Rigorous foundations for the renormalization of rational Lagrangians: minimal-singularity superpropagators and the locality of counterterms

The 1971 analysis of Delbourgo, Koller and Salam showed that four-dimensional scalar Lagrangians that are rational functions of the field, of the type $G\phi^{\nu_0}(1+\lambda\phi)^{-\nu_1}$, can be renormalized by a finite class of infinitely many counterterms whenever the Dyson index $\nu_0-\nu_1\le 3$, and become nonrenormalizable beyond that. The argument rests on two unproved supports: the minimal-singularity prescription for the superpropagator — the vacuum expectation value of a time-ordered product of two nonpolynomial vertex functions — rigorously justified only for the two-point exponential coupling by Lehmann and Pohlmeyer, and a locality postulate that discards the infinite-derivative ambiguity terms $\sum_{n=0}^{\infty}a_n(\partial^2)^n\delta^4(x)$ in the chronological products as nonlocal counterterms. The open problem is to put this renormalization theory on rigorous foundations, or to refute it: construct, or disprove the existence or uniqueness of, minimal-singularity superpropagators for general rational interactions; settle the consistency of the locality postulate; and establish the Dyson-index power-counting criterion, including the massive case and a complete classification of subgraph singularities, as a theorem — thereby fixing the renormalization status of the nonpolynomial forms arising from chiral, Yang–Mills and weak-interaction Lagrangians.

Background

Perturbative renormalization theory classifies polynomial interactions such as $\phi^4$ in four dimensions by power counting: loop integrals diverge at large momenta, but the divergences can be absorbed into a finite set of local counterterms — extra local polynomials added to the Lagrangian — leaving physical amplitudes finite. Interactions that are not polynomials in the field, such as rational functions $G\phi^{\nu_0}(1+\lambda\phi)^{-\nu_1}$ — where $G$ and $\lambda$ are coupling constants and $\nu_0,\nu_1$ are integers — or exponentials $e^{\lambda\phi}$, escape this classification: expanding them in powers of the field produces infinitely many vertices, and their behavior at large field values suggests, but does not prove, that ultraviolet divergences might be softened.

During 1965–1970 a program led by Efimov, Delbourgo, Salam and Strathdee developed the superpropagator method for such Lagrangians. Writing the interaction vertex as a Laplace transform in an auxiliary variable $\zeta$, $V(\phi)=\int d\zeta\,f(\zeta)e^{\zeta\phi}$, each internal line of a Feynman-like graph contributes the superpropagator $\langle 0|T\,V(\phi(x))V(\phi(y))|0\rangle$ (the vacuum expectation value of a time-ordered product of two vertex functions, $T$ denoting chronological ordering), whose short-distance singular behavior is controlled by the small-$\zeta$ behavior of $f$. The Dyson index — for the above rational family, the integer $\nu_0-\nu_1$ — plays the role that the mass dimension of a coupling plays for polynomial interactions and governs the ultraviolet degree of the whole theory. Because a distribution agreeing with the naive expression away from $x=y$ can be extended to the diagonal $x=y$ in many ways, the method needs a prescription; the standard one is minimal singularity: among all extensions, choose the one with the weakest singularity at coincident points. Lehmann and Pohlmeyer (1971) made this prescription rigorous for the two-point case of exponential coupling.

Within this framework Delbourgo, Koller and Salam (Annals of Physics 66, 1 (1971)) showed, at heuristic level, that rational Lagrangians with Dyson index at most 3 are renormalizable by a finite class of infinite counterterms, whose form they exhibited explicitly, while index 4 (which resembles a $g\phi^4$ theory at large field) is nonrenormalizable. Their analysis left four explicitly identified gaps: (i) the chronological products of composite nonpolynomial operators are defined only up to additive terms $\sum_{n=0}^{\infty}a_n(\partial^2)^n\delta^4(x)$, where $\delta^4$ is the Dirac delta supported at coincident points and $\partial^2$ is the d'Alembertian; this series sums to an infinite-derivative, hence nonlocal, object, which they discard by imposing a locality postulate, without rigorous general justification; (ii) the minimal-singularity ansatz for superpropagators is rigorously justified only in the two-point exponential case of Lehmann and Pohlmeyer; (iii) the renormalization status of the nonpolynomial forms obtained from chiral, Yang–Mills and weak-interaction Lagrangians by field redefinitions is only indicated; (iv) the power counting in the massive case rests on approximate superpropagator formulas and contour-pole dominance arguments, without controlled treatment of thresholds and subleading contributions or a systematic classification of subgraph singularities for general rational vertices.

Later work bears on these gaps in three distinct ways, none of which completes them. Epstein and Glaser (1972) constructed renormalized perturbation series for nonpolynomial Lagrangians in Jaffe's class — fields whose Fourier transforms decrease so rapidly at large momentum that exponentials of them are well-defined operator-valued distributions — treating the interaction as a formal power series $\sum_r t_r {:}A(x)^r{:}/r!$ in Wick-ordered monomials (normal ordering: creation operators placed to the left of annihilation operators, so that the vacuum expectation value of any single monomial vanishes) and preserving locality at every order; this settles the order-by-order construction but not the resummed rational theory of the source paper. The dedicated superpropagator literature (Blomer and Constantinescu; Vucetich; Houard and Irac-Astaud; Biswas, Malik and Sudarshan; Bollini and Giambiagi; Kapoor, 1971–1976) refined the exponential, zero-mass, higher-dimensional and internal-symmetry cases and then stopped. Finally, the effective-field-theory paradigm — for chiral interactions systematized by Gasser and Leutwyler (1984) — accepts infinitely many local counterterms organized by power counting around a cutoff scale rather than seeking fundamental renormalizability, and the modern rigorous renormalization framework of Brunetti and Fredenhagen (2000) likewise handles interactions order by order as formal series; a recent perturbative analysis of related non-polynomial scalar models by Santonocito and Zappalà (2023) finds them renormalizable but essentially free, or suppressed by an ultraviolet cutoff. The present problem asks for the settlement that none of these lines provides.

Problem Statement

Settle, at the level of rigor of constructive or axiomatic quantum field theory, the renormalization theory of rational interaction Lagrangians in four-dimensional Minkowski space initiated by Delbourgo, Koller and Salam (1971). The problem comprises the following pillars:

(1) Superpropagator construction. For general rational vertex functions $V(\phi)$ — in particular the class $G\phi^{\nu_0}(1+\lambda\phi)^{-\nu_1}$ and sums thereof — define the superpropagator $\langle 0|T\,V(\phi(x))V(\phi(y))|0\rangle$ as a distribution on Minkowski space by the minimal-singularity prescription, extending the rigorous two-point exponential treatment of Lehmann and Pohlmeyer; prove existence and uniqueness of this extension, or exhibit its failure or non-uniqueness; and construct, or prove impossible, the analogous higher multi-point chronological products of composite nonpolynomial operators.

(2) Locality of the counterterm prescription. The chronological products are fixed only up to additive terms $\sum_{n=0}^{\infty}a_n(\partial^2)^n\delta^4(x)$, which the source paper discards by a locality postulate as nonlocal infinite-derivative counter-Lagrangians. Prove that this prescription is consistent — with the axiomatic properties of the perturbative S-matrix (locality, causality, unitarity), and with the recovery of known polynomial renormalizations in the $\lambda\to 0$ limit — or prove that it is inconsistent and exhibit the obstruction.

(3) Power-counting theorem. Establish as a theorem, with a systematic classification of subgraph singularities for general rational vertices and a controlled treatment of thresholds and subleading contributions in the massive case, the Dyson-index criterion: rational Lagrangians with $\nu_0-\nu_1\le 3$ are renormalizable by a finite class of infinite counterterms, while those with $\nu_0-\nu_1>3$ (in particular $=4$) are nonrenormalizable; or demonstrate precisely where this criterion fails and replace it by the corrected statement.

(4) Physical scope. Apply the settled machinery to the nonpolynomial forms obtained from chiral, Yang–Mills and weak-interaction Lagrangians by field redefinitions, respecting their symmetry constraints, and decide their renormalization status under the same locality requirement.

References

1. R. Delbourgo, K. Koller, Abdus Salam, 'Renormalization of rational Lagrangians', Annals of Physics 66(1), 1–19 (1971), DOI 10.1016/0003-4916(71)90182-5, https://doi.org/10.1016/0003-4916(71)90182-5
2. G. V. Efimov, 'Formulation of a scalar quantum field theory with an essentially non-linear interaction', Nuclear Physics 74(3), 657–668 (1965), DOI 10.1016/0029-5582(65)90211-7, https://doi.org/10.1016/0029-5582(65)90211-7
3. R. Delbourgo, Abdus Salam, J. Strathdee, 'Infinities of Nonlinear and Lagrangian Theories', Physical Review 187(5), 1999–2007 (1969), DOI 10.1103/PhysRev.187.1999, https://doi.org/10.1103/PhysRev.187.1999
4. Abdus Salam, J. A. Strathdee, 'Momentum-space behavior of integrals in nonpolynomial Lagrangian theories', Physical Review D 1(12), 3296–3312 (1970), DOI 10.1103/PhysRevD.1.3296, https://doi.org/10.1103/PhysRevD.1.3296
5. H. Lehmann, K. Pohlmeyer, 'On the superpropagator of fields with exponential coupling', Communications in Mathematical Physics 20(2), 101–110 (1971), DOI 10.1007/BF01646529, https://doi.org/10.1007/BF01646529
6. R. J. Blomer, F. Constantinescu, 'On the zero-mass superpropagator', Nuclear Physics B 27(1), 173–192 (1971), DOI 10.1016/0550-3213(71)90089-7, https://doi.org/10.1016/0550-3213(71)90089-7
7. B. W. Keck, J. G. Taylor, 'Renormalization of nonpolynomial lagrangians', Journal of Physics A: General Physics 4(4), 441–455 (1971), DOI 10.1088/0305-4470/4/4/006, https://doi.org/10.1088/0305-4470/4/4/006
8. H. Epstein, V. Glaser, 'Renormalization of non polynomial Lagrangians in Jaffe's class', Communications in Mathematical Physics 27(3), 181–194 (1972), DOI 10.1007/BF01645691, https://doi.org/10.1007/BF01645691
9. H. Vucetich, 'Exponential superpropagator in ν-dimensional space', Lettere al Nuovo Cimento 6(9), 314–316 (1973), DOI 10.1007/BF02743607, https://doi.org/10.1007/BF02743607
10. J. C. Houard, M. Irac-Astaud, 'Some remarks on the exponential superpropagator', Lettere al Nuovo Cimento 7(12), 485–489 (1973), DOI 10.1007/BF02727860, https://doi.org/10.1007/BF02727860
11. S. N. Biswas, G. P. Malik, E. C. G. Sudarshan, 'Superpropagator for a nonpolynomial field', Physical Review D 7(10), 2884–2886 (1973), DOI 10.1103/PhysRevD.7.2884, https://doi.org/10.1103/PhysRevD.7.2884
12. C. G. Bollini, J. J. Giambiagi, 'On the exponential superpropagator', Journal of Mathematical Physics 15(1), 125–128 (1974), DOI 10.1063/1.1666489, https://doi.org/10.1063/1.1666489
13. A. K. Kapoor, 'On superpropagator for nonpolynomial Lagrangians with internal symmetry', Journal of Mathematical Physics 17(1), 61–62 (1976), DOI 10.1063/1.522783, https://doi.org/10.1063/1.522783
14. H. P. W. Gottlieb, 'Spontaneous breaking of gauge symmetries with non-linear realizations', Nuclear Physics B 54(2), 509–521 (1973), DOI 10.1016/0550-3213(73)90435-5, https://doi.org/10.1016/0550-3213(73)90435-5
15. H. P. W. Gottlieb, 'Massive gauge theories and the photon', Nuclear Physics B 73(2), 257–268 (1974), DOI 10.1016/0550-3213(74)90019-4, https://doi.org/10.1016/0550-3213(74)90019-4
16. M. La Camera, A. Wataghin, 'Nonpolynomial interactions at high energies', Il Nuovo Cimento A 17(4), 526–534 (1973), DOI 10.1007/BF02786832, https://doi.org/10.1007/BF02786832
17. T. Dass, R. Shyam, 'Conformal-invariant finite quantum electrodynamics', Physical Review D 15(6), 1580–1588 (1977), DOI 10.1103/PhysRevD.15.1580, https://doi.org/10.1103/PhysRevD.15.1580
18. J. Gasser, H. Leutwyler, 'Chiral perturbation theory to one loop', Annals of Physics 158(1), 142–210 (1984), DOI 10.1016/0003-4916(84)90242-2, https://doi.org/10.1016/0003-4916(84)90242-2
19. R. Brunetti, K. Fredenhagen, 'Microlocal analysis and interacting quantum field theories: renormalization on physical backgrounds', Communications in Mathematical Physics 208(3), 623–661 (2000), DOI 10.1007/s002200050004, arXiv:math-ph/9903028, https://arxiv.org/abs/math-ph/9903028
20. A. Santonocito, D. Zappalà, 'On the renormalization of non-polynomial field theories', European Physical Journal C 83(9), 807 (2023), DOI 10.1140/epjc/s10052-023-11986-x, arXiv:2307.06068, https://arxiv.org/abs/2307.06068