Modified Inertial Interpretation

• Modified inertial interpretation is a framework that reallocates anomalous dynamics from the gravitational to the inertial sector, retaining Newtonian gravity while altering the force–acceleration relation.
• It employs acceleration-dependent inertia, nonlocal kinetic functionals, and geometric modifications to explain phenomena like flat galactic rotation curves and the Tully–Fisher relation.
• This approach has broad applications across MOND, relativistic mechanics, quantum field theory, and fixed-point numerical methods, challenging conventional conservation and equivalence principles.

Modified inertial interpretation denotes a class of approaches in which anomalous or generalized dynamics are assigned to the inertial or kinetic sector rather than to the interaction sector. In the MOND literature, this means retaining Newtonian gravity or the Newtonian potential while modifying the relation between force and motion, often through acceleration-dependent inertia, field-dependent inertial mass, nonlocal kinetic functionals, or alternative spacetime kinematics (Pankovic et al., 2010, Milgrom, 2011, Milgrom, 2022). In other domains, the same interpretive shift appears in relativistic continuum mechanics, where inertial motion is carried by a center of mass and spin rather than by the naive center of mass (Medina et al., 2014), in quantum field theory on accelerated backgrounds, where inertial states acquire a universal thermal component for Rindler observers (Lochan et al., 2014), and in numerical fixed-point theory, where inertial extrapolation is introduced directly into the iteration map (Husain et al., 2021).

1. Conceptual scope and recurring motif

Across the cited literature, the common structure is a relocation of new physics from force generation to inertial response. In MOND, this is the contrast between modified gravity and modified inertia: the former changes the field equations, such as the nonlinear Poisson equation, whereas the latter keeps the gravitational field Newtonian and changes the kinetic term or the force–acceleration relation (Milgrom, 2011). In relativistic mechanics with spin, the same shift appears as a correction to what counts as the inertial worldline of an isolated system: not the ordinary center of mass, but a spin-corrected center (Medina et al., 2014). In QFT in non-inertial frames, it appears as a refinement of the meaning of an inertial state, because a Minkowski state acquires a thermal Rindler component plus state-dependent corrections (Lochan et al., 2014). In general relativity, a related move is to interpret inertial forces in uniformly accelerated systems as effects of a $\Lambda$-type stress tensor in a conformally flat anti de Sitter background (Culetu, 2014).

This recurring motif suggests an editor’s term, “inertial relocation,” for the strategy of moving explanatory burden from interaction laws to inertial structure. A plausible implication is that the modified inertial interpretation is less a single theory than a recurrent explanatory pattern. The pattern is strongest in MOND, where it is developed as a direct alternative to both dark matter and modified gravity, but the same logic recurs in relativistic kinematics, accelerated-frame QFT, and variational numerical analysis.

2. MOND as modification of inertia

Milgrom’s MOND paradigm introduces an acceleration scale $a_0$, with Newtonian behavior recovered for $a\gg a_0$ and deep-MOND behavior in the scale-invariant limit $a\ll a_0$ (Milgrom, 2011). In modified-inertia formulations, the Poisson equation for the gravitational potential is left unchanged, while the particle equation of motion is replaced by a nonstandard inertial functional. Milgrom’s general nonrelativistic form is

$$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$

rather than $\mathbf{a}=-\nabla\phi$, and for circular orbits in an axisymmetric potential this yields

$$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$

which directly implies asymptotically flat rotation curves and the baryonic Tully–Fisher relation $V_\infty^4=GMa_0$ (Milgrom, 2011).

A particularly explicit local realization is Panković and Kapor’s k-MOND, which preserves the Newtonian gravitational potential

$V_g=-\frac{Gm_gM}{r}$

but replaces the equality of inertial and gravitational mass by

$m_i(a)=m_g\,\frac{a^2}{a^2+a_0^2}.$

The kinetic term becomes

$a_0$0

so the Lagrangian depends on positions, velocities, and accelerations and must be varied with the Euler–Poisson equations

$a_0$1

For low accelerations and uniform circular motion, this construction yields

$a_0$2

so k-MOND is “identical to Milgrom’s MOND” in that regime (Pankovic et al., 2010).

A different modified-inertia realization makes inertial mass a function of the external Newtonian field rather than of instantaneous acceleration. In the weak-field regime,

$a_0$3

with interpolation

$a_0$4

so circular motion gives

$a_0$5

and again

$a_0$6

after identifying $a_0$7 with the MOND scale (1011.3618). In that framework, $a_0$8 is not strictly constant but varies slowly,

$a_0$9

which is proposed as a way to address bright-cluster and bright-galaxy discrepancies (1011.3618).

The most systematic modified-inertia models are time-nonlocal. In Fourier space they take the form

$a\gg a_0$0

or, in a simple subclass,

$a\gg a_0$1

These models preserve the salient MOND predictions, define nonlocal momentum, angular momentum, and energy, and reduce the general two-body problem in the deep-MOND regime to a single-body problem with a modified reduced mass (Milgrom, 2022). Milgrom emphasized that such modified-inertia theories are nonlocal in time, so MOND effects depend on the full trajectory rather than only on the instantaneous state; this is one of their main differences from modified-gravity theories (Milgrom, 2011).

3. Local, geometric, and kinematic realizations

Several works attempt to realize modified inertia through local higher-derivative actions or alternative spacetime geometries. A local Lagrangian proposal for a point particle takes

$a\gg a_0$2

with $a\gg a_0$3 at large accelerations and $a\gg a_0$4 in the deep-MOND regime (Costa et al., 2019). For gravitational circular motion this reproduces

$a\gg a_0$5

and $a\gg a_0$6, but perturbations activate higher-derivative sectors with exponentially unstable solutions. The exponentially unstable branches must be set to zero to match the very small scattering of the Tully–Fisher relation, while linearly growing modes remain phenomenologically acceptable for at least $a\gg a_0$7 billion years (Costa et al., 2019).

A more geometric realization uses Finsler spacetime. In that construction, low-acceleration inertia is quartic rather than quadratic in velocity, so energy scales as $a\gg a_0$8 and momentum as $a\gg a_0$9. For static isotropic quartic Finsler metrics, consistency with the Tully–Fisher relation yields a weak-field equation

$a\ll a_0$0

and for a point mass

$a\ll a_0$1

Thus a modified-inertia implementation in metric form forces a logarithmic gravitational potential $a\ll a_0$2, still linear in $a\ll a_0$3 but non-Newtonian in radius (Namouni, 2015). A plausible implication is that in this Finsler realization, modified inertia and modified gravity cease to be cleanly separable.

Other authors reconstruct modified inertia from frame symmetry. One proposal abandons Galilean invariance and introduces a Lorentz-type transformation between uniformly accelerated frames with invariant acceleration $a\ll a_0$4, yielding a factor

$a\ll a_0$5

a proper-time relation

$a\ll a_0$6

and a modified first law in which isolated bodies satisfy $a\ll a_0$7 rather than $a\ll a_0$8 (Alzain, 2017). Another proposal replaces $a\ll a_0$9 and $$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$0 by cosmological quantities using Sciama’s interpretation of Mach’s principle,

$$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$1

so that the MOND interpolation becomes a transformation between local and cosmic field intensities (Palomo, 2024).

A recent quantum version derives modified inertia from local short-time acceleration and de Sitter background broadening. The point-particle action acquires factors

$$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$2

leading to an effective acceleration

$$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$3

For circular galactic motion this gives

$$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$4

with $$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$5, so the MOND scale emerges from de Sitter curvature via $$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$6 (Luo, 16 Feb 2026).

4. Conservation laws, equivalence, and reformulation

Modified-inertia theories characteristically stress the inertial–gravitational distinction and therefore revisit conservation laws and equivalence principles. In k-MOND, $$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$7 for $$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$8, so the equality of inertial and gravitational mass is restored in the high-acceleration limit and low-acceleration deviations are confined to galactic-type environments (Pankovic et al., 2010). In field-dependent-inertia models, the weak equivalence principle is explicitly violated because $$\mathcal{A}[\{\mathbf{r}(t)\},t,a_0] = -\nabla\phi(\mathbf{r}(t)),$$9 depends on the external field $\mathbf{a}=-\nabla\phi$0, although Solar System behavior is recovered when $\mathbf{a}=-\nabla\phi$1 (1011.3618).

A technically important reformulation shows that modified inertia can be rewritten as an ordinary Newtonian equation $\mathbf{a}=-\nabla\phi$2, but now with an effective field

$\mathbf{a}=-\nabla\phi$3

derived from the Newtonian baryonic field $\mathbf{a}=-\nabla\phi$4. This field obeys

$\mathbf{a}=-\nabla\phi$5

so the equivalent Newtonian theory contains both an effective dark matter density

$\mathbf{a}=-\nabla\phi$6

and a generally nonconservative gravitational field (Shariati et al., 2021). In this picture, modified inertia is dynamically equivalent to Newtonian dynamics with a specific nonconservative gravity plus an induced dark component. For a binary in free fall inside a spheroidal galactic field, the tidal force becomes nonconservative and can perform net work over an orbital cycle, unlike in standard Newtonian gravity (Shariati et al., 2021).

In relativistic continuum mechanics with spin, conservation of total angular momentum similarly modifies what counts as inertial motion. If the energy–momentum tensor is non-symmetric, the ordinary center of mass

$\mathbf{a}=-\nabla\phi$7

does not move inertially in general. Defining the spin displacement

$\mathbf{a}=-\nabla\phi$8

one obtains the center of mass and spin

$\mathbf{a}=-\nabla\phi$9

which satisfies

$$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$0

The improved principle of inertia is therefore that the center of mass and spin of isolated systems moves with constant velocity (Medina et al., 2014). This is a modified inertial interpretation in a precise relativistic sense: inertia belongs to a spin-corrected centroid, not to the naive center of mass.

5. Accelerated frames, inertial states, and observer dependence

In quantum field theory, modified inertial interpretation arises not by changing equations of motion for particles but by refining what an inertial state means for accelerated observers. For a uniformly accelerated observer in the Rindler frame, the Minkowski vacuum produces the Unruh thermal spectrum

$$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$1

For an arbitrary inertial non-vacuum state, the expectation value of the Rindler number operator splits into a universal thermal component plus state-dependent corrections, determined by an effective field $$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$2 constructed from the inertial state’s mode amplitudes (Lochan et al., 2014). For all physically well-behaved normalizable inertial states, the correction terms decay with the Rindler-mode energy, so the high-frequency limit is dominated by thermal noise; non-normalizable states instead produce a constant contribution at high frequencies (Lochan et al., 2014). A similar dichotomy appears in the Unruh–DeWitt detector response.

This result refines the usual statement that “the inertial vacuum looks thermal in the Rindler frame.” The refinement is that any inertial state looks like a thermal component plus state-dependent spectral distortions, with the distortions asymptotically suppressed for normalizable states. A plausible implication is that the inertial sector, once referred to accelerated observers, has a universal thermal skeleton and a state-dependent remainder.

A different spacetime reinterpretation appears in a regular modified C-metric. In the weak-field limit with $$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$3, the metric becomes conformally flat,

$$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$4

with Ricci scalar

$$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$5

so the spacetime is anti de Sitter with

$$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$6

The stress tensor is of $$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$7-type with negative energy density,

$$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$8

and static observers have proper acceleration $$\frac{V^2}{R}\,\mu\!\left(\frac{V^2}{Ra_0}\right)=-\frac{\partial\phi}{\partial R},$$9 (Culetu, 2014). In that setting, inertial forces in uniformly accelerated systems are reinterpreted as gravitational effects of a negative cosmological constant, rather than as purely fictitious forces (Culetu, 2014).

6. Algorithmic and Hilbert-space usage

Outside fundamental physics, the term appears in fixed-point and monotone-inclusion theory as an algorithmic inertial modification. For the split monotone inclusion problem coupled to a fixed-point problem in real Hilbert space, the proposed inertial modified S-iteration is

$V_\infty^4=GMa_0$0

with

$V_\infty^4=GMa_0$1

Here the inertial term is the extrapolation $V_\infty^4=GMa_0$2, which plays the role of heavy-ball momentum inside a modified S-iteration (Husain et al., 2021).

The convergence theory is framed in standard fixed-point language. Under assumptions (D1)–(D4), for any $V_\infty^4=GMa_0$3,

$V_\infty^4=GMa_0$4

exists, and the asymptotic residuals vanish: $V_\infty^4=GMa_0$5 With Opial’s property, weak convergence to a point in $V_\infty^4=GMa_0$6 follows; with condition (B) or semicompactness, strong convergence is obtained (Husain et al., 2021). The same paper explicitly notes that the inertial modification is the first line of the algorithm and that, when $V_\infty^4=GMa_0$7, the scheme reduces to a non-inertial modified S-iteration (Husain et al., 2021). In this mathematical usage, “modified inertial interpretation” does not concern forces or mass, but the same structural idea remains: dynamical updating is altered by an inertial memory term.

7. Limitations and open issues

The modified inertial interpretation repeatedly encounters the same technical tensions. In MOND, action-based modified inertia is naturally time-nonlocal, and fully local realizations tend to require higher derivatives, with associated Ostrogradsky instabilities, runaway branches, or restricted validity to special sectors such as circular orbits (Milgrom, 2011, Costa et al., 2019). Many constructions recover $V_\infty^4=GMa_0$8 only in special limits, and several remain nonrelativistic or lack a complete lensing and cosmological framework (Namouni, 2015, Pankovic et al., 2010). Even when conservation laws can be defined, they are often nonlocal in time or re-expressed through effective dark components and nonconservative fields (Milgrom, 2022, Shariati et al., 2021).

Equivalence principles are also recurrently weakened or reformulated. k-MOND and field-dependent inertia separate inertial and gravitational mass at low acceleration (Pankovic et al., 2010, 1011.3618). The acceleration-invariant proposal replaces Galilean invariance by Lorentz-type transformations in acceleration space (Alzain, 2017). The quantum broadening approach explicitly requires a quantum equivalence principle extending equivalence from first moments to second-order fluctuations (Luo, 16 Feb 2026). In the relativistic spin case, inertial motion is reassigned to $V_\infty^4=GMa_0$9 rather than $V_g=-\frac{Gm_gM}{r}$0 (Medina et al., 2014).

These limitations suggest that modified inertial interpretation is best understood as a broad research program rather than a settled doctrine. Its strongest unifying claim is that discrepancies usually attributed to force-law modification, hidden sources, or observer artifacts can, in some regimes, be re-expressed as changes in inertia, kinetic structure, or inertial reference. The durability of that claim depends on whether local and nonlocal realizations, geometric reformulations, and observer-dependent constructions can be integrated into complete theories that retain the empirical successes of MOND, relativistic field theory, and many-body dynamics while avoiding instability, ambiguity, and loss of predictive control.