinor parameter, the constraints are first class, the $\gamma$-trace gauge fixing requires Faddeev–Popov spinor ghosts, and the physical states are the two helicities $\pm\tfrac32$. For $m\neq0$ the gauge symmetry is absent, the constraints are second class, the quantization uses the Dirac bracket, and the physical states are the four helicities $\pm\tfrac32,\pm\tfrac12$. The canonical quantization gives a fermionic mode expansion with four polarizations, anticommutators $\{a_r,a_s^\dagger\}=(2\pi)^3\delta_{rs}\delta^{(3)}$, and a completeness relation whose transverse projector is built from the same Clifford structure.

In the biquaternion framework the field is the tensor product of the material vector part $V$ — the adjoint-action triplet — with the left-ideal spinor carrier $S$, and the tensor product decomposes as $1\otimes\tfrac12=\tfrac32\oplus\tfrac12$ with $3\cdot2=4+2$. The $\gamma$-trace constraint removes the $\left(\tfrac12\right)$ component, and the surviving spin-$\tfrac32$ is the field. The structure is the exact analogue, one tensor factor down, of the graviton's symmetric trace-free square of the material vector: in each case a lower-spin carrier is tensored with the material vector and projected onto the top component by a projector built from the algebra's invariant form. The mass, the gauge choice, the propagator's detailed form, the odd coordinates, and any empirical content are imported; the decomposition and the projection are the framework's.

Summary of Notation

Symbol	Meaning
$\psi_\mu$	Rarita–Schwinger vector-spinor; material vector index $\mu$, spinor index suppressed
$(i\partial\!\!\!/-m)\psi_\mu=0$	Rarita–Schwinger field equation
$\gamma^\mu\psi_\mu=0$	Irreducibility constraint; removes the $\left(\tfrac12\right)$ component
$p^\mu\psi_\mu=0$	Transversality; follows from the constraint and the field equation
$\gamma^\mu \not p=2p^\mu-\not p\gamma^\mu$	Clifford identity used (verified)
$P^{\mu\nu}=g^{\mu\nu}I_4-\tfrac14\gamma^\mu\gamma^\nu$	$\gamma$-traceless projector; $P\gamma=0$, $\gamma P=0$, $P^2=P$ (verified)
$\mathrm{rank}\,P=12=16-4$	Verified by Gaussian elimination over $\mathbb{C}$
$V\otimes S$, $1\otimes\tfrac12=\tfrac32\oplus\tfrac12$	Triplet part of the carrier and its decomposition; $3\cdot2=4+2$
$\psi_\mu\mapsto\psi_\mu+\partial_\mu\epsilon$	Massless gauge symmetry; spinor parameter
$\pi^\mu$, $\{A,B\}_D$	Canonical momentum and Dirac bracket (second-class constraints, $m\neq0$)
$\{a_r(\mathbf{p}),a_s^\dagger(\mathbf{q})\}=(2\pi)^3\delta_{rs}\delta^{(3)}(\mathbf{p}-\mathbf{q})$	Fermionic mode algebra
$u_{\mu,r},v_{\mu,r}$, $r=1,\dots,4$	Polarization spinors; helicities $\pm\tfrac32,\pm\tfrac12$
$\Pi_{\mu\nu}$	Transverse $\gamma$-traceless (Rarita–Schwinger) projector; rank $8$, and $4$ on positive-energy solutions
$\sum_r u_{\mu,r}\bar{u}_{\nu,r}=-(\not p+m)\Pi_{\mu\nu}$	Completeness relation, carried by the transverse projector
$G_{\mu\nu}(p)$	Propagator; inverse of the quadratic form on the traceless subspace
$g=\mathrm{diag}(+1,-1,-1,-1)$, $\{\gamma^\mu,\gamma^\nu\}=2g^{\mu\nu}I_4$	Clifford metric (level 3); the $ict$ metric is $\eta=-g$