Four Fermi Theory in Four Dimensions is Renormalisable

This paper demonstrates that four-dimensional quantum field theories with elementary self-interacting Dirac fermions are renormalizable and predictive at leading order in the large-$N_{\rm f}$ limit, providing exact beta functions and scaling dimensions for extended four-fermion interactions including higher-derivative terms.

The Gist

The Big Picture: Fixing a Leaky Roof

Imagine you are a physicist trying to build a house (a theory of how particles interact). For decades, the standard rule was: "If your roof is made of materials that are too heavy or weak for the weather, the house will collapse."

In physics terms, this meant that certain types of interactions between particles (specifically, four particles bumping into each other at once) were considered "non-renormalizable." This is a fancy way of saying that if you tried to calculate what happens at very high energies (like the Big Bang), the math would blow up with infinite numbers, making the theory useless. Scientists thought these theories were only good as "temporary patches" (effective theories) and that a deeper, more fundamental theory must exist to fix them.

This paper says: "Actually, the house doesn't collapse. We just needed to install a better type of roof."

The authors, Charlie Cresswell-Hogg and Daniel Litim, have proven that these "broken" theories in our 4-dimensional universe (3 space + 1 time) are actually perfectly stable and predictable, provided you look at them through the right lens.

---

The Analogy: The Infinite Tower of Blocks

To understand how they fixed it, let's use an analogy of building a tower with blocks.

1. The Problem: The Wobbly Tower

Imagine you are stacking blocks to represent particle interactions.

• The Old View: You try to stack four blocks together (a "four-fermion" interaction). In a 4D world, the tower gets wobbly. As you add more layers (higher energies), the tower starts to shake violently, and the math predicts it will fall over (infinities).
• The Standard Solution: Usually, physicists say, "Okay, this tower is too wobbly. Let's stop building it here and say there's a hidden machine (new physics) underneath that holds it up."

2. The New Discovery: The Magic Glue

The authors looked at the tower again, but this time they imagined a giant crowd of people (a huge number of particle flavors, $N_f$) helping to build it.

They found that when you have enough people helping, the "wobbly" part of the tower doesn't break. Instead, it transforms.

• The Transformation: The simple four-block stack isn't just a stack anymore. It becomes a smart, self-adjusting structure.
• The Secret Ingredient: To make the math work, they had to add "higher-derivative interactions." In our analogy, this is like adding springs and shock absorbers between the blocks.
  • Instead of just rigid blocks touching, the blocks are connected by springs that can stretch and compress.
  • These springs represent "higher-derivative" terms. They allow the tower to absorb the shock of high energy without breaking.

3. The Result: A Self-Healing Structure

With these springs added, the tower doesn't just survive; it becomes predictable.

• Even though the tower looks complicated (with infinite springs), the authors proved that you only need to tune three specific dials (parameters) to make the whole thing work perfectly.
• The Dials:
  1. The strength of the basic connection.
  2. The stiffness of the springs (higher-derivative terms).
  3. A new type of connection involving eight blocks at once (an 8-fermion interaction).

Once you set these three dials, the tower is stable at any height. You don't need to invent new physics to explain why it doesn't fall. The theory is "UV Complete," meaning it works from the smallest scale to the highest energy imaginable.

---

Why This Matters: The "Asymptotic Safety" Concept

The paper relies on a concept called Asymptotic Safety (originally proposed by Stephen Weinberg).

• The Old Fear: We thought that at high energies, the forces between particles would get so strong that the theory would break.
• The New Reality: The authors show that at high energies, the particles find a "safe harbor" (a fixed point). It's like a car driving up a hill. Usually, if you drive too fast, you fly off the cliff. But in this theory, as you speed up, the road curves back down, keeping the car on the track.

The "quantum effects" (the help from the giant crowd of particles) act like a safety net. They turn interactions that should be dangerous (irrelevant) into interactions that are actually helpful (relevant or marginal).

The "Bottom-Up" vs. "Top-Down" Twist

Usually, scientists think in two ways:

1. Top-Down: We start with a perfect, fundamental theory and see how it looks at low energies.
2. Bottom-Up: We start with what we see (low energy) and assume there's a hidden theory above it.

This paper suggests a surprise: The "Bottom-Up" theory (the one we thought was just a temporary patch) is actually its own Top-Down theory.

It's like finding out that the "temporary patch" on your roof is actually a masterpiece of engineering that works perfectly on its own, without needing a secret attic above it. The "Fermi scale" (the energy where things get weird) isn't a limit where the theory breaks; it's just a crossover point where the physics changes style, but the rules remain consistent.

Summary for the General Audience

1. The Myth: Scientists thought theories involving four particles interacting at once were broken and useless in our 4D universe.
2. The Fix: By looking at them with a specific mathematical tool (large number of particle types), the authors found that these theories are actually robust.
3. The Mechanism: The theories naturally include "springs" (higher-derivative terms) and new connections (8-particle interactions) that absorb the chaos of high energy.
4. The Conclusion: These theories are Renormalizable. They are well-defined, predictable, and don't need new physics to fix them. They are as solid as the theories we already trust, like the Standard Model.

In short: The universe might be more flexible than we thought. What looked like a broken toy is actually a sophisticated machine that works perfectly, provided you know how to turn the right three knobs.

Technical Summary

1. Problem Statement

The central problem addressed is the renormalisability of four-fermion (4F) interactions in four spacetime dimensions ($d=4$).

• Canonical View: In standard perturbative power counting, 4F operators have a mass dimension of $[G] = 2-d = -2$ in $d=4$. Consequently, they are classified as "non-renormalisable" and are typically treated only as effective field theories (EFTs) valid below a specific energy scale (the Fermi scale), requiring an infinite number of counterterms for UV completion.
• Wilsonian View: Kenneth Wilson's notion of renormalisability suggests that interactions can be well-defined if their quantum scaling dimensions are $\leq d$, potentially allowing for interacting Ultraviolet (UV) fixed points (Asymptotic Safety).
• The Gap: While 4F theories are known to be renormalisable and UV-complete in $d=3$ (due to an interacting UV fixed point lowering the scaling dimension), rigorous proofs for $d=4$ have been lacking. Previous studies suggested the possibility of UV fixed points in $d=4$, but definitive conclusions were absent.

2. Methodology

The authors employ a large-$N_f$ expansion (where $N_f$ is the number of fermion flavors) to achieve strict control over the theory, including at strong coupling. Key methodological steps include:

• Template Theory: They use the Gross-Neveu (GN) model as a starting point:
  $$\mathcal{L}_{GN} = \bar{\psi}_a \not{\partial} \psi_a - \frac{1}{2N_f} G (\bar{\psi}_a \psi_a)^2$$
• Divergence Analysis: They systematically analyze UV divergences in the large-$N_f$ limit. Unlike in $d=3$, the bubble integrals in $d=4$ exhibit new divergence structures:
  • Quadratic Divergence: Momentum-independent ($\Lambda^2$).
  • Logarithmic Divergence: Momentum-dependent ($p^2 \ln \Lambda$).
  • Box Diagrams: New logarithmic divergences arising from eight-fermion (8F) vertices.
• Resummation: Instead of using auxiliary bosonic fields (bosonisation), they work directly with fermionic degrees of freedom. They perform geometric resummations of bubble chains to derive exact leading-order beta functions and vertex functions.
• Operator Uplift: To absorb the new momentum-dependent divergences, they "uplift" the point-like interaction to a quasi-local interaction involving an infinite tower of higher-derivative terms.

3. Key Contributions and Results

A. Identification of Necessary Building Blocks

The authors demonstrate that to render the theory finite in $d=4$, the fundamental action must be extended beyond the standard point-like 4F interaction. The renormalisable theory requires:

1. Elementary 4F and 8F Vertices: An elementary 8-fermion vertex is necessary to absorb divergences from box diagrams.
2. Higher-Derivative Interactions: The point-like coupling $G$ must be replaced by a differential operator $F(-\partial^2)$, generating an infinite tower of higher-derivative 4F and 8F operators (Mandelstam descendants).

B. Renormalisation of the 4F Vertex

• The bubble integral in $d=4$ diverges as $B(p^2) \sim a\Lambda^2 + b p^2 \ln \Lambda$.
• The authors introduce a quasi-local vertex function $F(p^2) = (Z p^2 + G^{-1})^{-1}$.
• Upon summing the geometric series of bubble diagrams, the resummed vertex becomes:
  $$V(p^2) = \frac{1}{Z p^2 + G^{-1} + B(p^2)}$$
• Result: The two types of divergences (quadratic and logarithmic) are absorbed entirely by redefining the two parameters $G^{-1}$ and $Z$. This implies that the counterterms for the infinite tower of higher-derivative 4F operators are fixed by just these two parameters.

C. Renormalisation of the 8F Vertex

• Box diagrams (Fig. 1, Diagram B) generate logarithmic divergences proportional to $\ln \Lambda$.
• The authors introduce an elementary 8F vertex with a coupling $\lambda$.
• The quasi-local 8F vertex takes the form:
  $$\mathcal{L}_{8F} \propto \frac{\lambda}{4! N_f^3} [F(-\partial^2)(\bar{\psi}\psi)]^4$$
• Result: The logarithmic divergence from the box diagram is absorbed by the dimensionless coupling $\lambda$.

D. The Extended Gross-Neveu Lagrangian

The final renormalisable Lagrangian is given by:
$$\mathcal{L}_{eGN} = Z_\psi \bar{\psi}_a \not{\partial} \psi_a - \frac{1}{2N_f} (\bar{\psi}_a \psi_a) F(-\partial^2) (\bar{\psi}_b \psi_b) + \frac{\lambda}{4! N_f^3} [F(-\partial^2)(\bar{\psi}_a \psi_a)]^4$$
The theory is defined by only three free parameters: $G$ (related to the Fermi scale), $Z$ (higher-derivative coefficient), and $\lambda$ (8F coupling).

E. Renormalisation Group (RG) and Scaling Dimensions

• Beta Functions: The exact large-$N_f$ beta functions are derived:
  • $\beta_Z = -1/(4\pi^2)$ and $\beta_\lambda = -3/\pi^2$ (universal, logarithmic running).
  • $\beta_g = 2g(1 - g/g^*)$ for the dimensionless coupling $g = \mu^2 G(\mu)$.
• UV Fixed Point: The theory possesses an interacting UV fixed point at $g = g^*$.
• Scaling Dimensions:
  • Classically, the 4F operator is irrelevant ($\Delta = 6$).
  • At the UV fixed point, quantum corrections lower the scaling dimension to $\Delta_{4F}|_* = 2 + O(1/N_f)$, making it relevant.
  • Similarly, the 8F operator becomes marginal ($\Delta_{8F}|_* = 4 + O(1/N_f)$).
• Universality: The critical exponents and the existence of the fixed point are universal, independent of the specific renormalisation scheme (though the value of $g^*$ depends on the scheme).

4. Significance and Implications

• Proof of Renormalisability: This work provides the first rigorous proof that 4F theories in 4D are renormalisable in the Wilsonian sense, provided they include specific higher-derivative and 8F interactions.
• UV Completeness: It demonstrates that "canonically non-renormalisable" theories can be UV-complete. The Fermi scale ($G_F^{-1/2}$) does not represent a breakdown of predictivity but rather a crossover from a weakly coupled IR regime to a strongly coupled UV regime governed by a conformal fixed point.
• Predictivity: Despite the presence of infinitely many higher-derivative operators, the theory remains predictive because all counterterms are determined by a finite number of parameters ($G, Z, \lambda$).
• Model Building: This opens new avenues for Beyond Standard Model (BSM) physics. It suggests that models previously dismissed as effective theories (e.g., Nambu–Jona-Lasinio models, Thirring models, or models without fundamental Higgs scalars) could be fundamental theories valid up to arbitrarily high energies if they possess such interacting UV fixed points.
• Asymptotic Safety: The results strongly support the Asymptotic Safety conjecture, showing that quantum effects can turn irrelevant operators into relevant ones, thereby stabilizing the theory in the UV.

In conclusion, Cresswell-Hogg and Litim establish that four-fermion theories in four dimensions are well-defined, renormalisable, and predictive quantum field theories, fundamentally altering the landscape of viable high-energy physics models.