1: \documentstyle[12pt,epsf,epsfig]{article}

2: \hoffset -0.2in 

3: \textwidth 6in

4: \textheight 8.5in

5: %\setcounter{page}{1} 

6: \parskip 7pt \openup1\jot \parindent=0.5in

7: \topmargin -0.5in

8: 

9: %       the stuff below defines \eqalign and \eqalignno in such a

10: %       way that they will run on Latex

11: \newfont{\thiplo}{msbm10 scaled\magstep 2}

12: \newfont{\gothic}{eufb10 scaled\magstep 2}

13: \newfont{\unc}{eurb10}  

14: \newskip\humongous \humongous=0pt plus 1000pt minus 1000pt

15: \def\caja{\mathsurround=0pt}

16: \def\eqalign#1{\,\vcenter{\openup1\jot \caja

17:         \ialign{\strut \hfil$\displaystyle{##}$&$

18:         \displaystyle{{}##}$\hfil\crcr#1\crcr}}\,}

19: \newif\ifdtup

20: \def\panorama{\global\dtuptrue \openup1\jot \caja

21:         \everycr{\noalign{\ifdtup \global\dtupfalse

22:         \vskip-\lineskiplimit \vskip\normallineskiplimit

23:         \else \penalty\interdisplaylinepenalty \fi}}}

24: \def\eqalignno#1{\panorama \tabskip=\humongous 

25:         \halign to\displaywidth{\hfil$\displaystyle{##}$

26:         \tabskip=0pt&$\displaystyle{{}##}$\hfil 

27:         \tabskip=\humongous&\llap{$##$}\tabskip=0pt

28:         \crcr#1\crcr}}

29: %	eqalignnoleft is eqalignno positioned flush left on the page

30: \def\eqalignnoleft#1{\panorama \tabskip=0pt

31:         \halign to\displaywidth{\hfil$\displaystyle{##}$

32:         \tabskip=0pt&$\displaystyle{{}##}$\hfil 

33:         \tabskip=\humongous&\llap{$##$}\tabskip=0pt

34:         \crcr#1\crcr}} 

35: %	\eqright causes display equation material between \eqright and \cr

36: %	to be positioned flush right on the page.  This is useful

37: %	in breaking long lines in a display equation.

38: %	\eqright is usually used in conjunction with \eqalignnoleft.

39: \def\eqright #1\cr{\noalign{\hfill$\displaystyle{{}#1}$}}

40: %	\eqleft causes display equation material between \eqleft and \cr

41: %	to be positioned flush left on the page.

42: \def\eqleft #1\cr{\noalign{\noindent$\displaystyle{{}#1}$\hfill}}

43: %       The oldref and fig macros are for formatting

44: %       references and figure lists at the end of the paper.

45: %       If you type \oldref{1}Dirac, P.A.M. you will get

46: %       [1] Dirac, P.A.M.

47: %       Same goes for \fig except you get Figure 2.1

48: \def\oldrefledge{\hangindent3\parindent}

49: \def\oldreffmt#1{\rlap{[#1]} \hbox to 2\parindent{}}

50: \def\oldref#1{\par\noindent\oldrefledge \oldreffmt{#1}

51:         \ignorespaces}

52: \def\figledge{\hangindent=1.25in}

53: \def\figfmt#1{\rlap{Figure {#1}} \hbox to 1in{}}

54: \def\fig#1{\par\noindent\figledge \figfmt{#1}

55:         \ignorespaces}

56: %

57: %       This defines et al., i.e., e.g., cf., etc.

58: \def\ie{\hbox{\it i.e.}{}}      \def\etc{\hbox{\it etc.}{}}

59: \def\eg{\hbox{\it e.g.}{}}      \def\cf{\hbox{\it cf.}{}}

60: \def\etal{\hbox{\it et al.}}

61: \def\dash{\hbox{---}}

62: %       common physics symbols

63: \def\tr{\mathop{\rm tr}}            

64: \def\Tr{\mathop{\rm Tr}}

65: \def\partder#1#2{{\partial #1\over\partial #2}}

66: \def\secder#1#2#3{{\partial~2 #1\over\partial #2 \partial #3}}

67: \def\bra#1{\left\langle #1\right|}

68: \def\ket#1{\left| #1\right\rangle}

69: \def\VEV#1{\left\langle #1\right\rangle}

70: \def\ME#1#2{\left\langle #1\right|\left. #2 \right\rangle}

71: \def\gdot#1{\rlap{$#1$}/}

72: \def\abs#1{\left| #1\right|}

73: \def\pr#1{#1~\prime}

74: \def\ltap{\raisebox{-.4ex}{\rlap{$\sim$}} \raisebox{.4ex}{$<$}}

75: \def\gtap{\raisebox{-.4ex}{\rlap{$\sim$}} \raisebox{.4ex}{$>$}}

76: % \contract is a differential geometry contraction sign _|

77: \def\contract{\makebox[1.2em][c]{

78:         \mbox{\rule{.6em}{.01truein}\rule{.01truein}{.6em}}}}

79: %	The command \sectioneq produces numbering of equations by section

80: \def\holdtheequation{\arabic}

81: \def\sectioneq{\def\theequation{\thesection.\arabic{equation}}{\let

82: \holdsection=\section\def\section{\setcounter{equation}{0}\holdsection}}}%

83: 

84: %	The commands \beginletts and \endletts delimit sections of

85: %	text in which successive equation numbers are distinguished by 

86: %	sequentially appending lower case letters 

87: \newcounter{holdequation}\def

88: 	\beginletts{\begingroup\setcounter

89: 	{holdequation}{\value{equation}}\addtocounter

90: 	{equation}{1}\edef

91: 	\holdtheequation{\theequation}\setcounter

92: 	{equation}{0}\def

93: 	\theequation{\holdtheequation\alph{equation}}}

94: \def\endletts{\endgroup\setcounter

95: 	{equation}{\value{holdequation}}\refstepcounter{equation}}

96: %	The command \num provides automatic numbering in LaTex when used in 

97: %	place of (equation number) in PlainTeX-style equations

98: \def\num{(\refstepcounter{equation}\theequation)}

99: %	\auto is shorthand for \eqno\num

100: \def\auto{\eqno(\refstepcounter{equation}\theequation)}

101: %	The commands \begineq and \endeq provide for one vertically 

102: %	centered automatic number for multiline equations

103: \def\begineq #1\endeq{$$ \refstepcounter{equation}\eqalign{#1}\eqno

104: 	(\theequation) $$}

105: %	The command \contlimit puts (a\rightarrow0) 

106: %	under \longrightarrow

107: \def\contlimit{\,{\hbox{$\longrightarrow$}\kern-1.8em\lower1ex

108: \hbox{${\scriptstyle (a\rightarrow0)}$}}\,}

109: %	The command \centeron#1#2 backs up #2 so that it is centered 

110: %	over #1.  \centerover and \centerunder work like \centeron,

111: %	except that they raise or lower #2 to place it over or under

112: %	#1.

113: \def\centeron#1#2{{\setbox0=\hbox{#1}\setbox1=\hbox{#2}\ifdim

114: \wd1>\wd0\kern.5\wd1\kern-.5\wd0\fi

115: \copy0\kern-.5\wd0\kern-.5\wd1\copy1\ifdim\wd0>\wd1

116: \kern.5\wd0\kern-.5\wd1\fi}}

117: %

118: \def\centerover#1#2{\centeron{#1}{\setbox0=\hbox{#1}\setbox

119: 1=\hbox{#2}\raise\ht0\hbox{\raise\dp1\hbox{\copy1}}}}

120: %

121: \def\centerunder#1#2{\centeron{#1}{\setbox0=\hbox{#1}\setbox

122: 1=\hbox{#2}\lower\dp0\hbox{\lower\ht1\hbox{\copy1}}}}

123: %	The commands \lsim and \gsim provide symbols for 

124: %	`less than of order' and `greater than of order'

125: \def\lsim{\;\centeron{\raise.35ex\hbox{$<$}}{\lower.65ex\hbox

126: {$\sim$}}\;}

127: \def\gsim{\;\centeron{\raise.35ex\hbox{$>$}}{\lower.65ex\hbox

128: {$\sim$}}\;}

129: %	The command \st (for stroke) puts a slash through the succeeding 

130: %	character in math mode

131: \def\st#1{\centeron{$#1$}{$/$}}

132: %	The command \newcases works like \cases except that 

133: %	the baselines and type size are the same as for

134: %	display equations

135: \def\newcases#1{\left\{\,\vcenter{\normalbaselines\openup1\jot \caja 

136: \ialign{\strut$\displaystyle{##}$\hfil

137: &\quad##\hfil\crcr#1\crcr}}\right.}

138: %	The command \super inserts the characters in its argument as a 

139: %	superscript with the correct spacefactor.

140: \def\super#1{\ifmmode \hbox{\textsuper{#1}}\else\textsuper{#1}\fi}

141: \def\textsuper#1{\newcount\holdspacefactor\holdspacefactor=\spacefactor

142: $^{#1}$\spacefactor=\holdspacefactor}

143: %	The command \supercite redefines \cite so that it makes superscripted 

144: %	citation numbers. It is to be used in conjunction with the 

145: %	\label command (for example, with one of the list-making 

146: %	environments).  The command \oldcite restores the original LaTeX 

147: %	\cite command.

148: \let\holdcite=\cite

149: \def\supercite{\def\cite{\newcite}}

150: \def\oldcite{\def\cite{\holdcite}}

151: \def\newcite#1{\super{\newcount\citenumber\citenumber=0\getcite#1,@, }}

152: \def\getcite#1,{\advance\citenumber by1

153: \def\getcitearg{#1}\def\lastarg{@}

154: \ifnum\citenumber=1

155: \ref{#1}\let\next=\getcite\else\ifx\getcitearg\lastarg\let\next=\relax

156: \else ,\ref{#1}\let\next=\getcite\fi\fi\next}

157: %	The command \nskip gives a vertical skip of the specified 

158: %	dimension (in braces) without including any extra \baselineskip 

159: %	or \parskip.

160: \def\nskip#1{\vbox{}\vskip-\baselineskip\vskip#1\vskip-\parskip\noindent}

161: %	The command \lskip skips vertically by one line, i.e., 

162: %	the current \baselineskip.  There is no indentation unless 

163: %	\indent is specified.

164: \def\lskip{\vskip\baselineskip\vskip-\parskip\noindent}

165: \def\np{Nucl.\ Phys.\ }

166: \def\pr{Phys.\ Rev.\ }

167: \def\prl{Phys.\ Rev.\ Lett.\ }

168: \def\pl{Phys.\ Lett.\ }

169: \def\arnps{Ann.\ Rev.\ Nucl.\ Part.\ Sci.\ }

170: \def\mn{\mu\nu}

171: \def\epm{e^+e^-}

172: \def\pom{{\rm P\kern -0.53em\llap I\,}}

173: \def\spom{{\rm P\kern -0.36em\llap \small I\,}}

174: \def\sspom{{\rm P\kern -0.33em\llap \footnotesize I\,}}

175: \def\gev{{\rm GeV}}

176: \def\mev{{\rm MeV}}

177: \def\parens#1{\left(#1\right)}

178: \relax

179: \def\contlimit{\,{\hbox{$\longrightarrow$}\kern-1.8em\lower1ex

180: \hbox{${\scriptstyle (a\rightarrow0)}$}}\,}

181: \def\upon #1/#2 {{\textstyle{#1\over #2}}}

182: \relax

183: \renewcommand{\thefootnote}{\fnsymbol{footnote}} 

184: \def\srf#1{$^{#1}$\ }

185: \def\mainhead#1{\setcounter{equation}{0}\addtocounter{section}{1}

186:   \vbox{\begin{center}\large\bf #1\end{center}}\nobreak\par}

187: \sectioneq

188: \def\subhead#1{\bigskip\vbox{\noindent\bf #1}\nobreak\par}

189: \def\rf#1#2#3{{\bf #1}, #2 (19#3)}

190: \def\autolabel#1{\auto\label{#1}}

191: \def\til#1{\centeron{\hbox{$#1$}}{\lower 2ex\hbox{$\char'176$}}}

192: \def\tild#1{\centeron{\hbox{$\,#1$}}{\lower 2.5ex\hbox{$\char'176$}}}

193: \def\sumtil{\centeron{\hbox{$\displaystyle\sum$}}{\lower

194: -1.5ex\hbox{$\widetilde{\phantom{xx}}$}}}

195: \def\sumtilt{\sum^{\raisebox{-.15mm}{\hspace{-1.75mm}$\widetilde{}$}}\ }

196: \def\gltext{$\raisebox{1mm}{\centerunder{$\scriptscriptstyle 

197: 	>$}{$\scriptscriptstyle <$}}$}

198: \def\intcent#1{\centerunder{$\displaystyle\int$}{\raisebox{-2.2mm}{$ #1 $}}}

199: \def\kbar{\underline{k}}

200: \def\qbar{\underline{q}}

201: \def\kbarsl{\underline{\st k}}

202: \def\qbarsl{\underline{\st q}}

203: \def\parens#1{\left(#1\right)}

204: \renewcommand{\arraystretch}{1.5}

205: \def\pbar{\underline{p}}

206: \def\pbarsl{\underline{\st p}}

207: \def\q{\unc q}

208: %\def\p{\unc p}

209: %-------------------------------------------------------------------

210: 

211: \newcommand{\bit}{\begin{itemize}}

212: \newcommand{\eit}{\end{itemize}}

213: \newcommand{\cl}{\centerline}

214: \newcommand{\beq}{\begin{equation}}

215: \newcommand{\eeq}{\end{equation}}

216: \newcommand{\beqa}{\begin{eqnarray}}

217: \newcommand{\eeqa}{\end{eqnarray}}

218: 

219: 

220: %\pagestyle{empty}

221: 

222: \begin{document} 

223: 

224: \begin{titlepage} 

225: 

226: \rightline{\vbox{\halign{&#\hfil\cr

227: &ANL-HEP-PR-01-017 \cr

228: &\today\cr}}} 

229: \vspace{0.25in} 

230: 

231: \begin{center} 

232:  

233: {\large\bf CHIRALITY VIOLATION IN THE }

234: 

235: {\large \bf QCD HIGH-ENERGY S-MATRIX}\footnote{Work 

236: supported by the U.S.

237: Department of Energy, Division of High Energy Physics, \newline Contracts

238: W-31-109-ENG-38 and DEFG05-86-ER-40272} 

239: \medskip

240: 

241: 

242: Alan. R. White\footnote{arw@hep.anl.gov }

243: 

244: \vskip 0.6cm

245: 

246: \centerline{High Energy Physics Division}

247: \centerline{Argonne National Laboratory}

248: \centerline{9700 South Cass, Il 60439, USA.}

249: \vspace{0.5cm}

250: 

251: \end{center}

252: 

253: \begin{abstract} 

254: 

255: In a previous paper it has been shown that the infra-red divergence

256: associated with the triangle graph axial anomaly 

257: can occur in triple-regge multi-reggeon

258: interactions due to unphysical asymptotic triple discontinuities. In 

259: this paper an asymptotic discontinuity analysis is applied to

260: high-order feynman diagrams to show that the anomaly exists in contributions

261: to the triple-regge nine-reggeon interaction. This implies that 

262: the anomaly occurs in the interactions of 

263: reggeon states that have the quantum numbers of 

264: the anomaly current and establishes a direct 

265: connection with the well-known U(1) problem.

266: 

267: \end{abstract} 

268: 

269: 

270: \renewcommand{\thefootnote}{\arabic{footnote}} \end{titlepage} 

271: 

272: %\setcounter{page}{2}

273: 

274: \mainhead{1. INTRODUCTION} 

275: 

276: Perhaps the most important property of a non-abelian gauge theory is the 

277: existence of non-perturbative euclidean clasical solutions with 

278: non-trivial topology. If the 

279: theory is quantized (in principle at least) via the euclidean path-integral,

280: such solutions produce additional  interactions that,

281: even if it is not well understood how to evaluate them, are believed to

282: modify properties of the theory significantly. In particular, 

283: the topological field configurations produce zero modes

284: of the Dirac operator which\cite{aj} 

285: prevent the gauge-invariant separation of 

286: massless fermion fields into  

287: chiral (right- and left-handed) components that create 

288: particles and antiparticles in a well-defined way. As a result,

289: the number of right- and left-handed Dirac particles is not separately 

290: conserved. The corresponding U(1) vector charge remains conserved,

291: but the axial charge conservation that is present in perturbation theory

292: is violated non-perturbatively, as allowed

293: by the anomaly in the U(1) axial current. We expect, therefore, that 

294: quark chirality transitions will play an important role    

295: within the QCD bound-state S-Matrix. (It is well-known 

296: that such processes can generate a mass\cite{gth} 

297: for a bound-state with the

298: quantum numbers of the $\eta'$.) Clearly, how such transitions contribute

299: must be a major component of a complete

300: non-perturbative definition of

301: the massless theory. 

302: 

303: In this paper we will provide evidence for a completely different 

304: argument that 

305: there should be non-conservation of the U(1) axial charge in the 

306: high-energy (multi-regge)

307: QCD bound-state S-Matrix. This argument makes no mention

308: of either the euclidean region, path-integral quantization, or zero modes.

309: Rather it is part of a program which has gradually taken shape over the 

310: years\cite{arw99,arw001} and

311: which we expect to carry out in detail in future papers. The aim of the 

312: program is to

313: construct the massless, multiparticle, multi-regge S-Matrix starting from

314: a spontaneously-broken theory in which gluons and quarks are massive,

315: and regge behavior and the unitarity

316: properties of reggeon diagrams

317: are well-established perturbatively\cite{fkl,jb}.

318: The massless theory that is our ultimate goal is, of course, 

319: infra-red divergent and it is widely believed 

320: that even if the divergences could be handled systematically the large

321: order behavior of the perturbation

322: expansion is so bad 

323: that it can not be used (by itself) to define the theory.

324: However, we plan to initially apply our construction to the case in which

325: the gauge coupling does not increase in the infra-red 

326: region\footnote{This requires an infra-red fixed point in the massless 

327: $\beta$-function which, in turn, is likely to require a special fermion 

328: content.}

329: and the divergence of the perturbation series is reduced considerably.

330: In addition, the multi-regge region  

331: may be a special situation. 

332: Because high-energy is involved, the S-Matrix should be close to

333: perturbation theory while, because low momentum transfers are also involved, 

334: $t$-channel unitarity properties

335: involving the physical spectrum must also be satisfied. 

336: 

337: If it is possible to construct the multi-regge behavior of a 

338: weak-coupling, massless, 

339: theory from the essentially perturbative 

340: starting point of massive reggeon diagrams,

341: without reference to the low-energy solution of the

342: theory, then there must be some element that

343: can produce the ``non-perturbative'' properties of 

344: confinement and chiral symmetry breaking in the spectrum. We expect that,

345: in our construction, this will be the triangle anomaly

346: ``chirality violation'' that we argued in \cite{arw99}

347: can occur\footnote{We will discuss precisely what 

348: we mean by ``chirality violation'' and ``the anomaly''

349: in this context shortly.} in certain reggeized gluon interactions due to a

350: quark loop.

351: In our future papers, we hope to demonstrate that a multi-regge

352: S-Matrix with all the desired properties is obtained by 

353: combining perturbative multi-regge

354: infra-red behavior with a treatment of the anomaly  

355: that clearly breaks the U(1) axial symmetry.

356: As a prelude to the full program, therefore, we must first establish 

357: that the U(1) anomaly (and also the chiral flavor anomaly - which plays a 

358: crucial role in the chiral symmetry breaking) does indeed appear in the 

359: framework of multiparticle multi-regge behavior. Our focus 

360: in this paper will be on explaining in detail the origin of 

361: this phenomenon. Our hope is that this explanation will add considerably 

362: to the arguments that we have already given\cite{arw99}. 

363: 

364: Reggeon diagrams contain reggeized gluon (or quark) propagators,

365: reggeon interaction vertices, and external couplings to the scattering states,

366: all of which are gauge invariant, even in perturbation theory. 

367: In general, many Feynman diagrams

368: give contributions to a single reggeon diagram and, therefore, to a 

369: single reggeon interaction. In addition, a reggeon 

370: interaction vertex has a significance that goes beyond it's perturbative 

371: description. As we will see, the essential ``non-perturbative'' element of the

372: calculations we describe is the role played by unphysical asymptotic

373: multiple discontinuities and the anomaly in contributions to

374: reggeon interaction vertices that they give. The anomaly in such contributions

375: can be non-perturbative in the sense that it

376: can be present in a reggeon vertex but not 

377: necessarily produce an effect in a perturbative 

378: amplitude in which it is contained. This is because

379: the structure of external couplings and additional symmetries of full

380: reggeon diagrams can be sufficient to produce a cancelation. 

381: Indeed, we argued in \cite{arw99} that when the scattering states are 

382: elementary quarks or gluons the anomaly always cancels in the full 

383: scattering amplitude. Conversely, we expect it to play a crucial role in 

384: the scattering of ``non-perturbative'' physical bound states.

385: 

386: Before enlarging further on the significance of unphysical asymptotic

387: multiple discontinuities and the results of this paper,

388: it will be helpful to briefly discuss the potential relationship 

389: between our work and the euclidean path-integral formalism. We note, first, 

390: that in Minkowski space the Dirac zero modes due to topological gauge fields 

391: are manifest\cite{aj} as the spectral flow of the eigenvalues of

392: the corresponding (gauge-dependent) ``Hamiltonian''. Since there 

393: is no complete non-perturbative Hamiltonian formalism for QCD, 

394: there is no well-developed understanding of what the general  

395: consequences of spectral flow might be. The common expectation is probably

396: that such phenomena will be overwhelmed

397: by strong-coupling effects of the kind usually assumed to be associated with

398: confinement. However, in a massless theory in which the gauge coupling remains 

399: small in the infra-red region, we should not expect this to be the case.

400: As a minimum, we anticipate that (in an appropriate background field)

401: zero energy fermion states identified initially 

402: as a particle (presumably within a boundstate) can evolve with time into 

403: a filled vacuum state of the corresponding Dirac sea 

404: and, similarly, filled vacuum states can evolve into particles. (The existence

405: of stable bound states and physical scattering processes in such an

406: environment is surely far from trivial!)

407: 

408: The U(1) anomaly can be interpreted in terms of infra-red

409: spectral flow as follows.

410: Associated with the anomalous divergence equation, 

411: the massless axial-vector triangle graph has\cite{cg} an infra-red divergence  

412: that involves a  zero four-momentum

413: propagator. Both the ``particle'' and ``antiparticle'' poles  

414: of the propagator contribute to the 

415: divergence. The coupling at one end of the propagator can be viewed 

416: as the vertex for production of the particle while simultaneously (and

417: symmetrically) that

418: at the other end describes the production of the antiparticle.

419: For the propagator to describe a physical zero momentum 

420: transition there must be 

421: spectral flow (due to a background gauge field) 

422: in that the production of the antiparticle (or the particle)

423: must be counted as the absorption of a particle (antiparticle). In this case, 

424: the transition becomes a ``chirality transition''.  

425: Consequently, in Minkowski space the U(1) divergence equation  

426: provides a connection between the topological 

427: structure of a background gauge field and the net 

428: infra-red ``spectral flow'' of zero momentum,

429: massless, Dirac particle and antiparticle states. 

430: 

431: In our analysis ``spectral flow'' is introduced by the appearance

432: of the triangle graph infra-red divergence (referred to above as 

433: the ``anomaly'') in reggeized gluon interactions.  

434: The triangle graph appears as an effective interaction generated by

435: particular multi-gluon interactions due to a quark loop. 

436: There are no axial-vector 

437: currents in the QCD interaction but, as we already described in \cite{arw99},

438: multi-regge effective interactions can contain  

439: components of an axial-vector interaction. In sufficiently high order,

440: interactions appear involving reggeon states with the quantum 

441: numbers of the anomaly (winding-number) current. Remarkably, perhaps,

442: we will establish in 

443: this paper that it is interactions of this last 

444: kind that have the contributions from unphysical multiple

445: discontinuities that we argued in \cite{arw99} are necessary for the

446: anomaly infra-red divergence to appear. As a result, the

447: U(1) problem is clearly encountered.

448: Indeed, we anticipate that the infra-red discussion we give 

449: is connected to ``ultra-violet'' problems (involving momenta flowing 

450: around the internal quark loop that are comparable in magnitude to 

451: the large external momenta) that reflect 

452: the usual relationship between infra-red and ultra-violet 

453: manifestations of the anomaly. We would expect  

454: short-distance interactions of the winding number current to appear

455: directly in this ultra-violet context. (It is, perhaps, 

456: unfortunate that the anomaly 

457: is a high-order, many gluon, phenomenon. However,

458: this is to be expected if the anomaly current, containing a 

459: product of three gluon fields, has to be  involved.)

460: 

461: We will call the basic process, in which a physical region

462: zero momentum propagator contributes to a triangle graph divergence, 

463: a chirality transition and will refer 

464: to the general phenomenon as 

465: ``chirality violation'', although we could

466: equally well call it spectral flow\footnote{Neither description is strictly

467: appropriate. Since we study only S-Matrix 

468: elements we can not define chirality via right and left-handed fields and 

469: since we do not have a hamiltonian we also can not define 

470: spectral flow in the normal manner. We also can not define the anomaly in 

471: terms of the divergence of an

472: axial current although we can, as we discussed in \cite{arw99},

473: relate it to the violation of reggeon Ward identities that normally are 

474: a consequence of gauge invariance.}. It will also be what we generically

475: refer to as ``the anomaly'', within our formalism.

476: In this paper, as in \cite{arw99},

477: we will concentrate on the feynman diagram amplitudes that

478: produce the anomaly in reggeon interactions 

479: and, apart from the brief description at the end of this Introduction, 

480: will not discuss the general program any further. 

481: In \cite{arw99} we distinguished

482: two methods for calculating multi-regge amplitudes - the direct calculation

483: of diagrams in light-cone co-ordinates and the calculation of multiple

484: asymptotic discontinuities with the subsequent use of an asymptotic dispersion

485: relation. We emphasized that the direct calculational method 

486: is impractical for the problem we are discussing. This is because of the 

487: large number of diagrams that could contribute

488: and because the complexity of the diagrams

489: makes a full discussion of 

490: whether or not integration contours are truly trapped, in the asymptotic  

491: limits involved, very difficult. Consequently the asymptotic dispersion 

492: relation method has to be used.

493: 

494: The form of the asymptotic dispersion relation for a given multi-regge 

495: process is determined by the  

496: asymptotic multiple discontinuities that satisfy the Steinmann relation

497: property that the discontinuities occur in non-overlapping invariant channels.

498: Such discontinuities are explicitly reflected in the analytic structure of 

499: asymptotic amplitudes and, conversely, 

500: using the dispersion relation, 

501: amplitudes can be calculated directly 

502: from the discontinuities\cite{jb,arw00}. 

503: The crucial feature of the high-order amplitudes that produce reggeon 

504: interactions containing the anomaly

505: is the presence of unphysical  multiple discontinuities that 

506: satisfy the Steinmann relation property.

507: Such discontinuities are present only in 

508: complex (imaginary momentum) parts of the asymptotic region for 

509: more complicated many-particle multi-regge 

510: processes, the simplest of which is the full triple-regge region\cite{gw}.

511: However, just because they

512: are in non-overlapping channels these discontinuities can 

513: (and must) consistently 

514: appear in the asymptotic amplitudes that describe also the 

515: real physical region behavior. 

516: 

517: The familiar amplitudes that appear in multi-regge

518: production processes (such as 

519: those that contribute to the BFKL equation\cite{fkl}) 

520: do not contain unphysical multiple discontinuities.

521: Rather they contain only multiple discontinuities

522: that are naturally interpreted as due to a succession of physical

523: region on-shell scattering processes\cite{jb}. (The necessity for a

524: physical time-ordering of such processes then determines the absence of

525: overlapping channel discontinuities.) Because the physical region multiple

526: discontinuities involve only physical amplitudes and physical intermediate

527: states, when they are calculated using the perturbative amplitudes

528: of the massless theory, they can not contain chirality transitions

529: associated with particle/antiparticle ambiguities. Therefore,

530: when only production processes are 

531: involved (i.e. at what we might call the BFKL level of multi-regge theory) 

532: there is no possibility for ``chirality violation''. 

533: In more elaborate scattering processes

534: the unphysical multiple discontinuities appear and they may, a-priori, 

535: contain potential chirality

536: transitions, even when calculated perturbatively. This is because

537: the discontinuities involved may contain what, in a physical region,

538: would be mixed combinations of forward-going particles and backward

539: going antiparticles (``mixed $\alpha$ singularities'', in the language

540: of S-Matrix Theory). Our purpose, in this paper, is to show that this 

541: phenomenon does occur 

542: and to demonstrate that, in the massless limit, the reggeon interactions

543: associated with the asymptotic amplitudes can contain 

544: the anomaly and so, potentially, can 

545: produce physical region infra-red divergences.

546: 

547: In our previous paper studying triple-regge interactions\cite{arw99} 

548: we already discussed why the anomaly could only appear in reggeized 

549: amplitudes containing unphysical triple discontinuities.

550: In practise, however, we only studied (what appeared to be) 

551: the lowest-order relevant  

552: diagrams, i.e. those that contain two gluons in each $t$-channel. 

553: We isolated the physical momentum configuration 

554: within ``maximally non-planar'' 

555: diagrams that, in the massless (quark)

556: limit, could potentially give an infra-red 

557: divergence associated with the anomaly, provided the appropriate 

558: on-shell propagators contribute to the asymptotic behavior. However,

559: although they are maximally non-planar these diagrams do not 

560: have the complexity required to  

561: contain the unphysical triple discontinuities that, according to our method

562: of analysis, would determine that these propagators do contribute.

563: We noted, nevertheless, that the necessary 

564: discontinuities did appear to be present in

565: the higher-order amplitudes that would give the reggeization of the gluons 

566: in the diagrams we studied. Therefore, we argued, the anomaly

567: configurations in the lowest-order diagrams

568: could  be required as (generalized) real parts

569: needed to accompany the higher-order unphysical

570: triple discontinuities. Paradoxically, perhaps,

571: we simultaneously suggested that there would 

572: be cancelations among diagrams 

573: such that the anomaly would survive only when reggeon states

574: with the quantum numbers of the winding-number current are involved. 

575: 

576: In this paper we will study

577: the high-order reggeization diagrams in detail and will find that the 

578: situation is actually simpler than we suggested. The anomaly infra-red

579: divergence is produced by a quark loop in which many propagators 

580: are on-shell and one 

581: propagator carries the zero momentum and energy that allows a chirality

582: transition. The on-shell conditions have to be associated with a 

583: triple discontinuity in such a way that a triplet

584: of the on-shell particles (each associated with a separate discontinuity) 

585: are all quarks (or all antiquarks). 

586: It is straightforward to see that this ``all quarks'' requirement

587: can not be satisfied by a physical discontinuity and that, in fact, it is 

588: very difficult to satisfy. Indeed, we find that 

589: the reggeization diagrams that we 

590: suggested in \cite{arw99} might contain the anomaly actually 

591: do not satisfy the all-quarks requirement. As we proceed to higher orders

592: we eventually find that this requirement is satisfied. However, a final

593: requirement that the spin structure that generates the anomaly also be 

594: present, further restricts the triple discontinuities that can

595: contribute. Eventually

596: we arrive at (a small class of) diagrams

597: that contain a triple discontinuity with all the right properties. 

598: However, this discontinuity is truly unphysical in that it 

599: occurs as a combination of three ``asymptotic pseudothresholds'' 

600: each of which contains particles, effectively,

601: going in opposite time directions. 

602: The reggeon interactions produced are also of sufficiently high order that 

603: the minimum circumstances in which they can occur (between color zero

604: reggeon states) is when each of the states involved carries

605: the quantum numbers of the U(1) anomaly current. Nevertheless, 

606: this establishes the essential result of this paper that the 

607: triangle anomaly does 

608: occur in reggeized gluon interactions extracted from unphysical multiple 

609: discontinuities and that the phenomenon we are

610: discussing is indeed the U(1) anomaly. The results of this paper

611: also imply that that the lower-order diagrams considered in \cite{arw99},

612: although valuable to discuss for illustrative processes, are 

613: essentially irrelevant. 

614: 

615: In this paper we are satisfied to simply demonstrate

616: that there are diagrams which generate a reggeon interaction 

617: in which the anomaly appears, and that the reggeon

618: states involved have the quantum numbers of the anomaly current. We

619: do not discuss whether there are cancelations that could occur.

620: We postpone this for the following papers that will lay out the details 

621: of the construction of the bound-state S-Matrix alluded to above.

622: For the moment we note only that triple-regge interactions of

623: the kind we consider here will contribute generally to the vertices 

624: and interactions of the reggeon bound states that emerge 

625: and refer to the brief discussion 

626: in \cite{arw99}, and also the outline in

627: \cite{arw001}, for more details. A brief, general,

628: description of the anticipated construction is as follows. 

629: 

630: We expect to obtain massless QCD from the massive theory 

631: in two stages. In the first stage, the  

632: (spontaneously-broken) gauge symmetry is restored to SU(2). 

633: The U(1) chiral symmetry is broken by the introduction of a ``wee-parton''  

634: condensate with anomaly current quantum numbers in 

635: scattering reggeon bound states.

636: Our expectation is that an anomaly infra-red divergence then appears

637: and determines the ``physical scattering amplitudes''. After the divergence

638: is factorized off, the condensate self-consistently

639: appears in all intermediate and final reggeon states.    

640: An essential ingredient will be to show\cite{arw99}

641: that if we regulate the anomaly ultra-violet divergences involved, 

642: the infra-red divergence gives a result that is independent of the

643: regularization used. (Note that,

644: by identifying, regulating, and organizing how chirality violation 

645: divergences produce reggeon states and scattering processes,

646: we constrain how, in the path-integral formalism, 

647: non-perturbative topological gauge fields must contribute to the 

648: massless theory. Indeed, if we succeed in our goals,

649: we will implicitly determine how the  

650: spectral flow of the Dirac sea must contribute if 

651: a unitary high-energy S-Matrix is to be obtained.)

652: 

653: U(1) chirality violation appears within interactions of

654: the pomeron and additional chirality violation, related to the 

655: anomaly in flavor current vertices, is responsible for the appearance 

656: of the pion and the ``nucleon''. (At this stage, 

657: the nucleon (to be) is a chiral Goldstone boson, just like the pion.)

658: The pomeron should be in a supercritical phase

659: of Reggeon Field Theory and the spectrum of 

660: bound-states should have both SU(2) confinement

661: and chiral symmetry breaking\footnote{It is possible, if not likely, 

662: that the role of the

663: Dirac sea in producing confinement in this context is related to that 

664: proposed by Gribov\cite{gr}.}. In the second stage,

665: the full gauge symmetry is to be restored

666: by the randomization of the SU(2) condensate within SU(3). The randomization

667: should correspond to a phase-transition within Reggeon Field Theory. 

668: We expect that the 

669: asymptotic freedom requirement that contact with perturbation theory

670: remain at short distances determines that the theory must be right 

671: at the critical point

672: associated with critical behavior for the pomeron\cite{cri}.

673: 

674: Finally, we note that

675: while we have not studied the issue in any detail, we believe that 

676: when the fermions involved are massive only the ``ultra-violet anomaly'' 

677: is present. As we remarked above, this will produce 

678: problems (for bound-state scattering amplitudes) 

679: in the  (internal) momentum region of 

680: reggeon interactions where the momenta are  

681: of the order of, or larger, than the external regge limit momenta and effective

682: interactions of the anomaly current should be 

683: directly involved. A-priori, we expect the presence of the anomaly, 

684: in this form,

685: to lead to the violation of reggeon Ward identities and to 

686: increased power behavior asymptotically for 

687: the amplitudes in which it is contained. A consequent 

688: violation of unitarity bounds by potential bound-state amplitudes

689: is therefore threatened and this is clearly 

690: where non-perturbative topological contributions

691: could be crucial. Of course, if the anomaly produces infra-red divergent

692: amplitudes in the massless theory, it could also imply that the 

693: reggeon diagram result for the asymptotic behavior is wrong and unitarity

694: bounds could be threatened. However, the implication of the infra-red 

695: divergence structure we envision is that the divergences both select

696: the physical states and can be absorbed into a redefinition of the

697: states that leaves reggeon asymptotic behavior intact.  

698: Thus, as a matter of both principle and practicality we believe

699: that the massless multi-regge amplitudes must be constructed first - by a 

700: procedure that regularizes the ultra-violet anomaly and allows the infra-red

701: behavior to dominate.  

702: We also believe that it is the relative simplicity of the infra-red structure 

703: of the anomaly, together with the unitarity properties of reggeon 

704: diagams and their relationship with Reggeon Field Theory, that will 

705: allow us to carry out such a procedure.

706: 

707: \newpage

708: 

709: \mainhead{2. MULTIPLE DISCONTINUITIES AND THE STEINMANN RELATIONS.} 

710: 

711: The Steinmann relations originated in axiomatic field theory\cite{ste}. They 

712: (essentially) describe the restrictions that the time-ordering of interactions

713: places on the combinations of intermediate states that can occur in a

714: scattering process. For on-shell S-Matrix amplitudes their significance is

715: most immediately appreciated in the approximation that we ignore higher-order

716: Landau singularities and consider only the normal threshold branch points (and

717: stable particle poles) that occur in individual channel invariants. The

718: Steinmann relations then say that simultaneous thresholds (and/or poles) can

719: not occur in overlapping channels. (Channels overlap if they contain a common

720: subset of external particles.) As a result an $N$-point amplitude has at most

721: $N-3$ simultaneous cuts (or poles) 

722: in distinct invariants. The possible combinations of

723: cuts can be described by tree diagrams with three-point vertices in which each

724: internal line corresponds to a channel invariant in which there is a cut due

725: to intermediate state thresholds - as illustrated in Fig.~2.1 for the 7-point

726: amplitude.

727: \begin{center}

728: \epsfxsize=4in

729: \epsffile{chir2.ps}

730: 

731: Fig.~2.1 A Tree Diagram Representing Simultaneous Invariant Cuts.

732: \end{center}

733: (As usual, $s_{12}=(P_1+P_2)^2~, s_{123}=(P_1+P_2+P_3)^2~,$

734: etc.) The set of all combinations of thresholds (and poles) allowed by the 

735: Steinmann

736: relations is the basic singularity structure of all scattering amplitudes. The

737: higher-order Landau singularities are believed\cite{arw00} 

738: to emerge from the normal

739: thresholds in a manner that for many purposes (including ours, as we discuss

740: shortly,) makes them a secondary effect.

741: 

742: Conversely, the combination of cuts represented 

743: by a particular tree diagram can

744: be directly associated with a set of physical scattering processes. As

745: illustrated in Fig.~2.2, this is the set of all processes (involving all the

746: external particles of the diagram as either ingoing or outgoing particles) in

747: which it is kinematically possible for all of the internal lines to be

748: replaced by physical 

749: multiparticle states\footnote{We do not distinguish processes in

750: which ingoing and outgoing particles are interchanged via CPT conjugation}.

751: \begin{center}

752: \epsfxsize=6in

753: \epsffile{chir4.ps}

754: 

755: Fig.~2.2 Physical Scattering Processes Corresponding to Fig.~2.1.

756: \end{center}

757: The hatched segments represent physical intermediate states that, if they are 

758: all placed on shell, give (essentially) the associated multiple discontinuity.

759: 

760: The Steinmann relations play a fundamental role in multi-regge theory. It is 

761: possible to show\cite{arw00} that in a physical multi-regge asymptotic region 

762: the analytic structure of scattering amplitudes can be treated as if only 

763: normal thresholds satisfying the Steinmann relations were present. In effect,

764: higher-order Landau singularities are suppressed. This has the very important

765: consequence that only the normal threshold cuts in individual channel 

766: invariants need be represented by multi-regge

767: asymptotic formulae. Furthermore, if we consider only the multi-regge limits

768: accessible in $2 \to M$ production processes, it can be shown that the maximal

769: number (M-1) of simultaneous thresholds is encountered asymptotically only in

770: physical regions. This is a generalization of the cut-plane analyticity

771: property familiar from elastic scattering.

772: 

773: If we consider the multi-regge regions of $M \to M'$ scattering amplitudes

774: ($M,M' \geq 3$) there is a significant change. To understand the point 

775: involved consider the simplest case of the tree diagram of

776: Fig.~2.3. At first sight this diagram corresponds only to the 

777: $2 \to 4$ production processes shown. 

778: \begin{center}

779: \epsfxsize=5in

780: \epsffile{chir3.ps}

781: 

782: Fig.~2.3 A Tree Diagram and Corresponding Physical Scattering Processes.

783: \end{center}

784: The three distinct scattering processes are distinguished 

785: by different constraints on the invariants, i.e.

786: $$

787: \eqalign{\hbox{i)}& ~\sqrt{s_{12}} ~> ~\sqrt{s_{34}}+\sqrt{s_{56}}~,~~ \cr

788: \hbox{ii)}&~ \sqrt{s_{34}} ~> ~\sqrt{s_{12}}+\sqrt{s_{56}}~,~~\cr

789: \hbox{iii)}&~ \sqrt{s_{56}} ~>~ \sqrt{s_{12}}+\sqrt{s_{34}} }

790: \auto\label{mag}

791: $$

792: We can also regard the three processes involved as distinguished by the 

793: selection of one

794: pair of particles as incoming, which then must have energy larger than the

795: sum of the subenergies of the other two pairs, which are necessarily in the 

796: outgoing state.

797: 

798: We may wonder about the symmetric asymptotic region in which

799: $$

800: \sqrt{s_{12}} ~\sim ~\sqrt{s_{34}}~\sim~ \sqrt{s_{56}} ~~\to~~ \infty

801: \auto\label{mag1}

802: $$

803: There are no physical scattering processes in this region. However,

804: the three processes of (\ref{mag}) are described by the same (analytically

805: continued) amplitude and so analytic continuation from each of the 

806: physical regions implies that such cuts must be present. 

807: It is, perhaps, natural 

808: that a triple discontinuity should exist that is 

809: symmetric with respect to the three processes of Fig.~2.3. Apparently, though,

810: the symmetry requirement could only be satisfied if all the external

811: particles are in the 

812: final, or initial, state. However, as we discuss further 

813: in Sections 4 and 5, 

814: if we allow particles to carry complex momenta, a positive 

815: value for a two-particle energy invariant can  

816: be achieved by a combination of an ``incoming'' and an 

817: ``outgoing'' particle in that they carry opposite sign, but imaginary, 

818: energies. Therefore, in the symmetric region it is possible for the 

819: three cuts of Fig.~2.3 

820: to be present if each is associated with such a combination.    

821: We will show in the following that

822: there are unphysical processes (with imaginary momenta) in this region that

823: do produce a triple discontinuity of this kind and we will refer to it as an 

824: ``unphysical triple disconinuity''.

825: 

826: Because the external particles for each cut are both  

827: ingoing and outgoing, intermediate states can also be produced that involve

828: such combinations. As a result, the triple discontinuity can contain the 

829: ``particle - antiparticle'' transitions that ultimately provide  

830: the massless chirality transitions that we are looking for. 

831: Moreover, since the complex momentum

832: part of (\ref{mag1}) is contained in the triple-regge

833: asymptotic region, the triple discontinuity 

834: must be present in triple-regge asymptotic formulae.

835: This is possible just because this 

836: combination of cuts has the Steinmann property. 

837: Moreover, because of the potential for chirality violation,

838: it is a natural context in which to see the anomaly appear.

839: The importance of the triple-regge region is that it is the simplest

840: multi-regge limit in which the vertices appear that provide the couplings

841: of bound-state regge

842: poles such as the pomeron or the pion. For higher-point $M \to M'$

843: amplitudes there is a wide range of unphysical multiple discontinuities 

844: satisfying the Steinmann relations. Bound-state scattering amplitudes can thus

845: appear in which the anomaly is a crucial element.

846: 

847: \newpage

848: 

849: \mainhead{3. THE PHYSICAL REGION ANOMALY AND THE 

850: TRIPLE-REGGE DISPERSION RELATION}

851: 

852: In our previous paper\cite{arw99} we studied the full triple-regge

853: limit\cite{gw} of 

854: three-to-three quark scattering. 

855: If we denote the initial momenta as $P_i~, ~i=1,2,3$, and the final momenta 

856: as $- P_{i'} = P_i + Q_i, ~i=1,2,3$,

857: the triple-regge limit can be realized, within the physical region, by taking 

858: each of $P_1,~P_2$ and $P_3$ large 

859: along distinct light-cones, with the momentum transfers $Q_1, Q_2$ and $Q_3$

860: kept finite, i.e.

861: \newline \parbox{3.1in}{ 

862: $$

863: \eqalign{ P_1~\to&~ P_{1^+}~= ~(p_1,p_1,0,0)~,~~p_1 \to \infty \cr

864: P_2~\to&~ {P_2^+}~= ~(p_2,0,p_2,0)~,~~p_2 \to \infty \cr

865: P_3~\to&~ {P_3^+}~= ~(p_3,0,0,p_3)~,~~p_3 \to \infty  }

866: $$}

867: \parbox{2.9in}{

868: $$ \eqalign{

869: ~~~q_1=Q_1/2~\to&~~ (\hat{q}_1,\hat{q}_1,q_{12},q_{13})\cr

870: ~~~q_2=Q_2/2~\to&~ ~(\hat{q}_2,q_{21},\hat{q}_2,q_{23})\cr

871: ~~~q_3=Q_3/2~\to&~~(\hat{q}_3,q_{31},q_{32},\hat{q}_3)}

872: \auto\label{np3}

873: $$}

874: Momentum conservation gives 

875: a total of five independent $q$ variables which, along  

876: with $p_1, p_2$ and $p_3$, give the necessary eight variables. The definition

877: of the triple-regge limit in terms of angular variables is 

878: given in \cite{arw99}. For our present purposes the above definition in 

879: terms of momenta will be sufficient. This will alow us to avoid the

880: extra complication of defining helicity angles, helicity-pole limits

881: etc. It will be important that the

882: asymptotic behavior involved must hold also for all complex values of the large momenta,

883: including the additional physical regions reached by reversing the 

884: signs of the $p_i$.

885: 

886: In \cite{arw99} we also studied feynman diagrams that contain a 

887: closed quark loop and generate

888: triple-regge reggeized gluon interactions containing the loop.

889: We considered the lowest-order amplitudes in which the anomaly could 

890: potentially appear and, in particular, studied maximally non-planar 

891: diagrams of the 

892: kind shown in Fig.~3.1(a). 

893: \begin{center}

894: \leavevmode

895: \epsfxsize=4.5in

896: \epsffile{chir9.ps}

897: 

898: Fig.~3.1 A maximally non-planar diagram and 

899: rhe triangle diagram reggeon interaction prouced.   

900: 

901: \end{center}

902: (As usual, the solid and wavy lines 

903: respectively represent a quark and a gluon.

904: We have reversed the direction of $P_3$ relative 

905: to the notation of \cite{arw99} in order to have a completely symmetric

906: notation.) The leading asymptotic contributions come from 

907: regions of gluon loop integrations where some of the 

908: propagators in the quark loop and the scattering quark systems

909: are on-shell. We discuss the determination of which propagators can be

910: on-shell below. For the moment we consider the possibility, discussed at 

911: length in \cite{arw99}, that the 

912: on-shell lines are those that are hatched in Fig.~3.1(a). We will eventually 

913: conclude that this combination of on-shell propagators can not produce a 

914: reggeon interaction with a physical region anomaly divergence, 

915: even though it does produce a triangle

916: diagram interaction. As we will see, the issue is not just which 

917: propagators are placed on-shell but also which pole (``particle'' or 

918: ``antiparticle'') is involved. (As the discussion in the previous 

919: Section suggested,

920: for the unphysical discontinuities, with which we will ultimately 

921: be concerned, the answer to this question is not necessarily unambiguous.)

922: In the following we initially 

923: ignore this subtlety. As it emerges in 

924: our discussion it will become clear that it is a vital part 

925: of the search for further diagrams which do produce an interaction 

926: containing the anomaly.

927: 

928: If the hatched on-shell propagators are used to carry out

929: light-like longitudinal momentum integrations the integrals over gluon loop

930: momenta reduce to two-dimensional ``transverse'' integrals over

931: spacelike momenta, as illustrated by Fig.~3.1(b). The transverse 

932: plane (and orthogonal light-like momenta)

933: can, in general, be chosen differently in each $t$-channel. 

934: If $Q_{i\perp}$ is the projection of $Q_i$ on the corresponding 

935: transverse plane, the leading asymptotic contribution then has the form

936: $$

937: \eqalign{ ~~~~~P_{1^+}~ P_{2^+}~ P_{3^+}~

938: \prod_{i=1}^3 \int & { d^2 k_{i1}d^2 k_{i2}\over  k_{ i1}^2  k_{i2}^2}  

939: ~~ \delta^2 (Q_{i\perp} -  k_{i1} -  k_{i2})~G^2_i(k_{i1},k_{i2},\cdots)

940: \cr &~~~~~~~~\times ~ R^6(Q_1,Q_2,Q_3,

941: k_{11}, k_{12}, \cdots )} \auto \label{211}

942: $$

943: where $ R^6(Q_1,Q_2,Q_3,k_{11}, k_{12}, \cdots )$  

944: can be identified with the ``reduced'',

945: or ``contracted'', feynman diagram of Fig.~3.1(c). If we write

946: $$

947: k_{i1} ~= ~q_i + k_i~, ~~~~ k_{i2} ~= ~q_i - k_i~,

948: \auto\label{dki}

949: $$

950: then (with a particular choice\cite{arw99} of transverse planes)

951: $$

952: \eqalign{ &R^6(q_1,q_2,q_3,k_1,k_2,k_3) ~=\cr

953: & \int d^4 k  {  Tr \{ 

954: \gamma_5 \gamma^{-,-,+} (\st{k}+ \st{k}_1 + \st{q}_2 +\st{k}_3) 

955: \gamma_5 \gamma^{-,-,-} ~\st{k}~ 

956: \gamma_5 \gamma^{-,-,-}(\st{k}- \st{k}_2 + \st{q}_1 + \st{k}_3 ) \}

957: \over  (k + k_1 + q_2 + k_3 )^2  

958: ~k^2 ~

959:  (k - k_2 + q_1 + k_3)^2 }  ~+ ~ \cdots }

960: \auto\label{580}

961: $$

962: where

963: $$

964: \gamma^{\pm,\pm,\pm} ~=~ \gamma^{\mu}\cdot n^{\pm,\pm,\pm}_{ \mu} ~,~~~~

965: n^{\pm,\pm,\pm \mu} ~= ~ (1,\pm1,\pm1,\pm1)

966: \auto\label{g64}

967: $$

968: The contributions to $R$ not shown explicitly in (\ref{580}) do not have

969: a $\gamma_5$ at all three vertices of the triangle diagram. The particular

970: $\gamma$-matrix projections appearing depend on the choice of transverse

971: co-ordinates. If the anomaly is present in $R$, however, we expect it to be

972: independent of this choice. We should

973: emphasize that while we have written (\ref{580}) as a function of

974: four-dimensional momenta, the $k_i$ are restricted to be 

975: two-dimensional spacelike momenta (plus longitudinal components

976: determined by the mass-shell conditions for the on-shell quarks) and 

977: the $q_i$ have the restricted form given by (\ref{np3}). These restrictions 

978: plays a crucial role in determining whether the anomaly can occur in a physical

979: region reggeon interaction.

980: 

981: A reggeon diagram amplitude 

982: that represents right-hand cuts in the unphysical triplet 

983: $\{s_{13'}, s_{32'}, s_{21'}\}$ and has two reggeons in each $t$-channel, 

984: each with trajectory $\alpha(t) = 1 +~O(g^2)$, has the form\cite{arw99}

985: $$

986: \eqalign{ &~~\prod_i\int { d^2k_i \over 

987: sin \pi \alpha (k_i^2)  sin \pi \alpha ((Q_i -k_i)^2)  } 

988: ~~~~\beta(k_1,k_2,k_3,Q_1,Q_2,Q_3)\cr 

989: &\biggl[ ~(s_{13'})^{[\alpha (k_1^2)+\alpha ((Q_1 -k_1)^2) + 

990: \alpha (k_3^2)+\alpha ((Q_3 -k_3)^2) -

991: \alpha (k_2^2)-\alpha ((Q_2 -k_2)^2) -1]/2} \cr

992: & ~~~~~~~~ (s_{32'})^{[\alpha (k_3^2)+\alpha ((Q_3 -k_3)^2) + 

993: \alpha (k_2^2)+\alpha ((Q_2 -k_2)^2) -

994: \alpha (k_1^2)-\alpha ((Q_1 -k_1)^2) -1]/2} \cr

995: & ~~~~~~~~~~~~(s_{21'})^{[\alpha (k_1^2)+\alpha ((Q_1 -k_1)^2) + 

996: \alpha (k_2^2)+\alpha ((Q_2 -k_2)^2) -

997: \alpha (k_3^2)-\alpha ((Q_3 -k_3)^2) -1]/2} \biggr]~ ~~\bigg/ \cr

998: &~~~~\biggl[[sin {\pi \over 2} [\hbox{${\scriptstyle\alpha (k_1^2)

999: +\alpha ((Q_1 -k_1)^2) + 

1000: \alpha (k_3^2)+\alpha ((Q_3 -k_3)^2) -

1001: \alpha (k_2^2)-\alpha ((Q_2 -k_2)^2)}$}] \cr

1002: &~~~~~~~~~~ sin {\pi \over 2} [\hbox{${\scriptstyle \alpha (k_3^2)

1003: +\alpha ((Q_3 -k_3)^2) + 

1004: \alpha (k_2^2)+\alpha ((Q_2 -k_2)^2) -

1005: \alpha (k_1^2)-\alpha ((Q_1 -k_1)^2)}$} ] \cr

1006: & ~~ ~~~~~~~~~~~~sin {\pi \over 2} [\hbox{${\scriptstyle \alpha (k_1^2)

1007: +\alpha ((Q_1 -k_1)^2) + 

1008: \alpha (k_2^2)+\alpha ((Q_2 -k_2)^2) -

1009: \alpha (k_3^2)-\alpha ((Q_3 -k_3)^2)}$} ] \biggr]}

1010:  \auto \label {2ra1}

1011: $$

1012: $$

1013: \centerunder{$\sim$}{\raisebox{-6mm}{$ g^2 \to 0$}}~

1014: (s_{13'})^{1/2}(s_{32'})^{1/2}(s_{21'})^{1/2}

1015: ~\prod_i\int{ d^2k_i \over 

1016: k_i^2  (Q_i -k_i)^2  }

1017: ~~\beta_(k_1,k_2,k_3,Q_1,Q_2,Q_3) ~~

1018: \auto\label{2ra}

1019: $$

1020: Taking the triple discontinuity in $s_{13'}$, $s_{32'}$

1021: and $s_{23'}$ removes the poles due to the sine factors in the square 

1022: bracket, but leaves the $g^2 \to 0$ limit unchanged. Since   

1023: the triple discontinuity is unphysical and

1024: of the kind discussed in the previous Section,

1025: according to the discussion in

1026: \cite{arw99}, the ``six-reggeon interaction vertex''

1027: $\beta_(k_1,k_2,k_3,Q_1,Q_2,Q_3)$ could contain the anomaly.

1028: 

1029: Writing

1030: $$

1031: P_{1^+} P_{2^+} P_{3^+}~\equiv ~(s_{13'})^{1/2}(s_{32'})^{1/2}(s_{21'})^{1/2}

1032: \auto\label{p+inv}

1033: $$

1034: and comparing with (\ref{2ra}) we see that (\ref{211}) 

1035: could be identified as a lowest-order contribution to such a

1036: reggeon diagram amplitude if the reduced feynman

1037: diagram amplitude of Fig.~3.1(c) is identified as a reggeon vertex, i.e.

1038: $$

1039: R^6(Q_1,Q_2,Q_3, k_{1},Q_1- k_{1},\cdots)~\equiv ~

1040: \beta(k_1,k_2,k_3,Q_1,Q_2,Q_3)

1041: \auto\label{6rv}

1042: $$

1043: Therefore, if (\ref{211}) does represent a contribution to the

1044: asymptotic behavior of the feynman diagram in Fig.~3.1(a) it could   

1045: contribute to a triple-regge amplitude 

1046: of the form of (\ref{2ra1}). Note that while the right-side of (\ref{p+inv})

1047: clearly has a triple discontinuity in $\{s_{13'}, s_{32'}, s_{21'}\}$,

1048: the left-side does not. The equivalence of the two sides is only determined

1049: if higher-order terms in (\ref{2ra1}) appear and add to (\ref{211}) in 

1050: the appropriate manner. Such terms are contributed by what we refer to as 

1051: reggeization diagrams, both in the Introduction and in the following.   

1052: 

1053: As we have emphasized, the amplitude (\ref{580}) representing Fig.~3.1(c) 

1054: is the full four-dimensional

1055: triangle diagram amplitude except that special $\gamma$-matrices

1056: appear at the vertices and

1057: only combinations of (essentially) two-dimensional transverse  

1058: momenta flow through the diagram.  As discussed in \cite{arw99}, the $\gamma$-matrix

1059: couplings are appropriate to produce the anomaly but 

1060: whether the necessary 

1061: momentum configuration can occur within a physical region 

1062: and provide a physical region inra-red divergence is a non-trivial

1063: and subtle question that depends crucially on the choice of propagator

1064: poles used to put lines on-shell, as we now describe.

1065: 

1066: The divergence of the (massless) triangle diagram occurs\cite{cg,arw99} 

1067: when a single light-like momentum flows through the diagram and

1068: all other momenta are spacelike and scaled to zero.

1069: Such a momentum configuration for the reggeon interaction $R$ appears to be 

1070: (essentially uniquely) realized by that of the 

1071: full feynman diagram shown in Fig.~3.2(a).

1072: \newline \parbox{4.3in}{

1073: \begin{center}

1074: \epsfxsize=4.1in

1075: \epsffile{chir6.ps}

1076: \newline (a)

1077: \end{center}}

1078: \parbox{1.7in}{

1079: \begin{center}

1080: \leavevmode

1081: \epsfxsize=1.5in

1082: \epsffile{chir60.ps}

1083: \newline (b)

1084: \end{center}}

1085: \begin{center}

1086: Fig.~3.2 The basic anomaly process.

1087: \end{center}

1088: If we label the momenta entering the reggeon interaction 

1089: as in Fig.~3.2(b),

1090: an explicit configuration for Fig.~3.2(a) is

1091: $$

1092: q_1-k_1~=~(2l,2l,0,0)~, ~~~

1093: q_2-k_2~=~(-2l,0,-2l,0)

1094: \auto\label{chm1}

1095: $$ 

1096: together with

1097: $$

1098: \hat{q}_1=- \hat{q}_2 =l ~~~~q_{13}=-q_{23} ~~~~ q_{12}=q_{21}=0

1099: \auto\label{chm2}

1100: $$

1101: This determines $k_1$ and $k_2$ and also gives

1102: $$

1103: q_3~=~- (q_1+q_2)~= ~(0,-l,l,0)

1104: \auto\label{chm3}

1105: $$

1106: If we then take

1107: $$

1108: k_3~=~l(0, 1 -2~ \cos{\theta}~, 1 - 2~\sin{\theta}~, 0) 

1109: \auto\label{chm4}

1110: $$

1111: the light-cone momentum 

1112: $$

1113: -~2l(1, \cos{\theta},\sin{\theta}, 0)

1114: \auto\label{chm30}

1115: $$

1116: flows along the two vertical non-hatched lines in Fig.~3.2(b). 

1117: It is straightforward

1118: to check that all three of the hatched lines are on mass-shell.

1119: If spacelike momenta of $O(q)$ are added to the momentum 

1120: configuration (\ref{chm1})-(\ref{chm30}) and the limit $q \to 0$ is taken

1121: the presence of the anomaly divergence will lead to the behavior

1122: $$

1123: R^6(l,\theta,q) ~~\centerunder{$\sim$}{\raisebox{0.3mm}{$q \to 0$}}  

1124: ~~ {(1 - \cos{\theta} - \sin{\theta})^2

1125: ~l^2 \over q } 

1126: \auto\label{anomd}

1127: $$

1128: 

1129: Apart from the reversal of direction for $P_3$,  the process represented by 

1130: Fig.~3.2(a) 

1131: is what we called ``the basic anomaly process'' in \cite{arw99}.

1132: The scattering should be thought of as taking place with the time axis

1133: vertical on the page. The space axes are not, of course, represented 

1134: accurately since each of the external quarks is traveling along orthogonal

1135: directions. The dashed lines indicate light-like (``wee

1136: parton'') gluons with finite momenta parallel to the 

1137: incoming/outgoing quark that they are emitted/absorbed by.

1138: A zero-momentum quark (indicated by the open line) 

1139: is emitted by the incoming wee-parton gluon and  

1140: is absorbed by the outgoing wee-parton gluon. 

1141: The first and last gluon 

1142: interactions of the antiquark rotate it's incoming/outgoing

1143: lightlike momentum to/from the

1144: triangle light-like momentum associated with the infra-red divergence.

1145: The two intermediate gluons carry equal but opposite transverse 

1146: momenta. Their combination provides a 

1147: forward scattering of the antiquark that, most importantly,

1148: includes a spin flip (the momentum factor for which reduces what would be a $1/q^2$

1149: factor in (\ref{anomd}) to $1/q$). 

1150: 

1151: The zero momentum quark is produced by one wee

1152: gluon and absorbed by the other, allowing the chirality

1153: transition produced by the anomaly

1154: to compensate for the spin flip of the antiquark.  

1155: Note, however, that when the wee gluons are massless,

1156: the scattering processs represented by Fig.~3.2 is physical only when the 

1157: quark and antiquark involved are also massless. In addition, as we 

1158: noted in the Introduction, the 

1159: anomaly infra-red divergence involves both poles of 

1160: the zero momentum quark propagator. According to the helicity analysis

1161: of \cite{cg} the vertices coupling to the propagator 

1162: should be symmetrically interpreted

1163: as describing either the simultaneous production of the two states 

1164: in the propagator or their simultaneous 

1165: absorption. Therefore, if (the infra-red divergence analysis that we 

1166: ultimately employ to define physical states and amplitudes should

1167: require that) we interpret 

1168: the massless scattering as entering the physical region

1169: with the time ordering implied by Fig.~3.2, we are implicitly

1170: assuming the presence of a non-perturbative background gauge field. The

1171: background field would be needed to

1172: produce the necessary spectral flow at one vertex 

1173: that is required to interpret the process as a chirality transition. 

1174:  

1175: While the mass-shell

1176: conditions are indeed satisfied by (\ref{chm1})-(\ref{chm30}), we must now discuss

1177: the important subtlety as to which propagator pole is chosen. 

1178: With the momenta given by  (\ref{chm1})-(\ref{chm30}),

1179: the energy component of each of the three hatched lines in Fig.~3.2(b)

1180: has the same sign.

1181: Since the exchanged gluons carry only spacelike momenta, it is clear that 

1182: this must be the case. Equivalently, the

1183: on-shell states in the loop must be treated symmetrically in

1184: that, if the zero momentum state is an antiquark (quark), 

1185: all hatched lines must be quarks (antiquarks).

1186: We refer to this as  the ``all quarks requirement''. 

1187: As we already remarked on in the Introduction, and as is discussed at length 

1188: in \cite{arw99}, the only practicable calculational method

1189: to determine whether

1190: a given combination of on-shell lines contributes to the triple-regge behavior

1191: (after all diagrams are added) is the dispersion relation method that 

1192: we outline very briefly below. In this 

1193: approach all on-shell lines 

1194: in a reggeon interaction result directly from the taking of a 

1195: triple asymptotic 

1196: discontinuity. ``Real part'' interactions with the 

1197: same on-shell lines may be

1198: generated when the full dispersion relation is written

1199: or, equivalently, multi-regge theory is used\cite{arw99} 

1200: to convert the triple discontinuity to a full amplitude. 

1201: 

1202: In fact, the ``all quarks requirement''  

1203: is (as we shall see from the examples we

1204: discuss below) very difficult to obtain in a reggeon interaction derived 

1205: from a multiple discontinuity. We believe, however, that it is an essential

1206: requirement for a reggeon interaction to contain a physical region divergence

1207: produced by the anomaly, i.e. some variant of the ``basic anomaly process''

1208: must be involved. We recognized in \cite{arw99} that 

1209: the necessary triple discontinuity is not present in the diagram of Fig.~3.1

1210: but we suggested that nevertheless it may be present in the higher-order 

1211: reggeization

1212: diagrams that produce the reggeization of the gluons and so the basic anomaly 

1213: process of Fig.~3.2 may be required as a real part interaction

1214: (via the equivalence (\ref{p+inv})).

1215: In fact, we will show in the remaining part of

1216: this paper that this is not the case. Instead,

1217: to satisfy the all quarks requirement,

1218: there must be at least two wee gluons (instead of one) either emitted, or 

1219: absorbed, in the basic anomaly process. Ultimately this implies that reggeon

1220: interactions with the quantum numbers of the winding number current must

1221: be involved. 

1222: As we emphasized in the Introduction (and also discussed in \cite{arw99}) 

1223: we do not expect the anomaly divergence to be present in the scattering of

1224: elementary quarks and/or gluons after all diagrams are summed. Rather,

1225: as we briefly comment on in Section 6, we 

1226: expect it to be present when  

1227: the basic process is generalized to describe the scattering of 

1228: the particular multi-regge states that ultimately form

1229: bound states. The corresponding $G_i$ will then appear in 

1230: a generalization of (\ref{211}) and the 

1231: wee partons involved will be a crucial characteristic of  

1232: scattering states. Also the chirality transitions produced (and the implicit 

1233: spectral flow) will be an essential part of scattering processes.

1234: 

1235: In general, an asymptotic dispersion relation\cite{arw00} gives the leading 

1236: multi-regge behavior of an amplitude as

1237: a sum over multiple discontinuity contributions 

1238: allowed by the Steinmann relations. For the particular case (described in

1239: detail in \cite{arw99}) of 

1240: the triple-regge 

1241: behavior of a six-point amplitude we can write 

1242: $$

1243: M_6(P_1,P_2,P_3,Q_1,Q_2,Q_3)~ =~ 

1244: \sum_{\cal C} M_6^{\cal C}(P_1,P_2,P_3,Q_1,Q_2,Q_3)

1245: ~+~M_6^0~,\auto\label{dis}

1246: $$

1247: where $M_6^0$ contains all non-leading 

1248: triple-regge behavior, double-regge behavior, etc. and the sum is

1249: over all triplets ${\cal C}$ of 

1250: asymptotic 

1251: cuts in non-overlapping (large) invariants. For each triplet ${\cal C}$, 

1252: say ${\cal C}= (s_1,s_2,s_3)$, we can write 

1253: $$  

1254: \eqalign{M_6^{\cal C}(P_1,P_2,P_3,Q_1,Q_2,Q_3)~=~{1\over (2\pi i)^{3}}  &~~\int

1255: ds'_1 ds'_2 ds'_{3} ~~{\Delta^{\cal C} \over

1256: (s'_1-s_1)(s'_2-s_2)(s'_{3}-s_{3})} }

1257: \auto\label{dis2}

1258: $$

1259: where $\Delta^{\cal C}$ is the triple discontinuity.

1260: 

1261: The triple discontinuities are of three

1262: kinds corresponding to the three tree diagrams of Fig.~3.3. 

1263: There are 24 corresponding to Fig.~3.3(a),

1264: 12 corresponding to Fig.~3.3(b), 

1265: and 12 of the  Fig.~3.3(c) kind - including those described by Fig.~2.3. 

1266: Those of Fig.~3.3(a) and (b), occur in the physical

1267: regions, while those corresponding to Fig.~3.3(c) are all unphysical 

1268: triple discontinuities of the kind discussed in the last Section.

1269: \begin{center}

1270: \leavevmode

1271: \epsfxsize=2.2in

1272: \epsffile{520.ps}

1273: 

1274: Fig.~3.3 Tree Diagrams for triple discontinuities.

1275: \end{center}

1276: 

1277: As we discussed in \cite{arw99}, the diagram of Fig.~3.1(a) has physical region

1278: triple discontinuities of both the Fig.~3.3(a) and (b) kinds, although 

1279: neither gives leading triple-regge behavior. Unphysical discontinuities

1280: are more complicated to discuss. If the usual cutting rules hold,

1281: the diagram of Fig.~3.1(a) has no asymptotic triple 

1282: discontinuities corresponding to Fig.~3.3(c), but rather has

1283: only double discontinuities. To see this, 

1284: consider cutting the diagram as in Fig.~3.4, superficially giving an

1285: $\{s_{13'}, s_{32'}, s_{21'} \}$  triple discontinuity.

1286: \begin{center}

1287: \epsfxsize=2in

1288: \epsffile{chtr1.ps}

1289: 

1290: Fig.~3.4 An unphysical triple discontinuity?

1291: \end{center}

1292: In fact, 

1293: taking any double discontinuity clearly cuts all the available lines, implying 

1294: that there is no independent third discontinuity that can be taken.

1295: 

1296: It is not clear a-priori 

1297: that the cutting rules do apply to unphysical discontinuities. However, we will

1298: show directly in Section 5 that there is no symmetric

1299: triple discontinuity present (giving the desired common energy 

1300: component sign) in the diagram of Fig.~3.1. 

1301: Therefore, as we described above, whether there is an anomaly contribution

1302: from diagrams of this kind depends on whether the necessary triple

1303: discontinuities are present when  

1304: reggeization effects appear.

1305: In \cite{arw99} we noted only that such discontinuities 

1306: appeared to be present in reggeization diagrams 

1307: but did not discuss the structure of such diagrams in any detail. 

1308: 

1309: As an example of a diagram that should produce reggeization, 

1310: consider that shown in Fig.~3.5  

1311: \begin{center}

1312: \epsfxsize=4.5in

1313: \epsffile{chtr2.ps}

1314: 

1315: Fig.~3.5 A diagram with an unphysical triple discontinuity.

1316: \end{center} 

1317: in which one of the gluons in the diagram of Fig.~3.4 is replaced by 

1318: two-gluon exchange - potentially  

1319: giving the one-loop contribution to the trajectory

1320: function of the original gluon. The thin lines again indicate how an

1321: unphysical $\{s_{13'}, s_{32'}, s_{21'} \}$ discontinuity would be taken.

1322: In such a contribution  

1323: the corresponding six reggeon interaction, together with a remnant 

1324: seven reggeon interaction, would be generated by putting the cut

1325: lines on-shell. The discontinuity is clearly not symmetric and, in addition,

1326: if the particles put on-shell by a 

1327: discontinuity must be either all ``incoming'' or all ``outgoing'' 

1328: (this is the ``$+\alpha$'' condition that is part of  

1329: the normal cutting rules) then the energy components of

1330: the on-shell lines in the quark loop (apart from that potentially giving

1331: the reggeization contribution) can not have the common sign 

1332: required for the anomaly. This is because 

1333: both the $s_{2'3}$ and $s_{3'1}$ cuts involve two 

1334: of the relevant quark loop lines which must, therefore, be either incoming or

1335: outgoing quark/antiquark pairs. This requirement then 

1336: eliminates the possibility that such lines are all antiquarks, or all quarks.

1337: 

1338: In the next Section we will confirm by direct calculation that the diagram of

1339: Fig.~3.5 does not have the triple discontinuity needed to give the anomaly.

1340: Consequently, the reggeon interactions generated do not contain the anomaly.

1341: Although a more complete analysis of all diagrams should be given, this 

1342: essentially determines that the anomaly process of Fig.~3.2 is not 

1343: generated as a ``real part interaction'' when higher-order 

1344: reggeization effects are included. 

1345: 

1346: To obtain a symmetric triple discontinuity 

1347: in which the normal cutting rules could potentially give the anomaly amplitude 

1348: associated with  Fig.~3.2 , we consider the

1349: high-order diagram shown in Fig.~3.6(a)

1350: \newline \parbox{1.7in}{

1351: \begin{center}

1352: $~$

1353: \newline $~$

1354: \newline $~$

1355: \newline $~$

1356: \newline \epsfxsize=1.2in

1357: \epsffile{chir1.ps}

1358: \newline $~$

1359: \newline $~$

1360: \newline(a)

1361: \end{center}}

1362: \parbox{4.3in}{

1363: \begin{center}

1364: \epsfxsize=3.8in

1365: \epsffile{chtr.ps}

1366: \newline $~$

1367: \newline(b)

1368: \end{center}}

1369: \begin{center}

1370: Fig.~3.6 (a) A diagram with a symmetric 

1371: unphysical triple discontinuity

1372: \newline (b) expected reggeon interactions.

1373: \end{center}

1374: in which there are three gluons in each

1375: $t$-channel. A triple discontinuity in $\{s_{1'2}, s_{2'3}, s_{3'1} \}$ 

1376: is obtained by cutting the diagram as indicated in Fig.~3.6(b). 

1377: The closed loops involving two-gluon exchange could give

1378: both one loop contributions to the corresponding one reggeon trajectory

1379: function and the leading contribution of a two reggeon state. A-priori,

1380: therefore, we expect the diagram to contribute to the six-, seven-,

1381: eight- and nine-reggeon interaction as illustrated.

1382: Since the triple discontinuity is manifestly symmetric we again might 

1383: expect the anomaly to appear in the six-reggeon interaction, just

1384: as anticipated by the lower-order amplitude of Fig.~3.1(a).

1385: 

1386: For consistency with our previous discussion,

1387: the anomaly should not (and does not) appear quite so simply.

1388: As will be clear after we carry out the explicit 

1389: evaluation of asymptotic discontinuities in Section 5,  

1390: the triple discontinuity of Fig.~3.6(b) requires a 

1391: particular routing of the internal loop momenta. For this routing 

1392: the numerators of the cut quark propagators do not give the combination of

1393: $\gamma_5$ interactions needed for the anomaly. 

1394: In fact, the anomaly does occur within 

1395: a reggeon interaction generated by the diagram of Fig.~3.6(a) but  

1396: only when the unphysical discontinuities are actually taken 

1397: as shown in Fig.~3.7. It then occurs in the

1398: nine-reggeon interaction obtained, as illustrated, by putting lines on-shell.

1399: \begin{center}

1400: \epsfxsize=3.2in

1401: \epsffile{chir111.ps}

1402: 

1403: Fig.~3.7 Another cutting of Fig.~3.6(a).

1404: \end{center}

1405: Note that the discontinuity lines in Fig.~3.7 cross each other. This is 

1406: possible because the particles contributing to each discontinuity do not

1407: all have the same time direction. 

1408: To evaluate a multiple discontinuity of this kind

1409: we must develop direct methods to

1410: compute asymptotic discontinuities.

1411: 

1412: In the reggeon interaction of Fig.~3.7  there are 

1413: three reggeons in each $t$-channel and each reggeon state is 

1414: ``vector-like'' in that it has (close to) unit angular momentum 

1415: and appears in odd-signature amplitudes.

1416: As discussed in \cite{arw99},

1417: to avoid cancelation of the anomaly by transverse momentum integrations

1418: each reggeon state should have abnormal parity. 

1419: Therefore, according to the above discussion, the simplest

1420: interaction in which 

1421: the anomaly could appear is the nine-reggeon interaction in which 

1422: each reggeon state

1423: is vectorlike, composed of (at least) three gluons, and has abnormal parity.

1424: If, in addition, each reggeon state has zero color 

1425: then all three states carry anomaly current quantum numbers.

1426: It is surely remarkable that we are led directly to 

1427: the anomaly current by looking for the anomaly within   

1428: reggeon interactions.

1429: 

1430: In fact, the analyticity properties of amplitudes require \cite{arw99} 

1431: that the anomaly appears only when 

1432: signature conservation is also satisfied, which it is not

1433: not if all three reggeon states carry odd signature. Therefore, to avoid

1434: cancelation when further diagrams are added, an additional (reggeized) gluon

1435: must be present in one $t$-channel. This gives 

1436: only a relatively trivial modification of Fig.~3.7 and the 

1437: analysis that follows. In this paper, we are not interested

1438: in determining when the anomaly ultimately

1439: survives after all diagrams 

1440: are summed. We are satisfied just to find diagrams in which our asymptotic 

1441: discontinuity analysis determines that the anomaly is definitively present

1442: in the extracted reggeon interaction.

1443: This already requires that we go to the complexity of Fig.~3.7.

1444: 

1445: \newpage

1446: 

1447: \mainhead{4. LIGHT-CONE ANALYSIS OF ASYMPTOTIC DISCONTINUITIES}

1448: 

1449: In the next Section we will analyse triple-regge asymptotic discontinuities

1450: and will use a generalization of the simple light-cone analysis

1451: that we develop in this Section.

1452: Consider the box-diagram illustrated in Fig.~4.1. 

1453: \begin{center} 

1454: \leavevmode

1455: \epsfxsize=2.5in

1456: \epsffile{chbox1.ps}

1457: 

1458: Fig.~4.1 The box diagram.

1459: \end{center}

1460: Initially we ignore the role played by 

1461: numerators and so we consider, in the notation shown,

1462: $$

1463: \eqalign{             

1464: I(s,t,m^2) = &\int d^4k \left[k^2-m^2+i\epsilon\right]^{-1}

1465: \left[\left(p-{q\over 2}+k\right)^2-m^2+i\epsilon\right]^{-1}\cr

1466: &\times \left[(q-k)^2-m^2+i\epsilon\right]^{-1}\left[\left(

1467: p' +{q\over 2}-k\right)^2-m^2+i\epsilon\right]^{-1}.}

1468: \auto\label{lcan1}

1469: $$

1470: This integral is, of course, a function of invariants only even though it is

1471: specified using four momenta. Indeed, we can evaluate the integral using

1472: complex, unphysical, momenta that give physical values of the invariants,

1473: provided we are careful to define the integral via analytic continuation from

1474: the appropriate physical momentum region. Our purpose in this Section is to

1475: discuss momentum dependence of this kind for the simplifying

1476: case of the leading asymptotic behavior, in a manner that we can apply 

1477: to much more complicated diagrams in Section 5.

1478: 

1479: For illustrative purposes we set both $q=0$ and $m=0$ in (\ref{lcan1}) 

1480: and ignore infra-red divergences. We can then write

1481: $$

1482: I(s) ~= ~\int d^4k \left[k^2+i\epsilon\right]^{-2}

1483: \left[\left(p +k\right)^2 +i\epsilon\right]^{-1}

1484: \left[\left(p'  -k\right)^2+i\epsilon\right]^{-1}

1485: \auto\label{lcan2}

1486: $$

1487: We choose a particular Lorentz frame and introduce light-cone co-ordinates 

1488: such that 

1489: $$

1490: \eqalign{

1491: p&~= ~\left({P_+ \over 2}~,{P_+ \over 2}~,~\til{0}\right) 

1492: ~ +O\left({1\over  s}\right), ~~~~~P_+ ~\sim ~ s ~\to ~ \infty \cr

1493: p'&~=~  \left({P_+'+ P_-' \over 2}~, {P_+'- P_-' \over 2}~,~

1494: \underline{p}_{\perp}'~\right)   }

1495: \auto\label{lcan3}

1496: $$

1497: so that $s~ = P_+P_-'~[1 + O(1/ s)]~$. We can then write

1498: $$

1499: \eqalign{I(s)~\centerunder{$\large\sim$} {\raisebox{-3mm} 

1500: {$\scriptstyle s\to \infty$}}~~

1501: {1\over 2}\int &                                        

1502: d^2\underline{k}_{\,\perp} dk_+dk_- \left[ k_+k_- -k^2_{\,\perp} + 

1503: i\epsilon\right]^{-2}~

1504: \bigg[ \left( k_+ + P_+ \,\right) k_-

1505: - \underline{k}^2_{\,\perp}

1506: +i\epsilon\bigg]^{-1}  \cr

1507: & \times~

1508: \left[\left(k_+ -P_+' \right)\left(k_- -P_-' \right)- 

1509: (\underline{k}_{\,\perp} - \underline{p}_{\,\perp}')^2

1510: +i\epsilon\right]^{-1} }

1511: \auto\label{lcan4}

1512: $$

1513: 

1514: To obtain a non-zero answer by closing the $k_+$ contour, with $k_-$ and 

1515: $k_{\perp}$ fixed, the three poles given by the three square brackets of 

1516: (\ref{lcan4}) must

1517: not be on the same side of the contour. This requires $ 0 <k_-< P_-'$ and,

1518: in this case, the $k_+$ contour can be closed to pick up only

1519: the pole in the last bracket. This gives 

1520: $$

1521: k_+ ~=~ P_+' +

1522: {(\underline{k}_{\,\perp} - \underline{p}_{\,\perp}')^2

1523: - i\epsilon \over \left(k_- -P_-' \right)}

1524: \auto\label{lcan40}

1525: $$

1526: which is finite and so can be neglected compared to $P_+$. Note also that 

1527: $$

1528: k_- \sim 0,~~ {k_{\perp}}^2 \sim 0 ~~=> ~~k_+~\sim ~2k_0~ \sim 

1529: ~{{p'}^2 \over P_-'}

1530: \auto\label{5an}

1531: $$

1532: (we will need this approximation for the analysis of 

1533: Section 5). We thus obtain, 

1534: $$

1535: I(s)~\centerunder{$\large\sim$} {\raisebox{-3mm} 

1536: {$\scriptstyle s\to \infty$}}~~\pi i\int 

1537: d^2\underline{k}_{\perp} \left[ -k^2_{\perp} + 

1538: i\epsilon \right]^{-2} ~\int_0^{P_-'} dk_-  

1539: \left[k_- -P_-' \right]^{-1}\left[ P_+ k_-

1540: - \underline{k}^2_{\perp}

1541: +i\epsilon \right]^{-1} 

1542: \auto\label{lcan5}

1543: $$

1544: 

1545: We are specifically interested in the leading real and imaginary parts of 

1546: (\ref{lcan5}). They are given by the logarithm generated by 

1547: the pole factor containing $P_+$ as it approaches the $k_- = 0$ end-point of 

1548: the integration. If we keep only the integration over $0 < k_- < \lambda P_-'$ 

1549: and take $\lambda << 1$ so that we can 

1550: make the approximation $k_- / P_-'~ \sim 0$ we obtain

1551: $$

1552: \eqalign{ I(s)~& \centerunder{$\large\sim$} {\raisebox{-3mm} 

1553: {$\scriptstyle s\to \infty$}}~~

1554: \pi i\int 

1555: d^2\underline{k}_{\perp} \left[ -\underline{k}^2_{\perp} + 

1556: i\epsilon \right]^{-2} ~

1557: {1 \over P_-'} ~\int_0^{\lambda P_-'} dk_-  

1558: \left( P_+ k_-

1559: - \underline{k}^2_{\perp}

1560: +i\epsilon \right)^{-1} \cr 

1561: &~\sim ~{1 \over P_+ P_-'} ~[\log{(P_+P_-'\lambda  

1562:  -\underline{k}^2_{\perp} + i\epsilon]~J_1(0)} \cr 

1563: &~\sim ~{1 \over s} ~[\log{(s\lambda + i\epsilon]~J_1(0)}

1564: ~ \sim ~{1 \over s} 

1565: ~[\log{s} + i\pi]~J_1(0) }

1566: \auto\label{lcan60}

1567: $$

1568: where $J_1(0) ~\sim ~ \int 

1569: d^2\underline{k}_{\perp} \left[ -\underline{k}^2_{\perp} + 

1570: i\epsilon \right]^{-2}$ is infinite, but would be finite if we added a 

1571: mass to the particle propagators. 

1572: 

1573: As we have indicated, the sign of the imaginary part in (\ref{lcan60}) arises 

1574: directly from the $i\epsilon$ prescription. To obtain the leading imaginary 

1575: part or, equivalently, the leading behavior of the discontinuity in $s$,

1576: it suffices to keep the $i\epsilon$ dependence while dropping the 

1577: $ -\underline{k}^2_{\perp}$ dependence in the $k_-$ integral. 

1578: (\ref{lcan60}) is, of course, independent of $\lambda$. It will, however, be 

1579: useful to note the role of $\lambda$ with respect to the 

1580: analytic structure of $I(s)$ in the $s$-plane. As illustrated in Fig.~4.2,

1581: the finite end of the branch-cut asociated with the logarithm in 

1582: (\ref{lcan60}) moves out as $\lambda \to 0$.

1583: \begin{center}

1584: \leavevmode

1585: \epsfxsize=1.6in

1586: \epsffile{chir42.ps}

1587: 

1588: Fig.~4.2 $\lambda$-dependence of the branch cut.

1589: 

1590: \end{center}

1591: This is irrelevant to the asymptotic behavior and the 

1592: ``asymptotic discontinuity'' clearly remains unchanged. We will, nevertheless, 

1593: be able to exploit this 

1594: simple feature in evaluating multiple discontinuities in the next Section.

1595: Also, although (\ref{lcan60}) is an 

1596: invariant result, for

1597: our purposes it will be useful to keep the dependence on both $P_+$ and $P_-'$

1598: and discuss the dependence of the phase on $P_+$.

1599: 

1600: The initial $k_-$ integration contour for (\ref{lcan60})

1601: is as shown in Fig.~4.3(a) with the pole at 

1602: $ k_- = \underline{k}_{\perp}^2/ P_+$ indicated by a dot.

1603: \begin{center}

1604: \leavevmode

1605: \epsfxsize=4.5in

1606: \epsffile{conts.ps}

1607: 

1608: Fig.~4.3 Integration Contours for (a) (\ref{lcan5})~ (b) 

1609: ~$P_+ \to~ e^{2\pi i}P_+$ ~(c) Fig.~4.4.

1610: \end{center}

1611: As $P_+$ (and therefore $s$)

1612: completes a circle in the complex plane the pole 

1613: moves around the end-point as illustrated in Fig.~4.3(b). The result is 

1614: that the phase of the logarithm in (\ref{lcan60}) changes from $\pi$ 

1615: to $-\pi$ and there is a net discontinuity of $2\pi i / s$, as is given 

1616: directly by (\ref{lcan60}).

1617: This is also the result that would be 

1618: obtained by applying directly the standard

1619: cutting rules to Fig.~4.1, cut by the thin line, if 

1620: the $k_+$ and $k_-$ integrations are used to put the vertical lines 

1621: on shell. The above discussion is simply an asymptotic analysis of 

1622: how the two cut propagators pinch the integration region to generate 

1623: a branch-point in $s$. Introducing $\lambda$ limits the integration 

1624: region for the original integral such that the pinching only takes place for

1625: $s \sim P_+ ~> \lambda$. Note also that the residue function $J_1(0)$,

1626: multipying the logarithm in (\ref{lcan60}),

1627: is directly obtained from the original box

1628: diagram by putting the cut lines giving the discontinuity on-shell using 

1629: the longitudinal momentum integrations. This is a very simple example 

1630: (the simplest) of the relationship between a discontinuity and asymptotic

1631: behavior.  

1632: 

1633: In evaluating unphysical (multiple) discontinuities in 

1634: Section 5 we will not assume 

1635: that the standard cutting rules apply. Instead we will directly analyse 

1636: the discontinuities produced by logarithms. To understand how a discontinuity

1637: generated by a logarithm can provide leading asymptotic behavior 

1638: we note that the twisted diagram of Fig.~4.4, for

1639: $q=0$, differs from that of Fig.~4.1 only by $P_+ \to -P_+$.

1640: \begin{center}

1641: \leavevmode

1642: \epsfxsize=2in

1643: \epsffile{chbox2.ps}

1644: 

1645: Fig.~4.4 The twisted box diagram.

1646: \end{center}

1647: As a result,

1648: the integration contour and pole position of Fig.~4.3(a) is replaced by 

1649: that of Fig.~4.3(c). In this case a discontinuity is generated 

1650: for $s< 0$. For

1651: $s>0$ there is no phase generated by Fig.~4.4 and only the real

1652: logarithms cancel when this diagram is added to that of Fig.~4.1. 

1653: The leading behavior 

1654: of the discontinuity in $s$, i.e the imaginary part, produced by the

1655: diagram of Fig.~4.1 remains. 

1656: This cancelation of the logarithms is very well-known, of course. 

1657: It is also well known

1658: that the cancelation fails when a non-abelian symmetry group is present

1659: and that a consequence is the reggeization of the gluon.

1660: 

1661: We can briefly summarize the effect of adding numerators to (\ref{lcan1})

1662: as follows.

1663: First we note that the numerator of the internal fermion propagator 

1664: carrying $P_+$ gives an additional $P_+$ factor of the form $\gamma_- P_+$. 

1665: As a consequence, in (\ref{lcan60}), there is the replacement 

1666: $$

1667: ~\int_0 dk_- ~

1668: \left( P_+ k_- + \cdots \right)^{-1} ~~ \to ~\gamma_- P_+ ~ 

1669: \int_0 dk_-  \left( P_+ k_- + \cdots \right)^{-1} ~\sim~ \log{P_+}

1670: \auto\label{lcan61}

1671: $$

1672: and there is no inverse power of $P_+$. Also, each

1673: coupling to a gluon gives a $\gamma$ matrix factor and since the external 

1674: fermion lines are on-shell we can use the asymptotic form of 

1675: the Dirac equation (i.e. $ \gamma_- P_+ \psi ~\sim m ~\psi $) to write

1676: $$

1677: \eqalign{ <P_+|\gamma_{\mu}\gamma_- \gamma_{\nu}|P_+>~& \sim ~

1678: <P_+|{\gamma_- P_+ \over m} ~\gamma_{\mu}\gamma_- \gamma_{\nu}~

1679: {\gamma_- P_+ \over m}|P_+> \cr

1680: &  =~ <P_+|P_+ \gamma_- P_+ |P_+> / m^2~~

1681: \sim P_+ ~/m }

1682: \auto\label{coup}

1683: $$

1684: This gives another power of $P_+$ ($\sim s$) provided that the corresponding

1685: factor of $P_-'$ is present in the finite momentum part of the scattering 

1686: process. Not surprisingly this factor emerges from that part 

1687: which would dominate if $P_-'$ were large. However, we want to 

1688: emphasize that this selection is made only by the need to form a Lorentz 

1689: invariant amplitude from the non-invariant large momentum process.

1690: 

1691: Finally we note that the above analysis goes through with very little

1692: modification if we take both $m^2$ and $q$ to

1693: be non-zero so that (\ref{lcan2}) will not be infra-red divergent. 

1694: 

1695: \newpage

1696: 

1697: \mainhead{5. UNPHYSICAL TRIPLE DISCONTINUITIES AND HIGHER-ORDER GRAPHS}

1698: 

1699: In this Section we generalize the analysis of the last Section to asymptotic 

1700: triple discontinuities. The essential idea is that there is a well-defined 

1701: leading-log result for each triple discontinuity, just as there was for

1702: the single discontinuity in $s$ in the last Section, and that this can be found

1703: from the leading-log calculation of an amplitude by keeping the $i \epsilon$

1704: dependence of all logarithms. 

1705: 

1706: We begin by considering again the maximally non-planar graph

1707: shown in Fig.~3.1. To understand 

1708: how asymptotic discontinuities of the kind we are interested in 

1709: arise, we first consider

1710: a physical region discontinuity. To this end we interchange 

1711: $P_1$ and $P_{1'}$ in (\ref{np3}) so that $P_{1'}$ and $P_{2}$ are the

1712: momenta of incoming particles. For simplicity, we also set 

1713: $Q_i=0,~ i=1,2,3$. This could cause confusion as to which invariants

1714: discontinuities actually occur in. However, for the discontinuities

1715: that interest us, we will be able to

1716: avoid this issue. (As in the previous Section, adding both transverse 

1717: momenta and masses to our discussion would not change the essential features 

1718: of the analysis, but would eliminate gluon 

1719: infra-red divergences. We will discuss,

1720: at some points, the general effect of adding transverse momenta.)

1721: Therefore we write, asymptotically,  

1722: $$

1723: \eqalign{ P_{1'}~\to &~- P_{1}~= ~(p_{1'},p_{1'},0,0)~,~~p_{1'} \to \infty \cr

1724: P_2~\to &~- P_{2'}~= ~(p_2,0,p_2,0)~,~~p_2 \to \infty \cr

1725: P_3~\to &~ -P_{3'}~= ~(p_3,0,0,p_3)~,~~p_3 \to \infty  }

1726: \auto\label{pas}

1727: $$

1728: 

1729: Since we will ultimately be looking for a symmetric triple discontinuity,

1730: we consider only routes for the 

1731: internal loop momenta of Fig.~3.1 that are completely symmetric with 

1732: respect to the three external loops. There is essentially only one possibility.

1733: The two apparently distinct possibilities   

1734: illustrated in Fig.~5.1 are related by interchanging the primed and 

1735: unprimed external momenta.

1736: We will also want to make a symmetric choice for 

1737: the quark lines we place on shell. Although we will not discuss the anomaly 

1738: in detail until the next Section, 

1739: we note that a product of three orthogonal $\gamma$-matrices

1740: must be associated with the process of putting on-shell 

1741: each internal quark line.

1742: To achieve this it is necessary to put on-shell, symmetrically,  

1743: the internal lines in Fig.~5.1(a) 

1744: along which a single loop momentum flows. Therefore, we 

1745: consider only such lines in the following.

1746: \newline \parbox{3in}{ 

1747: \begin{center}

1748: \epsfxsize=2.2in

1749: \epsffile{chir33.ps}

1750: \newline (a)

1751: \end{center}}

1752: \parbox{3in}{ 

1753: \begin{center}

1754: \epsfxsize=2.2in

1755: \epsffile{chir44.ps}

1756: \newline (b)

1757: \end{center}}

1758: \newline \centerline{Fig.~5.1 Routing Loop Momenta for Fig.~3.2.}

1759: 

1760: Using the momentum routing of Fig.~5.1(a)

1761: and the analysis of the previous Section

1762: we consider logarithms generated by the $k_1$ and $k_2$ 

1763: integrations.

1764: The $k_1$ and $k_2$ loops are shown in Fig.~5.2.

1765: \begin{center}

1766: \epsfxsize=4in

1767: \epsffile{chir520.ps}

1768: 

1769: Fig.~5.2 (a) The $k_1$ Loop (b) The $k_2$ Loop.

1770: \end{center}

1771: For the moment, we omit the propagators in the sloping lines

1772: and all propagator numerators. (The omitted propagators will, nevertheless,

1773: play an important role below. They are also

1774: relevant if we wish to consider the other kinds of discontinuities

1775: that appear in Fig.~3.3.)

1776: In this case, the two loops differ only in the light-cone direction of

1777: $P_1'$ and $P_2$.

1778: 

1779: We consider Fig.~5.2(a) first. We 

1780: can directly apply the discussion following (\ref{lcan4})

1781: if we identify $P_{1'}$ with $p$, 

1782: $q$ with $p'$, 

1783: $k_1$ with $k$, and consider the propagator pole at 

1784: $(k_1+q)^2 = 0$. We then obtain 

1785: $$

1786: \eqalign{ I(p_{1'}q_{1^-})~\sim~&i\int 

1787: d^2\underline{k}_{1\perp} \left[ -k^2_{1\perp} + 

1788: i\epsilon \right]^{-2} ~\int_0^{\lambda q_{1^-}} dk_{1^-}  

1789: \left[k_{1^-} -q_{1^-} \right]^{-1}\left[ p_{1'} k_{1^-}

1790: - \underline{k}^2_{1\perp}

1791: +i\epsilon \right]^{-1} \cr 

1792:  \sim~&

1793: {1 \over p_{1'}q_{1^-}}~\log{[p_{1'}\lambda q_{1^-} + i\epsilon]} } 

1794: \auto\label{l50}

1795: $$

1796: We have used the notation (used extensively in the following) that for any

1797: four-momentum $k$

1798: $$

1799: k_{i^-}~=~k_0 - k_i ~~~~~\underline{k}_{i\perp}~=~(k_j,k_k)~~j\neq k \neq i

1800: ~~~~~ ~~i,j,k~ = 1,2,3 

1801: \auto\label{not-}

1802: $$

1803: The $q_{1^-}$ 

1804: dependence indicates that the logarithm is a reflection of a threshold

1805: in the invariant $P_{1'}.q$ . This dependence plays an important 

1806: role in the following discussion. We also retain the $\lambda$-dependence, 

1807: for technical reasons that will become apparent later. The final

1808: result will be independent of $\lambda$, as it must be. 

1809: From Fig.~5.2(b) we analagously obtain

1810: $$

1811: I(p_2q_{2^-})~\sim~~

1812: {1 \over p_{2}q_{2^-}}~\log{[- p_{2}\lambda q_{2^-} + i\epsilon]}

1813: \auto\label{l51}

1814: $$

1815: The minus sign appears relative to (\ref{l50}) because of the 

1816: opposite direction of $P_2$.

1817: 

1818: Next we consider how the logarithmic branch cuts

1819: generated by the  $k_1$ and $k_2$ integrations can trap the internal loop

1820: integration over $q$ to produce an overall

1821: discontinuity in $s_{1'2} \sim p_{1'}p_{2}$. 

1822: For simplicity, we consider the region where

1823: $$

1824: \underline{k}_{i \perp}^2 ~\sim~ q^2~~\sim ~ 0 ~~~~~i~=~1,2,3

1825: \auto\label{5an0}

1826: $$

1827: Appealing to (\ref{5an}) we can then, for our present purposes, effectively 

1828: ignore the remaining 

1829: $k_i$ dependence of the quark loop (including the propagators that

1830: we ignored in the above discussion). If we parameterize $q$ as 

1831: $$

1832: q~=~\biggl(q_0,~q_{1^-},~q_{2^-},~q_{3^-} \biggr)

1833: \auto\label{l52}

1834: $$

1835: we can treat the $q_{i^-}$ as independent variables, 

1836: with $q_0$ essentially determined by the constraint $q^2 \sim 0$.

1837: The logarithmic cuts of (\ref{l50}) and (\ref{l51}) appear, respectively,

1838: in the $q_{1^-}$ and $q_{2^-}$ planes and if we make a further change of 

1839: variables to

1840: $$

1841: q_{1^-}~=~x_2x_3~~, ~~~~q_{2^-}~=~x_3x_1 ~ ~, ~~~~~q_{3^-}~=~x_1x_2

1842: \auto\label{l53}

1843: $$

1844: the two branch points appear in the $x_3$-plane, for fixed, positive, 

1845: $x_1,x_2$, 

1846: as illustrated in Fig.~5.3(a).

1847: \begin{center}

1848: \epsfxsize=5.8in

1849: \epsffile{chir530.ps}

1850: \newline Fig.~5.3 Contours in the $x_3$-plane (a) the initial contour 

1851: (b) $p_{2} \to e^{2\pi i}p_{2}$ 

1852: \newline (c)  the discontinuity   

1853: (d) the discontinuity as a line integral. 

1854: \end{center} 

1855: (The branch points also appear, separately, 

1856: in the $x_2$ and $x_1$ planes. To focus on the

1857: $s_{1'2}$ discontinuity and avoid any complication

1858: from discontinuities involving a

1859: logarithm of $p_3$ in these planes 

1860: we can take the $\lambda$ for this logarithm to be much smaller.) 

1861: The propagator poles that are not on-shell, that we ignored in the 

1862: above discussion, combine to give a multiple pole at

1863: $q^2=0$ (on both sides of the contour, as determined by

1864: the presence of $i\epsilon$ in all propagators). 

1865: If we continue to ignore propagator

1866: numerators then the factors of $1/q_{1^-}$ and  $1/q_{2^-}$, 

1867: in (\ref{l50}) and (\ref{l51}) respectively, will also contribute poles at 

1868: $x_3=0$ (that will partly be compensated by the jacobian due to 

1869: the change of variables).

1870: However, in the anomaly contribution we will ultimately consider, these poles 

1871: will be directly canceled by numerator factors.

1872: 

1873: The threshold we are interested in occurs when the two branch points collide

1874: (at $x_3 = 0$ for $\epsilon = 0$). To extract the discontinuity we

1875: consider a full-plane rotation of $p_{2}$, with $p_{1'}$ fixed, so that  

1876: the logarithmic branch-cut

1877: (\ref{l51}) deforms the contour as shown in Fig.~5.3(b) 

1878: - the dashed line indicates that the contour is on the 

1879: second sheet of the branch-point (\ref{l50}). (We have omitted the poles at

1880: $x_3=0$.) Note that 

1881: the continuation path we have chosen isolates the discontinuity

1882: around the $s_{1'2}$ branch cut, since it avoids the pinching of the 

1883: integration contour with the singularity at 

1884: $q^2=0$ that would give other discontinuities. 

1885: The desired discontinuity is obtained by adding 

1886: the original contour in the opposite direction, as shown in Fig.~5.3(c).

1887: Combining both contours 

1888: we obtain Fig.~5.3(d) which, as illustrated can be written as a line integral 

1889: between the two branch 

1890: points of the double-discontinuity due to both cuts. As $\epsilon \to 0$, 

1891: or in the asymptotic

1892: limit $p_{1'}, p_2 \to \infty$, the branch points approach each other

1893: and the result is a closed contour integral around the 

1894: singularity at $q^2=0$ which is independent of the position of the end points 

1895: and remains finite in the asymptotic limit. This is the asymptotic 

1896: discontinuity and the singularity at $q^2 = 0$ 

1897: is clearly crucial in producing a non-zero result. 

1898: 

1899: In Fig.~5.4(a) we have illustrated

1900: the effect of adding 

1901: (external and internal) transverse momenta in the 

1902: the foregoing analysis. 

1903: The integral between the branch points, of the double discontinuity,

1904: is still obtained, while the 

1905: singularity at $q^2 =0$ separates into a set of poles at both 

1906: positive and negative $x_3$.

1907: \begin{center}

1908: \epsfxsize=5in

1909: \epsffile{chir540.ps}

1910: 

1911: Fig.~5.4 (a) The $x_3$

1912: contour with finite transverse momenta (b) equivalence of the

1913: asymptotic contour to the original contour.  

1914: \end{center} 

1915: In Fig.~5.4(b) we have  shown the asymptotic discontinuity.

1916: Since the branch points are

1917: logarithmic, the double discontinuity involved is simply $4 \pi^2$ and so 

1918: no longer contains either branch cut.

1919: Consequently, the asymptotically finite integral around 

1920: the poles to the left can be opened up to give the original contour, 

1921: as illustrated. (If there is a singularity at $x_3=0$, the contour is 

1922: constrained to pass through this point although, as we noted above, for the

1923: anomaly contribution to graphs, this will not be the case). 

1924: The final result shown in Fig.~5.4(b) is just what

1925: would be given by the normal cutting rules for a discontinuity in $s_{1'2}$ ,

1926: i.e. the original integral with the four propagators involved in generating 

1927: the discontinuity placed on-shell. Note that the same result 

1928: is obtained if the discontinuity is evaluated by 

1929: varying $p_{1'}$. An integral around the positive $x_3$ poles appears at the

1930: intermediate stage, which can then 

1931: be opened up to give the same final contour as in Fig.~5.4(b).  

1932: 

1933: An obvious, but essential, requirement in the

1934: origin of the asymptotic discontinuity, which we want to emphasize, 

1935: is that the branch-cuts due to the logarithms

1936: in $p_{1'}$ and $p_2$ must lie on opposite sides of the $x_3$ contour.

1937: In a physical region this requirement 

1938: is normally straightforward for a loop integration producing a threshold

1939: due to two massive states since the loop momentum will flow oppositely

1940: through the two states and the $i\epsilon$ prescription will place

1941: the states on opposite sides of the energy integration contour. In the 

1942: variables we are using the generation of the threshold is a little more 

1943: subtle.

1944: Note, for example, that when $x_1 < 0$ the branch-point (\ref{l50}) appears 

1945: in the upper half-plane (moving through infinity as $x_1$ moves through zero)

1946: and there is no discontinuity. Therefore, the signs of the $x_{i}$ play an 

1947: essential role in the occurrence of the discontinuity. 

1948: A further requirement, which clearly holds in the case just discussed, 

1949: is that the trapping (pinching)

1950: of the contour that we have discussed 

1951: must combine with the pinching associated with the logarithms

1952: to give a complete cut through the diagram. That is to say, the complete set of

1953: pinchings must correspond to an overall invariant cut.

1954: 

1955: We consider next the unphysical discontinuities that are our 

1956: principal interest.

1957: According to the discussion in Section 3, we are looking for a triple 

1958: discontinuity of the form of Fig.~3.4 that treats the three cut lines of the

1959: quark loop symmetrically so that, in a  

1960: physical region, the sign of the energy component can be 

1961: the same for all three on-shell states. We will, therefore, confine 

1962: our discussion to a search for a symmetric triple discontinuity. 

1963: As we noted, if the normal cutting rules apply 

1964: there is no triple discontinuity (symmetric or not) of the Fig.~3.4  kind.

1965: We consider whether the direct evaluation of discontinuities 

1966: gives the same result. 

1967: 

1968: The discontinuity we 

1969: discussed above occurred in a physical region that is unsymmetric 

1970: in that $P_2$ is the momentum of an incoming particle while 

1971: $P_1$ is the momentum of an outgoing particle.

1972: To look for a symmetric discontinuity we will 

1973: use an analysis that treats the complete graph symmetrically throughout.

1974: To this end, we will start in the symmetric asymptotic region (\ref{np3})

1975: where all momenta are real and 

1976: $$

1977: s_{i'j}~\sim ~-p_i p_j ~~<~0

1978: \auto\label{ninv}

1979: $$

1980: In this region, the diagram is defined by the usual $i\epsilon$ prescription.

1981: Since all three invariants must be positive, the triple discontinuity of

1982: Fig.~3.4 can only be present in the triple-regge limit if we allow the large

1983: momenta involved to be unphysical. A symmetric way to do this is to start 

1984: from the real physical region and take

1985: $$

1986: p_i ~\to ~ e^{-i\pi /2} p_i~= i p_i~,~~i=1,2,3 ~~~~~

1987: => s_{i'j}~\sim ~ (- ip_i)(i p_j) ~~>~0  

1988: \auto\label{pinv}

1989: $$

1990: 

1991: Given the symmetry of the present discussion, 

1992: it is immediately apparent that there will not be a (symmetric)

1993: triple discontinuity, as we now show.

1994: Using the above analysis, logarithms will be generated by each of the $k_i$ 

1995: integrations. If we consider again the region where the 

1996: transverse momenta are close to zero then, from (\ref{5an}), the requirement 

1997: that the energy component of each on-shell line in the loop have the same sign

1998: is equivalent to requiring that the $q_{i^-}$ all have the same sign. 

1999: This, in turn, requires 

2000: that the $x_i$ should all have the same sign.

2001: However, in the symmetric real physical region, 

2002: if $x_1$ and $x_2$ have the same sign, 

2003: the logarithmic branch cuts in $P_1$ and $P_2$ lie on the same side 

2004: of the $x_3$ contour as illustrated in Fig.~5.5.

2005: \begin{center}

2006: \epsfxsize=2in

2007: \epsffile{chir54.ps}

2008: 

2009: Fig.~5.5 The Symmetric Location of Branch-Cuts in the $x_3$-plane. 

2010: \end{center}

2011: Since the continuation

2012: (\ref{pinv}) is symmetric they will remain on the same side after the

2013: continuation. 

2014: As a consequence, in the symmetric $x_i$ region, 

2015: the contour will not be trapped and distorted as 

2016: one branch point moves aound the other, as it was in Fig.~5.3, and no

2017: discontinuity will result. We conclude therefore that, 

2018: for the graph we are discussing, 

2019: discontinuities can only be generated in asymmetric 

2020: regions of the $x_i$ that can 

2021: not provide the symmetric triple discontinuity that we are looking for.

2022: 

2023: The foregoing analysis also precludes the occurence of a triple discontinuity,

2024: that is appropriately symmetric, in the diagram of Fig.~3.5. 

2025: To obtain a symmetric triple discontinuity we look for a graph that has the

2026: appropriate overall symmetry and also, for each $i \neq j \neq k$,

2027: has logarithmic branch

2028: cuts on both sides of the $x_i$ contour

2029: in a symmetric region of $x_j$ and $x_k$. With these requirements in mind, 

2030: an obvious graph to consider is that of Fig.~3.7. To discuss this graph we 

2031: continue, for simplicity, to take $Q_1=Q_2=Q_3=0$. Two symmetric (distinct)

2032: routes for the internal momenta are shown in Fig.~5.6. 

2033: For a threshold corresponding to the cutting of particular

2034: lines of the internal quark loop to be generated the  

2035: external loop momentum generating the relevant logarithms

2036: must pass through at least one of the lines. With this constraint,

2037: only the routing shown in 

2038: Fig.~5.6(a) will give both discontinuities of the kind we are looking for

2039: and the $\gamma$-matrix structure for on-shell contributions

2040: that we show, in the next Section, gives the anomaly.

2041: The routing of Fig.~5.6(b) 

2042: would be appropriate for discussing 

2043: the triple discontinuity of Fig.~3.6. However, in this case the 

2044: $\gamma$-matrix structure needed to generate the anomaly does not 

2045: appear in the on-shell contributions. Therefore, the triple discontinuity 

2046: of Fig.~3.6 does not contain the anomaly.

2047: \newline \parbox{3in}{ 

2048: \begin{center}

2049: \epsfxsize=2.4in

2050: \epsffile{chir11.ps}

2051: \newline (a)

2052: \end{center}}

2053: \parbox{3in}{ 

2054: \begin{center}

2055: \epsfxsize=2.4in

2056: \epsffile{chir22.ps}

2057: \newline (b)

2058: \end{center}}

2059: \begin{center}

2060: Fig.~5.6 Labeling Momenta for Fig.~3.7.

2061: \end{center}

2062: 

2063: Using the momentum routing of Fig.~5.6(a) 

2064: we consider the logarithms generated by both the $k_i$ and $k_i'$ loop 

2065: integrations. Extracting all logarithms places on-shell all the hatched lines

2066: of Fig.~5.6(a), and 

2067: gives leading behavior of the form of (\ref{211}) multiplied by double 

2068: logarithms of each of the $P_{i^+}$.

2069: At the diagrammatic level (i.e. temporarily 

2070: discussing diagrams rather than discontinuities), we anticipate that

2071: existing calculations can be adapted to show that 

2072: the double logs are canceled by adding diagrams of the kind illustrated

2073: in Fig.~5.7. That is, we add diagrams containing twists relative to Fig.~3.7,

2074: as in Figs.~5.7(a) and (b) together with diagrams, 

2075: such as that in Fig.~5.7(c), that 

2076: produce the well-known cancelations necessary for reggeization.

2077: \begin{center}

2078: \epsfxsize=5in

2079: \epsffile{chir8.ps}

2080: \newline Fig.~5.7 Diagrams with (a) Twisted $k_i'$ Loops (b) Twisted $k_i$ 

2081: Loops 

2082: (c) Reggeization Cancelations

2083: \end{center}

2084: 

2085: After the double logs 

2086: are canceled, the remaining single logs should go into 

2087: reggeization contributions, in analogy with Fig.~3.6, with 

2088: the remaining terms providing new lowest-order

2089: reggeon interactions. As we have emphasized repeatedly, to discuss this 

2090: systematically we consider multiple asymptotic discontinuities 

2091: rather than the behavior of full diagrams. We do this, as above, 

2092: by keeping the $q$-dependence of all logarithms together with all

2093: $i\epsilon$ dependence.

2094: We consider specifically the logarithms generated by

2095: the $k_1$ and $k_1'$ loops, but the symmetry of the diagram 

2096: obviously determines that the others can be treated identically. The loops,

2097: extracted from Fig.~5.6, are shown in Fig.~5.8.

2098: The $k_1$ loop is identical to those of Fig.5.2 and 

2099: can be evaluated analagously.

2100: \begin{center}

2101: \epsfxsize=3in

2102: \epsffile{chir7.ps}

2103: \newline Fig.~5.8 (a) The $k_1$ Loop (b) The $k_1'$ Loop.

2104: \end{center}

2105: Using a similar analysis, the $k_1'$ loop gives an integral of the form

2106: $$

2107: \int_0^{(k_2+q)_{1^-}}~dk_{11^-}' ~~\cdots

2108: \auto\label{l54}

2109: $$

2110: 

2111: If we again go to the region where all transverse momenta are close 

2112: to zero then,

2113: using (\ref{5an}), it follows that after the $k_2$ integration

2114: $$

2115: k_{21^-} ~ \sim ~k_{20} ~\sim ~q^2/q_{2^-}~ << q_{1^-}

2116: \auto\label{l55}

2117: $$

2118: Therefore, we can take the upper end-point in (\ref{l54}) to be $q_{1^-}$. In 

2119: this case both the $k_1$ and $k_1'$ integrations give logarithms with $q_{1^-}$

2120: in the argument - but with opposite signs.

2121: We then have branch-cuts located as in Fig.~5.9(a) in each of the 

2122: $x_1,x_2$ and $x_3$ planes.

2123: We have included poles at $q^2=0$ and $x_i=0$

2124: and have used different $\lambda_i$ and $\lambda_i'$

2125: for each branch-cut to allow us to separate the branch points in our 

2126: discussion. 

2127: 

2128: With values of the $\lambda_i$ and $\lambda_i'$ implied 

2129: by Fig.~5.9(a),

2130: we could clearly obtain a discontinuity in $s_{jk'}$ (due to the two

2131: closest branch points) by repeating the 

2132: discussion illustrated by Fig.~5.3. The discontinuity would similarly be an 

2133: integral between the two branch points involved, as in Fig.~5.3(d), 

2134: but because of the additional

2135: branch points that are present, the contour could not be opened up as in 

2136: Fig.~5.4. Therefore, having taken $x_j, x_k > 0$ so that the branch 

2137: cuts lie  as in Fig.~5.9(a),

2138: the discontinuity would involve only pure imaginary or negative 

2139: real part values of $x_i$.

2140: \begin{center}

2141: \epsfxsize=5.5in

2142: \epsffile{chir550.ps}

2143: \newline Fig.~5.9 (a) Branch Points in the $x_i$-plane 

2144: (b) $p_i ~\to ~ e^{-i\pi /2} p_i~= i p_i~,~~i=1,2,3$ 

2145: \end{center}

2146:  Consequently,

2147: any further discontinuity obtained by the collision of 

2148: branch points in the $x_j$ or $x_k$ planes would have to involve mixed 

2149: real part signs for

2150: the $x_i$. We conclude (not surprisingly) that 

2151: in the physical region a triple discontinuity can not be obtained that 

2152: involves only positive values of all three $x_i$.

2153: 

2154: This brings us to the central point of the paper. If

2155: we go to the unphysical

2156: region (\ref{pinv}), where we expect to encounter

2157: an unphysical triple discontinuity, the last analysis changes in a crucial

2158: manner. The 

2159: resulting location of branch cuts is now as shown in Fig.~5.9(b), allowing

2160: the integration contour to be rotated as illustrated. 

2161: In Fig.~5.9(b) we have also, for emphasis, 

2162: chosen significantly different values of the $\lambda_i$ and $\lambda_i'$.

2163: If we again determine the discontinuity associated with the collision of the 

2164: two nearest branch points, as above, the result will be the contour 

2165: integral of the double discontinuity shown in Fig.~5.10.

2166: \begin{center}

2167: \epsfxsize=4in

2168: \epsffile{chir560.ps}

2169: 

2170: Fig.~5.10 The unphysical region discontinuity.

2171: \end{center}

2172: Now the integral involves positive real values of $x_i$ and, as illustrated,

2173: the asymptotic limit gives a loop integral over just positive values.

2174: The contour integral can not be opened up, however, since the other branch cuts

2175: remain.

2176: 

2177: Having derived a first discontinuity from two branch points in the 

2178: $x_i$ plane, as in Fig.~5.10, it is straightforward

2179: to keep the remaining branch points and move on to the $x_j$ and $x_k$ planes

2180: where, in each case, only two branch cuts now appear. In both planes,

2181: discontinuities of the form of Fig.~5.10 occur, provided the 

2182: $x_i$ integration is restricted to positive real values. Therefore, we obtain 

2183: a triple discontinuity in which each of the $x_i$, $x_j$ and $x_k$

2184: integrations is consistently over positive values and the asymptotic 

2185: contour is obtained as illustrated by the first two contours in Fig.~5.11.

2186: \begin{center}

2187: \epsfxsize=4in

2188: \epsffile{chir511.ps}

2189: 

2190: Fig.~5.11 Contours for the $x_i$, $x_j$ and $x_k$ integrations.

2191: \end{center}

2192: Since all logarithmic branch cuts are now removed, all three contours can 

2193: be opened up to obtain the last contour of Fig.~5.11 which is, once again the

2194: original contour of integration for each of $x_i$, $x_j$ and $x_k$.

2195: We thus obtain a triple discontinuity which, at firat sight, corresponds to

2196: the usual cutting rules since all cut lines are on-shell. However, there 

2197: is a subtlety.

2198: 

2199: If we consider the discontinuity arising from the pinching of logarithms 

2200: of $p_1\lambda_1$ and $p_2\lambda_2'$, for example, then the lines  

2201: put on-shell in the discontinuity are those that have thick hatches in

2202: Fig.~5.12(a).

2203: \begin{center}

2204: \epsfxsize=4.5in

2205: \epsffile{chir512.ps}

2206: 

2207: Fig.~5.12 On-shell lines for (a) an $s_{12'}$ discontinuity (b) a potential

2208: $s_{13'}$ discontinuity.

2209: \end{center}

2210: These lines are only a subset of those required to obtain a 

2211: complete cut of the diagram. This implies that the corresponding

2212: pinching does not, by itself, give a singularity of the complete integral and

2213: a-priori the integration contour could be deformed away from the pinched

2214: region.

2215: To obtain a complete cut we must add the lines that have thin hatches in

2216: Fig.~5.12(a). When 

2217: these lines are on shell the pinching does give an overall singularity.

2218: But, if we require a common sign for the $x_i$

2219: the two thin-hatched lines actually have the wrong $i\epsilon$ 

2220: prescription to straightforwardly combine with the asymptotic

2221: pinching to give what would be a physical sheet 

2222: ``asymptotic normal threshold''. However,

2223: each of the two thin hatched lines is separately placed on shell 

2224: by one of the additional discontinuities. Therefore, a triple discontinuity

2225: of the kind we have found does correspond to the triplet 

2226: $\{s_{12'},s_{23'},s_{32'}\}$ of invariant cuts. 

2227: 

2228: Note that if we consider instead 

2229: the discontinuity arising from the pinching of logarithms 

2230: of $p_1\lambda_1$ and $p_3\lambda_3'$ then the lines put on shell are 

2231: those hatched in Fig.~5.12(b). In this case there is no simple way to 

2232: include additional lines and obtain an invariant cut. Therefore, this pinching

2233: can not be extended to a complete cut of the diagram.

2234: We conclude that the triple discontinuity in $\{s_{12'},s_{23'},s_{32'}\}$

2235: that is illustrated in Fig.~3.7 is the only combination that exists, 

2236: as an extension of the above analysis. It

2237: is symmetric, with each of the internal quark lines 

2238: that are put on shell by $k_i$ integrations treated symmetrically. 

2239: All three of these lines contribute to each invariant cut but, as we 

2240: have just discussed, two of them always have the 

2241: wrong $i\epsilon$ prescription, relative to the third,

2242: to give a physical normal threshold.

2243: Singularities associated with combinations of forward and backward 

2244: going particles

2245: (as the mixture of $i\epsilon$ prescriptions implies is the case) are

2246: ``mixed-$\alpha$'' solutions of the Landau equations\cite{arw00} and

2247: are referred to as pseudothresholds. In general, pseudothresholds are not 

2248: singular on the physical sheet, just because of the conflicting 

2249: $i\epsilon$ prescriptions. However, they are generally singular on 

2250: unphysical sheets and can appear in multiple 

2251: discontinuities. For the unphysical multiple 

2252: discontinuity we are discussing, a combination of 

2253: ``asymptotic pseudothresholds'' can contribute when the same combination of 

2254: normal thresholds can not.

2255: 

2256: \newpage

2257: 

2258: \mainhead{6. THE TRIANGLE ANOMALY }

2259: 

2260: In this Section we give a brief discussion of how the anomaly occurs in the

2261: triple discontinuity of Fig.~3.7. A complete discussion would be obtained 

2262: by a straightforward generalization, to include the minor

2263: additional complexities, of

2264: the lengthy analysis of Fig.~3.1 in \cite{arw99}.    

2265: 

2266: All the cut lines of Fig.~3.7 are on-shell, as described in the last Section.

2267: We begin by 

2268: adding in the numerator dependence that we essentially ignored in the 

2269: previous Section. For the external lines, additional powers of the external

2270: momenta are generated as in (\ref{lcan61}) and (\ref{coup}). As a result,

2271: inverse external momentum factors, such as ${p_{1'}}^{-1}$ in (\ref{l50}) and 

2272:  ${p_2}^{-1}$ in (\ref{l51}) are eliminated and the 

2273: factor of $P_{1^+}P_{2^+}P_{3^+}$ that appears in (\ref{211}) is produced.

2274: Also, if we use the natural transverse momenta

2275: given by (\ref{not-}), the light-like 

2276: $\gamma$-matrix couplings that appear at each of the vertices of the internal loop

2277: (after the triple-regge limit is taken) are as illustrated in Fig.~6.1(a). 

2278: For the hatched lines that appear in both

2279: Fig.~6.1(a) and (b), we keep the $\gamma$ matrices shown. 

2280: These are the ``local couplings'' (see \cite{arw99}) that appear when that 

2281: part of the associated numerator is kept that cancels the 

2282: internal momentum factors, such as ${q_{1^-}}^{-1}$ in (\ref{l50}) and 

2283:  ${q_{2^-}}^{-1}$ in (\ref{l51}), that arise from the longitudinal

2284: loop momentum integrations. 

2285: \begin{center}

2286: \epsfxsize=5.8in

2287: \epsffile{chir61.ps}

2288: 

2289: Fig.~6.1 $\gamma$-matrix structure for the reggeon interaction extracted from

2290: Fig.~3.7.

2291: \end{center}

2292: The resulting asymptotic behavior then has the form

2293: $$

2294: \eqalign{ ~~~~~P_{1^+}~ P_{2^+}~ P_{3^+}~

2295: \prod_{i=1}^3 \int & { d^2 k_{i1}d^2 k_{i2} d^2 k_{i3}\over  

2296: k_{ i1}^2  k_{i2}^2 k_{i3}^2 }  

2297: ~~ \delta^2 (Q_{i\perp} -  k_{i1} -  k_{i2}  - k_{i3})~G^3_i(k_{i1},k_{i2},k_{i3} 

2298: \cdots) \cr 

2299: &~~~~~~~~\times ~ R^9(Q_1,Q_2,Q_3,

2300: k_{11}, k_{12},k_{13} \cdots )} \auto \label{611}

2301: $$

2302: where $R^9$ is the triangle diagram illustrated in Fig.~6.1(c).

2303: 

2304: By comparing with the three-reggeon version of (\ref{2ra1}) and (\ref{2ra})

2305: we can extract $R^9$ as a nine-reggeon interaction which, if we now write 

2306: $$

2307: k_{i1} ~= ~q_i + k_i~, ~~~~ k_{i2} ~= ~q_i - k_i -k_i'~, ~~~~

2308: k_{i3} ~= k_i'~,

2309: \auto\label{dki6}

2310: $$

2311: can be written (very similarly to (\ref{580})) as 

2312: $$

2313: \eqalign{ &R^9(q_1,q_2,q_3,k_1,k_2,k_3,k_1',k_2',k_3') ~=\cr

2314: & \int d^4 k  {  Tr \{ 

2315: \gamma_5 \gamma^{1^-3^+2^-}(\st{k}+ \st{k}_1 + \st{q}_2 +\st{k}_3) 

2316: \gamma_5 \gamma^{2^-1^+3^-} \st{k} 

2317: \gamma_5 \gamma^{3^-2^+1^-}(\st{k}- \st{k}_2 + \st{q}_1 + \st{k}_3 )\} 

2318: \over  (k + k_1 + q_2 + k_3 )^2  

2319: ~k^2 ~

2320:  (k - k_2 + q_1 + k_3)^2 } }

2321: \auto\label{612}

2322: $$

2323: where 

2324: $$ 

2325: \eqalign{\gamma^{1^-3^+2^-}~&=~

2326: \gamma_{1^-}\gamma_{3^+}\gamma_{2^-} ~=~\gamma^{-,-,-}~-~ i~

2327: \gamma^{-,-,+}  ~\gamma_5 \cr

2328: \gamma^{2^-1^+3^-}~&=~\gamma_{2^-}\gamma_{1^+}\gamma_{3^-}

2329:  ~=~\gamma^{-,-,-}~-~ i~

2330: \gamma^{+,-,-}  ~\gamma_5 \cr

2331: \gamma^{3^-2^+1^-}~&=~\gamma_{3^-}\gamma_{2^+}\gamma_{1^-}

2332:  ~=~\gamma^{-,-,-}~-~ i~

2333: \gamma^{-,+,-}  ~\gamma_5 }

2334: \auto\label{g63}

2335: $$

2336: and $\gamma^{\pm,\pm,\pm}$ is defined by (\ref{g64}).

2337: 

2338: Because of the symmetric choice of co-ordinates and the completely symmetric

2339: manner in which we 

2340: have evaluated the triple discontinuity, the anomaly appears in a 

2341: slightly different way to that discussed for Fig.~3.1 in \cite{arw99}.

2342: To obtain the anomaly divergence

2343: we must have a component of the axial-vector triangle diagram

2344: tensor  $\Gamma^{\mu\nu\lambda}$  with

2345: $\mu= \nu $ having a lightlike projection and $\lambda $ 

2346: having an orthogonal spacelike projection. There must also be 

2347: a transverse momentum (scaled to zero) in the remaining orthogonal

2348: spacelike direction. If we choose the $\gamma_5$

2349: component from all three vertices, 

2350: the first requirement is not met. However, if we choose the 

2351: $\gamma_5$ component from one of the three 

2352: vertices in Fig.~6.1(c), and choose the vector coupling from the other

2353: two vertices, it is met.

2354: The finite light-like momentum involved must then have a projection on 

2355: $n^{-,-,- \mu}$ and the orthogonal spacelike momentum must be distinct

2356: in each case.

2357: The three distinct possibilities for the anomaly to occur are associated 

2358: with the three distinct hexagraphs described in \cite{arw99}, 

2359: and hence with three distinct 

2360: helicity amplitudes. We will discuss this relationship further in our later 

2361: papers.

2362: 

2363: As we discussed at length in \cite{arw99}, while the triple discontinuity

2364: giving the interaction of Fig.~6.1 occurs in an unphysical region, 

2365: the interaction will, nevertheless, provide a ``real'' reggeon interaction

2366: in physical regions. Because the discontinuity has the symmetry property 

2367: that we emphasized in previous Sections,

2368: the anomaly infra-red divergence can occur in the physical-region

2369: configuration shown in Fig.~6.2. (The large dots indicate that a local

2370: interaction is involved.)

2371: The $\gamma_5$ interaction is at the intermediate vertex and the 

2372: light-like momenta are as in(\ref{chm1})-(\ref{chm30}).  

2373: Fig.~6.2 can then be identified with

2374: the basic anomaly process of Fig.~3.2 except

2375: that, as anticipated in Section 3, there is an additional wee gluon involved.

2376: There are also additional gluons with finite transverse momentum.

2377: \begin{center}

2378: \epsfxsize=2.5in

2379: \epsffile{chir62.ps}

2380: 

2381: Fig.~6.2 Physical region configuration for the anomaly divergence in Fig.~6.1.

2382: \end{center}

2383: 

2384: If there are no

2385: reggeization logarithms of the same order that appear accompanying the anomaly, 

2386: as our discussion in Sections 3 and 5 implies, then that

2387: part of the 

2388: triple discontinuity interaction given by Fig.~6.1 that contains the anomaly

2389: will appear as the leading triple-regge coupling of the three three-reggeon

2390: states. All other diagrams that contribute will then have a similar triple 

2391: discontinuity. The discussion in Section 5 shows that such diagrams must have 

2392: ``right-hand  and left-hand'' cuts in each $x_i$-plane, suggesting that only

2393: diagrams having the same structure as that of Fig.~3.7, 

2394: but with incoming and outgoing lines switched, can contribute. If this is the case,

2395: signature conservation will occur, requiring an additional reggeon in at least one

2396: channel. In this paper we will not introduce color factors, except to note that

2397: we expect every reggeon state coupling to the anomaly

2398: to carry anomalous color parity. This will ensure that the anomaly does not occur in

2399: the scattering of elementary quarks and gluons - as we have anticipated.

2400: Instead the scattering states must have an essential ``non-perturbative'' wee-parton

2401: content that ensures they can scatter by 

2402: exchanging reggeon states coupling to the anomaly.

2403: 

2404: We will postpone all further discussion of cancelations to 

2405: our following papers. The purpose of this paper

2406: has been to establish that a class of diagrams contain  

2407: a multiple discontinuity that does generate the reggeon interaction anomaly. 

2408: For the moment we note only that when the SU(3) gauge 

2409: symmetry of QCD is broken to SU(2) the infra-red divergence that involves the

2410: anomaly and that actually dominates bound-state interactions

2411: occurs in diagrams that are very similar to the ones we have discussed.

2412: An example, corresponding to a triple-regge multi-pomeron interaction,

2413: is shown in Fig.~6.3.

2414: \begin{center}

2415: \epsfxsize=2.5in

2416: \epsffile{chir63.ps}

2417: 

2418: Fig.~6.3 The anomaly configuration in bound state interactions. 

2419: \end{center}

2420: The scattering states are bound-states and the solid, wavy, lines

2421: are, reggeized, massive gluon states that are SU(2) 

2422: singlets. The dashed lines represent massless 

2423: gluons carrying zero transverse momentum. In this situation the three

2424: multi-reggeon (pomeron) states that are interacting through the anomaly 

2425: all have a wee-parton component that participates in the divergence.

2426: 

2427: \newpage

2428: 

2429: \begin{thebibliography}{99}

2430: 

2431: \bibitem{aj} M.~F.~Atiyah, in {\it Anomalies, Geometry

2432: and Topology}, edited by W.~A.~Bardeen and A.~R.~White (World Scientific, 

2433: Singapore, 1985); R.~Jackiw, ibid. and in

2434: {\it Effects of Dirac's Negative Energy Sea Quantum Numbers},

2435: Dirac Prize lecture, March 1999, hep-th/9903255. 

2436: 

2437: \bibitem{gth} G.~'t~Hooft, {\it Phys. Rep.} {\bf 142}, 357 (1986).

2438: 

2439: \bibitem{arw99} A.~R.~White,  {\it Phys. Rev.} {\bf D63} 016007, (2001), 

2440: {\bf D58}, 074008 (1998).

2441: 

2442: \bibitem{arw001} A.~R.~White, {\it Nucl. Phys. Proc. Suppl.}

2443: {\bf 96}, 277-286 (2001) - hep-ph/0008267; see 

2444: also Lectures in the Proceedings of the Theory Institute on Deep-Inelastic

2445: Diffraction, Argonne National Laboratory (1998).

2446: 

2447: \bibitem{fkl} E.~A.~Kuraev, L.~N.~Lipatov, V.~S.~Fadin, {\it Sov. Phys.

2448: JETP} {\bf 45}, 199 (1977);

2449: J.~B.~Bronzan and R.~L.~Sugar, {\it Phys. Rev.} {\bf D17}, 

2450: 585 (1978), this paper organizes into reggeon diagrams the results from 

2451: H.~Cheng and C.~Y.~Lo, Phys. Rev. {\bf D13}, 1131 (1976), 

2452: {\bf D15}, 2959 (1977); 

2453: V.~S.~Fadin and V.~E.~Sherman, Sov. Phys. JETP {\bf 45}, 

2454: 861 (1978);

2455: V.~S.~Fadin and L.~N.~Lipatov, {\it Nucl. Phys.} {\bf B477},

2456: 767 (1996) and further references therein;

2457: A.~R.~White, {\it Int. J. Mod. Phys.} {\bf A8}, 4755 (1993). 

2458: 

2459: \bibitem{jb} J.~Bartels, {\it Z. Phys.} {\bf C60}, 471 (1993) and further

2460: references therein.

2461: 

2462: \bibitem{cg} S.~Coleman and B.~Grossman, {\it Nucl. Phys. }

2463: {\bf B203}, 205 (1982); see also 

2464: T.~Banks, Y.~Frishman, A.~Schwimmer and S.~Yankielowicz,

2465: {\it Nucl. Phys.} {\bf B177}, 157 (1981).

2466: 

2467: \bibitem{gr} V.~N.~Gribov, Orsay Lectures on  Confinement (III),

2468: hep-ph/9905285.

2469: 

2470: \bibitem{cri}  A.~A.~Migdal, A.~M.~Polyakov and K.~A.~Ter-Martirosyan, 

2471: {\it Zh. Eksp. Teor.  Fiz.} {\bf 67}, 84 (1974); 

2472: H.~D.~I.~Abarbanel and J.~B.~Bronzan, {\it Phys. Rev.} {\bf D9}, 2397 (1974).

2473: 

2474: \bibitem{arw00} A.~R.~White, hep-ph/0002303 

2475: - {\it The Past and Future of S-Matrix Theory}

2476: in {\it Scattering}, edited by E.~R.~Pike and P.~Sabatier 

2477: (to be published by Academic Press, London) and references therein.

2478: 

2479: \bibitem{gw} P.~Goddard and A.~R.~White, {\it Nucl. Phys.} {\bf B17}, 1, 45

2480: (1970). As in \cite{arw99} we note that the full triple-regge

2481: limit is distinct from the

2482: ``triple-regge'' limit of the one-particle inclusive 

2483: cross-section that is a ``non-flip helicity-pole'' limit.

2484: 

2485: \bibitem{ste} O.~Steinmann, {\it Helv. Phys. Acta.} {\bf 33} 257, 347 (1960);

2486: H. Epstein, in  

2487: {\it Structural Analysis Of Collision Amplitudes} edited by R.~Balian

2488: and D.~Iagolnitzer (Amsterdam 1976), and references therein.

2489: 

2490: \end{thebibliography}

2491: 

2492: 

2493: \end{document}

2494: 

2495: