The guiding questionWhich low-energy observations require a quantum mediator, and which would distinguish different microscopic theories of spacetime?
Starting knowledge: Quantum mechanics, special relativity, and the meaning of spacetime curvature. The effective-theory argument is developed below.
- The physical question
- What is known, and what is open
- Build the minimum theory
- Worked example: a phase that cannot be assigned to either mass
- Questions to investigate
- First investigation and its stopping point
- Where this leads
- Read with a purpose
The physical question
General relativity describes gravity through a metric: a field that determines distances, clock readings, and the paths of freely falling bodies. Matter determines curvature, and curvature affects matter. Quantum theory allows matter to occupy coherent alternatives. If a mass is in two different places, should its gravitational field also occupy two alternatives? A classical metric sourced only by the average mass distribution is one possible prescription, but it has different consequences from a quantum field correlated with each alternative.
There are two questions here. Can gravity carry quantum information? This concerns relatively weak fields and could, in principle, admit laboratory tests. What replaces our description at arbitrarily short distances or inside a classical singularity? This concerns the microscopic completion of the theory. An answer to the first need not identify the answer to the second.
What is known, and what is open
Gravity already has a useful quantum description at low energies. In an effective field theory, we retain the degrees of freedom accessible at a chosen resolution and represent shorter-distance effects by additional interactions. At any fixed accuracy, only finitely many coefficients are needed. The nonrenormalizability of the Einstein action means that it cannot, by itself, predict all energies with finitely many parameters; it does not make low-energy quantum calculations meaningless. [1]
String theory supplies candidate microscopic degrees of freedom; certain gauge/gravity dualities relate a gravitational system to a quantum theory without gravity on its boundary. The best-developed examples use anti-de Sitter boundary conditions, associated with negative cosmological curvature, and particular matter theories. These are major structural achievements. Their application to our expanding universe is an additional problem. [2]
No experimentally selected microscopic theory of quantum spacetime follows from these results. Other programs quantize geometric variables, seek a high-energy fixed point, or sum over discrete causal structures. A fixed point means that dimensionless couplings stop changing with resolution; it could make high-energy behavior predictive. A proposal must still recover known low-energy physics and offer a controlled definition of observables.
Build the minimum theory
Use units in which the speed of light $c$ and the reduced Planck constant $\hbar$ equal one. The reduced Planck mass is $M_{\rm P}=(8\pi G)^{-1/2}$, where $G$ is Newton's constant. A schematic action is
$$S=\int d^4x\sqrt{-g}\left[\frac{M_{\rm P}^2}{2}R-\rho_\Lambda+c_1R^2+c_2R_{\mu\nu}R^{\mu\nu}+\cdots\right].$$
An action is the functional whose stationary points give classical equations and whose phase weights quantum histories. Here $g$ is the determinant of the metric, $R$ its scalar curvature, $R_{\mu\nu}$ the Ricci curvature tensor, and $\rho_\Lambda$ a constant energy density. The coefficients $c_1,c_2$ describe shorter-distance physics. The omitted terms have still more derivatives or fields. This ordering is useful when characteristic momenta and curvatures are sufficiently small compared with the corresponding high-energy scales.
The mechanism of predictivity is suppression: a term with additional derivatives contributes additional powers of a small ratio of energies. Massless propagation can also produce nonanalytic momentum dependence, such as logarithms. A local interaction is polynomial in momentum and cannot freely imitate that dependence. This is why some long-distance quantum contributions can be predicted without knowing every coefficient. [1]
Worked example: a phase that cannot be assigned to either mass
Prepare two masses in equal superpositions of paths labelled $0$ and $1$. Hold the four branch separations $r_{ij}$ approximately constant for time $t$, with negligible motion and unwanted interactions. Newton's potential energy is $V_{ij}=-Gm_1m_2/r_{ij}$. Quantum evolution supplies the dimensionless phase
$$\phi_{ij}=-V_{ij}t/\hbar=\frac{Gm_1m_2t}{\hbar r_{ij}},\qquad |\psi\rangle=\frac12\sum_{i,j=0}^1 e^{i\phi_{ij}}|ij\rangle.$$
A phase added to path $i$ of the first mass changes an entire row; a phase added to path $j$ of the second changes an entire column. Both disappear from
$$\Phi=\phi_{00}+\phi_{11}-\phi_{01}-\phi_{10}.$$
The $2\times2$ matrix of state amplitudes describes a product state exactly when its determinant vanishes. Here that condition is $\Phi=0$ modulo $2\pi$. Thus a nonzero invariant phase can entangle initially independent path degrees of freedom. The calculation is exact for this prescribed two-path interaction, not for a complete experimental apparatus.
Insight. The relevant signature is a joint correlation that independent local phases cannot produce. Inferring a nonclassical gravitational mediator further assumes that gravity is the only coupling and that the proposed class of classical mediators cannot transmit the required quantum information. This does not discriminate string theory from every other quantum-gravity completion. [3]
Questions to investigate
Calibration questions have established answers. Diagnostics ask what follows inside a specified model or comparison. Extensions are bounded proposals whose novelty and feasibility need a focused literature check. Each answer should record its assumptions and limits.
QG-1 · Quantum gravityCalibration
Which gravitational phases are physically joint?
Do: Derive the determinant criterion above and compute each mass's reduced-state eigenvalues as functions of $\Phi$. Produce an entropy-versus-phase plot.
Check: Entropy vanishes at $\Phi=0,2\pi$ and is maximal at $\Phi=\pi$. Changing any local row or column phase leaves it unchanged.
Meaning: Establishes the precise witness that a proposed weak-field experiment must preserve. It is a known two-qubit result, not a new theory of gravity.
QG-2 · Quantum gravityCalibration
When is a curvature correction small?
Do: For a background curvature radius $L$, estimate the ratio of $c_1R^2$ to the Einstein term using $R\sim L^{-2}$. Repeat after including a large dimensionless $c_1$.
Check: The ratio scales as $c_1/(M_{\rm P}L)^2$, up to geometric factors. Identify the actual breakdown scale rather than assuming it is always $M_{\rm P}$.
Meaning: Turns “low energy” into a condition on an observable regime and its coefficients.
QG-3 · Quantum gravityDiagnostic
Which correction is robust against unknown local interactions?
Do: In Donoghue's low-momentum scattering amplitude, separate polynomial terms from logarithms and other nonanalytic terms. Fourier-transform representative terms at nonzero separation.
Check: Contact terms and long-range tails must be distinguished; compare amplitudes or explicitly defined observables before comparing coordinate-dependent potentials.
Meaning: Produce a table of universal predictions and coefficients requiring measurement. This is a literature-guided calculation with established benchmarks.
QG-4 · Quantum gravityDiagnostic
How much decoherence destroys the phase witness?
Do: Apply independent dephasing channels to the two-path state. Calculate the negative eigenvalues, if any, of its partial transpose, obtained by transposing only one subsystem's indices.
Check: Recover the pure-state result at zero noise and a separable mixture at complete dephasing. Separate state entanglement from the sensitivity of a particular witness.
Meaning: Gives a quantitative noise budget within the model, not an apparatus design or a detection claim.
QG-5 · Quantum gravityExtension
Does a proposed classical mediator actually satisfy its stated restrictions?
Do: Choose one published classical-channel model. Write its allowed operations and noise explicitly, then test the QG-1 preparation under that dynamics.
Check: Account for any hidden shared quantum resource, direct nonlocal interaction, or measurement feedback. Reproduce its published result first.
Meaning: A defensible output is an assumption-dependent distinction between model classes. Novelty requires checking the specific mediator literature; the general witness argument is established.
First investigation and its stopping point
Start with QG-1 and QG-4 together. They produce one transparent two-qubit calculation, one noise model, and one statement of what an observation would imply. Stop short of a microscopic-gravity claim unless a second observable actually separates competing completions.
Where this leads
Black holes and information · Local quantum fields and operational access · Vacuum energy and cosmic acceleration
Read with a purpose