D-theory: Dual Facets in Gravity & Algebra

• D-theory is a multifaceted concept encompassing modified gravity, operator algebras, and string theoretical dualities.
• In modified gravity, it introduces curvature-dependent disformal metrics that map C-theory to nonlocal quadratic actions.
• In operator algebra and string theory, it formulates discrete analogues of E-theory and organizes dual descriptions from D-branes and D-particles.

“D-theory” does not denote a single universally fixed construction across theoretical physics and mathematics. In the literature represented here, it refers most specifically to a proposed disformal generalization of C-theory in modified gravity, where the metric defining the connection is related to the physical metric by curvature-dependent conformal and disformal terms (Sandstad et al., 2013). In a mathematically distinct usage, Thomsen’s D-theory is a discrete analogue of Connes–Higson E-theory for separable C$^*$-algebras (Hunger, 2020). In string- and brane-based settings, the term also appears in broader, more interpretive senses connected with D-branes, D-particles, doubled formalisms, and emergent higher-dimensional dynamics (Yoneya, 2016).

1. Terminological scope

In modified-gravity research, “D-theory” arises as an extension of C-theory. The letter “D” stands for disformal, in contrast with the purely conformal relation that defines C-theories. In this usage, the central object is a two-metric framework in which the auxiliary metric governing the connection is allowed to depend explicitly on curvature tensors, not only on scalar curvature (Sandstad et al., 2013).

In operator-algebraic topology, Thomsen’s D-theory has a different meaning altogether. It is a bivariant functor $D(A,B)$ built from discrete asymptotic homomorphisms, and it is explicitly presented as a “discrete” analogue of E-theory. Its objects, products, and comparison maps are formulated entirely within the category of separable C$^*$-algebras (Hunger, 2020).

In string and M-theory, the label is less uniform. Some works use “D-theory” in a broad “dimensional origin” or D-brane-centered sense, where lower-dimensional gauge or matrix degrees of freedom encode higher-dimensional dynamics. Other nearby constructions, such as doubled descriptions of D-branes, are not themselves named D-theory in a strict sense, but they contribute to the same conceptual landscape by making D-brane dualities and emergent geometry structurally explicit (Douglas, 2010).

2. Disformal metric–affine D-theories in modified gravity

The most explicit definition of D-theory in the supplied literature appears in the context of nonlocal modified gravity. The starting point is C-theory, where two metrics are used: the physical metric $g_{\mu\nu}$, which defines distances and appears in the action, and an auxiliary metric $\hat g_{\mu\nu}$, whose Levi–Civita connection defines $\hat R_{\mu\nu}$. In C-theory these metrics are related conformally,

$$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$

and the action is

$S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$

Because $\mathcal C$ depends on $\mathcal R$, while $D(A,B)$0 itself depends on $D(A,B)$1, the relation between $D(A,B)$2 and the Ricci scalar $D(A,B)$3 of $D(A,B)$4 is recursive and generically infinite-order in derivatives (Sandstad et al., 2013).

A key result is that certain C-theories can be mapped exactly to nonlocal actions of the form

$D(A,B)$5

with analytic, nontruncating form factor

$D(A,B)$6

The mapping becomes exact when the recursion for $D(A,B)$7 can be reduced to a pure $D(A,B)$8-operator structure. For the power-law ansatz

$D(A,B)$9

the recursion simplifies to

$^*$0

so that formally

$^*$1

For quadratic

$^*$2

this yields

$^*$3

with

$^*$4

The exact mapping to the Biswas–Mazumdar–Siegel type bilinear nonlocal action holds only for quadratic $^*$5; higher-than-quadratic choices generate products such as $^*$6, which do not reorganize into a single $^*$7 term (Sandstad et al., 2013).

Within this framework, D-theories are proposed by replacing the purely conformal relation with a curvature-disformal one,

$^*$8

where $^*$9 and $g_{\mu\nu}$0 may depend on curvature invariants such as $g_{\mu\nu}$1 and $g_{\mu\nu}$2. The paper does not supply a complete D-theory action or field equations; it introduces the structure as a geometric generalization motivated by the nonlocal mapping of C-theories and by the need to access more general quadratic curvature sectors (Sandstad et al., 2013).

3. Nonlocality, renormalisability, and the status of the gravity programme

The gravity motivation for D-theories is tied directly to the distinction between scalar and spin-2 modifications of the gravitational sector. The mapped C-theories reproduce nonlocal actions based on the scalar curvature, but the paper stresses that conformal theories of the form $g_{\mu\nu}$3 alone cannot fully address the ultraviolet problems of Einstein gravity. To modify the graviton propagator in a way relevant for potential renormalisability, one needs Weyl-type terms and, more generally, quadratic invariants involving $g_{\mu\nu}$4 and $g_{\mu\nu}$5 (Sandstad et al., 2013).

This is the point at which the disformal extension is introduced. A purely conformal relation between $g_{\mu\nu}$6 and $g_{\mu\nu}$7 is too restrictive to geometrize the full nonlocal quadratic sector. By contrast, a relation involving $g_{\mu\nu}$8 suggests a route to induced nonlocal operators acting not only on $g_{\mu\nu}$9, but also on tensorial curvature components. The paper therefore presents D-theories as a possible geometric embedding of nonlocal renormalisable gravity, rather than as an already developed renormalisable model (Sandstad et al., 2013).

The concrete state of the proposal remains preliminary. The supplied analysis leaves open the precise mapping from D-theories to fully general nonlocal actions, the conditions for ghost freedom, the well-posedness of the corresponding infinite-order equations, and any explicit power-counting or loop analysis establishing renormalisability. A common misconception is therefore that D-theory in this sense is already a complete alternative formulation of nonlocal gravity. The paper supports a narrower statement: D-theories are a proposed curvature-disformal generalization of C-theory whose connections to nonlocal renormalisable gravity “remain to be explored” (Sandstad et al., 2013).

The same paper gives a worked 4D example on the C-theory side,

$\hat g_{\mu\nu}$0

which maps to a nonlocal model with

$\hat g_{\mu\nu}$1

Its linearised spectrum about flat space contains two additional propagator poles,

$\hat g_{\mu\nu}$2

and in the degenerate case

$\hat g_{\mu\nu}$3

the poles coalesce into a double pole of Pais–Uhlenbeck type, the residue vanishes, and for $\hat g_{\mu\nu}$4 the model can be made stable. This result pertains to the mapped nonlocal model, not to a fully developed D-theory, but it explains why the authors regard the broader disformal generalization as potentially useful (Sandstad et al., 2013).

4. Thomsen’s D-theory as a discrete analogue of E-theory

In operator algebra, D-theory has a precise and unrelated definition. Let

$\hat g_{\mu\nu}$5

and

$\hat g_{\mu\nu}$6

The corresponding asymptotic algebras are

$\hat g_{\mu\nu}$7

An asymptotic homomorphism is a $\hat g_{\mu\nu}$8-homomorphism $\hat g_{\mu\nu}$9, while a discrete asymptotic homomorphism is a $\hat R_{\mu\nu}$0-homomorphism $\hat R_{\mu\nu}$1. Thomsen’s D-theory is then defined by

$\hat R_{\mu\nu}$2

with $\hat R_{\mu\nu}$3 and $\hat R_{\mu\nu}$4 (Hunger, 2020).

A technically important reformulation uses the kernel

$\hat R_{\mu\nu}$5

the sequentially trivial asymptotic algebra. The paper proves a natural $\hat R_{\mu\nu}$6-isomorphism

$\hat R_{\mu\nu}$7

so that

$\hat R_{\mu\nu}$8

This model supports products

$\hat R_{\mu\nu}$9

and mixed products with E-theory,

$$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$0

all compatible with the corresponding asymptotic composition operations (Hunger, 2020).

The main computation in the paper concerns

$$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$1

For any separable C$$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$2-algebra $$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$3,

$$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$4

Thus D-theory captures asymptotic sequences of $$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$5-classes modulo eventual triviality. The same work also compares the E-theoretic Kronecker pairing

$$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$6

with the D-theoretic pairing induced by

$$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$7

and proves that the two constructions agree after identifying

$$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$8

In this sense, D-theory is a sequence-sensitive refinement of E-theory rather than a reformulation of geometric gravity or brane dynamics (Hunger, 2020).

5. D-brane, KK-theoretic, and doubled-formalism usages

A further usage of “D-theory” appears in noncommutative geometry and D-brane topology. In this setting, D-branes are modeled by geometric K-cycles $$\hat g_{\mu\nu}=\mathcal C(\mathcal R)\,g_{\mu\nu},\qquad \mathcal R=g^{\mu\nu}\hat R_{\mu\nu},$$9 or, more generally, by Kasparov classes in $S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$0. Here KK-theory furnishes a bivariant category in which separable C$S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$1-algebras represent backgrounds or open-string algebras and KK-classes represent D-branes, correspondences, and T-duality morphisms. Noncommutative Poincaré duality, K-orientation, Grothendieck–Riemann–Roch, and cyclic cohomology then provide a framework for RR charge formulae and duality on noncommutative spaces (Szabo, 2008).

In that literature, T-duality is naturally expressed through KK-equivalence. A KK-equivalence $S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$2 induces isomorphisms

$S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$3

and geometric correspondences implement transforms such as the Fourier–Mukai map for torus duality. The framework also accommodates continuous-trace algebras $S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$4 for $S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$5-flux backgrounds, twisted K-theory $S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$6, and the noncommutative RR charge

$S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$7

which generalizes the Minasian–Moore expression $S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$8 (Szabo, 2008).

Adjacent to this, though not presented there under the name D-theory itself, is the doubled-formalism description of open strings and D-branes. In Hull’s doubled formalism, open-string target coordinates are doubled,

$S=\int d^d x\,\sqrt{-g}\,f(\mathcal R).$9

with $\mathcal C$0-invariant metric

$\mathcal C$1

generalized metric

$\mathcal C$2

and self-duality constraint

$\mathcal C$3

D-branes are encoded by Neumann and Dirichlet projectors satisfying null, orthogonality, and integrability conditions in doubled space, and a single doubled object can represent a T-dual pair of conventional D-branes, termed a double D-brane (Albertsson et al., 2011).

The corresponding one-loop open-string analysis yields a doubled DBI-type effective action

$\mathcal C$4

with $\mathcal C$5, and an explicitly $\mathcal C$6-covariant “master” form,

$\mathcal C$7

This suggests a broader D-brane-centered meaning of D-theory in which duality-covariant master actions, noncommutative correspondences, and bivariant charge formalisms all serve as organizational structures for D-brane physics, even though the underlying definitions remain distinct (Albertsson et al., 2011).

6. Emergent higher-dimensional theory from D-particles, 5D MSYM, and nonrelativistic branes

In a broader “dimensional origin” sense, D-theory can denote the idea that lower-dimensional or discrete degrees of freedom encode a higher-dimensional continuum theory. One example is the proposal that 5D maximally supersymmetric Yang–Mills theory can serve as a nonperturbative definition of the 6D $\mathcal C$8 theory compactified on a circle. The compactification relation is

$\mathcal C$9

and the 5D instanton particle with mass $\mathcal R$0 is identified with the KK mode of the sixth dimension. The paper explicitly presents this as a D-theory-like picture in which the extra dimension is encoded in nonperturbative solitons rather than added kinematically (Douglas, 2010).

A more explicit D-particle realization appears in the covariantized Matrix theory for D-particles. There the basic variables are generalized 11D vectors

$\mathcal R$1

with $\mathcal R$2 Hermitian matrix parts, higher gauge fields, and a discretized Nambu 3-bracket

$\mathcal R$3

The bosonic action is Lorentz-covariant in 11D and scale-invariant; the 11D length scale emerges through a super-selection rule fixing the conserved invariant

$\mathcal R$4

In the light-front gauge with DLCQ compactification it reduces to the usual BFSS formulation, while in the time-like gauge with ordinary spatial compactification it becomes a non-Abelian Born–Infeld-like theory that approaches BFSS in the large-$\mathcal R$5 limit (Yoneya, 2016).

A further instance of emergent higher-dimensional structure is provided by nonrelativistic D-brane dualization. Nonrelativistic D$\mathcal R$6-brane actions are obtained from a scaling limit of relativistic DBI+CS actions in string Newton–Cartan geometry, with DBI sector

$\mathcal R$7

Dualizing the worldvolume $\mathcal R$8 on a D2-brane produces a scalar $\mathcal R$9 that is interpreted as an eleventh coordinate, and the resulting dual action is that of a nonrelativistic M2-brane moving in membrane Newton–Cartan geometry with codimension-three foliation. In the D1 and D3 sectors, analogous dualizations realize nonrelativistic versions of $D(A,B)$00-strings and S-duality structures, while the D4 analysis points toward a nonrelativistic M5 sector (Ebert et al., 2021).

These examples do not define a single common formalism. They do, however, display a shared pattern: D-particles, instantons, or D-brane worldvolume gauge fields act as carriers of dimensions or dual descriptions that are absent in the naive classical variables. This suggests a broad D-theory motif in which nonperturbative or discrete objects furnish the substrate for an emergent higher-dimensional continuum.

7. Conceptual synthesis

A common misconception is that D-theory names one coherent theory spanning gravity, operator algebras, and string theory. The literature represented here supports the opposite conclusion. In gravity, D-theory is a proposal for curvature-disformal metric relations intended to generalize C-theories and perhaps geometrize nonlocal renormalisable gravity (Sandstad et al., 2013). In C$D(A,B)$01-algebra theory, D-theory is already a well-defined bivariant functor based on discrete asymptotic morphisms (Hunger, 2020). In string and M-theory, the label functions more as an organizing perspective tied to D-branes, D-particles, duality, and emergent dimensions than as a universally standardized construction (Douglas, 2010).

Another recurrent point of confusion concerns completeness. Thomsen’s D-theory is technically complete enough to support products, suspension, stability, and explicit computations such as

$D(A,B)$02

By contrast, the gravity D-theory of curvature-disformal metrics is explicitly prospective: its mapping to the full nonlocal quadratic sector, its ghost conditions, and its renormalisability analysis are left open. Likewise, D-particle and nonrelativistic brane formulations are structurally rich but belong to specific sectors—Matrix-theoretic, duality-theoretic, or Newton–Cartan—rather than to a single universal D-theory (Yoneya, 2016).

The most defensible unifying characterization is therefore methodological rather than doctrinal. Across these literatures, “D-theory” marks attempts to replace a direct formulation of the target theory by a more structural one: a disformal two-metric relation in gravity, a discrete asymptotic bivariant category in operator algebra, a KK-theoretic category of D-branes, a doubled master action for T-duality, or a matrix/solitonic substrate from which higher-dimensional physics emerges. This suggests a family resemblance centered on reformulation, dual description, and emergent structure, but not a single shared definition.