S-duality and Montonen–Olive

S-duality is the proposed equivalence between a gauge theory at coupling g and a theory at a coupling of order 1/g in which electrically charged particles and magnetic monopoles exchange roles, so that a strongly coupled theory has a weakly coupled description. It grows out of a symmetry of Maxwell’s equations in empty space, which are unchanged when the electric and magnetic fields are exchanged. Montonen and Olive proposed the quantum version in 1977. The checkpoint records it as exact in maximally supersymmetric Yang–Mills theory, one of the few theories in which it has passed non-trivial tests, and Seiberg’s related dualities supply the “magnetic quarks” of a dual description. In string theory S-duality sits with T-duality inside a larger U-duality. The program uses S-duality as a structural comparison, because it acts on the coupling through the same matrix as Born reciprocity, and it carries a standing caution: outside special supersymmetric theories the duality is an expectation, not an established symmetry.

Status: Established — [T] — the checkpoint cites Montonen–Olive 1977 and Seiberg 1995 as [18†], from model memory, so the citations await the † literature battery. The round-3 audit checked the citations of the withdrawn #14 to Montonen–Olive, Osborn 1979 and Harvey (hep-th/9603086) and found that they support the exactness caveat (Step 1). The explanations below that go beyond the documents are textbook physics [T], and the references they add are marked †.

Duality in Maxwell’s equations [T]#

Without charges, Maxwell’s equations keep their form when the fields are exchanged as E → B, B → −E (units with c = 1). The exchange is a quarter-turn in the plane of (E, B), and applying it twice reverses both fields:

(E, B) → (B, −E) → (−E, −B)

Electric charges spoil the symmetry, because Gauss’s law for E has a source and Gauss’s law for B does not. Adding magnetic charges and currents restores it, with the electric and magnetic sources exchanged together with the fields. A particle carrying magnetic charge alone is a magnetic monopole (magnetic monopoles).

In quantum mechanics the two kinds of charge are not independent. Dirac showed in 1931 that a monopole of charge g can coexist with an electric charge e only if their product is quantized; the checkpoint writes the condition e·g ~ 2π (#11 §11.2; Dirac quantization and the DSZ pairing, where DSZ is Dirac–Schwinger–Zwanziger). A small electric charge therefore forces a large magnetic one. Exchanging the electric and magnetic descriptions of a theory exchanges weak coupling with strong coupling, which is why the quantum form of the duality is also called strong–weak duality. #11 §11.1 abbreviates it as “electric–magnetic (E↔B, g → 1/g)”.

The coupling and the generator S [T]#

A gauge theory’s coupling g and its vacuum angle θ combine into the complexified coupling τ_c = θ/2π + 4πi/g² [T]; the notation is #11’s, and τ_c is unrelated to the dark-energy field τ (notation). Duality transformations act on τ_c through the modular group SL(2, ℤ), generated by T: τ_c → τ_c + 1, the shift of θ by 2π, and S = [[0, 1], [−1, 0]]: τ_c → −1/τ_c, the exchange of electric and magnetic charges [T]. At θ = 0, S sends τ_c = 4πi/g² to ig²/4π, the same form with g replaced by 4π/g [T]. #11 §11.3 states the action the program uses: “S-duality acts on the complexified coupling by τ_c → −1/τ_c, generated by the same [[0,1],[−1,0]] rotation as Born’s ρ and our Z → −Z⁻¹”. The matrix is also the quarter-turn of the Maxwell exchange above and of Born reciprocity, (x, p) → (p, −x); the program’s use of the coincidence is the subject of the symplectic charge lattice.

Montonen–Olive duality [T]#

The simplest gauge theory with monopoles has an SU(2) symmetry broken to U(1) by a Higgs field in the adjoint representation. Its spectrum contains electrically charged vector bosons W± of mass g·v, where v is the symmetry-breaking scale, and magnetically charged ’t Hooft–Polyakov monopoles of mass M = 4πf v/g, with f = 1 in the Bogomolny–Prasad–Sommerfield (BPS) limit of a vanishing Higgs self-coupling (#14 §2, §3 give both formulas for the hidden sector). In that limit the exchange g → 4π/g turns the W mass into the monopole mass and back [T].

Montonen and Olive took the symmetry of the mass spectrum as a sign of an exact duality: a dual theory, at coupling 4π/g, in which the monopoles are the elementary gauge particles and the W bosons are the solitons. Their title states the conjecture as a question, “Magnetic monopoles as gauge particles?” (#11 ref [18]†).

The proposal has a defect in the bosonic theory, which the v3 worklog states: “the bosonic spectrum mismatches (W spin 1, monopole spin 0)”. The W is a vector particle and the monopole is not, so the two cannot be exchanged by an exact symmetry. In N = 4 super Yang–Mills theory (SYM), which has the most supersymmetry (SUSY) a four-dimensional gauge theory without gravity can have, the monopoles come in multiplets that include spin 1, so the mismatch disappears; #14 cites Osborn 1979 for this setting (#14 ref [8]). The checkpoint’s summary is that Montonen–Olive duality is “exact in maximally supersymmetric Yang–Mills” (#11 §11.2 [T]).

How well tested the exactness is depends on the theory. Harvey’s lectures, as the round-3 referee quoted them after checking the source: the duality has been “put to non-trivial tests in finite N = 4 Super Yang-Mills theory and special finite N = 2 theories”, while for the non-supersymmetric theory “there is no reason to think that the duality of the spectrum should be maintained by quantum corrections” (review/round3/referee/audit-notes.md, Step 1).

Seiberg duality and magnetic quarks [T]#

Seiberg found a related equivalence in N = 1 supersymmetric versions of quantum chromodynamics (QCD), the theory of quarks and gluons. A theory with a given number of colors and quark flavors has, at long distances, the same physics as a second gauge theory with a different number of colors, whose charged matter fields are called “magnetic quarks” [T, Seiberg 1995 [18†]]. In part of the range of flavors, the second description is weakly coupled where the first is strongly coupled. The equivalence holds for the long-distance limit of the two theories, not at every scale [T].

#11 cites it for the vocabulary: “magnetic quarks” is “standard vocabulary via Seiberg duality [18†]” (§11.2). The checkpoint’s italic opening dates the idea “proposed in 1977, named in 1994”; its reference list dates Seiberg’s paper 1995, the journal year, while the preprint number hep-th/9411149 dates from November 1994 [T]. The popular article #12 says that “‘magnetic quark’ has been standard vocabulary in field theory since the 1990s” (the popular articles).

T-, S- and U-duality in string theory [T]#

String theory has dualities of two kinds [T]:

  • T-duality relates a closed string theory on a circle of radius R to the same or a related theory on a circle of radius 1/R in string units. A closed string on a circle carries two whole numbers, its momentum around the circle, as a wave on a ring does, and the number of times it winds; T-duality exchanges them. The doubled geometry that makes the exchange manifest is the subject of double field theory.
  • S-duality in string theory relates strong and weak string coupling, and in some compactifications reduces to Montonen–Olive duality of the four-dimensional gauge theory.

Hull and Townsend (1995)† proposed that in the type II superstrings on a torus both are parts of a single larger discrete symmetry, U-duality. #11 §11.3 describes the situation as “T- and S-duality cohabiting inside U-duality in string theory [T]”, and uses it as the model for its conjecture that the program’s two polarization choices are parts of one structure (open problem 8; OP8 is item 8 of #11 §14).

The exactness burden#

Any argument that needs S-duality as an exact symmetry outside supersymmetric theories inherits the caution above. The burden arose in the program through the soliton manuscript #14, whose second version fixed the hidden sector’s coupling by a conjectured action of reciprocity as the modular S map (Obituary: the coupling α_d ≈ 34).

  • Round 2 posed it. Among the valid hits pending for v3: “SUSY-exactness burden for Montonen–Olive (either reciprocity supplies the rigidity or the dual sector is approximately SUSY)” (HANDOFF.md §6; review rounds 1 and 2).
  • The worklog answered it as item R1, marked “ready”: “Exact S-duality has been tested only in N=4 SYM and special finite N=2 theories. For non-SUSY theories there is “no reason to think” it survives quantum corrections, and the bosonic spectrum mismatches (W spin 1, monopole spin 0). That is an expectation, not a no-go [Harvey hep-th/9603086; Osborn 1979]. v3 must state it. If F2 is adopted, S no longer carries α_d, and the burden shrinks to the §2 “full mirror spectrum” remark.” (review/v3-worklog.md, ‘Round-2 pending list’).
  • F2 was adopted. Finding F2 separated the hidden sector’s weak elementary coupling from its monopoles’ strong coupling, and Paul decided on 2026-10-19 that “S-duality is no longer what fixes α_d” (‘Findings neither referee raised’). The conjecture that reciprocity acts on couplings as S is superseded.
  • v3 stated it. “Exact electric-magnetic duality is also not assumed. It has been tested non-trivially only in N=4 super-Yang-Mills and certain finite N=2 theories, and there is no reason to expect it to survive quantum corrections in a non-supersymmetric theory [8]. Nothing here requires it: the monopole physics used is semiclassical, at alpha_W’ ~ 1/34.” (#14 §2). The manuscript was withdrawn after round 3 over its abundance arithmetic (Obituary: superheavy conjugate solitons).
  • The checkpoint today cites Montonen–Olive as literature for the magnetic mirror-family, stating exactness only for maximally supersymmetric Yang–Mills (§11.2), and uses S only for its action on τ_c (§11.3).

How the program uses it#

  • As established physics behind founding supposition 11. P.R.’s supposition of “A full magnetic mirror-family, quarks included” is “alive as established physics (Montonen–Olive, Seiberg, dual superconductor)” (#11 App. A item 11; the dual superconductor).
  • As the third appearance of one matrix. S’s matrix is the matrix of ρ and of the modular inversion Z → −Z⁻¹ (reciprocity as modular inversion); #11 calls this “the third appearance of one matrix in this program” (the symplectic charge lattice).
  • As the model for open problem 8. The U-duality inclusion is the analogy for one enlarged structure containing both polarization choices (open problem 8).
  • Not as an explanation of the missing monopoles. The audit that killed that explanation turned on the point that electric–magnetic duality, not Born reciprocity, is the duality that relates magnetic charges to electric ones (Obituary: conjugate monopoles explain the null result).

Standard references#

As the documents cite them:

  • C. Montonen, D. Olive, “Magnetic monopoles as gauge particles?” Phys. Lett. B 72, 117 (1977) — #11 [18]†; #14 [8].
  • N. Seiberg, “Electric–magnetic duality in supersymmetric non-abelian gauge theories,” Nucl. Phys. B 435, 129 (1995), hep-th/9411149 — #11 [18]†.
  • H. Osborn, Phys. Lett. B 83, 321 (1979) — #14 [8].
  • J. A. Harvey, “Magnetic monopoles, duality and supersymmetry,” hep-th/9603086 — #14 [8]; conference proceedings, not a journal article (review/v3-worklog.md, ‘Verified references for v3’).

Added by the wiki from model memory (†); the program’s documents do not cite them:

  • A. Sen (1994)†: evidence for S-duality in N = 4 theory from bound states of monopoles.
  • N. Seiberg and E. Witten (1994)†: the exact low-energy solution of N = 2 Yang–Mills theory, with monopole condensation after a breaking to N = 1.
  • C. M. Hull and P. K. Townsend (1995)†: U-duality.

Used by#

Sources#

  • papers/broken-born-checkpoint-v2.md §11.2, §11.3, ref [18] — the statement [T], the action on τ_c, the U-duality analogy, and the citations (†); §11 italic opening — “proposed in 1977, named in 1994”.
  • review/v3-worklog.md ‘Round-2 pending list’ (R1), ‘Findings neither referee raised’ (F2), ‘Verified references for v3’ — the exactness burden, the decision that S no longer fixes α_d, Harvey as proceedings.
  • review/round3/referee/audit-notes.md Step 1 — Harvey’s wording, checked against the source.
  • notes/knot-dark-matter-standalone.md §2, §3, ref [8] — “Exact electric-magnetic duality is also not assumed”; the W′ and monopole masses; Montonen–Olive, Osborn and Harvey as cited (withdrawn manuscript).
  • HANDOFF.md §6 — the round-2 framing of the burden.
  • popsci/universes-other-half.md ‘Dark matter as the other polarization’ — “magnetic quark” since the 1990s (not a status source).
Backward The modular group SL(2, ℤ) The dual superconductor Forward