arXiv:1804.05588v1 [hep-th] 16 Apr 2018 lt ewe h asesFezPuiato n h Curtr the and [ action action Fierz-Pauli massless the between ality aehsbe icse eetyol in only recently discussed been has partially-ma case the while gravitons, massive and massless for (A)dS Hoˇrava-Lifshitz g of context [ the ity in used been also has setup (A)dS In action. eli order by — duality di electric-magnetic inating by related fiel whose are — equations actions second-order equivalent two obtains one salsmn fa qiaec ewe hois o whic for o theories, between equivalence an of establishment ed a tde n[ in studied was fields es fFaknadTeti [ Tseytlin and Fradkin of sense pctm o iersdgaiyi [ in in gravity studied linearised been for has spacetime (self-)duality, pseudo named often try, field. gauge mixed-symmetry a featuring principle action dimension arbitrary spin of massless backgrounds (partially) (A)dS and in massive theories of formulations dual Introduction 1. duality electric-magnetic of treatment unified a provide We Abstract asesa ela atal-asesgaiospropagat gravitons partially-massless as well as massless n h dnicto ftepyia ere ffedmand partially-massl freedom the of In [ degrees straightforward. physical helicities the the mak thereby of of smooth, identification are limits the massless ing and flat buil mani we the a that that in actions such and dual level The action way. Lorentz-invariant the festly at formulations dual furnish we aueo h ulgaio steeoecaie o l val all for clarified therefore is graviton discus dual previously the been of has nature duality the the in t where freedom case, analyses of massless degrees previous physical of complete counting we direct enable fields, massless and massive ersr oteprn-cintcnqeepoe ntepa- the in employed technique [ parent-action pers the to resort We c 14 ff eatmnod ´sc eAtsEegıs nttt eCi de Energ´ıas, Instituto Altas F´ısica de de Departamento a ttelvlo h edeutos sl-daiysymme- (self-)duality equations, field the of level the At on analyses previous extend and complete we note, this In h cinpicpe htw rsn etr ohteorigi the both feature present we that principles action The constant, cosmological the of and mass the of values all For seldaiyrlto sncsay nfltsaeie the spacetime, flat In necessary. is relation duality -shell rued ´cnqee rvtto,Ui fTertclan Theoretical of Unit Gravitation, M´ecanique et de Groupe ,w banfrtefis ieada,mnfsl covariant manifestly dual, a time first the for obtain we ], 13 1 4 ]. – suodaiysmer o atal asesspin-2 massless partially for symmetry pseudo-duality , 20 10 ff , massive and partially-massless massless, for actions Dual nodrt eieeuvln,da cin nthe in actions dual equivalent, derive to order in ] 21 rn eso ed rmacmo prn”first- “parent” common a from fields of sets erent a rvni h eiso ok [ works of series the in proven was ] 18 n , hs ehiushv enemployed been have techniques these b 19 ntttfu hoeicePyi,EHZrc,Wolfgang-P Zurich, ETH Physik, f¨ur Theoretische Institut .Teeaefis tp oad the towards steps first are These ]. 11 .I re,i hsframework this in brief, In ]. ioa Boulanger Nicolas 15 , 16 n ;seas [ also see ]; = [ 3 nisNcers–UA,Crut xeirs Exterior Circuito UNAM, – Nucleares encias 12 2 .Tesame The ]. – ahmtclPyis nvri´ eMn MN,2 plac 20 UMONS, – Mons Universit´e de Physics, Mathematical d [email protected] 4 [email protected] [email protected] ]. a nraCampoleoni Andrea , s case ess 17 n ssless e ftems n ftecsooia constant. cosmological the of and mass the of ues .In ]. > n nmxmlysmercsaeie faydimension any of spacetimes symmetric maximally in ing are t an h rav- ight tteato ee n ihmnfs oet naine f invariance, Lorentz manifest with and level action the at , du- flat e ncvratfr nya h ee fteeutoso mot of equations the of level the at only form covariant in sed a n aseslmt h aeteteti xeddt the to extended is treatment same The limit. massless and flat m- . 3 -2 d - - - a s aetato ehiusb iigda descriptio dual giving by techniques parent-action use hat 1 n n[ in and etdi hsnt ed otefloigresults: following the to leads note this in sented omltoso iersdgaiy ntepeec fsour of [ presence in the given been in have gravity, linearised duality-inv of manifest formulations d enough, Interestingly electric-magnetic symmetry. manifest ity nearly a exhibiting mulation atal-asescs,seas [ also see case, partially-massless rpriso ult-ymti ytm eesuidi [ in studied were systems duality-symmetric of properties g (A)dS linearised Lorentz-covaria around of ity manifestly formulation non double-potential same, a the framework, In therein. ences ed fmxmldph h ae [ paper the depth, maximal of fields re rudfltsaeiei h aes[ co- papers the spacetime in manifest spacetime of flat loss higher-spin around and the spin-2 ories is massless for pay explained as to variance, not the price inside does fields the demo- fields magnetic a tion, and dual such electric For of of way. pair appearance duality-symmetric cratic the a in hand, Lorentz-invari action other the manifestly enter the a in On dual, its fashion. and field spin-2 nal nmr eal,teuie ramn fsi- ult pre- duality spin-2 of treatment unified the details, more In • • • uiSrse2,89 Z¨urich, Switzerland 8093 27, auli-Strasse etr.Teeoe u cin di mohfltlmtin limit flat smooth a both admit actions our scalar and Therefore, spin-1 sectors. spin-2, dual the representing fields gauge limit; flat admittin smooth formulation a Stueckelberg its to therein obtained ntecs famsls rvtni Ad,w complete we (A)dS, [ in in graviton sketched massless programme a the of case the In ult rmaped oagnieo genuine the a elevating to thereby pseudo a a level, obtain from action we duality the (A)dS, at in description field dual spin-2 massless partially For ntemsiecs n()S ecaiytefltlmto the [ of in limit flat presented the model clarify we dual (A)dS, in case massive the In 27 b gai Cortese Ignacio , lcrcadmgei formulations. magnetic and electric for ] n n > / ,C.Uiestra 41 iddd M´exico, Mexico de Ciudad 04510 Universitaria, Cd. n, pctm a tde n[ in studied was spacetime sfrwa ocrspartially-massless concerns what for As . 4 29 n na lentv a n[ in way alternative an in and ] rvtn n(A)dS in gravitons c 6 nta ehv Stueckelberg have we that in ] 10 31 28 uPr,70 os Belgium Mons, 7000 Parc, du e ylnigteda action dual the linking by ] .Fnly h integrability the Finally, ]. rvdsa o an provides ] ff selduality; -shell 22 25 – , 24 26 n n refer- and ] rmassive, or ff for ] 30 > selfor- -shell partially- o.The ion. ;i the in ]; For . 3 sthat ns n ariant 32 rav- the- ual- ces, = ac- ]. ant nt 4 g In sections 2 and 3 we recall the main features of the first- When m = 0 the fields of spin one and zero decouple from order description of massive spin-two fields, that we use as a the spin-two sector and one recoversthe usual first-order formu- parent action. Section 4 collects our original results on dual lation of linearised gravity in (A)dS. At µ = 0, the sole scalar formulations for spin-2 fields in (A)dS. sector decouples and one obtains a first-order description ofa partially-massless graviton, propagating helicities two and one in the flat limit. With the manifestly unitary conventions used 2. The parent action in (1)–(3), one can set µ to zero by tuning the mass m R only in dS (σ = 1). In section 4.2 we shall show that partially-∈ We consider as parent action the first-order Stueckelberg ac- − tion describing, for generic values of the parameters, the propa- massless fields in AdS can be described in this formalism at the gation of a massive spin-2 field in a constant curvature back- price of flipping the sign of the spin-one kinetic term, which ground [33]. It is obtained by coupling the free actions for makes their lack of unitarity manifest. massless fields of spin two, one and zero. It thus comprises Eq. (1) can be expressed in terms of the field [2] the kinetic terms for these fields,1 bc a a bc ab cd ac bd Y | = ω | + g ωd g ωd , (9) | − | (2) ǫabcp[n 3] ab c 1 a qb c p[n 3] which is antisymmetric in its first two indices and transforms as = − ω h + ω qω e¯ H − , (1) L −2(n 3)! ∇ n 2 ! 2 − − bc bc [b c]d (n 2)µ [b c] δY |a = aΛ + 2¯ea dΛ − e¯a ξ . (10) ∇ ∇ − n 1 (1) ǫabc[n 2] ab 1 k l c[n 2] = − F A Fkl e¯ e¯ H − , (2) − L 2(n 2)! ∇ − 4 ! The spin-2 kinetic term can then be cast in the form (from now − on we will omit the integration measure dnx √ g brought by ǫ a1 an a1 an n − (0) ab[n 1] a 1 k b[n 1] e¯ e¯ = det(¯e)ǫ ··· d x) = − π ϕ πk e¯ H − , (3) L (n 1)! ∇ − 2 ! ··· − (2) a bc 1 bc a 1 ab c = bhc Y |a Y | Yab c + Y |bYac , (11) together with cross couplings and mass terms: L ∇ | − 2 | n 2 | ! − while the cross couplings and mass terms read ǫabc[n 2] ab a d b a b cross = − (n 1)m ω A + mF d h e¯ + µπ A e¯ L (n 1)! − = 2m ab ab a −  cross Y |bAa mF ha b µπ Aa 2 2 (4) L − n 2 − | − (n 2)µ a b a b m 2 a b c[n 2] (12) − h h m µ ϕ h e¯ ϕ e¯ e¯ H − . 2 − 2 (n 2)µ b a 2 nm 2 − 4 − − (n 2)  − ha bh | h + m µ h ϕ + ϕ , − − 4(n 1) | − n 2 2 (s) −   − The full action is the integral of = = + cross and it is = a L s 0 L L where h ha denotes the trace of the linearised vielbein. invariant under the gauge symmetries | bc P As recalled in section 3, eliminating the auxiliary fields Y |a, Fab and πa from the parent action one obtains a second-order a a ab 2m a L δh = ξ Λ e¯b + ǫ e¯ , (5a) description of a massive spin-2 field in terms of the linearised ∇ − n 2 − metric and the fields Aµ and ϕ, which reduces to the Fierz-Pauli µ2 δωab = Λab + e¯[aξb] , (5b) action for m = 0. In section 4 we will instead show how elim- ∇ n 1 b inating the fields ha , Aa and ϕ leads to its dual description, − | and involving mixed-symmetry fields for generic values of n.

a ab ab δA = ǫ m ξ e¯a , δF = 2m Λ , (6) ∇ − 3. Electric reduction δϕ = µǫ, δπa = m µ ξa . (7) bc ab a − − The equations of motion for Y |a, F and π arising from [h, Y, A, F,ϕ,π] allow to solve for them algebraically. E.g. For later convenience, we introduced the constants m and µ, L even if the action actually depends only on a single mass pa- Yab c = ch[a b] ah(b c) + bh(a c) | ∇ | − ∇ | ∇ | (13) rameter (besides the (A)dS radius). Gauge invariance requires d + 2gc[a hb] d b]h + 2mAb] . ∇ | − ∇ 2(n 1)  ab a µ2 = − 2m2 + σ(n 2)λ2 . (8) By plugging this and the similar expressions for F and π n 2 − −   into the parent Lagrangian , the latter reduces, modulo a total derivative, to the second-orderL Stueckelberg Lagrangian for a 1We denote the background vielbein bye ¯a , while is the Lorentz-covariant symmetric spin-2 field [34, 35]: ∇2 c = 2 c b derivative on (A)dSn. In our conventions, it satisfies V σ λ e¯ e¯b V , = 1 a (b c) + c (b a) = = ∇ − ∧ [h, A, ϕ] ah(b c) h | ah(b c) h | so that σ 1 in AdSn and σ 1 in dSn. We define the Levi-Civita sym- L − 2 ∇ | ∇ ∇ | ∇ − 2 bol ǫa a such that ǫ01 n 1 = 1 and we adopt the mostly-plus convention 1 a (a b) (n 1)σλ (a b) 2 1 n + ah h ah bh | − 2h(a b)h | h for the··· metric. In the following··· − − we omit wedge products and we substitute 2 2 ∇ ∇ − ∇ ∇ −  | −  groups of antisymmetrised indices with a label denoting the total number of A [aAb] (n 1)σ λ2A Aa 1 ϕ aϕ a[k] a a a a [a b] a 2 a (14) indices. For instance, we introduce the k-form H H 1··· k = e¯ 1 e¯ k . − ∇ ∇ − − − ∇ ∇ ≡ ··· a (a b) a Indices enclosed between square brackets are antisymmetrised, and dividing by 2mAa h bh | + µ ϕ aA the number of terms involved is understood (strength-one convention). Finally, − ∇ − ∇ ∇ 2  (a b) 2  nm2 2 repeated indices also denote an antisymmetrisation, e.g., AaBa A[a Ba ]. m h h | h + ϕ + m µ h ϕ. ≡ 1 2 (a b) n 2 −  | −  − 2 The resulting action is invariant under the gauge transforma- mixed-symmetry fields display less gauge symmetries in (A)dS tions (5a) for h(a b), to be identified with the linearised metric, than in flat space. This is the Brink-Metsaev-Vasiliev (BMV) | together with (6) and (7) for the Stueckelberg fields Aa and ϕ . mechanism conjectured in [36], provedfor AdSn in [37–39] and The antisymmetric part of the vielbein, h[a b] , enters the reduced for dSn in [40]. It is also discussed in [41] from the point of Lagrangian only through a total derivative,| consistently with the view of reducibility conditions. As a result, in the flat limit, shift symmetry it enjoys under Lorentz transformations. mixed-symmetrygauge fields decompose in multiplets of gauge The first two lines of (14) gives the Fierz-Pauli Lagrangian fields. In this case, in the limit λ 0 the field T decomposes for a massless spin-2 field in (A)dS. For µ = 0 one obtains a into a “proper”[n 3, 1] dual graviton→ plus an additional field of description of a partially-massless spin-2 field in dS in terms type [n 2, 1] that− does not carry any local degrees of freedom. of the Stueckelberg coupling of the Fierz-Pauli and Proca La- This− phenomenon can be described by introducing a suitable a grangians. The field Aa can be gauged away using ξ , and the set of Stueckelberg fields. In the current example, following resulting action is invariant under [42] one can introduce a new field, antisymmetric in its first three indices and traceless, implementing the shift 1 2 δh(a b) = (a b)ǫ + λ gabǫ . (15) | λ bc bc 1 bcd abc ∇ ∇ Yˆ |a Yˆ |a + dW |a , W |c 0 , (19)   → λ ∇ ≡ In this context, the partially-massless gauge symmetry thus follows because gauge transformations (5a) and (6) with either in the parent action (16) or in (17). This leads to the Lagrangian aǫ m ξa = 0 preserve the gauge fixing Aa = 0. ∇ − = 1 1 abc d e + ˆ ab c e 0[Y, W] λ2 2 cW | Wdbe a λ Y | Wcbe a L ∇ ∇ | ∇ | (20) 4. Magnetic reduction σh ab c d λ2 ab c + bYˆ | Yˆcd a + Yˆ | Yˆac b , 2(n 2) ∇ ∇ | 2 | 4.1. Massless case − i that is invariant up to total derivatives under3 When m = 0 the fields of spin one and zero decoupleand one can consider the parent Lagrangian ˆ bc = bcd + Λbc + 2 [b Λc]d + (n 3)λ bc δY |a dζ |a a n 1 e¯a d −σ χ a , (21) ∇ ∇ − ∇ = (2) (n 2)σλ2 b a 2 bcd = bcde + bcd 3 [b cd]e bcd 0[h, Y] − 2 ha bh | h , (16) δW |a eυ |a aχ n 2 e¯a eχ λζ |a . (22) L L −  | −  ∇ ∇ − − ∇ − with (2) given in (11). Its gauge symmetries are obtained by Note that the new field can be gauged away using the shift L bcd a setting m = 0in(5a)and(10). Contrary to flat space [4] (where symmetry generated by the traceless ζ | , while it also brings ff bcde a it enters the action linearly), in (A)dS the linearised vielbein is its own di erential symmetries generated by υ | (which is an auxiliary field thanks to the mass term in (16): it can thus be traceless and antisymmetric in the first four indices) and by the abc eliminated through its own equation of motion [5]. This leads fully antisymmetric χ . bc = 1 cde to an action depending only on the traceless projection of Y |a : Introducing the Hodge dual Ca[n 3] b 3! ǫa[n 3]cdeW |b − | − cde (that is a GL(n) Young-projected [n 3, 1] field, since W |b ˆ bc = bc + 2 [b c]d =− b Y |a Y |a n 1 e¯a Y |d . (17) is traceless) and denoting C′a[n 4] Ca[n 4]b together with − b − − | T ′a[n 3] = Ta[n 3]b , one obtains the dual Lagrangian bc a − − | After the elimination of ha b , the trace of Y | indeed con- tributes to the action only via| a boundaryterm, consistently with [C, T] = 1 [C] + + σ [T] , (23) L0 − 2λ2(n 3)! L Lcross (n 2)2 I the shift symmetry generated by ξa in (10), which is still present − h − i b for m = 0. One can cast the resulting Lagrangian in the form where (denoting antisymmetrisations with repeated indices) = a c[n 3] b a c b[n 3] σ cd ab 2 ba c [C] aCc[n 3] b C − | aCb[n 3] C − |c (24) = ˆ ˆ + ˆ ˆ − | − | 0[Y] 2(n 2)λ2 a Y |b c Y |d σλ Ybc a Y | , (18) L ∇ ∇ − ∇ ∇ L − ∇ ∇ | a b[n 4] c a b[n 3] h i (n 3) aC′b[n 4] C′ − + bCab[n 4] C − |c − − ∇ − ∇ ∇ − | ∇ in agreement with the result obtained by eliminating the viel- n h a b b[n 4] c b[n 4] 2( 1) Cb[n 3] a C′ − (n 4) bC′cb[n 5] C′ − , bein from the linearised McDowell-Mansouri action [10]. − − ∇ − | ∇ − − ∇ − ∇ = 1 ˆ cd i Introducing the Hodge dual Ta[n 2] b 2 ǫa[n 2]cdY |b (which a[n 2]bc = − | − ab = satisfies ǫ − Ta[n 2] b 0 on account of Yˆ |b 0), one = c b a[n 3] b a[n 3] − | cross 2λ Ta[n 3]b C − |c T ′a[n 3] C − |b obtains a dual description of a massless graviton in (A)dSn . The L − | ∇ − − ∇ h n a a[n 4] field T, however, has the same structure as a massive graviton b + ( 1) (n 3) T ′a[n 3] C′ − , (25) − − − ∇ in flat space [1]; when λ = 0, the dual of a massless spin-two i field is instead a GL(n) Young-projected2 [n 3, 1] field [2–4]. and ff − As discussed in [10], the di erent nature of the dual graviton [T] = [T] + σ(n 2)λ2 (26) in (A)dS and flat space can be explained as follows: massless I L − × a[n 2] b a[n 3] T − | Ta[n 2] b (n 2)T ′ − T ′a[n 3] . × h − | − − − i 2Two-column, GL(n)-irreducible fields are denoted by [p, q], where p and q stand for the lengths of the first and second column of the corresponding Young 3If one implements the Stueckelberg shift (19) already in the parent action a b = n 3 abc tableau, respectively. (16), the vielbein acquires the new transformation δχh | (n −2)λ cχ . − ∇ 3 The expression for [T] is obtained from [C]in (24) by re- massless graviton in flat space [4], rather than those presented L L placing everywhere in the latter expression the symbols C and in sections 4.1 and 4.3. The constraint a b = 0 is solved by n by T and n + 1 , respectively. C | bc 1 bcd σ bc [b c]d Lagrangian (23) is invariant, up to total derivatives, under Y | = W | F + 2¯e F , (35) a λ d a 2m a a d ∇ − ∇ ∇  n 1 δTa[n 2] b = ( 1) − (n 2) aζ˜a[n 3] b + (n 3)σλ gbaχ˜a[n 3] bcd − | − − ∇ − | − − where W |a has the same structuree as the field introduced in n 1 h + ( 1) − Λ˜ + n 1 Λ˜ the Stueckelberg shift (19). In particular, it is traceless. Substi- −n 1 b a[n 2] ( 1) − a a[n 3]b , (27) − ∇ − − ∇ −  i tuting (35)in(29), one obtains n 1 δCa[n 3] b = ( 1) − (n 3) aυ˜ a[n 4] b λ ζ˜a[n 3] b − | − − ∇ − | − − | 1 bcd a e + n 3 + n PM[W] = dW | Wabe c n−2 bχ˜a[n 3] ( 1) aχ˜a[n 4]b , (28) L − 2λ2 ∇ ∇ | − ∇ − − ∇ − (36)
+ ab c c ab + 4σm abd c a F Fbc Fbc F λ Fbc dW | . where the parameters ζ˜a[n 3] b, Λ˜ a[n 2] ,υ ˜a[n 4] b andχ ˜a[n 3] are ∇ ∇ − ∇ ∇ − | − − | −  e  the Hodge duals of those entering the transformations (21) and This Lagrangian actually depends only on the field Wbcd a: Fab (22) (the dualisation always involves only the group of antisym- | contributes only via a total derivative consistently with the shift metrised indices). In the limit λ 0 the field T decouples and symmetry (34). It is still invariant under does not propagate any degrees→ of freedom, while one retains the gauge field C , the dual graviton in flat space [4]. bcd bcde a[n 3] b δW |a = eυ |a , (37) In a spacetime of− dimension| D > n, the action (23) — fea- ∇ turing one of the two possible BMV couples of fields includ- while the other differential symmetry that was present in the ing T — would give a non-unitary propagation in dS. This is massless case (cf. (22)) is absent. bcd a manifested by the σ-dependent relative sign between the ki- All gauge symmetries that the field W | and, consequently, netic terms that we obtained. In this specific case, the relative its Hodge dual would display in flat space can be recovered by sign is irrelevant because Y is a topological field in flat space implementing the Stueckelberg shift and, indeed, the massless theory is unitary for any value of the 1 cosmological constant. bcd bcd + m− bcd 3 [b cd]e W |a W |a aU n 2 e¯a eU . (38) → n 3 ∇ − − ∇ e−   4.2. Partially-massless case Substituting in (36) one obtains the Lagrangian Partially-massless spin-2 fields exist for any non-vanishing = 1 bcd a e + σ abd c values of the cosmological constant, although they are unitary PM[W, U] 2 dW | Wabe c Uabc dW | L − 2λ ∇ ∇ | m ∇ only in dS [43]. To exhibit these facts, in this subsection we σ abc d λ2 abc (39) 2 cU Uabd 2 UabcU e, slightly modify our conventions, multiplying (1) by σ . With − 2(n 2)m ∇ ∇ − 2m L − this choice the factor σ in (8) is replaced by 1, so that one can which is invariante up to total derivativese under reach the point µ = 0 in both dS and AdS.− This leads to the parent Lagrangian bcd = bcde + bcd 3 [b cd]e σλ2 bcd δW |a eυ |a aχ n 2 e¯a eχ m ρ a , (40) ∇ ∇ − − ∇ − = a b + σ b a abc abcd abc e PM[h, Y, A, F] ha b | Aa b | δU = dρ (n 3)m χ . (41) L | C m ∇ C (29) ∇ − − 1 bc a + 1 ab c σ ab 2 Y | Yab c n 2 Y |bYace 4 FabF , − | − | − The contribution in ρ in (40)e (that was absent in (22)) is neces-   sary because, contrary to the massless case, the field U does not where we defined enter the action only via its divergence. As in the massless case, the sign of one of the two kinetic a b = ac b ab = n 2 | cY | mF , m λ − . (30) σ C ∇ − ± r 2 terms depends on . This is consistent with the observation that, after Hodge dualisation, one obtains a BMV couple of In the conventions adoptede in thise subsection, the gauge sym- fields which is unitary only in dS [40]. However, in this case metries of the action are both fields propagate in the flat limit: the [n 3, 1]dualof W car- ries the spin-2 helicities, while the [n 3] dual− of U carries the b b b 2m b δha = aξ +Λa + e¯a ǫ, (31) − | n 2 spin-1 helicities. Consequently, the sign flip of a kinetic terms ∇ −e bc bc [b c]d does matter: recovering the BMV couple of fields that is not- δY |a = aΛ + 2¯ea dΛ , (32) ∇ ∇ unitary in AdS is just another way to see that partially-massless δAa = aǫ + σ m ξa , (33) fields are not unitary in AdS. ∇ ab = Λab Using the dual field Ca[n 3] b defined as in the massless case δF 2 σ m e . (34) − | = 1 bcd − and introducing the Hodge dual field Aa[n 3] 3! ǫa[n 3]bcdU , the Stueckelberg Lagrangian we obtain− for the dual− partially In (29) we stressed that thee fields ha b and Aa are both La- grange multipliers when µ = 0 (although| the constraint im- massless spin-2 field in (A)dSn is posed by the latter field is not independent). The analysis of 2 2 = 1 [C] 2σλ [A] + 4σλ , (42) the partially-massless case therefore follows closely that of a PM 2(n 3)!λ2 (n 2)m2 m cross L − − hL − − L L i 4 e e e where [C] is given in (24), as a result, it would introduce second-order kinetic terms mix- L abc ab c ing U with the fully antisymmetric projection of Y | . On a b[n 3] a c[n 4]b [A] = aAb[n 3] A − (n 3) Ac[n 4]a bA − the other hand, the shifts L ∇ − ∇ − − ∇ − ∇ 2 a[n 3] + 3σλ Aa[n 3]A − , (43) − bc a bc a 1 bcd a m abc Y | Y | + dW | U , (49a) → µ ∇ − µ and the cross terms are ab ab 1 abc F F + cU (49b) → µ ∇ a[n 3] b n 1 cross = A − bCa[n 3] + ( 1) − (n 3) aC′a[n 2] . (44) L ∇ − | − − ∇ − do not modify the equation of motion for ha b and therefore they   | Thee action is invariant under cannot introduce any mixed kinetic term. The elimination of the fields ha b, Aa and ϕ then proceeds as above and one obtains the n 1 σλ2 | δCa[n 3] b = ( 1) − (n 3) aυ˜a[n 4] b gba ρ˜a[n 4] sum of the kinetic terms − | − − ∇ − | − m − n 3  n  + − bχ˜a[n 3] + ( 1) aχ˜a[n e4]b , (45) 1 1 abc d e n 1 ab c d n 2 = W | W + Yˆ | Yˆ − ∇ − − ∇ − µ2 2 c dbe a n−2 b cd a
K ∇ ∇ | − ∇ ∇ | (50) = n 1 h1 abc d 1 ab c δAa[n 3] (n 3) ( 1) − aρ˜a[n 4] m χ˜a[n 3] , (46) + cU Uabd + bF Fac , − − − ∇ − − − 4 2   ∇ ∇ ∇ ∇ i where the parametersυ ˜a[n 4] b,χ ˜a[n 3] andρ ˜ea[n 4] are the Hodge with the cross couplings duals of those entering the− transformations| − (40−)and (41). (1) = 1 ˆ acd b + m abd c cross µ Yab c dW | µ Uabc dW | 4.3. Massive case L | ∇ ∇ (51) h2(n 1)m bc a 1 abc − F Yˆ + F U We now consider the full Stueckelberg action presented in (n 2)µ ab c | 2 ab c − − ∇ ∇ i section 2. The elimination of the fields ha b, Aa and ϕ has been considered in [6, 9]. In the spirit of our discussion| of the special and mass-like terms = µ = points m 0and 0, we complementthese works by exhibit- (2) = 1 ˆ ab c ˆ + m ˆ ab c cross 2 Y | Yac b µ Y | Uabc ing a dual description with a smooth massless and flat limit. For L | (52) m2 abc 1 (n 1)m2 ab generic values of m, ha b is an auxiliary field and it can be elimi- U U + − F F . | 2µ2 abc 4 (n 2)µ2 ab nated through its equation of motion as in section 4.1. The field −  − −  Aa is instead a Lagrange multiplier enforcing the constraint This Lagrangian is invariant up to total derivatives under the following gauge transformations:4 ba 2m ab a = bF n 2 Y |b µπ 0 , (47) ∇ − − − bcd = bcde + bcd 3 [b cd]e a δW |a eυ |a aχ n 2 e¯a eχ which can be solved by expressing π in terms of the other ∇ ∇ − − ∇ bcd bcd fields. The equation of motion of ϕ does not bring any new in- µζ |a m ρ a , (53) − − formation, since it is not independenton account of the Noether ˆ bc = bcd + Λbc + 2 [b Λc]d identity associated with the gauge symmetry generated by ǫ . δY |a dζ |a a n 1 e¯a d ∇ ∇ − ∇ a (n 3)µ bc bc Substituting the on-shell values of ha b and π in the Stueck- + − χ a m ψ a , (54) | 2(n 1) elberg Lagrangian leads to [6] − − abc = abcd abc + 2(n 3)m abc δU dρ µ ψ n−2 χ , (55) 2 ∇ − = 1 n 1 Yˆ ab c dYˆ + µ Yˆ ab cYˆ + 1 Fab cF − µ2 n−2 b | cd a 2 | ac b 2 b ac ab abc ab L  − ∇ ∇ | | ∇ ∇ δF = cψ + 2m Λ . (56) 2 2 ∇ 2(n 1)m ˆ bc a + µ (n 1)m ab n−2 Fab cY | 4 −n 2 FabF , (48) − − ∇  − −   The action involving the Hodge duals of the previous fields ff where we recall that the parameters m and µ are related by (8). now admits a smooth flat and massless limit, in which di er- One can then consider the Hodge duals of the fields Yˆ and F ent helicities decouple. The spin-two ones are carried by the [n 3, 1] Hodge dual of W (as discussed in section 4.1), while and obtain a dual theory for a massive graviton in terms of the − Stueckelberg coupling of a massless spin-2 field (accounted by spin-one and zero helicities are carried, respectively, by the the [n 2, 1] dual of Yˆ) with a Proca field (accounted by the fully-antisymmetric Hodge duals of U and F . We refrain from [n 2]− dual of F). Its gauge symmetries are those inherited displaying this action explicitly, as it can straightforwardly be from− (6)and (10) after dualisation. obtained by expressing all fields in terms of their Hodge duals In orderto obtain a smooth massless and flat limit, one should in (50)–(52). One can also check that the appropriate Curtright bcd gauge symmetries are recovered from (53)–(56) together with introduce two additional fields: the traceless W |a that we al- ready encountered in section 4.1 and a 3-form Uabc . This will their gauge-for-gauge symmetries. allow to recover all Curtright gauge symmetries for the Hodge dual of Yˆ bc and the usual gauge symmetry for the massless |a 4 (n 2)-form which is the Hodge dual of Fab . Due to the cou- Implementing the shift (49) before the elimination of ha b etc. from the parent action does not modify the gauge transformations of A |, ϕ and πa. The − Fabh a pling a b in (12), introducing the 3-form via a Stueckelberg variation of ha b takes instead the same form as in the massless case and acquires ab| | a b = n 3 abc shift of F would modify the equation of motion for ha b and, a contribution δχh | (n −2)µ cχ . | − ∇ 5 Acknowledgments [24] M. Henneaux, S. H¨ortner, A. Leonard, Twisted self-duality for higher spin gauge fields and prepotentials, Phys. Rev. D94 (10) (2016) 105027, [Erratum: Phys. Rev.D97,no.4,049901(2018)]. arXiv:1609.04461. We are grateful to Th. Basile for discussions and collabo- [25] B. Julia, J. Levie, S. Ray, Gravitational duality near de Sitter space, JHEP ration at an early stage of this work. We thank X. Bekaert, 11 (2005) 025. arXiv:hep-th/0507262. J. A. Garc´ıa and L. Traina for fruitful discussions. We per- [26] R. G. Leigh, A. C. Petkou, Gravitational duality transformations on (A)dS(4), JHEP 11 (2007) 079. arXiv:0704.0531. formed or checked several computations with the package [27] S. H¨ortner, A deformation of the Curtright action, Phys. Rev. D95 (2) xTras [44] of the suite of Mathematica packages xAct. N.B. (2017) 024039. arXiv:1610.07225. thanks ETH Z¨urich and the Institut Denis Poisson (Univer- [28] S. Deser, A. Waldron, PM = EM: Partially Massless Duality Invariance, sit´ede Tours), while A.C. thanks the Universit`adi Firenze and Phys. Rev. D87 (2013) 087702. arXiv:1301.2238. [29] C. W. Bunster, S. Cnockaert, M. Henneaux, R. Portugues, Monopoles INFN (sezione di Firenze), for kind hospitality during the com- for gravitation and for higher spin fields, Phys. Rev. D73 (2006) 105014. pletion of this paper. The stay of N.B. at the Institut Denis Pois- arXiv:hep-th/0601222. son in Tours was funded by a grant of the Universit´ede Mons [30] G. Barnich, C. Troessaert, Manifest spin 2 duality with electric and mag- (UMONS). The work of A.C. has been supported in part by netic sources, JHEP 01 (2009) 030. arXiv:0812.0552. [31] K. Hinterbichler, R. A. Rosen, Partially Massless Monopoles and the NCCR SwissMAP, funded by the Swiss National Science Charges, Phys. Rev. D92 (10) (2015) 105019. arXiv:1507.00355. Foundation. [32] G. Barnich, C. Troessaert, Duality and integrability: Electromagnetism, linearized gravity and massless higher spin gauge fields as bi-Hamiltonian [1] T. L. Curtright, P. G. O. Freund, Massive Dual Fields, Nucl. Phys. B172 systems, J. Math. Phys. 50 (2009) 042301. arXiv:0812.4668. (1980) 413–424. [33] Yu. M. Zinoviev, Frame-like gauge invariant formulation for massive high [2] P. C. West, E(11) and M theory, Class. Quant. Grav. 18 (2001) 4443– spin particles, Nucl. Phys. B808 (2009) 185–204. arXiv:0808.1778. 4460. arXiv:hep-th/0104081. [34] Yu. M. Zinoviev, On massive high spin particles in AdS, [3] P. C. West, Very extended E(8) and A(8) at low levels, grav- arXiv:hep-th/0108192. ity and supergravity, Class. Quant. Grav. 20 (2003) 2393–2406. [35] Yu. M. Zinoviev, On massive spin 2 interactions, Nucl. Phys. B770 (2007) arXiv:hep-th/0212291. 83–106. arXiv:hep-th/0609170. [4] N. Boulanger, S. Cnockaert, M. Henneaux, A note on spin s duality, JHEP [36] L. Brink, R. R. Metsaev, M. A. Vasiliev, How massless are 06 (2003) 060. arXiv:hep-th/0306023. massless fields in AdS(d), Nucl. Phys. B586 (2000) 183–205. [5] A. S. Matveev, M. A. Vasiliev, Dual formulation for higher arXiv:hep-th/0005136. spin gauge fields in (A)dSd, Phys. Lett. B609 (2005) 157–166. [37] N. Boulanger, C. Iazeolla, P. Sundell, Unfolding Mixed-Symmetry Fields arXiv:hep-th/0410249. in AdS and the BMV Conjecture: I. General Formalism, JHEP 07 (2009) [6] Yu. M. Zinoviev, On dual formulations of massive tensor fields, JHEP 10 013. arXiv:0812.3615. (2005) 075. arXiv:hep-th/0504081. [38] N. Boulanger, C. Iazeolla, P. Sundell, Unfolding Mixed-Symmetry Fields [7] Yu. M. Zinoviev, On dual formulation of gravity, in AdS and the BMV Conjecture. II. Oscillator Realization, JHEP 07 arXiv:hep-th/0504210. (2009) 014. arXiv:0812.4438. [8] B. Gonzalez, A. Khoudeir, R. Montemayor, L. F. Urrutia, Duality for [39] K. B. Alkalaev, M. Grigoriev, Unified BRST description of AdS gauge massive spin two theories in arbitrary dimensions, JHEP 09 (2008) 058. fields, Nucl. Phys. B835 (2010) 197–220. arXiv:0910.2690. arXiv:0806.3200. [40] T. Basile, X. Bekaert, N. Boulanger, Mixed-symmetry fields in de [9] A. Khoudeir, R. Montemayor, L. F. Urrutia, Dimensional reduction as a Sitter space: a group theoretical glance, JHEP 05 (2017) 081. method to obtain dual theories for massive spin two in arbitray dimen- arXiv:1612.08166. sions, Phys. Rev. D78 (2008) 065041. arXiv:0806.4558. [41] A. Campoleoni, D. Francia, Maxwell-like Lagrangians for higher spins, [10] T. Basile, X. Bekaert, N. Boulanger, Note about a pure spin-connection JHEP 03 (2013) 168. arXiv:1206.5877. formulation of general relativity and spin-2 duality in (A)dS, Phys. Rev. [42] N. Boulanger, O. Hohm, Non-linear parent action and dual gravity, Phys. D93 (12) (2016) 124047. arXiv:1512.09060. Rev. D78 (2008) 064027. arXiv:0806.2775. [11] E. S. Fradkin, A. A. Tseytlin, Quantum Equivalence of Dual Field Theo- [43] A. Higuchi, Symmetric Tensor Spherical Harmonics on the N Sphere ries, Annals Phys. 162 (1985) 31. and Their Application to the De Sitter Group SO(N,1), J. Math. Phys. [12] D. Galviz, A. Khoudeir, Partially Massless and Self Duality in three di- 28 (1987) 1553, [Erratum: J. Math. Phys.43,6385(2002)]. mensions, arXiv:1712.01050. [44] T. Nutma, xTras : A field-theory inspired xAct package for mathematica, [13] I. Cortese, J. A. Garc´ıa, Electric-magnetic duality in linearized Hoˇrava- Comput. Phys. Commun. 185 (2014) 1719–1738. arXiv:1308.3493. Lifshitz gravity, Phys. Rev. D90 (6) (2014) 064029. arXiv:1405.6536. [14] S. Deser, R. I. Nepomechie, Gauge Invariance Versus Masslessness in De Sitter Space, Annals Phys. 154 (1984) 396. [15] C. M. Hull, Strongly coupled gravity and duality, Nucl. Phys. B583 (2000) 237–259. arXiv:hep-th/0004195. [16] C. M. Hull, Duality in gravity and higher spin gauge fields, JHEP 09 (2001) 027. arXiv:hep-th/0107149. [17] X. Bekaert, N. Boulanger, Massless spin two field S duality, Class. Quant. Grav. 20 (2003) S417–S424. arXiv:hep-th/0212131. [18] K. Hinterbichler, Manifest Duality Invariance for the Partially Massless Graviton, Phys. Rev. D91 (2) (2015) 026008. arXiv:1409.3565. [19] D. Cherney, S. Deser, A. Waldron, G. Zahariade, Non-linear duality in- variant partially massless models?, Phys. Lett. B753 (2016) 293–296. arXiv:1511.01053. [20] T. Curtright, Generalized gauge fields, Phys. Lett. 165B (1985) 304–308. [21] C. S. Aulakh, I. G. Koh, S. Ouvry, Higher Spin Fields With Mixed Sym- metry, Phys. Lett. B173 (1986) 284–288. [22] M. Henneaux, C. Teitelboim, Duality in linearized gravity, Phys. Rev. D71 (2005) 024018. arXiv:gr-qc/0408101. [23] C. Bunster, M. Henneaux, S. H¨ortner, Twisted Self-Duality for Lin- earized Gravity in D dimensions, Phys. Rev. D88 (6) (2013) 064032. arXiv:1306.1092. 6