Scalar Fields in Modern Physics

• Scalar fields are invariant quantities that serve as minimal carriers of physical information across various theoretical frameworks.
• They underpin models ranging from the Higgs mechanism in particle physics to perturbative and emergent formulations in cosmology and gravity.
• Scalars enable geometric and computational reformulations in lattice gauge theory and machine learning, unifying diverse physical and numerical approaches.

A scalar field is a field transforming trivially under Lorentz transformations, but current research uses the term more broadly for several distinct yet technically related objects: Higgs-sector degrees of freedom in particle physics, gauge-invariant bilinears and emergent links in lattice gauge theory, scalar perturbations in cosmology, scalar charges in modified gravity, structure scalars in gravitating fluids, and scalar invariants that parameterize equivariant maps or numerical wave solvers (Pedro, 2016, Wetterich, 2012, Uggla et al., 2011, Anderson et al., 2019, Sharif et al., 2013, Villar et al., 2021). The modern literature therefore treats “scalar” not as a single construction, but as a family of roles played by quantities that organize dynamics without carrying the tensorial transformation content of vectors or higher-rank fields.

1. Scalar sectors in relativistic field theory

In the Standard Model, the scalar sector contains one complex $SU(2)_L$ doublet $H$ of hypercharge $Y=1/2$, with

$$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$

For $\mu^2>0$ and $\lambda>0$, the minimum occurs at $\langle \Phi\rangle=(0,v/\sqrt2)^T$ with $v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$, implying $m_W=g v/2$, $m_Z=\sqrt{g^2+g'^2}\,v/2$, and fermion masses $H$0 after electroweak symmetry breaking (Pedro, 2016). Extended scalar sectors enlarge this structure by adding singlets, doublets, or triplets. The overview of beyond-the-Standard-Model constructions emphasizes real and complex singlet extensions, two-Higgs-doublet models, the Inert Doublet Model, and triplet sectors such as Type-II seesaw and Georgi–Machacek; motivations include dark matter, new CP violation, and modified electroweak phase transitions (Robens et al., 29 Jul 2025).

A geometric reformulation makes the scalar sector less dependent on field coordinates. The general two-derivative scalar-sector Lagrangian can be written as

$H$1

with observables determined by geometric invariants of the scalar manifold $H$2, such as $H$3, $H$4, and Killing vectors. In this language, the $H$5-matrix is invariant under nonsingular local field redefinitions, and the renormalizability of the Standard Model is tied to the flatness of $H$6, not to whether the Higgs fields are written linearly or non-linearly (Alonso et al., 2016).

This geometric perspective extends to non-perturbative state construction. In two-Higgs-doublet models, gauge-invariant asymptotic states are built from composite operators such as $H$7 and $H$8; expanding around the vacuum yields poles and masses coincident with those of the perturbative Higgs and electroweak gauge fields up to small higher-order corrections (Pedro, 2016). That correspondence makes scalar sectors simultaneously kinematic, geometric, and spectroscopic objects.

2. Scalars as emergent gauge-theory variables

Scalar lattice gauge theory removes fundamental lattice link variables and starts instead from site scalars $H$9 transforming under local $Y=1/2$0. Gauge-invariant mesons are

$Y=1/2$1

while composite link variables

$Y=1/2$2

transform exactly like ordinary lattice links. Plaquette invariants built from $Y=1/2$3 reproduce Wilson and Polyakov loop observables, and after a generalized Hubbard–Stratonovich transformation one obtains an equivalent link–scalar model. In the limit $Y=1/2$4 with $Y=1/2$5, the theory becomes standard lattice gauge theory with Wilson action, $Y=1/2$6, and $Y=1/2$7 in $Y=1/2$8; confinement and continuum observables coincide with those of the usual non-abelian gauge theory (Wetterich, 2012).

The same paper makes the emergent-gauge interpretation explicit in the continuum. Defining

$Y=1/2$9

the composite gauge field

$$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$0

transforms as a connection,

$$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$1

In this formulation, gauge bosons arise as collective excitations of scalars, and their masslessness is enforced by local gauge symmetry rather than by introducing fundamental gauge fields ab initio (Wetterich, 2012).

A conceptually distinct construction starts from Euclidean $$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$2 gauge fields and defines scalar variables directly as projections of the gauge vectors onto a gauge-derived orthonormal basis. The scalar matrix elements obey

$$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$3

and can be organized into charged and neutral combinations with emergent hypercharge assignments. In that framework, spontaneous symmetry breaking occurs in one of four scalar fields, the heavy-vector masses satisfy

$$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$4

and the spectrum contains nine massive Higgs particles—one neutral triplet at $$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$5 and three charged conjugate pairs at $$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$6—together with seven massless quasi-Goldstone scalars (Stingl, 2017).

A third nonstandard scalar formalism is scalar supersymmetry, where the anticommuting transformation parameter is a Lorentz scalar rather than a spinor. Implemented on inhomogeneous differential forms or Dirac–Kähler bi-spinors, it mixes bosonic and fermionic diforms, and the algebra closes on the Dirac–Kähler operator $$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$7 rather than on a pure translation. The basic free $$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$8 transformation law,

$$\mathcal L \supset |D_\mu H|^2 - V(H),\qquad V(H)=-\mu^2 H^\dagger H+\lambda (H^\dagger H)^2.$$9

realizes a supersymmetry that is related to, but not reducible to, standard supersymmetry (Wang et al., 2012).

3. Scalar perturbations in cosmology

Scalar cosmological perturbations admit several gauge-invariant formulations. Around FLRW, the 2011 review develops a dimensionless framework based on gauge invariants associated with the metric and stress-energy tensor, using the normalization $\mu^2>0$0. In this language, one obtains compact first-order systems for the density perturbation $\mu^2>0$1, velocity potential $\mu^2>0$2, and the gauge-invariant curvature variables $\mu^2>0$3 and $\mu^2>0$4, together with conservation laws such as

$\mu^2>0$5

so that $\mu^2>0$6 is conserved on large scales for adiabatic perturbations (Uggla et al., 2011).

At second order, scalar perturbations are sourced by quadratic combinations of first-order scalars. For a perfect fluid in conformal Newtonian gauge, the induced Bardeen potential obeys

$\mu^2>0$7

with $\mu^2>0$8 built from $\mu^2>0$9, $\lambda>0$0, and projected quadratic spatial derivatives. Exact analytic kernels are derived for radiation domination and matter domination, and the late-time envelopes scale as $\lambda>0$1, $\lambda>0$2, $\lambda>0$3 in RD, while in MD one finds $\lambda>0$4 for deep subhorizon modes (Inomata, 2020).

Anisotropic inflation provides a scalar spectrum with directional dependence. In the model with a scalar inflaton coupled to a $\lambda>0$5 gauge kinetic term, the curvature spectrum is parameterized by

$\lambda>0$6

and the explicit computation yields $\lambda>0$7. A value $\lambda>0$8 is obtained when the vector energy density is about $\lambda>0$9–$\langle \Phi\rangle=(0,v/\sqrt2)^T$0 of the scalar energy density during inflation, while the scalar–tensor correlation remains smaller than the tensor–tensor contribution in that regime (Gumrukcuoglu et al., 2010).

Large primordial scalar amplitudes also induce gravitational waves. With local non-Gaussianity

$\langle \Phi\rangle=(0,v/\sqrt2)^T$1

the scalar–tensor sector receives a new hybrid scalar–tensor contribution at $\langle \Phi\rangle=(0,v/\sqrt2)^T$2. In the monochromatic limit the Gaussian scalar–tensor term vanishes for $\langle \Phi\rangle=(0,v/\sqrt2)^T$3, whereas the hybrid scalar–tensor term vanishes only for $\langle \Phi\rangle=(0,v/\sqrt2)^T$4, producing a distinctive high-$\langle \Phi\rangle=(0,v/\sqrt2)^T$5 knee in the induced gravitational-wave spectrum (Picard et al., 2024).

4. Scalars in gravitation and modified gravity

In massless scalar–tensor gravity, compact objects acquire scalar charges that directly control deviations from general relativity. For neutron stars, the asymptotic scalar field is

$\langle \Phi\rangle=(0,v/\sqrt2)^T$6

and the effective scalar charge is

$\langle \Phi\rangle=(0,v/\sqrt2)^T$7

Two additional charges,

$\langle \Phi\rangle=(0,v/\sqrt2)^T$8

enter binary-pulsar timing and slow-rotation observables. For sufficiently negative coupling, spontaneous scalarization appears at approximately $\langle \Phi\rangle=(0,v/\sqrt2)^T$9, and in the non-scalarizing regime one finds the scaling relations

$v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$0

which permit rapid interpolation across theory space and neutron-star equations of state (Anderson et al., 2019).

Scalarization can also be dynamical in strong-field mergers. In Einstein–Maxwell–Scalar theory with

$v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$1

the scalar equation linearized around $v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$2 contains the effective mass

$v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$3

For electrically dominated configurations, $v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$4, so sufficiently large positive $v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$5 gives $v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$6 and triggers scalar hair. Numerical-relativity simulations of charged head-on binaries show that strong coupling and nonzero remnant charge lead to a scalarized final black hole, whereas weak coupling or charge cancellation yields dynamical descalarization (Díaz et al., 19 Jun 2026).

A different gravitational use of “scalar” is purely diagnostic. In charged plane symmetry, orthogonal splitting of the Riemann tensor produces four structure scalars,

$v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$7

with

$v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$8

$v=\sqrt{\mu^2/\lambda}=246\ \mathrm{GeV}$9

Here $m_W=g v/2$0 sources the Raychaudhuri equation, $m_W=g v/2$1 governs shear evolution, and $m_W=g v/2$2 acts as the inhomogeneity factor (Sharif et al., 2013).

Modified-gravity perturbation theory introduces yet another scalar notion: propagating scalar modes. In general $m_W=g v/2$3 models on FLRW, there are generically two dynamical scalar perturbation fields. One branch is luminal, while the other obeys a quartic dispersion relation,

$m_W=g v/2$4

If the coefficient is positive, the group velocity grows linearly with $m_W=g v/2$5 and becomes superluminal at short scales; if negative, the mode is violently unstable (0907.5378).

5. Geometry, symmetry, and scalarized reformulations

The geometry of scalar sectors can itself be promoted to the central dynamical principle. In the Standard Model written in non-linear Callan–Coleman–Wess–Zumino variables, the scalar manifold remains flat; equivalently $m_W=g v/2$6. By contrast, SMEFT and HEFT describe curved scalar manifolds, and the HEFT Lagrangian can be rewritten in SMEFT form if and only if $m_W=g v/2$7 possesses an $m_W=g v/2$8-invariant fixed point. Observable deviations in longitudinal gauge-boson scattering are then controlled by sectional curvatures of $m_W=g v/2$9 (Alonso et al., 2016).

A separate classification problem considers single-field $m_Z=\sqrt{g^2+g'^2}\,v/2$0 theories with enlarged classical symmetries. To leading order in derivatives, there are exactly three flat-space families with continuous symmetries beyond Poincaré invariance: Dirac–Born–Infeld, Cuscuton, and Scaling theories. The general scaling-invariant family takes the form

$m_Z=\sqrt{g^2+g'^2}\,v/2$1

while the Cuscuton action is

$m_Z=\sqrt{g^2+g'^2}\,v/2$2

and enjoys the symmetry $m_Z=\sqrt{g^2+g'^2}\,v/2$3 on any background admitting a Killing vector $m_Z=\sqrt{g^2+g'^2}\,v/2$4 (Grall et al., 2019).

Lorentzian Cofinsler constructions push this further by generating the spacetime metric from a scalar on the cotangent bundle and the differential of a scalar field on the base manifold. With

$m_Z=\sqrt{g^2+g'^2}\,v/2$5

one can choose $m_Z=\sqrt{g^2+g'^2}\,v/2$6 so that the induced metric is FLRW with

$m_Z=\sqrt{g^2+g'^2}\,v/2$7

Evaluating Horndeski Lagrangians on this geometry produces “hidden scalar” Lagrangians in $m_Z=\sqrt{g^2+g'^2}\,v/2$8, with the quartic and quintic terms reducible to second-order form up to total derivatives (Horndeski, 2019). A related construction based on a “true” scalar–scalar Lagrangian gives

$m_Z=\sqrt{g^2+g'^2}\,v/2$9

with vacuum solutions

$H$00

These solutions are used to build self-inflating universes, explosive beginnings with $H$01 and $H$02 for $H$03, and multiverse constructions with non-Hausdorff topology (Horndeski, 2022).

6. Scalars as invariants and computational primitives

In equivariant machine learning, scalar quantities are treated as universal coordinates for symmetry-respecting maps. The first fundamental theorem for $H$04 implies that an invariant scalar function of vectors $H$05 can be written as $H$06, with $H$07. An $H$08-equivariant vector polynomial admits the form

$H$09

where each $H$10 is an invariant scalar polynomial. For general inputs containing scalars, vectors, and tensors, the paper proposes the scalar-universal parameterization

$H$11

with scalar invariants $H$12 and canonical equivariant basis tensors $H$13. This construction extends to Euclidean, Lorentz, Poincaré, and permutation symmetries (Villar et al., 2021).

“SCALAR” is also the name of an adaptive-mesh Schrödinger solver for ultralight-axion or fuzzy-dark-matter simulations. Built within RAMSES, it evolves the Schrödinger–Poisson system on AMR grids using a Lie–Trotter split integrator, a third-order Taylor treatment of the kinetic operator, a conservative continuity-equation corrector, high-order prolongation for complex wavefunctions, and optional artificial viscosity near coarse–fine boundaries. In the test suite, the continuity corrector improves mass conservation from $H$14 to $H$15, and gravity-enabled runs reach about $H$16 cell updates per second on the reported CPU configuration (Mina et al., 2019).

This computational use is closely aligned with the invariant-theoretic one. In both settings, scalar objects are the minimal carriers of information needed to reconstruct or evolve more complicated structures: invariant coefficients modulate equivariant tensor bases in one case, and complex scalar wavefunctions generate density, phase, velocity, and gravitational potential in the other (Villar et al., 2021, Mina et al., 2019).

Taken together, these literatures show that scalar variables are not confined to a single ontological status. They can be fundamental fields, collective excitations, gauge-invariant observables, perturbative modes, symmetry generators, geometric data on field space, or computational state variables. The persistent theme is that scalar objects often provide the lowest-complexity quantities from which nontrivial dynamics, symmetry realization, and observable structure can be reconstructed.