Quantum Gravity: Why We Already Found The Answer

Professor John Donoghue argues that quantum mechanics and general relativity are not fundamentally incompatible — the popular narrative of their conflict is misleading. Gravity is a field (the metric), and it can be quantized exactly like QCD using path integral methods, as Feynman and DeWitt showed decades ago. The real issue is not incompatibility but hidden assumptions — such as the demand for grand unification, naturalness, or strict causality at all scales — that may be human biases rather than physical requirements.

• Effective field theory (EFT) is the key framework: All our theories, including the Standard Model and general relativity, are effective field theories valid within a certain energy range. Wilson’s insight was that non-renormalizable theories are perfectly acceptable as EFTs — quantum effects generate higher-dimensional terms in the Lagrangian, and these can be handled systematically without knowing the high-energy completion.
  • Donoghue’s 1994 calculation of the leading quantum correction to the Newtonian potential demonstrated this explicitly: the long-distance corrections are fully determined by known low-energy physics (general relativity), with no need to know the UV theory.
• The geometric picture of GR may be a historical obstacle: The mantra “gravity is geometry, not a force” makes gravity seem fundamentally different from other interactions. But GR can equally be formulated as a field theory with exchange particles (gravitons), and in that form its quantization proceeds identically to QCD. The path integral approach is essential — it is the only practical framework for non-Abelian gauge theories like QCD, and gravity should be treated the same way.

Quadratic gravity is a renormalizable theory of gravity that Donoghue has worked on extensively (with collaborator Gabriel Menezes). It extends Einstein’s action by adding curvature-squared terms (Ricci scalar squared and Weyl tensor squared), making the theory renormalizable — high-energy divergences are tamed because propagators fall off as 1/E⁴ instead of 1/E².

• The trade-off is a potential violation of causality at high energies: Higher-derivative theories generically involve four derivatives of the metric, which conflicts with the usual axiomatic requirement of analyticity (tied to causality). Donoghue’s analysis suggests the theory is unitary and stable but contains a heavy ghost particle that propagates with the opposite arrow of causality — it moves backward in time for short periods during intermediate stages of scattering.
  • This is the “dueling arrows of causality”: the usual massless graviton propagates forward in time (via e^(iS)), while the heavy ghost propagates via e^(-iS), corresponding to reversed causal direction. The S-matrix remains well-defined because only normal particles exist in the asymptotic past and future; the ghost appears only as an unstable intermediate state.
  • The theory has been verified at one-loop order; higher-loop calculations remain incomplete, and lattice simulations are being explored to test its consistency non-perturbatively.
• Quadratic gravity is not necessarily the final theory — Donoghue compares it to the Standard Model: a renormalizable quantum theory valid up to some high scale (perhaps inflation), but likely embedded in a deeper theory that explains the origin of all interactions. It does, however, offer possible non-singular black hole solutions, unlike classical GR.

Donoghue is sympathetic to “random dynamics” (Holger Nielsen’s idea) as an alternative to grand unification — the anti-unification perspective. Instead of all forces merging into a single simple group at high energies, random dynamics posits that everything happens at high energies, and only certain structures survive to low energies because they are protected by symmetries (gauge symmetry for bosons, chirality for fermions). This naturally explains why we see U(1), SU(2), SU(3) and generally covariant gravity.

• Grand unification may be a field-wide bias with no empirical support: Donoghue points out that historical “unifications” (electricity and magnetism, electroweak) were really identifications of a single underlying theory, not mergers of distinct ones. There is no evidence that the three gauge groups of the Standard Model actually converge at high energy. The LHC’s failure to find supersymmetry — which was motivated by naturalness arguments to protect the electroweak scale from GUT-scale radiative corrections — suggests that naturalness itself may be a flawed guiding principle.
  • The naturalness problem: GUTs require the electroweak scale to be enormously smaller than the GUT scale, which is unnatural unless some mechanism (like supersymmetry) protects it. The absence of such mechanisms at the LHC undermines the motivation for GUTs.

Donoghue’s broader methodological stance is to question hidden assumptions at every stage — the assumption that GR and QM are incompatible turned out to be wrong; the assumption of strict causality at all scales may be wrong; the assumption of unification may be wrong; the assumption of naturalness may be wrong. He encourages students to read widely across subfields, to look for unstated premises in arguments, and to remain open to ideas that seem radical if they are empirically grounded.

• On string theory: Donoghue does not see it as in conflict with EFT — it is one possible UV completion. However, he cautions that its apparent uniqueness (e.g., finite graviton amplitudes) relies on assumptions like exact causality at all scales, which may not hold. Other UV completions may exist.
• On the anthropic principle: Donoghue was an early contributor (1998) but has set it aside not because he rejects it, but because he ran out of productive calculations to do with it. He notes that the anthropic route gained motivation from the failure of naturalness — the parameter space compatible with stars and atoms is so small that a unique “theory of everything” predicting all couplings seems unlikely.