Constructing P(φ)₂ and φ⁴₃ Models

The scalar models $P(\phi)_2$ and $\phi^4_3$ are genuine constructive QFT successes, but their proofs and conclusions are dimension-specific. In two dimensions Wick ordering and volume control suffice for stable polynomial interactions; in three dimensions an ultraviolet cutoff and divergent local counterterms must be controlled. Neither theorem is a template that proves four-dimensional existence.

Required background. Interacting measures, stability, and Wick ordering supplies finite-volume densities; cluster expansions and correlation inequalities supplies one route to uniform limits.

Helpful background. The constructive program and cutoff removal organizes the limiting steps; the OS reconstruction theorem explains the relativistic conclusion.

Two-dimensional polynomial interactions

Let $P$ be a real polynomial bounded below, $C=(-\Delta+m_0^2)^{-1}$ on $\mathbb R^2$, and $\Lambda$ a bounded rectangle. The finite-volume measure

$$d\mu_\Lambda=Z_\Lambda^{-1}\exp\!\left[-\lambda\int_\Lambda :P(\phi(x)):_C d^2x\right]d\mu_C, \qquad \lambda\ge0,$$

is well defined. For sufficiently small $\lambda/m_0^2$, cluster expansions give weak convergence as $\Lambda\uparrow\mathbb R^2$, OS axioms, exponential clustering, and an isolated one-particle mass. Correlation inequalities extend existence results for suitable even ferromagnetic polynomials without the same small-coupling restriction. The quantifiers matter: the result ranges over bounded-below $P$ for the finite-volume construction, but the strongest spectral conclusions use additional weak-coupling and phase hypotheses. Summers 2016, §3.1, pp. 10–14 separates these results.

Non-Gaussianity can be checked by a connected four-point function or by the interacting field equation; it is not established merely because $\lambda$ occurs in the density. OS reconstruction supplies a Wightman model only after reflection positivity, Euclidean invariance, regularity, and clustering have been verified for the limiting Schwinger hierarchy.

Three-dimensional quartic construction

In $d=3$, introduce an ultraviolet covariance $C_\kappa$ suppressing momenta above $\kappa$ and a region $\Lambda$. A schematic regulated action is

$$V_{\Lambda,\kappa}=\int_\Lambda \left(\lambda:\!\phi_\kappa^4\!:+\delta m^2_\kappa:\!\phi_\kappa^2\!:+E_\kappa\right)d^3x.$$

The mass counterterm and vacuum-energy term diverge with $\kappa$ in the prescribed way. Phase-cell and cluster estimates control the remaining effective interactions across scales. For weak coupling the Schwinger functions converge as $\kappa\to\infty$ and $\Lambda\uparrow\mathbb R^3$, satisfy the OS axioms, and reconstruct a Wightman theory. Other methods remove the weak-coupling restriction for existence. Feldman and Osterwalder’s primary theorem proves the Wightman axioms and a mass gap for weakly coupled $\phi^4_3$; Feldman and Osterwalder 1976, pp. 80–135. The regulator and counterterm form, simultaneous limiting claim, and later extensions are summarized in Summers 2016, §3.2, pp. 16–17.

The proof mechanism can be read scale by scale. Split $C_\kappa$ into finite-range or momentum-shell covariances. Integrating the highest shell changes the relevant mass and constant terms and generates irrelevant local interactions. Counterterms tune the relevant coordinates, while a norm contracts the irrelevant remainder. Iteration yields bounds uniform in the number of shells; the final long-distance integration supplies volume control. This is far more than perturbative cancellation at each fixed order.

The first worked application is precisely $\phi^4_3$: finite volume, smooth momentum cutoff, tuned mass and vacuum terms, and convergence of non-Gaussian Schwinger functions. Its status belongs with rigorous status, construction, and open problems, where it must remain distinct from unsolved four-dimensional claims.

Dimension is a theorem hypothesis

Power counting assigns the quartic coupling dimension $4-d$. It is positive in $d=2,3$, zero in $d=4$, and negative above four. The finite list of relevant ultraviolet counterterms in the superrenormalizable cases underlies their constructive control. In $d=4$, marginal flow and possible triviality change the problem. Using the three-dimensional counterterms in four dimensions leaves logarithmic divergences and supplies no uniform continuum estimate. Summers 2016, §3.3, pp. 17–20 explains this dimensional boundary.

An independent check is dimensional: $[\phi]=(d-2)/2$, so $[\lambda]=d-4[\phi]=4-d$. A proposed $d=3$ formula whose mass counterterm has the wrong engineering dimension cannot be correct. A second check takes $\lambda\to0$: connected correlations above order two must tend to zero and the two-point function to the massive covariance after the chosen mass normalization.

Adversarial dimensional test

Replace $d=3$ by $d=4$ but keep only the $\phi^4_3$ mass and vacuum counterterms. The power-counting hypothesis has changed, wave-function and coupling renormalization enter, and the scale norm no longer contracts in the same way. The surviving conclusion is only that each regulated finite-volume integral exists. No interacting four-dimensional continuum QFT follows.

This dimensional check must precede any comparison of formulas or perturbative coefficients.

Exercises

1. Relevant coupling. Compute $[\lambda]$ for $\lambda\phi^4$ in $d=2,3,4$.

Solution

Since $[\phi]=(d-2)/2$, $[\lambda]=4-d$: respectively $2,1,0$. Positive dimension marks the two superrenormalizable cases; zero marks the marginal four-dimensional case.

2. Non-Gaussian check. Why does a nonzero connected four-point function rule out a Gaussian limiting measure?

Solution

All Gaussian cumulants of order greater than two vanish by Wick’s theorem. A finite nonzero fourth cumulant therefore separates the limit from every Gaussian law with the same two-point function.

References

• Feldman, Joel, and Konrad Osterwalder. “The Wightman Axioms and the Mass Gap for Weakly Coupled $\phi^4_3$ Quantum Field Theories.” Annals of Physics 97 (1976): 80–135. DOI.
• Glimm, James, and Arthur Jaffe. “A $\lambda\phi^4_2$ Quantum Field Theory Without Cutoffs. II.” Annals of Mathematics 91 (1970): 362–401. DOI.
• Summers, Stephen J. “A Perspective on Constructive Quantum Field Theory.” arXiv:1203.3991, revised 2016. Open PDF.