PHYSICAL REVIEW LETTERS 125, 231601 (2020)

Forced Soliton Equation and Semiclassical Soliton Form Factors

*
Ilarion V. Melnikov
Department of Physics and Astronomy, James Madison University, 901 Carrier Drive, Harrisonburg, Virginia 22807, USA
†
Constantinos Papageorgakis
CRST and School of Physics and Astronomy, Queen Mary University of London, Mile End Road, London E1 4NS, United Kingdom
‡
Andrew B. Royston
Department of Physics, Penn State Fayette, The Eberly Campus, 2201 University Drive, Lemont Furnace, Pennsylvania 15456, USA
(Received 21 August 2020; accepted 19 October 2020; published 1 December 2020)

We show that the leading semiclassical behavior of soliton form factors at arbitrary momentum transfer
is controlled by solutions to a new wavelike integro-differential equation that describes solitons undergoing
acceleration. We work in the context of two-dimensional linear σ models with kink solitons for
concreteness, but our methods are purely semiclassical and generalizable.

DOI: 10.1103/PhysRevLett.125.231601

Introduction.—Solitons feature prominently in quantum carrying out perturbative (semiclassical) computations of

field theories relevant to our current understanding of
nature: the Abrikosov and Nielsen-Olesen vortices in the
theories of superconductors [1] and dual strings [2],
baryons in the Skyrme model of nucleonic interactions
[3], and magnetic monopoles in grand unified theories
[4,5]. Furthermore, through semiclassical analysis, solitons
provide a direct and appealing connection between
solutions to nonlinear classical equations of motion and
related quantum field-theoretical observables.
In classical field theory on Minkowski spacetime,
topological solitons arise when the field equations admit
a set of topologically distinct boundary conditions at spatial
infinity, preserving finiteness of the energy. For example,
kink solitons exist in two-dimensional models of a single
scalar field, ϕ ¼ ϕðt; xÞ, with potential VðϕÞ, when the set
of global minima of the potential has multiple components.
The kink is a time-independent energy-minimizing solution
to the equations of motion, ϕ ¼ ϕ0 ðxÞ, that interpolates
between two distinct minima as x → ∞.
These disjoint sectors in the space of finite-energy field
configurations lead to orthogonal sectors of the Hilbert
space of quantum states, and a classical soliton solution
corresponds to a one-particle state in a topologically
nontrivial sector [6]. In the two-dimensional scalar field
example, the soliton state is fully specified by an on-shell
momentum P and will be denoted jΨP ⟫. The framework
for defining soliton states in quantum field theory, and for
observables involving soliton states, was developed in the
mid 1970s [6–13]. For classic reviews, see Refs. [14–16].
Despite the maturity of the subject, however, seemingly
simple yet profound questions remain unanswered.
Consider the case of soliton form factors—matrix elements
of field operators between initial and final soliton states,
such as ⟪ΨPf jϕ̂ðt; xÞjΨPi ⟫. The leading semiclassical
behavior of this form factor for generic momentum transfer,
k ≡ Pf − Pi , is not known in nonintegrable models. For
momentum transfers much less than the soliton mass,
k ≪ M, the semiclassical form factor is given by the
Fourier transform of the classical soliton profile ϕ0 [6].
But naive attempts to extrapolate this result to momentum
transfers k ∼ OðMÞ fail upon comparing to results from
integrable models, such as sine-Gordon [17].
In the intervening years, especially after Ref. [18], most
work on solitons in nonintegrable quantum field theories
has centered on quantum-exact results at leading order in
the time-derivative, or adiabatic, expansion. With some
notable exceptions [19–21], very little progress has been
made in the opposite limit of high-momentum transfer but
small coupling.
The importance—and the underlying difficulty—of the
form factor computation is illuminated by considering its role
with respect to crossing symmetry of the quantum field theory.
On the one hand, the soliton form factor determines the
amplitude, AðPi ; k → Pf Þ, for a soliton to absorb a pertur-
bative particle with momentum k created by the field ϕ:

Published by the American Physical Society under the terms of

the Creative Commons Attribution 4.0 International license. ið2πÞ2 δð2Þ ðk þ Pi − Pf ÞAðPi ; k → Pf Þ
Further distribution of this work must maintain attribution to Z
μ
the author(s) and the published article’s title, journal citation, ¼ d2 xeikμ x ⟪ΨPf jϕ̂ðt; xÞjΨPi ⟫: ð1Þ
and DOI. Funded by SCOAP3.

0031-9007=20=125(23)=231601(6) 231601-1 Published by the American Physical Society

 PHYSICAL REVIEW LETTERS 125, 231601 (2020)

On the other hand, by crossing symmetry, this amplitude is We assume that the minima of V 0 are gapped and
equal to the amplitude for a perturbative particle to create a associated to a spontaneously broken discrete symmetry;
virtual soliton-antisoliton pair: the parameter m0 controls the mass gap to the perturbative
spectrum. The coupling g is a parameter in the potential,
AðPi ; k → Pf Þ ¼ Aðk → Pf ; −P̄i Þ: ð2Þ and we consider potentials with the scaling property
1
Hence, if one had access to the soliton form factor at V 0 ðm0 ; ϕÞ ¼ Ṽ 0 ðm0 ; ϕ̃Þ; ð4Þ
g2
momentum transfers k ≳ OðMÞ, and could analytically
continue in k to the kinematically allowed region for the where Ṽ 0 does not depend on g and ϕ̃ ¼ gϕ.
pair creation process, then one could determine the pair In such theories there exists a standard renormalization
creation amplitude. This, in turn, would allow one to scheme in which, to all orders in perturbation theory, only
quantitatively address deep and difficult questions in the mass parameter is renormalized:
quantum field theory, such as, what is the leading
contribution of soliton-antisoliton pairs running in loops m20 ¼ m2 þ Δm2 ; ð5Þ
to processes involving perturbative particles? See
Ref. [22] for a recent discussion of this issue, which with m2 a finite mass parameter and Δm2 the coefficient of
has received renewed attention in light of various con- the mass counterterm V Δm2 ðϕÞ. In this scheme the renor-
jectures and computations in certain maximally super- malized potential takes the form
symmetric gravitational and higher-dimensional gauge
theories [23–27]. VðϕÞ ¼ V 0 ðm; ϕÞ þ V Δm2 ðϕÞ; ð6Þ
In this Letter, we report on progress in determining
semiclassical soliton form factors at arbitrary momentum where V Δm2 is Oðg2 Þ [30]. Hence, when we speak of the
transfer. We follow the phase space path integral formalism classical soliton profile ϕ0 ðxÞ, we mean a time-independent
of Ref. [12] and carry out a saddle-point analysis in the one- solution to the equation of motion:
soliton sector, keeping all time derivatives. This results in
dV 0
an expansion around solutions to the forced soliton ∂ 2x ϕ − ðm; ϕÞ ¼ 0: ð7Þ
equation, a novel second-order wavelike integro-differ- dϕ
ential equation for an accelerating soliton.
An example to keep in mind is ϕ4 theory, where
Although we do not currently know of explicit time-
dependent solutions to this equation, we can nevertheless 8
1
make progress by expanding around a hypothetical >
< V0 ¼ g2
ðg2 ϕ2 − 14 m2 Þ2
solution, by analogy with standard collective coordinate ϕ4 theory∶ ð8Þ
>
: ϕ0 ¼ mffiffi
2g tanh ( 2 ðx − XÞ):
m p
reductions for solitons at small velocity [28].
Using this approach, we construct the generating func-
tional for semiclassical soliton form factors depending on a The constant X represents the kink position.
source FðtÞ. The generating functional is built from a The renormalized Hamiltonian for any theory in this
solution to the forced soliton equation, where the soliton class takes the form
momentum is dictated by Newton’s second law, with the
Z  
source playing the role of the force, P_ ¼ F. The semi- 1 1
classical limit of the soliton form factor of any local H ¼ dx π 2 þ ð∂ x ϕÞ2 þ VðϕÞ ; ð9Þ
2 2
operator Ô is given by the action of a functional derivative
operator, f O ¼ f O ½δ=δF, acting on the generating func- in terms of which the phase-space path-integral represen-
tional. For any local O we determine the function f O tation of matrix elements is
explicitly in terms of the classical soliton profile.
This result demonstrates that the forced soliton equation ⟪Ψf jÔjΨi ⟫
is the key to unlocking the semiclassical limit of soliton Z R R
_
form factors at arbitrary momentum transfer. An expanded ¼ ½DϕDπΨf ½ϕ Ψi ½ϕei dtð dxπϕ−HÞ O½π; ϕ; ð10Þ
version of this Letter can be found in Ref. [29].
Saddle-point approximation in the soliton sector.—We
work in the class of two-dimensional linear σ models with where Ô is any local operator inserted at some spacetime
classical action: point ðt; xÞ, and Ψi;f ½ϕ are initial and final state wave
functionals.
Z   To define soliton states, we must work in variables
2 1 μ
S¼ d x − ∂ μ ϕ∂ ϕ − V 0 ðm0 ; ϕÞ : ð3Þ appropriate for the soliton sector. This is achieved by a
2 canonical transformation,

231601-2
 PHYSICAL REVIEW LETTERS 125, 231601 (2020)

ðπ; ϕÞ ↦ ðP; ϖ; X; φÞ; ð11Þ 1

ΨP ½X; φ ¼ pffiffiffiffiffiffi eiPX Ψ0 ½φ; ð19Þ
2π
where the kink position X has been promoted to a
dynamical variable (the collective coordinate), P is its where Ψ0 ½φ is the ground state wave functional in the φ-ϖ
conjugate momentum, and ðϖ; φÞ encapsulate the remain- sector. The Hamiltonian (17) is nonlocal in space, but local
ing field-theoretical degrees of freedom. The latter are in time, and a diagrammatic Feynman perturbation theory
constrained to satisfy was developed for it in Ref. [12] under the assumption of
small soliton velocity, P=M 0 ∼ OðgÞ.
Z Z
In this Letter our objective is instead to carry out a
dxψ 0 ϖ ¼ 0 and dxψ 0 ðφ − ϕ0 Þ ¼ 0; ð12Þ saddle-point analysis in the φ-ϖ sector, treating PðtÞ as an
arbitrary background field. The classical equations of
where motion in this sector are the constraints
Z hψ 0 jφ − ϕ0 i ¼ 0 and hψ 0 jϖi ¼ 0; ð20Þ
1
ψ 0 ≔ pffiffiffiffiffiffiffi ∂ x ϕ0 ; M0 ≔ dxð∂ x ϕ0 Þ2 ð13Þ
M0
together with
are the normalized zero-mode and classical soliton mass. φ_ ¼ ϖ þ νψ 0 þ βφ0 ;
Explicitly, the canonical transformation is [12,13,29]
ð1Þ
_ ¼ −λψ 0 þ βϖ 0 þ φ00 − V 0 ðφÞ − ψ 00 β2 hψ 0 jφ0 i:
ϖ ð21Þ
ðP þ hϖjφ0 iÞ
πðt; xÞ ¼ − ψ 0 (x − XðtÞ) þ ϖ(t; x − XðtÞ); ð1Þ
The V 0 ðφÞ denotes ∂V 0 ðm; φÞ=∂φ, and we have intro-
hψ 0 jφ0 i
duced the soliton velocity functional:
ϕðt; xÞ ¼ φ(t; x − XðtÞ): ð14Þ
P þ hϖjφ0 i
Here, β½φ; ϖ; P ≔ : ð22Þ
hψ 0 jφ0 i2
Z
The moniker is apt since Hamilton’s equations in the full
hfjgi ≔ dρfðt; ρÞ gðt; ρÞ; ð15Þ
theory give ∂HT =∂P ¼ β.
Let ðϖ; λ; φ; νÞ ¼ ðϖ̄; λ̄; φ̄; ν̄Þ denote a solution to
where ρ ≔ x − XðtÞ is the comoving coordinate, and Eqs (20) and (21). We find that, on a solution, the velocity
φ0 ≡ ∂ ρ φ, φ_ ≡ ∂ t φ. functional is
In terms of the new variables, the matrix element (10)
takes the form P þ hφ̄_ jφ̄0 i
β½φ̄; ϖ̄; P ¼ ≕ β̄½φ̄; P; ð23Þ
Z Z hφ̄0 jφ̄0 i
⟪Ψf jÔjΨi ⟫ ¼ ½DXDP ½DφDϖDνDλ and that
× Ψf ½X; φ Ψi ½X; φ
R P_ d
_ _ ν̄ ¼ −β̄hψ 0 jφ̄0 i; λ̄ ¼ − ðβ̄hψ 0 jφ̄0 iÞ;
× ei dtðPXþhϖjφi−H TÞ
O½P; ϖ; X; φ; ð16Þ 0
hψ 0 jφ̄ i dt
ϖ̄ ¼ φ̄_ − β̄ðφ̄0 − hψ 0 jφ̄0 iψ 0 Þ; ð24Þ
with “total Hamiltonian”
while φ̄ is a solution to the forced soliton equation:
ðP þ hϖjφ0 iÞ2
HT ¼ λhψ 0 jφ − ϕ0 i þ νhψ 0 jϖi þ _
2hψ 0 jφ0 i2 ð1Þ PðtÞψ 0 ðρÞ
Z   ð∂ t − β̄½φ̄;P∂ ρ Þ2 φ̄ − φ̄00 þ V 0 ðφ̄Þ þ ¼ 0; ð25Þ
1 1 hψ 0 jφ̄0 i
þ dρ ϖ 2 þ φ02 þ VðφÞ ; ð17Þ
2 2 additionally satisfying the constraint hψ 0 jφ̄ − ϕ0 i ¼ 0. The
left-hand side of Eq. (25) takes values in the orthogonal
where λðtÞ, νðtÞ are Lagrange multipliers enforcing the complement of Spanfφ̄0 g with respect to the inner product
constraints (12). A soliton state takes the form h; i, and hence the system is not overconstrained.
If one assumes that P is constant, then it is consistent
jΨP ⟫ ¼ jP⟫ ⊗ jΨ0 ⟫ ð18Þ with the equations of motion to also assume that
ðϖ̄; λ̄; φ̄; ν̄Þ are time independent. The solution to
and has wave functional Eq. (25) is then the boosted soliton profile,

231601-3
 PHYSICAL REVIEW LETTERS 125, 231601 (2020)
R Z
φ̄ðρÞ ¼ ϕ0 (γ̄ðρ − ρ0 Þ); ð26Þ −i dtH eff ½P
e ≔ ½DφDϖDνDλ
R
 i dtðhϖjφi−H
_ TÞ
with the Lorentz factor related to the relativistic momentum × Ψ0;f ½φ Ψ0;i ½φe :
as expected [12]:
ð31Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
γ̄ ¼ 1 þ ðP=M0 Þ2 : ð27Þ The effective Hamiltonian admits an expansion in g of the
form

The integration constant ρ0 is fixed by the λ constraint ð−2Þ ð0Þ

H eff ¼ Heff þ Heff þ OðgÞ ≡ Hsc þ OðgÞ; ð32Þ
(ρ0 ¼ 0 in ϕ4 theory).
The assumption of constant P is valid for the transition such that Hsc captures the tree-level Oðg−2 Þ and one-loop
amplitude—translational invariance guarantees it—but not Oðg0 Þ contributions. These in turn can be evaluated in terms
in the presence of operator insertions: setting P constant of the solution, φ̄, to the forced soliton equation. We find that
leads to results for form factors that are only valid at leading
order in a momentum-transfer (k=M0 ) expansion. Z  
ð−2Þ 1 1 2
To continue, we will assume that, for a given PðtÞ, there Heff ¼ 2 02 _
dρ ð1 þ β̄ Þφ̄ − φ̄ þ V 0 ðm; φ̄Þ : ð33Þ
2 2
is a unique solution to the initial value problem associated
to the system (25) with constraint hψ 0 jφ̄ − ϕ0 i ¼ 0. The Gaussian path integral for the one-loop contribution Heff
ð0Þ
Specifically, under the assumption that PðtÞ is constant is regularized and evaluated in Ref. [29]; we will not need the
for early enough t with Pðti Þ ¼ Pi , the initial data will be details here. In the limit of constant P, Hsc is consistent with
given by the time-independent solution for Pi just the relativistic energy expanded through one loop.
discussed. The boundary conditions for the soliton as Semiclassical soliton form factors.—Equation (29) may
ρ → ∞ will be the standard ones that follow from be stated succinctly as
finiteness of the energy.
Now we expand the Lagrangian, hϖjφi _ − HT , as well as
⟪ΨPf jÔ½π̂; ϕ̂jΨPi ⟫ ¼ ⟪Pf jÔsc ½P̂; X̂jPi ⟫½1 þ OðgÞ; ð34Þ
the insertion O, in fluctuations around the solution:
for a Weyl-ordered operator Ôsc ½P̂; X̂. The dependence of
ðϖ; λ; φ; νÞ ¼ ðϖ̄ þ δϖ; λ̄ þ δλ; φ̄ þ δφ; ν̄ þ δνÞ: ð28Þ Osc on the spacetime insertion point ðt; xÞ occurs in a rather
distinct way. While t appears as the argument of PðtÞ, XðtÞ
The expansion in fluctuations is a g expansion: one in the insertion, x occurs only through the combination
observes that Eqs. (24) and (25) are consistent with all x − X, as dictated by the canonical transformation (14)
background fields being Oðg−1 Þ, with order one veloc- evaluated on the background solution ðϖ; φÞ ¼ ðϖ̄; φ̄Þ.
ities β̄. We are thus led to consider a general class of quantum-
The leading-in-g result for the form factor (16) reduces to mechanical matrix elements ⟪Pf jf̂½P̂; x − X̂jPi ⟫ for a
a quantum-mechanical matrix element and is compactly Weyl-ordered operator f̂ (and corresponding phase-space
expressed in terms of a semiclassical effective Hamiltonian function f). This motivates the definition of the generating
for the soliton and a semiclassical insertion: functional,
Z
Z 1
1 F Pf ;Pi ½K; fF; xg ≔ ½DXDPeiðPi Xi −Pf Xf Þ
⟪ΨPf jÔjΨPi ⟫ ¼ ½DXDP 2π
2π Rt
_
i f dt0 fPX−H sc ½P−PK−ðx−XÞFg
R × e ti ;
_ sc ½PÞ
× eiðPi Xi −Pf Xf Þ ei dtðPX−H
Osc ½P; X
× ½1 þ OðgÞ: ð29Þ in terms of which

⟪Pf jf̂ðP̂; x − X̂ÞjPi ⟫

The insertion is given by evaluating O on the solution,    
δ δ
¼ f i ;i F Pf ;Pi ½K; fF; xg : ð36Þ
δKðtÞ δFðtÞ
Osc ½P; X ≔ O½P; ϖ̄; X; φ̄: ð30Þ K;F¼0

We refer to F as the generator of semiclassical soliton form

Meanwhile, Hsc is determined by a one-loop saddle-point factors. Since Hsc is X independent, the path integral (35)
approximation to the soliton effective Hamiltonian Heff ½P can be evaluated. The X integral enforces Newton’s second
defined through law, via δ½P_ − F, leading to

231601-4
 PHYSICAL REVIEW LETTERS 125, 231601 (2020)

F Pf ;Pi ½K;fF;xg A natural first step in studying solutions to Eq. (25) is to

 Z t  Rt 0 consider perturbation theory in small P, _ where the com-
−iðPf −Pi Þx −i ti dt fH sc ½P̄þP̄Kg
f f
0 0
¼ δ Pf −Pi − Fðt Þdt e e : plete diagonalization of the linearized problem around a
ti constant P solution obtained in Ref. [29] should be useful.
ð37Þ It would be interesting to see if the framework of Ref. [31]
can shed light on the structure of solutions to the forced
Here, all F dependence is contained in P̄ðtÞ—obtained as a soliton equation. The methods and results presented here
solution to the second law—and the single remaining δ are generalizable to other classes of theories with solitons.
function imposes the impulse-momentum theorem. In the An important technical prerequisite is the exact canonical
presence of this δ function, the solution P̄ðtÞ can be written as transformation from perturbative to soliton sector phase-
space variables. The latter is known for the magnetic
Z Z  monopole in Yang-Mills-Higgs theory [32].
0 1 1 t0 tf
P̄ðt Þ ¼ ðPi þ Pf Þ þ − dt̃Fðt̃Þ; ð38Þ
2 2 ti t0 We thank Jeff Harvey for comments on the manuscript.
We acknowledge partial financial support from the NSF
implying the useful fact PHY-1914505, the Royal Society UF120032, and the
STFC ST/P000754/1.
δP̄ðtÞ
¼ 0: ð39Þ
δFðtÞ
*
melnikix@jmu.edu
In order to apply Eq. (37) to evaluate the semiclassical †
c.papageorgakis@qmul.ac.uk
form factor, Eq. (36), we need to investigate the functional ‡
abr84@psu.edu
derivatives of F with respect to K and F. The dependence [1] A. Abrikosov, Sov. Phys. JETP 5, 1174 (1957).
of F on K is simple and, thanks to Eq. (38), allows for an [2] H. B. Nielsen and P. Olesen, Nucl. Phys. B61, 45 (1973).
explicit evaluation of all K derivatives, resulting in [29] [3] T. Skyrme, Nucl. Phys. 31, 556 (1962).
[4] G. ’t Hooft, Nucl. Phys. B79, 276 (1974).
[5] A. M. Polyakov, JETP Lett. 20, 194 (1974).
⟪Pf jf̂ðP̂; x − X̂ÞjPi ⟫
    [6] J. Goldstone and R. Jackiw, Phys. Rev. D 11, 1486 (1975).
Pi þ Pf δ [7] R. F. Dashen, B. Hasslacher, and A. Neveu, Phys. Rev. D
¼ f ;i F Pf ;Pi ½0; fF; xg : ð40Þ
2 δFðtÞ F¼0
10, 4114 (1974).
[8] R. F. Dashen, B. Hasslacher, and A. Neveu, Phys. Rev. D
10, 4130 (1974).
This result is significant: for phase-space functions of the [9] J.-L. Gervais and B. Sakita, Phys. Rev. D 11, 2943
form f O ¼ Osc , we can use the constant P ¼ 12 ðPi þ Pf Þ (1975).
solution for ϖ̄, φ̄. Thus the differential operator [10] C. G. Callan, Jr. and D. J. Gross, Nucl. Phys. B93, 29 (1975).
f O ½12 ðPi þ Pf Þ; iðδ=δFÞ appearing on the right-hand side [11] N. Christ and T. Lee, Phys. Rev. D 12, 1606 (1975).
of Eq. (40) will be known explicitly, provided the standard [12] J.-L. Gervais, A. Jevicki, and B. Sakita, Phys. Rev. D 12,
soliton solution is known. 1038 (1975).
For example, in the case Ô ¼ ϕ̂, f O is given by the [13] E. Tomboulis, Phys. Rev. D 12, 1678 (1975).
boosted soliton profile (26), with Lorentz factor expressed [14] R. Jackiw, Rev. Mod. Phys. 49, 681 (1977).
in terms of the momentum P ¼ 12 ðPi þ Pf Þ, and spatial [15] L. Faddeev and V. Korepin, Phys. Rep. 42, 1 (1978).
[16] R. Rajaraman, Solitons and Instantons (North-Holland
argument ρ ¼ x − X replaced with the derivative operator
Publishing Co., Amsterdam, 1982), ISBN 0-444-86229-3.
i½δ=δFðtÞ. All nontrivial dependence of the form factor on [17] P. Weisz, Phys. Lett. 67B, 179 (1977).
the momentum transfer is contained in the generating [18] N. Seiberg and E. Witten, Nucl. Phys. B426, 19 (1994).
functional F Pf ;Pi ½0; fF; xg, which depends on the solution [19] A. K. Drukier and S. Nussinov, Phys. Rev. Lett. 49, 102
to the forced soliton equation with time-dependent (1982).
PðtÞ ¼ P̄ðtÞ, given by Eq. (38). It can be shown that [20] C. Bachas, Teor. Mat. Fiz. 95, 187 (1993) [Theor. Math.
Eq. (40) reproduces known results in the low-momentum Phys. 95, 491 (1993)].
transfer limit, jkj ≪ M0 [29]. [21] S. V. Demidov and D. G. Levkov, J. High Energy Phys. 11
Conclusions.—We have obtained a novel equation—the (2015) 066.
forced soliton equation. We have shown that understanding [22] C. Papageorgakis and A. B. Royston, J. High Energy Phys.
09 (2014) 128.
the solutions of the forced soliton equation is the key to
[23] M. R. Douglas, J. High Energy Phys. 02 (2011) 011.
understanding the semiclassical behavior of soliton form [24] N. Lambert, C. Papageorgakis, and M. Schmidt-
factors away from the low momentum-transfer limit. Thus Sommerfeld, J. High Energy Phys. 01 (2011) 083.
we have, after 40 years, a concrete starting point to address [25] T. Banks, arXiv:1205.5768.
profound questions concerning the nonperturbative struc- [26] G. Bossard and A. Kleinschmidt, J. High Energy Phys. 01
ture of quantum field theory. (2016) 164.

231601-5
 PHYSICAL REVIEW LETTERS 125, 231601 (2020)

[27] G. Bossard and A. Kleinschmidt, J. High Energy Phys. 03
(2018) 100.
[28] N. Manton, Phys. Lett. 110B, 54 (1982).
[29] I. V. Melnikov, C. Papageorgakis, and A. B. Royston,
companion paper, Phys. Rev. D 102, 125002 (2020).
[30] A. Rebhan and P. van Nieuwenhuizen, Nucl. Phys. B508,
449 (1997).
[31] T. J. Hughes, T. Kato, and J. E. Marsden, Arch. Ration.
Mech. Anal. 63, 273 (1977).
[32] E. Tomboulis and G. Woo, Nucl. Phys. B107, 221 (1976).

231601-6