Path 3: Asymptotic Safety

Gravity as a renormalizable quantum field theory, saved by a high-energy fixed point.

Inventory

Rationale: In standard perturbative quantization, gravity’s infinities can’t be tamed by a finite number of counterterms – in technical terms, Einstein’s gravity is non-renormalizable. However, Steven Weinberg conjectured that maybe a non-perturbative approach could work: perhaps as the energy scale grows, the gravitational interaction strength approaches a fixed, finite value (an UV fixed point of the renormalization group). If so, the theory would be “asymptotically safe” – well-behaved at arbitrarily high energies, with only a finite number of measurable parameters needed. Extensive evidence from functional renormalization group calculations suggests such a fixed point might exist (including contributions of curvature-squared terms, etc.) – making gravity effectively renormalizable in a generalized sense. This path thus treats gravity as a high-dimensional quantum field theory that self-regulates at the Planck scale. It doesn’t introduce new particles or extra dimensions; it modifies the way we sum up quantum effects.

Prerequisites: Quantum field theory (especially renormalization group concepts and effective field theory), perturbation theory and Feynman diagrams (to understand why naive gravity fails), and techniques of functional integration. Knowledge of classical gravity is needed to set up the initial action, and familiarity with critical phenomena in statistical mechanics can be insightful (as the approach often draws analogies to how systems reach fixed points).

Dependencies: Asymptotic safety is a standalone consistency requirement – it can apply to pure gravity or gravity + matter. It complements Path 1 in that string theory can be thought of as one way to achieve finiteness, whereas asymptotic safety achieves finiteness by summing gravitational interactions in a self-consistent way. In fact, some suspect string theory (if it has a unique vacuum) might automatically be asymptotically safe when all heavy modes are included – but this is speculative. There is also overlap with Path 4 (CDT): lattice studies of gravity (like CDT) can provide evidence of fixed-point behavior (some CDT results hint at second-order transitions consistent with asymptotic safety). Asymptotic safety calculations assume no new symmetries like supersymmetry, but one could incorporate those too (combining with Path 1/Path 6 ideas to see if SUSY QG is asymptotically safe, etc.).

Signs of Progress: A major milestone would be a calculation showing a UV fixed point in a sufficiently complete truncation (including matter fields) such that low-energy parameters (like the ratio of proton to electron mass or the cosmological constant) could be predicted. Already, truncated RG flows find fixed points with finite Newton’s constant and cosmological constant. If asymptotic safety is true, one “prediction” is that at extremely high energies, gravity becomes effectively two-dimensional (a phenomenon called dimensional reduction) – if some future experiment or simulation (perhaps cosmic ray scattering or black hole production if ever observed at small scales) sees signs of that, it would be supportive. Another sign is internal consistency: no unphysical divergences in observables and agreement with known gravitational physics at accessible scales. Ultimately, the payoff would be to show quantitative predictions: for example, asymptotic safety might predict the value of the cosmological constant or relationships between coupling constants. Progress is measured by more sophisticated computations (e.g. higher loops, broader truncations) continuing to support the existence of the UV fixed point and by connecting these results to observable consequences (perhaps in early-universe cosmology or black hole behavior).

Base Camp 3.1: Quantum Field Theory and Renormalization Group Basics

Scope: Master the general notions of renormalization in quantum field theory – especially the concept of a renormalization group (RG) flow in the space of couplings and the idea of fixed points (like asymptotic freedom in QCD). Understand perturbative renormalizability vs. non-renormalizability with simple examples, and effective field theory (EFT) reasoning.

Stepping Stones: (1) Renormalize a simple QFT: e.g. $\lambda \phi^4$ in 4D – compute one-loop beta function, see that coupling grows with energy (Landau pole). Then $\phi^4$ in $4-\varepsilon$ dimensions shows a non-trivial fixed point (Wilson-Fisher fixed point); note how in 4D strictly it’s trivial. (2) Contrast with QCD: asymptotic freedom (negative beta function yields a UV fixed point at zero coupling). (3) Discuss what it means for a theory to be non-renormalizable (like $\phi^6$ in 4D or gravity in 4D by naive power counting) – using EFT, explain that it’s still predictive at low energies with a cutoff, but needs infinitely many parameters for high precision as energy grows. (4) Introduce Wilson’s idea: maybe “non-renormalizable” just means there’s a non-zero UV fixed point (instead of trivial Gaussian one) such that infinite couplings become functions of a few parameters (the relevant directions at that fixed point). (5) Get comfortable with functional integrals and generating functionals for multi-point functions – necessary for understanding the functional renormalization group later.

Base Camp 3.2: General Relativity as an Effective Field Theory

Scope: Treat Einstein’s gravity as a QFT at low energies (well below Planck scale): show that one-loop graviton corrections are finite up to known renormalizable interactions (like those involving matter) but introduce higher-dimension terms (like $R^2$) requiring counterterms. Understand why gravity is perturbatively non-renormalizable (increasing divergences with each loop, needing higher powers of curvature to cancel). However, note that as an EFT, we can predict at low energy using finite number of terms if we truncate at a given order in energy expansion.

Stepping Stones: (1) Write the Einstein-Hilbert action with leading higher-order terms: $S = \frac{1}{16\pi G}\int d^4x\sqrt{-g}(2\Lambda - R) + \alpha R^2 + \beta R_{\mu\nu}R^{\mu\nu} + \dots$ – see that beyond $G$ and $\Lambda$, terms with more derivatives appear. (2) Power-count the superficial divergence of a graviton loop: show that in $D=4$, each loop effectively brings two more powers of momentum in numerator than in renormalizable theory, making divergences worse at higher loops (non-renormalizability). (3) Consider gravity coupled to scalar or other fields: one-loop divergences generate $R^2$ terms (cite or derive known results from ’t Hooft & Veltman 1974 calculation: pure gravity 1-loop is finite on-shell, but with matter yields divergences requiring $R^2$ terms). (4) Emphasize that non-renormalizable doesn’t mean non-predictive: it means if you only consider energy $E \ll M_{\rm Planck}$, higher-dim terms are suppressed by $(E/M_P)^2$ etc. We can make predictions with a finite number of parameters to any given precision (as long as $E$ is low enough). (5) Mention that Weinberg’s asymptotic safety conjecture suggests maybe as $E \to M_P$ all these couplings approach a fixed finite set of values (with finite number of free parameters describing the approach).

Base Camp 3.3: Functional Renormalization Group (FRG) and Asymptotic Safety

Scope: Learn the functional RG approach (Wetterich equation) which allows a continuous interpolation from high-energy to low-energy by integrating out momentum modes gradually. In practice, get familiar with the concept of a “flowing action” $\Gamma_k$ which at scale $k$ includes effects of modes above scale $k$. Write down the Wetterich equation: $k \partial_k \Gamma_k = \frac{1}{2}{\rm Tr}\left[(\Gamma_k^{(2)} + R_k)^{-1} k \partial_k R_k\right]$. Understand how fixed points are found by looking for scale-independent dimensionless couplings.

Stepping Stones: (1) Derive or accept the Wetterich/Morris equation for a scalar field as an example: see how a mass/coupling flow can be computed. (2) Set up the ansatz for gravitational effective average action: e.g. $\Gamma_k = \frac{1}{16\pi G_k}\int \sqrt{g}(-R + 2\Lambda_k) +$ higher terms. Explain that $G_k$ and $\Lambda_k$ become scale-dependent couplings and we seek a fixed point $(G_*, \Lambda_*)$ such that their beta functions vanish when made dimensionless. (3) Discuss truncation: in practice one keeps a finite number of terms (like just $R$ and $\Lambda$, or adds $R^2$ terms) and solves flow equations for those couplings – hoping the truncation still captures the existence of a fixed point. (4) Present the result: typically one finds a UV fixed point with positive $\Lambda_*$ and $G_*$, with a finite number of unstable (relevant) directions (so predictive). (5) Note the evidence: many successive truncations (adding $R^2$, $R^3$, …) still show a fixed point (though couplings shift). Also mention that including matter fields hasn’t spoiled the fixed point in various studies.

Base Camp 3.4: Evidence and Extensions of the Fixed Point

Scope: Examine the evidence accumulated for the asymptotic safety scenario: e.g. the stability of the UV fixed point under inclusion of higher order terms (like $R^2$, $R^3$ or involving Ricci tensor/tensor invariants), inclusion of matter fields (find bounds on number/types of matter for which fixed point exists), etc. Also look at extensions like scalar-tensor theories (relevant for inflation), or supersymmetric gravity (does it improve convergence?).

Stepping Stones: (1) Summarize results: Einstein-Hilbert truncation yields ~2 UV-attractive directions (so two free parameters, basically $G$ and $\Lambda$ are determined, leaving an arbitrary ratio i.e. one parameter, plus other relevant like matter couplings). (2) Add $R^2$ and $R_{\mu\nu}R^{\mu\nu}$ terms: show that the fixed point still exists, shifts moderately, and one more direction may be relevant or irrelevant depending (studies usually find it remains with 2 relevant directions). (3) Consider heavy matter content: mention that beyond some threshold (e.g. ~$N_{\rm scalars}$ or $N_{\rm fermions}$ large), fixed point might disappear or change – thus asymptotic safety can give predictions like upper bounds on fermion flavors or gauge group sizes that fit our world. (4) Mention results from an $\varepsilon$-expansion around $2+\varepsilon$ dimensions: an analytic check (Weinberg’s idea also considered gravity in 2 dimensions plus small $\varepsilon$, where gravity is renormalizable and a fixed point can be found perturbatively – indeed a UV fixed point at $G_* \sim \varepsilon$ appears). (5) Note criticisms or open issues.

Base Camp 3.5: Asymptotic Safety in Cosmology and Particle Physics

Scope: Explore how the running of $G(k)$ and $\Lambda(k)$ might impact cosmology (early universe, inflation, etc.) or low-energy observables. Also how asymptotic safety might provide a UV completion for inflation (e.g. Starobinsky $R^2$ inflation, which itself has a near-UV fixed point behavior) or solve hierarchy problems.

Stepping Stones: (1) Derive or use the RG-improved Friedmann equation: replace $G$ and $\Lambda$ with $G(k), \Lambda(k)$, and relate $k$ to cosmological scales (e.g. $k \sim 1/t$ or $k\sim (\rho)^{1/4}$). Solve qualitatively: find if there’s a phase where $\Lambda$ is large and $G$ small such that the universe could start in a near de Sitter state and then smoothly connect to classical FRW. (2) Check Starobinsky inflation ($R^2$ term dominated early universe): how asymptotic safety with $R^2$ fixed point might justify the initial conditions for Starobinsky inflation or predict its parameters. (3) Particle physics: mention asymptotic safety approach by Shaposhnikov et al. where gravity+Higgs fixed point can yield a prediction for the Higgs mass (which was ~126 GeV – they predicted near 126, which indeed happened, albeit their original reasoning was more subtle). (4) Could discuss whether AS can solve the triviality of Higgs coupling by providing a fixed point for it, and similarly with the $U(1)$ hypercharge which in the SM alone has a Landau pole.

← Back to Gasherbrum I – Quantum Gravity