1: \input harvmac

2: 

3: 

4: 

5: \let\includefigures=\iftrue

6: \let\useblackboard=\iftrue

7: \newfam\black

8: 

9: %Figure Stuff

10: \includefigures

11: \message{If you do not have epsf.tex (to include figures),}

12: \message{change the option at the top of the tex file.}

13: \input epsf

14: \def\figin{\epsfcheck\figin}\def\figins{\epsfcheck\figins}

15: \def\epsfcheck{\ifx\epsfbox\UnDeFiNeD

16: \message{(NO epsf.tex, FIGURES WILL BE IGNORED)}

17: \gdef\figin##1{\vskip2in}\gdef\figins##1{\hskip.5in}% blank space instead

18: \else\message{(FIGURES WILL BE INCLUDED)}%

19: \gdef\figin##1{##1}\gdef\figins##1{##1}\fi}

20: \def\DefWarn#1{}

21: \def\figinsert{\goodbreak\midinsert}

22: \def\ifig#1#2#3{\DefWarn#1\xdef#1{Fig.~\the\figno}

23: \writedef{#1\leftbracket Fig.\noexpand~\the\figno}%

24: \figinsert\figin{\centerline{#3}}\medskip\centerline{\vbox{

25: \baselineskip12pt\advance\hsize by -1truein

26: \noindent\footnotefont{\bf Fig.~\the\figno:} #2}}

27: \bigskip\endinsert\global\advance\figno by1}

28: %%%

29: \else

30: \def\ifig#1#2#3{\xdef#1{Fig.~\the\figno}

31: \writedef{#1\leftbracket Fig.\noexpand~\the\figno}%

32: %\figinsert\figin{\centerline{#3}}\medskip

33: %\centerline{\vbox{\baselineskip12pt

34: %\advance\hsize by -1truein\noindent

35: %\footnotefont{\bf Fig.~\the\figno:} #2}}

36: %\bigskip\endinsert

37: \global\advance\figno by1}

38: \fi

39: %

40: 

41: 

42: \def\doublefig#1#2#3#4{\DefWarn#1\xdef#1{Fig.~\the\figno}

43: \writedef{#1\leftbracket Fig.\noexpand~\the\figno}%

44: \figinsert\figin{\centerline{#3\hskip1.0cm#4}}\medskip\centerline{\vbox{

45: \baselineskip12pt\advance\hsize by -1truein

46: \noindent\footnotefont{\bf Fig.~\the\figno:} #2}}

47: \bigskip\endinsert\global\advance\figno by1}

48: 

49: 

50: 

51: %%BLACKBOARD FONT STUFF

52: \useblackboard

53: \message{If you do not have msbm (blackboard bold) fonts,}

54: \message{change the option at the top of the tex file.}

55: \font\blackboard=msbm10 scaled \magstep1

56: \font\blackboards=msbm7

57: \font\blackboardss=msbm5

58: \textfont\black=\blackboard

59: \scriptfont\black=\blackboards

60: \scriptscriptfont\black=\blackboardss

61: \def\Bbb#1{{\fam\black\relax#1}}

62: \else

63: \def\Bbb{\bf}

64: \fi

65: % *************************************

66: %\draft

67: %

68: \def\subsubsec#1{\bigskip\noindent{\it{#1}} \bigskip}

69: \def\yboxit#1#2{\vbox{\hrule height #1 \hbox{\vrule width #1

70: \vbox{#2}\vrule width #1 }\hrule height #1 }}

71: \def\fillbox#1{\hbox to #1{\vbox to #1{\vfil}\hfil}}

72: \def\ybox{{\lower 1.3pt \yboxit{0.4pt}{\fillbox{8pt}}\hskip-0.2pt}}

73: %

74: %

75: %%MATH MACROS

76: %Greek letters and their bars

77: \def\ep{\epsilon}

78: \def\bep{\bar\epsilon}

79: \def\blam{\bar\lambda}

80: \def\bsig{\bar\sigma}

81: \def\bpsi{\bar\psi}

82: \def\bphi{\bar\phi}

83: \def\bp{\bar\partial}

84: \def\bthet{\overline \theta}

85: \def\Dbar{\overline D}

86: \def\bgam{\bar\gamma}

87: \def\xit{\xi}

88: 

89: %More bars

90: \def\bi{\bar i}

91: \def\jb{\bar j}

92: \def\Qbar{\overline Q}

93: 

94: %\def\l{\left}

95: %\def\r{\right}

96: \def\comments#1{}

97: \def\cc{{\rm c.c.}}

98: \def\tM{\tilde M}

99: \def\bM{\bar M}

100: \def\QC{\Bbb{C}}

101: \def\QH{\Bbb{H}}

102: \def\QM{\Bbb{M}}

103: \def\QR{\Bbb{R}}

104: \def\QX{\Bbb{X}}

105: \def\QZ{\Bbb{Z}}

106: \def\p{\partial}

107: \def\tilp{\tilde\partial}

108: \def\eps{\epsilon}

109: \def\half{{1\over 2}}

110: \def\Tr{{{\rm Tr~ }}}

111: \def\tr{{\rm tr\ }}

112: \def\Re{{\rm Re\hskip0.1em}}

113: \def\Im{{\rm Im\hskip0.1em}}

114: \def\even{{\rm even}}

115: \def\odd{{\rm odd}}

116: \def\lcm{{\rm lcm}}

117: \def\diag{{\rm diag}}

118: \def\bra#1{{\langle}#1|}

119: \def\ket#1{|#1\rangle}

120: \def\bbra#1{{\langle\langle}#1|}

121: \def\kket#1{|#1\rangle\rangle}

122: \def\vev#1{\langle{#1}\rangle}

123: \def\Dslash{\rlap{\hskip0.2em/}D}

124: \def\CA{{\cal A}}

125: \def\CC{{\cal C}}

126: \def\CD{{\cal D}}

127: \def\CE{{\cal E}}

128: \def\CF{{\cal F}}

129: \def\CG{{\cal G}}

130: \def\CT{{\cal T}}

131: \def\CM{{\cal M}}

132: \def\CN{{\cal N}}

133: \def\CO{{\cal O}}%AEL

134: \def\CP{{\cal P}}

135: \def\CL{{\cal L}}

136: \def\CV{{\cal V}}

137: \def\CS{{\cal S}}

138: \def\CW{{\cal W}}

139: \def\CX{{\cal X}}%AEL

140: \def\ad#1#2{{\delta\over\delta\sigma^{#1}(#2)}}

141: \def\ppt{{\partial\over\partial t}}

142: \def\comment#1{[#1]}

143: \def\nl{\hfill\break}

144: \def\a{\alpha}

145: \def\ta{\tilde\alpha}

146: \def\ap{\alpha'}

147: \def\sqap{\sqrt{\alpha'}}

148: \def\bA{\bar A}

149: \def\floor#1{{#1}}

150: \def\sgn{{\rm sgn\ }}

151: \def\I{{\rm I}}

152: \def\IA{{\rm IA}}

153: \def\II{\relax{I\kern-.10em I}}

154: \def\IIa{{\II}a}

155: \def\IIb{{\II}b}

156: \def\TeV{{\rm TeV}}

157: \def\hk{hyperk\"ahler\  }

158: \def\Hk{Hyperk\"ahler\  }

159: \def\cascade{{\cal A}}

160: %

161: \def\imp{$\Rightarrow$}

162: %\def\IZ{\relax\ifmmode\mathchoice

163: %{\hbox{\cmss Z\kern-.4em Z}}{\hbox{\cmss Z\kern-.4em Z}}

164: %{\lower.9pt\hbox{\cmsss Z\kern-.4em Z}}

165: %{\lower1.2pt\hbox{\cmsss Z\kern-.4em Z}}

166: %\else{\cmss Z\kern-.4emZ}\fi}

167: %\def\IZ{{\bf Z}}

168: \def\IZ{\relax{\rm Z\kern-.34em Z}}

169: \def\IB{\relax{\rm I\kern-.18em B}}

170: \def\IC{{\relax\hbox{$\inbar\kern-.3em{\rm C}$}}}

171: \def\ID{\relax{\rm I\kern-.18em D}}

172: \def\IE{\relax{\rm I\kern-.18em E}}

173: \def\IF{\relax{\rm I\kern-.18em F}}

174: \def\IG{\relax\hbox{$\inbar\kern-.3em{\rm G}$}}

175: \def\IGa{\relax\hbox{${\rm I}\kern-.18em\Gamma$}}

176: \def\IH{\relax{\rm I\kern-.18em H}}

177: \def\II{\relax{\rm I\kern-.18em I}}

178: \def\IK{\relax{\rm I\kern-.18em K}}

179: \def\IP{\relax{\rm I\kern-.18em P}}

180: %\def\IX{\relax{\rm X\kern-.01em X}}

181: %this doesn't work

182: \def\IX{{\bf X}}

183: %

184: \def\barom{\overline{\Omega}}

185: \def\barA{\bar{A}}

186: \def\jb{{\bar \jmath}}

187: \def\Hom{{\rm Hom}}

188: \def\inbar{\,\vrule height1.5ex width.4pt depth0pt}

189: \def\mod{{\rm\; mod\;}}

190: \def\ndt{\noindent}

191: \def\p{\partial}

192: \def\pab{\pb_{\bar A} }

193: \def\pb{{\bar \p}}

194: \def\pgp{\pb g g^{-1}}

195: \font\cmss=cmss10 \font\msss=cmss10 at 7pt

196: \def\IR{\relax{\rm I\kern-.18em R}}

197: \def\pbar{\bar{\p}}

198: \def\qmvw{\CM_{\vec \zeta}(\vec v, \vec w) }

199: \def\sdtimes{\mathbin{\hbox{\hskip2pt\vrule

200: height 4.1pt depth -.3pt width .25pt\hskip-2pt$\times$}}}

201: %\def\simlt{\hskip0.05in\relax{

202: %\raise3.0pt\hbox{ $<$

203: %{\lower5.0pt\hbox{\kern-1.05em $\sim$}} } \hskip0.05in}

204: \def\simgt{\hskip0.05in\relax{

205: \raise3.0pt\hbox{ $>$

206: {\lower5.0pt\hbox{\kern-1.05em $\sim$}} }} \hskip0.05in}

207: \def\ker{{\rm ker\ }}

208: \def\cok{{\rm cok\ }}

209: \def\im{{\rm im\ }}

210: \def\ind{{\rm ind\ }}

211: \def\ub{{\bar{u}}}

212: \def\vol{{\rm vol}}

213: \def\Vol{{\rm Vol}}

214: \def\wb{{\bar{w}}}

215: \def\zb {{\bar{z}}}

216: \def\one {{\bf 1}}

217: \def\dpr{^{\prime\dagger}}

218: %

219: \def\BR{\IR}

220: \def\BZ{Z} % for now

221: \def\BP{\IP}

222: \def\BR{\IR}

223: \def\BC{\IC}

224: \def\BM{\QM}

225: \def\BH{\QH}

226: \def\BX{\QX}

227: %

228: \def\ls{l_s}

229: \def\ms{m_s}

230: \def\gs{g_s}

231: \def\lp10{\ell_p^{10}}

232: \def\lp11{\ell_p^{11}}

233: \def\R11{R_{11}}

234: \def\mb{{m_{\rm brane}}}

235: \def\vb{{v_{\rm brane}}}

236: %

237: \def\brs{Q_{{\rm B}}}

238: \def\brsb{\bar{Q}_{{\rm B}}}

239: \def\dels{\delta_{{\rm S}}}

240: \def\delb{\delta_{{\rm B}}}

241: \def\zb{\bar{z}}

242: \def\frac#1#2{{#1 \over #2}}

243: \def\epb{\bar{\epsilon}}

244: \def\tts{{\rm tt}{}^\ast}

245: \def\ch{{\rm ch}}

246: \def\td{{\rm Td}}

247: \def\ahat{\hat{A}}

248: \def\cft#1{{\rm CFT_{#1}}}

249: \def\dS{\partial \Sigma}

250: \def\ni{\noindent}

251: \def\imt{{\rm Im} \tau}

252: \def\mfl{ (-1)^{F_L}}

253: \def\mfr{ (-1)^{F_R}}

254: 

255: %identity operator from doyon-fonseca

256: \def\Ione{\hbox{$1\hskip -1.2pt\vrule depth 0pt height 1.53ex width 0.7pt

257:                   \vrule depth 0pt height 0.3pt width 0.12em$}}

258: 

259: 

260: \def\thetafunction#1#2#3#4{

261: \vartheta\left[\matrix{ #1 \cr #2} \right] (#3 |#4)}

262: 

263: 

264: %% from the topological vertex paper

265: 

266: 

267: %%                              TABLEAUX.TEX

268: %%      This  macro file is for producing a ``Young Tableau'' which is

269: %%      an array of little squares sometimes used in mathematical physics.

270: %%      For instance, the command $\tableau{6 3 2}$ will produce a tableau

271: %%      with 6 squares in the top row, 3 in the next, and 2 in the last.

272: %%                                  OOOOOO

273: %%      This tableau will look like OOO    but made of squares instead of O's.

274: %%                                  OO

275: %%      Any number of rows may be present, each having a nonzero number of

276: %%      squares.

277: %%

278: %%      A tableau is math mode material, so use $ or $$ to enclose it.

279: %%

280: %%      The size and line-thickness of the little boxes are controlled by the

281: %%      dimension parameters --

282: %%              \tableauside=1.0ex              %(size)

283: %%              \tableaurule=0.4pt              %(line-thickness)

284: %%      Change them if you want.

285: %%

286: %%                                                      -- Doug Eardley 9/19/8%%

287: %%

288: \newdimen\tableauside\tableauside=1.0ex

289: \newdimen\tableaurule\tableaurule=0.4pt

290: \newdimen\tableaustep

291: \def\phantomhrule#1{\hbox{\vbox to0pt{\hrule height\tableaurule width#1\vss}}}

292: \def\phantomvrule#1{\vbox{\hbox to0pt{\vrule width\tableaurule height#1\hss}}}

293: \def\sqr{\vbox{%

294:   \phantomhrule\tableaustep

295:   \hbox{\phantomvrule\tableaustep\kern\tableaustep\phantomvrule\tableaustep}%

296:   \hbox{\vbox{\phantomhrule\tableauside}\kern-\tableaurule}}}

297: \def\squares#1{\hbox{\count0=#1\noindent\loop\sqr

298:   \advance\count0 by-1 \ifnum\count0>0\repeat}}

299: \def\tableau#1{\vcenter{\offinterlineskip

300:   \tableaustep=\tableauside\advance\tableaustep by-\tableaurule

301:   \kern\normallineskip\hbox

302:     {\kern\normallineskip\vbox

303:       {\gettableau#1 0 }%

304:      \kern\normallineskip\kern\tableaurule}%

305:   \kern\normallineskip\kern\tableaurule}}

306: \def\gettableau#1 {\ifnum#1=0\let\next=\null\else

307:   \squares{#1}\let\next=\gettableau\fi\next}

308: 

309: \tableauside=1.0ex

310: \tableaurule=0.4pt

311: 

312: 

313: %% from shiraz

314: 

315: 

316: \font\smallrm=cmr8

317:  \def\app#1#2{\global\meqno=1\global\subsecno=0\xdef\secsym{\hbox{#1.}}

318:  \bigbreak\bigskip\noindent{\bf Appendix.}\message{(#1. #2)}

319:  \writetoca{Appendix {#1.} {#2}}\par\nobreak\medskip\nobreak}

320:  %

321:  %       \eqn\label{a+b=c}       gives displayed equation, numbered

322:  %                               consecutively within sections.

323: %     \eqnn and \eqna define labels in advance (of eqalign?)

324:  %

325:  \def\eqnn#1{\xdef #1{(\secsym\the\meqno)}\writedef{#1\leftbracket#1}%

326:  \global\advance\meqno by1\wrlabeL#1}

327:  \def\eqna#1{\xdef #1##1{\hbox{$(\secsym\the\meqno##1)$}}

328:  \writedef{#1\numbersign1\leftbracket#1{\numbersign1}}%

329:  \global\advance\meqno by1\wrlabeL{#1$\{\}$}}

330:  \def\eqn#1#2{\xdef #1{(\secsym\the\meqno)}\writedef{#1\leftbracket#1}%

331:  \global\advance\meqno by1$$#2\eqno#1\eqlabeL#1$$}

332: 

333: %\def\IZ{\relax\ifmmode\mathchoice

334: % {\hbox{\cmss Z\kern-.4em Z}}{\hbox{\cmss Z\kern-.4em Z}}

335: % {\lower.9pt\hbox{\cmsss Z\kern-.4em Z}}

336: % {\lower1.2pt\hbox{\cmsss Z\kern-.4em Z}}\else{\cmss Z\kern-.4em Z}\fi}

337: 

338: 

339: \global\newcount\itemno \global\itemno=0

340: \def\itemized{\global\itemno=0}

341: %\bigbreak}

342: %\bigskip\noindent}

343: \def\itemaut#1{\global\advance\itemno by1\noindent\item{\the\itemno.}#1}

344: 

345: %\itemized

346: %\itemaut{First this.}

347: %\itemaut{Then that.}

348: 

349: 

350: \def\bs{\backslash}

351: \def\bfe{{\bf e}}

352: 

353: 

354: 

355: 

356: %%ENGLISH MACROS

357: \def\eg{{\it e.g.}}

358: \def\ie{{\it i.e.}}

359: \def\cf{{\it c.f.}}

360: \def\etal{{\it et. al.}}

361: \def\etc{{\it etc.}}

362: \hyphenation{Di-men-sion-al}

363: 

364: %%REFERENCING MACROS

365: \def\np{{\it Nucl. Phys.}}

366: \def\prl{{\it Phys. Rev. Lett.}}

367: \def\prev{{\it Phys. Rev.}}

368: \def\pl{{\it Phys. Lett.}}

369: \def\atamp{{\it Adv. Theor. Math. Phys.}}

370: \def\cqg{{\it Class. Quant. Grav.}}

371: \def\mpl{{\it Mod. Phys. Lett.}}

372: \def\cmp{{\it Comm. Math. Phys.}}

373: \def\annm{{\it Ann. Math.}}

374: \def\jhep{{\it J. High Energy Phys.}}

375: \def\ijmp{{\it Int. J. Mod. Phys.}}

376: %%

377: 

378: 

379: 

380: 

381: \lref\othervacuarefs{%\MartinecWG

382: E.~J.~Martinec and G.~W.~Moore, ``On decay of K-theory,'' arXiv:hep-th/0212059;

383: %%CITATION = HEP-TH 0212059;%%

384: G.~W.~Moore and A.~Parnachev,

385: ``Localized tachyons and the quantum McKay correspondence,''

386: JHEP {\bf 0411}, 086 (2004)

387: [arXiv:hep-th/0403016];

388: %%CITATION = HEP-TH 0403016;%%

389: I.~Melnikov and M.~R.~Plesser, ``The 

390: Coulomb Branch in Gauged Linear Sigma Models,''

391: arXiv:hep-th/0501238.}

392: %I. Melnikov, talk at SLAC fall 2004.}

393: 

394: %\AdamsSV

395: \lref\AdamsSV{

396: A.~Adams, J.~Polchinski and E.~Silverstein,

397: ``Don't panic! Closed string tachyons in ALE space-times,''

398: JHEP {\bf 0110}, 029 (2001)

399: [arXiv:hep-th/0108075].

400: %%CITATION = HEP-TH 0108075;%%

401: }

402: 

403: %\RohmAQ

404: \lref\RohmAQ{ R.~Rohm, ``Spontaneous Supersymmetry Breaking In Supersymmetric String Theories,'' Nucl.\ Phys.\ B

405: {\bf 237}, 553 (1984).

406: %%CITATION = NUPHA,B237,553;%%

407: }

408: 

409: 

410: 

411: 

412: 

413: \lref\stromcon{

414: %\StromingerCZ

415: A.~Strominger,

416: ``Massless black holes and conifolds in string theory,'' Nucl.\ Phys.\ B {\bf 451}, 96 (1995)

417: [arXiv:hep-th/9504090].

418: %%CITATION = HEP-TH 9504090;%%

419: }

420: 

421: \lref\dealwisetal{

422: %\deAlwisPR

423: S.~P.~de Alwis, J.~Polchinski and R.~Schimmrigk, ``Heterotic Strings With Tree Level Cosmological Constant,''

424: Phys.\ Lett.\ B {\bf 218}, 449 (1989);

425: %%CITATION = PHLTA,B218,449;%%

426: %\MyersFV

427: R.~C.~Myers, ``New Dimensions For Old Strings,'' Phys.\ Lett.\ B {\bf 199}, 371 (1987).

428: %%CITATION = PHLTA,B199,371;%%

429: }

430: 

431: %\DaCunhaFM

432: \lref\DaCunhaFM{

433: %\PolchinskiFN

434: %\lref\PolchinskiFN{

435: J.~Polchinski,

436: ``A Two-Dimensional Model For Quantum Gravity,''

437: Nucl.\ Phys.\ B {\bf 324}, 123 (1989);

438: %%CITATION = NUPHA,B324,123;%%

439: B.~C.~Da Cunha and E.~J.~Martinec, ``Closed string tachyon condensation

440: and worldsheet inflation,'' Phys.\ Rev.\ D {\bf 68}, 063502 (2003) [arXiv:hep-th/0303087];

441: %%CITATION = HEP-TH 0303087;%%

442: %\MartinecKA

443: E.~J.~Martinec,

444: ``The annular report on non-critical string theory,''

445: arXiv:hep-th/0305148.

446: %%CITATION = HEP-TH 0305148;%%

447: }

448: 

449: 

450: 

451: %\WittenGJ

452: \lref\WittenGJ{ E.~Witten, ``Instability Of The Kaluza-Klein Vacuum,'' Nucl.\ Phys.\ B {\bf 195}, 481 (1982).

453: %%CITATION = NUPHA,B195,481;%%

454: }

455: 

456: %\ShankarCM

457: \lref\ShankarCM{ R.~Shankar and E.~Witten, ``The S Matrix Of The Supersymmetric Nonlinear Sigma Model,'' Phys.\

458: Rev.\ D {\bf 17}, 2134 (1978).

459: %%CITATION = PHRVA,D17,2134;%%

460: }

461: 

462: %\AhnGN

463: \lref\AhnGN{ C.~Ahn, D.~Bernard and A.~LeClair, ``Fractional Supersymmetries In Perturbed Coset Cfts And

464: Integrable Soliton Theory,'' Nucl.\ Phys.\ B {\bf 346}, 409 (1990).

465: %%CITATION = NUPHA,B346,409;%%

466: }

467: 

468: %\FabingerJD

469: \lref\FabingerJD{ M.~Fabinger and P.~Horava, ``Casimir effect between world-branes in heterotic M-theory,''

470: Nucl.\ Phys.\ B {\bf 580}, 243 (2000) [arXiv:hep-th/0002073].

471: %%CITATION = HEP-TH 0002073;%%

472: }

473: 

474: 

475: 

476: 

477: \lref\openconfine{

478: %\YiHD

479: P.~Yi, ``Membranes from five-branes and fundamental strings from Dp branes,'' Nucl.\ Phys.\ B {\bf 550}, 214

480: (1999) [arXiv:hep-th/9901159].

481: %%CITATION = HEP-TH 9901159;%%

482: %\BergmanXF

483: O.~Bergman, K.~Hori and P.~Yi, ``Confinement on the brane,'' Nucl.\ Phys.\ B {\bf 580}, 289 (2000)

484: [arXiv:hep-th/0002223].

485: %%CITATION = HEP-TH 0002223;%%

486: %\SenMD

487: A.~Sen, ``Supersymmetric world-volume action for non-BPS D-branes,'' JHEP {\bf 9910}, 008 (1999)

488: [arXiv:hep-th/9909062].

489: %%CITATION = HEP-TH 9909062;%%

490: }

491: 

492: %\SenMG

493: \lref\sen{

494: A.~Sen, ``Non-BPS states and branes in string theory,'' arXiv:hep-th/9904207.

495: %%CITATION = HEP-TH 9904207;%%

496: }

497: 

498: \lref\tachtime{

499: %\OkudaYD

500: T.~Okuda and S.~Sugimoto, ``Coupling of rolling tachyon to closed strings,'' Nucl.\ Phys.\ B {\bf 647}, 101

501: (2002) [arXiv:hep-th/0208196];

502: %%CITATION = HEP-TH 0208196;%%

503: A.~Maloney, A.~Strominger and X.~Yin,

504: ``S-brane thermodynamics,''

505: JHEP {\bf 0310}, 048 (2003)

506: [arXiv:hep-th/0302146];

507: %%CITATION = HEP-TH 0302146;%%

508: %\LambertZR

509:  N.~Lambert, H.~Liu and J.~Maldacena, ``Closed strings from decaying D-branes,'' arXiv:hep-th/0303139;

510: %%CITATION = HEP-TH 0303139;%%

511: %\GaiottoRM

512: %\lref\GaiottoRM{

513: D.~Gaiotto, N.~Itzhaki and L.~Rastelli,

514: ``Closed strings as imaginary D-branes,''

515: Nucl.\ Phys.\ B {\bf 688}, 70 (2004)

516: [arXiv:hep-th/0304192].

517: %%CITATION = HEP-TH 0304192;%%

518: }

519: 

520: 

521: 

522: 

523: 

524: \lref\otherRG{

525: %\VafaRA

526: %C.~Vafa, ``Mirror symmetry and closed string tachyon condensation,''

527: %arXiv:hep-th/0111051;

528: %%CITATION = HEP-TH 0111051;%%

529: %\DavidVM

530: J.~R.~David, M.~Gutperle, M.~Headrick and S.~Minwalla, ``Closed string tachyon condensation on twisted

531: circles,'' JHEP {\bf 0202}, 041 (2002) [arXiv:hep-th/0111212];

532: %%CITATION = HEP-TH 0111212;%%

533: %\HeadrickHZ

534: M.~Headrick, S.~Minwalla and T.~Takayanagi, ``Closed string tachyon condensation: An overview,'' Class.\ Quant.\

535: Grav.\  {\bf 21}, S1539 (2004) [arXiv:hep-th/0405064];

536: %%CITATION = HEP-TH 0405064;%%

537: }

538: 

539: \lref\davetal{

540: %\MorrisonFR

541: D.~R.~Morrison, K.~Narayan and M.~R.~Plesser, ``Localized tachyons in C(3)/Z(N),'' JHEP {\bf 0408}, 047 (2004)

542: [arXiv:hep-th/0406039];

543: %%CITATION = HEP-TH 0406039;%%

544: %\MorrisonJA

545: D.~R.~Morrison and K.~Narayan, ``On tachyons, gauged linear sigma models, and flip transitions,''

546: arXiv:hep-th/0412337.

547: %%CITATION = HEP-TH 0412337;%%

548: }

549: 

550: \lref\chicago{

551: %\HarveyWM

552:  J.~A.~Harvey, D.~Kutasov, E.~J.~Martinec and G.~W.~Moore, ``Localized tachyons and RG flows,''

553: arXiv:hep-th/0111154;

554: %%CITATION = HEP-TH 0111154;%%

555: }

556: 

557: \lref\FQS{

558: %\FriedanXQ

559: D.~Friedan, Z.~Qiu and S.~H.~Shenker, ``Conformal Invariance, Unitarity And Two-Dimensional Critical

560: Exponents,'' Phys.\ Rev.\ Lett.\  {\bf 52}, 1575 (1984).

561: %%CITATION = PRLTA,52,1575;%%

562: }

563: 

564: \lref\KMS{

565: %\KastorEF

566: D.~A.~Kastor, E.~J.~Martinec and S.~H.~Shenker, ``RG Flow In N=1 Discrete Series,'' Nucl.\ Phys.\ B {\bf 316},

567: 590 (1989).

568: %%CITATION = NUPHA,B316,590;%%

569: }

570: 

571: \lref\earlierSStach{%\KachruED

572: S.~Kachru, J.~Kumar and E.~Silverstein, ``Orientifolds, RG flows, and closed string tachyons,'' Class.\ Quant.\

573: Grav.\  {\bf 17}, 1139 (2000) [arXiv:hep-th/9907038].

574: %%CITATION = HEP-TH 9907038;%%

575: }

576: 

577: 

578: %\ColemanBU

579: \lref\ColemanBU{ S.~R.~Coleman, ``Quantum Sine-Gordon Equation As The Massive Thirring Model,'' Phys.\ Rev.\ D

580: {\bf 11}, 2088 (1975).

581: %%CITATION = PHRVA,D11,2088;%%

582: }

583: 

584: %\GreeneYB

585: \lref\GreeneYB{ B.~R.~Greene, K.~Schalm and G.~Shiu, ``Dynamical topology change in M theory,'' J.\ Math.\

586: Phys.\  {\bf 42}, 3171 (2001) [arXiv:hep-th/0010207];

587: %%CITATION = HEP-TH 0010207;%%

588: }

589: 

590: 

591: 

592: \lref\classtop{

593: %\AspinwallNU

594: P.~S.~Aspinwall, B.~R.~Greene and D.~R.~Morrison, ``Calabi-Yau moduli space, mirror manifolds and spacetime

595: topology  change in string theory,'' Nucl.\ Phys.\ B {\bf 416}, 414 (1994) [arXiv:hep-th/9309097];

596: %%CITATION = HEP-TH 9309097;%%

597: %\WittenYC

598: E.~Witten, ``Phases of N = 2 theories in two dimensions,'' Nucl.\ Phys.\ B {\bf 403}, 159 (1993)

599: [arXiv:hep-th/9301042];

600: %%CITATION = HEP-TH 9301042;%%

601: %\DistlerMK

602: J.~Distler and S.~Kachru, ``(0,2) Landau-Ginzburg theory,'' Nucl.\ Phys.\ B {\bf 413}, 213 (1994)

603: [arXiv:hep-th/9309110].

604: %%CITATION = HEP-TH 9309110;%%

605: %\DistlerBC

606: J.~Distler and S.~Kachru, ``Duality of (0,2) string vacua,'' Nucl.\ Phys.\ B {\bf 442}, 64 (1995)

607: [arXiv:hep-th/9501111].

608: %%CITATION = HEP-TH 9501111;%%

609: }

610: 

611: \lref\trapping{

612: %\SilversteinHF

613: %\KofmanYC

614: L.~Kofman, A.~Linde, X.~Liu, A.~Maloney, L.~McAllister and E.~Silverstein, ``Beauty is attractive: Moduli

615: trapping at enhanced symmetry points,'' JHEP {\bf 0405}, 030 (2004) [arXiv:hep-th/0403001];

616: %\MohauptPQ

617: T.~Mohaupt and F.~Saueressig, ``Effective supergravity actions for conifold transitions,'' arXiv:hep-th/0410272.

618: %%CITATION = HEP-TH 0410272;%%

619: %\JarvQY

620: L.~Jarv, T.~Mohaupt and F.~Saueressig, ``M-theory cosmologies from singular Calabi-Yau compactifications,'' JCAP

621: {\bf 0402}, 012 (2004) [arXiv:hep-th/0310174].

622: %%CITATION = HEP-TH 0310174;%%

623: %%CITATION = HEP-TH 0403001;%%

624: E.~Silverstein and D.~Tong, ``Scalar speed limits and cosmology: Acceleration from D-cceleration,'' Phys.\ Rev.\

625: D {\bf 70}, 103505 (2004) [arXiv:hep-th/0310221];

626: %%CITATION = HEP-TH 0310221;%%

627: %\LukasDU

628: %\lref\LukasDU{

629:   A.~Lukas, E.~Palti and P.~M.~Saffin,

630:  ``Type IIB conifold transitions in cosmology,''

631:   Phys.\ Rev.\ D {\bf 71}, 066001 (2005)

632:   [arXiv:hep-th/0411033].

633:   %%CITATION = HEP-TH 0411033;%%

634: }

635: 

636: \lref\quantop{

637: %\GreeneHU

638: B.~R.~Greene, D.~R.~Morrison and A.~Strominger, ``Black hole condensation and the unification of string vacua,''

639: Nucl.\ Phys.\ B {\bf 451}, 109 (1995) [arXiv:hep-th/9504145];

640: %%CITATION = HEP-TH 9504145;%%

641: %\CandelasJS

642: P.~Candelas and X.~C.~de la Ossa, ``Comments On Conifolds,'' Nucl.\ Phys.\ B {\bf 342}, 246 (1990);

643: %%CITATION = NUPHA,B342,246;%%

644: %\KachruRS

645: S.~Kachru and E.~Silverstein, ``Chirality-changing phase transitions in 4d string vacua,'' Nucl.\ Phys.\ B {\bf

646: 504}, 272 (1997) [arXiv:hep-th/9704185].

647: %%CITATION = HEP-TH 9704185;%%

648: %\GukovZG

649: S.~Gukov, J.~Sparks and D.~Tong, ``Conifold transitions and five-brane condensation in M-theory on Spin(7)

650: %manifolds,''

651: Class.\ Quant.\ Grav.\  {\bf 20}, 665 (2003) [arXiv:hep-th/0207244].

652: %%CITATION = HEP-TH 0207244;%%

653: }

654: 

655: \lref\MandelstamHB{

656: S.~Mandelstam,

657: ``Soliton Operators For The Quantized Sine-Gordon Equation,''

658: Phys.\ Rev.\ D {\bf 11}, 3026 (1975).

659: %%CITATION = PHRVA,D11,3026;%%

660: }

661: %\MandelstamHB

662: %\ColemanBU

663: \lref\ColemanBU{

664: S.~R.~Coleman,

665: ``Quantum Sine-Gordon Equation As The Massive Thirring Model,''

666: Phys.\ Rev.\ D {\bf 11}, 2088 (1975).

667: %%CITATION = PHRVA,D11,2088;%%

668: }

669: 

670: %\KosterlitzXP

671: \lref\KosterlitzXP{

672: J.~M.~Kosterlitz and D.~J.~Thouless,

673: ``Ordering, Metastability And Phase Transitions In Two-Dimensional  Systems,''

674: J.\ Phys.\ C {\bf 6}, 1181 (1973).

675: %%CITATION = JPCBA,C6,1181;%%

676: }

677: 

678: %\KogutSN

679: \lref\KogutSN{

680: J.~B.~Kogut and L.~Susskind,

681: ``Vacuum Polarization And The Absence Of Free Quarks In Four-Dimensions,''

682: Phys.\ Rev.\ D {\bf 9}, 3501 (1974).

683: %%CITATION = PHRVA,D9,3501;%%

684: }

685: 

686: %\PolyakovFU

687: \lref\PolyakovFU{

688: A.~M.~Polyakov,

689: ``Quark Confinement And Topology Of Gauge Groups,''

690: Nucl.\ Phys.\ B {\bf 120}, 429 (1977).

691: %%CITATION = NUPHA,B120,429;%%

692: }

693: 

694: %\AhnUQ

695: \lref\AhnUQ{ C.~Ahn, ``Complete S Matrices Of Supersymmetric Sine-Gordon Theory And Perturbed Superconformal

696: Minimal Model,'' Nucl.\ Phys.\ B {\bf 354}, 57 (1991).

697: %%CITATION = NUPHA,B354,57;%%

698: }

699: 

700: %\ZamolodchikovXM

701: \lref\ZamolodchikovXM{ A.~B.~Zamolodchikov and A.~B.~Zamolodchikov, ``Factorized S-Matrices In Two Dimensions As

702: The Exact Solutions Of  Certain Relativistic Quantum Field Models,'' Annals Phys.\  {\bf 120}, 253 (1979).

703: %%CITATION = APNYA,120,253;%%

704: }

705: 

706: 

707: 

708: %\PolchinskiFN

709: \lref\PolchinskiFN{

710: J.~Polchinski,

711: ``A Two-Dimensional Model For Quantum Gravity,''

712: Nucl.\ Phys.\ B {\bf 324}, 123 (1989).

713: %%CITATION = NUPHA,B324,123;%%

714: }

715: 

716: \lref\Xiao{

717: S.~Hellerman and X.~Liu,

718: ``Dynamical dimension change in supercritical string theory,''

719: arXiv:hep-th/0409071.

720: %%CITATION = HEP-TH 0409071;%%

721: }

722: 

723: %\SaltmanJH

724: \lref\SaltmanJH{

725: A.~Saltman and E.~Silverstein,

726: ``A new handle on de Sitter compactifications,''

727: arXiv:hep-th/0411271.

728: %%CITATION = HEP-TH 0411271;%%

729: }

730: 

731: %\ScherkTA

732: \lref\ScherkTA{

733: J.~Scherk and J.~H.~Schwarz,

734: ``Spontaneous Breaking Of Supersymmetry Through Dimensional Reduction,''

735: Phys.\ Lett.\ B {\bf 82}, 60 (1979).

736: %%CITATION = PHLTA,B82,60;%%

737: }

738: 

739: %\HollowoodEX

740: \lref\HollowoodEX{

741: T.~J.~Hollowood and E.~Mavrikis,

742: ``The N = 1 supersymmetric bootstrap and Lie algebras,''

743: Nucl.\ Phys.\ B {\bf 484}, 631 (1997)

744: [arXiv:hep-th/9606116].

745: %%CITATION = HEP-TH 9606116;%%

746: }

747: 

748: %\BajnokDK

749: \lref\BajnokDK{

750: Z.~Bajnok, C.~Dunning, L.~Palla, G.~Takacs and F.~Wagner,

751: ``SUSY sine-Gordon theory as a perturbed conformal field theory and  finite

752: size effects,''

753: Nucl.\ Phys.\ B {\bf 679}, 521 (2004)

754: [arXiv:hep-th/0309120].

755: %%CITATION = HEP-TH 0309120;%%

756: }

757: 

758: %\FerraraJV

759: \lref\FerraraJV{

760: S.~Ferrara, L.~Girardello and S.~Sciuto,

761: ``An Infinite Set Of Conservation Laws Of The Supersymmetric Sine-Gordon

762: Theory,''

763: Phys.\ Lett.\ B {\bf 76}, 303 (1978).

764: %%CITATION = PHLTA,B76,303;%%

765: }

766: 

767: 

768: 

769: %\WittenYC

770: \lref\WittenYC{

771: E.~Witten,

772: ``Phases of N = 2 theories in two dimensions,''

773: Nucl.\ Phys.\ B {\bf 403}, 159 (1993)

774: [arXiv:hep-th/9301042].

775: %%CITATION = HEP-TH 9301042;%%

776: }

777: 

778: 

779: %\VafaRA

780: \lref\VafaRA{

781: C.~Vafa,

782: ``Mirror symmetry and closed string tachyon condensation,''

783: arXiv:hep-th/0111051.

784: %%CITATION = HEP-TH 0111051;%%

785: }

786: 

787: %\MorrisonJA

788: \lref\MorrisonJA{

789: D.~R.~Morrison and K.~Narayan,

790: ``On tachyons, gauged linear sigma models, and flip transitions,''

791: arXiv:hep-th/0412337.

792: %%CITATION = HEP-TH 0412337;%%

793: }

794: 

795: %\BuscherQJ

796: \lref\BuscherQJ{

797: T.~H.~Buscher,

798: ``Path Integral Derivation Of Quantum Duality In Nonlinear Sigma Models,''

799: Phys.\ Lett.\ B {\bf 201}, 466 (1988).

800: %%CITATION = PHLTA,B201,466;%%

801: }

802: 

803: %\RocekPS

804: \lref\RocekPS{

805: M.~Rocek and E.~Verlinde,

806: ``Duality, quotients, and currents,''

807: Nucl.\ Phys.\ B {\bf 373}, 630 (1992)

808: [arXiv:hep-th/9110053].

809: %%CITATION = HEP-TH 9110053;%%

810: }

811: 

812: 

813: 

814: 

815: %\MorrisonYH

816: \lref\MorrisonYH{

817: D.~R.~Morrison and M.~R.~Plesser,

818: ``Towards mirror symmetry as duality for two dimensional abelian gauge

819: theories,''

820: Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 46}, 177 (1996)

821: [arXiv:hep-th/9508107].

822: %%CITATION = HEP-TH 9508107;%%

823: }

824: 

825: 

826: 

827: 

828: %\HoriKT

829: \lref\HoriKT{

830: K.~Hori and C.~Vafa,

831: ``Mirror symmetry,''

832: arXiv:hep-th/0002222.

833: %%CITATION = HEP-TH 0002222;%%

834: }

835: 

836: %\MorrisonYH

837: \lref\MorrisonYH{ D.~R.~Morrison and M.~R.~Plesser, ``Towards mirror symmetry as duality for two dimensional

838: abelian gauge theories,'' Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 46}, 177 (1996) [arXiv:hep-th/9508107].

839: %%CITATION = HEP-TH 9508107;%%

840: }

841: 

842: 

843: 

844: %\AharonySX

845: \lref\AharonySX{

846: O.~Aharony, J.~Marsano, S.~Minwalla, K.~Papadodimas and M.~Van Raamsdonk,

847: ``The Hagedorn/deconfinement phase transition in weakly coupled large N gauge

848: theories,''

849: arXiv:hep-th/0310285.

850: %%CITATION = HEP-TH 0310285;%%

851: }

852: 

853: %\ZamolodchikovGT

854: \lref\ZamolodchikovGT{

855: A.~B.~Zamolodchikov,

856: ``'Irreversibility' Of The Flux Of The Renormalization Group In A 2-D Field

857: Theory,''

858: JETP Lett.\  {\bf 43}, 730 (1986)

859: [Pisma Zh.\ Eksp.\ Teor.\ Fiz.\  {\bf 43}, 565 (1986)].

860: %%CITATION = JTPLA,43,730;%%

861: }

862: 

863: %\BanksQS

864: \lref\BanksQS{

865: T.~Banks and E.~J.~Martinec,

866: ``The Renormalization Group And String Field Theory,''

867: Nucl.\ Phys.\ B {\bf 294}, 733 (1987).

868: %%CITATION = NUPHA,B294,733;%%

869: }

870: 

871: 

872: 

873: %\CostaNW

874: \lref\CostaNW{

875: M.~S.~Costa and M.~Gutperle,

876: ``The Kaluza-Klein Melvin solution in M-theory,''

877: JHEP {\bf 0103}, 027 (2001)

878: [arXiv:hep-th/0012072].

879: %%CITATION = HEP-TH 0012072;%%

880: }

881: 

882: 

883: %\GutperleMB

884: \lref\GutperleMB{

885: M.~Gutperle and A.~Strominger,

886: ``Fluxbranes in string theory,''

887: JHEP {\bf 0106}, 035 (2001)

888: [arXiv:hep-th/0104136].

889: %%CITATION = HEP-TH 0104136;%%

890: }

891: 

892: 

893: %\DavidVM

894: %\lref\DavidVM{

895: %J.~R.~David, M.~Gutperle, M.~Headrick and S.~Minwalla,

896: %``Closed string tachyon condensation on twisted circles,''

897: %JHEP {\bf 0202}, 041 (2002)

898: %[arXiv:hep-th/0111212].

899: %%%CITATION = HEP-TH 0111212;%%

900: %}

901: 

902: %\GutperleBP

903: \lref\GutperleBP{

904: M.~Gutperle,

905: ``A note on perturbative and nonperturbative instabilities of twisted circles,''

906: Phys.\ Lett.\ B {\bf 545}, 379 (2002)

907: [arXiv:hep-th/0207131].

908: %%CITATION = HEP-TH 0207131;%%

909: }

910: 

911: 

912: %\AspinwallRB

913: \lref\AspinwallRB{

914: P.~S.~Aspinwall, D.~R.~Morrison and M.~Gross,

915: ``Stable singularities in string theory,''

916: Commun.\ Math.\ Phys.\  {\bf 178}, 115 (1996)

917: [arXiv:hep-th/9503208].

918: %%CITATION = HEP-TH 9503208;%%

919: }

920: 

921: 

922: \lref\LindeSK{ A.~D.~Linde, ``Hard art of the universe creation (stochastic

923: approach to tunneling and baby

924: universe formation),'' Nucl.\ Phys.\ B {\bf 372}, 421 (1992)

925: [arXiv:hep-th/9110037].

926: %%CITATION = HEP-TH 9110037;%%

927: }

928: 

929: 

930: 

931: %\ArkaniHamedSP

932: \lref\ArkaniHamedSP{

933: N.~Arkani-Hamed, H.~Georgi and M.~D.~Schwartz,

934: ``Effective field theory for massive gravitons and gravity in theory space,''

935: Annals Phys.\  {\bf 305}, 96 (2003)

936: [arXiv:hep-th/0210184].

937: %%CITATION = HEP-TH 0210184;%%

938: }

939: 

940: %\VafaUE

941: \lref\VafaUE{

942: C.~Vafa,

943: ``c Theorem And The Topology Of 2-D Qfts,''

944: Phys.\ Lett.\ B {\bf 212}, 28 (1988).

945: %%CITATION = PHLTA,B212,28;%%

946: }

947: 

948: 

949: \lref\RobbinsHX{

950: D.~Robbins and S.~Sethi,

951: ``A barren landscape,''

952: arXiv:hep-th/0405011.

953: }

954: 

955: 

956: %\PolchinskiHB

957: \lref\PolchinskiHB{

958: J.~Polchinski,

959: ``Cosmic superstrings revisited,''

960: arXiv:hep-th/0410082.

961: %%CITATION = HEP-TH 0410082;%%

962: }

963: 

964: \lref\joequote{

965: ``I had started to include a list,

966: but decided that it would be distracting

967: and inflammatory.  It was rather long.'' 

968: \PolchinskiHB.

969: }

970: 

971: %\FriedanJF

972: \lref\friedan{

973: D.~Friedan,

974: ``Nonlinear Models In $2 + \epsilon$ Dimensions,''

975: Phys.\ Rev.\ Lett.\  {\bf 45}, 1057 (1980).

976: %%CITATION = PRLTA,45,1057;%%

977: }

978: %\CappelliYC

979: \lref\Cappelli{

980: A.~Cappelli, D.~Friedan and J.~I.~Latorre,

981: ``c-Theorem And Spectral Representation,''

982: Nucl.\ Phys.\ B {\bf 352}, 616 (1991).

983: %%CITATION = NUPHA,B352,616;%%

984: }

985: 

986: \lref\ricciflow{

987: G.~Perelman, ``The entropy formula for the Ricci flow

988: and its geometric applications," arXiv:math.DG/0211159;

989: ``Ricci flow with surgery on three-manifolds,"

990: arXiv:math.DG/0303109.

991: For a review of earlier work, see 

992: H.-D.~Cao, B.~Chow, ``Recent Developments on the Ricci Flow,"

993: arXiv:math.DG/9811123.}

994: 

995: 

996: %\HeadrickYU

997: \lref\HeadrickYU{

998: M.~Headrick,

999: %``Decay of C/Z(n): Exact supergravity solutions,''

1000: JHEP {\bf 0403}, 025 (2004)

1001: [arXiv:hep-th/0312213].

1002: %%CITATION = HEP-TH 0312213;%%

1003: }

1004: 

1005: %\HeadrickPS

1006: \lref\HeadrickPS{

1007: M.~Headrick and J.~Raeymaekers,

1008: %``The large N limit of C/Z(N) and supergravity,''

1009: arXiv:hep-th/0411148.

1010: %%CITATION = HEP-TH 0411148;%%

1011: }

1012: %\OkawaRH

1013: \lref\OkawaRH{

1014: Y.~Okawa and B.~Zwiebach,

1015: %``Twisted tachyon condensation in closed string field theory,''

1016: JHEP {\bf 0403}, 056 (2004)

1017: [arXiv:hep-th/0403051].

1018: %%CITATION = HEP-TH 0403051;%%

1019: }

1020: 

1021: 

1022: 

1023:  

1024: 

1025:  

1026: 

1027: \lref\timeprobes{

1028: R.~Gregory and J.~A.~Harvey,

1029: ``Spacetime decay of cones at strong coupling,''

1030: Class.\ Quant.\ Grav.\  {\bf 20}, L231 (2003)

1031: [arXiv:hep-th/0306146]; 

1032: %%CITATION = HEP-TH 0306146;%%

1033: M.~Headrick,

1034: ``Decay of C/Z(n): Exact supergravity solutions,''

1035: JHEP {\bf 0403}, 025 (2004)

1036: [arXiv:hep-th/0312213];

1037: %%CITATION = HEP-TH 0312213;%%

1038: M.~Headrick and J.~Raeymaekers,

1039: ``The large N limit of C/Z(N) and supergravity,''

1040: arXiv:hep-th/0411148;

1041: %%CITATION = HEP-TH 0411148;%%

1042: Y.~Okawa and B.~Zwiebach,

1043: ``Twisted tachyon condensation in closed string field theory,''

1044: JHEP {\bf 0403}, 056 (2004)

1045: [arXiv:hep-th/0403051].

1046: %%CITATION = HEP-TH 0403051;%%

1047: }

1048: 

1049: \lref\wormholerefs{ 

1050: S.~R.~Coleman,

1051: ``Black Holes As Red Herrings: Topological Fluctuations And The Loss Of

1052: Quantum Coherence,''

1053: Nucl.\ Phys.\ B {\bf 307}, 867 (1988);

1054: %%CITATION = NUPHA,B307,867;%%

1055: ``Why There Is Nothing Rather Than Something: A Theory Of The Cosmological

1056: Constant,''

1057: Nucl.\ Phys.\ B {\bf 310}, 643 (1988);

1058: %%CITATION = NUPHA,B310,643;%%

1059: S.~B.~Giddings and A.~Strominger, ``Loss Of Incoherence And Determination Of Coupling

1060: Constants In Quantum Gravity,'' Nucl.\ Phys.\ B {\bf 307}, 854 (1988).

1061: %%CITATION = NUPHA,B307,854;%%

1062: }

1063: 

1064: 

1065:  %\BarbonDI

1066: \lref\Barbon{

1067:   J.~L.~F.~Barbon and E.~Rabinovici,

1068:  ``Closed-string tachyons and the Hagedorn transition in AdS space,''

1069:   JHEP {\bf 0203}, 057 (2002)

1070:   [arXiv:hep-th/0112173];

1071:   %%CITATION = HEP-TH 0112173;%%

1072:  %\BarbonNW

1073: %\lref\BarbonNW{

1074: %  J.~L.~F.~Barbon and E.~Rabinovici,

1075:  ``Remarks on black hole instabilities and closed string tachyons,''

1076:   Found.\ Phys.\  {\bf 33}, 145 (2003)

1077:   [arXiv:hep-th/0211212];

1078:   %%CITATION = HEP-TH 0211212;%%

1079: %}

1080: %\BarbonDD

1081: %\lref\BarbonDD{

1082: %  J.~L.~F.~Barbon and E.~Rabinovici,

1083:  ``Touring the Hagedorn ridge,''

1084:   arXiv:hep-th/0407236;

1085:   %%CITATION = HEP-TH 0407236;%%

1086: %}

1087:  %\BarbonJR

1088: %\lref\BarbonJR{

1089: %  J.~L.~F.~Barbon and E.~Rabinovici,

1090:  ``Topology change and unitarity in quantum black hole dynamics,''

1091:   arXiv:hep-th/0503144.

1092:   %%CITATION = HEP-TH 0503144;%%

1093: }

1094: 

1095: 

1096:  

1097: 

1098:  

1099: 

1100: 

1101: 

1102: 

1103: 

1104: \Title{\vbox{\baselineskip12pt\hbox{hep-th/0502021} \hbox{SU-ITP-05/06}\hbox{SLAC-PUB-11011}\hbox{HUTP-05/A0006}}}

1105: {\vbox{

1106: %\centerline{Tachyon Attack: Topology Change}\centerline{and Baby Universes from Tachyon Condensation}

1107: %\centerline{A New Handle}\centerline{ on Topology Change and Baby Universes} }

1108: %\centerline{Handle With Care}

1109: %\centerline{Tachyonic Scalpels and Stringy Handles}

1110: %\centerline{The Incredible Shrinking Handle}

1111: %\centerline{Handling Bulk Tachyons}

1112: %\centerline{Handling Bulk Tachyons and Baby Universes}

1113: %\centerline{Sewing Spacetime Tachyonic Pants}

1114: %\centerline{Tachyonic Pants and Baby Universes}

1115: %\centerline{Tachyonic Handles and Baby Universes}

1116: %\centerline{Bulky Tachy Spacetime Pants}

1117: %\centerline{Faster, Tachyon, Kill, Kill!}

1118: %\centerline{The Rise and Fall of the Bulk Tachyon}

1119: %\centerline{Death and Tachyons}

1120: %\centerline{Asexual Reproduction of String Vacuua}

1121: %\centerline{Watermusic for Riemann Surfaces}

1122: %\centerline{Beyond Good and Evil}

1123: %\centerline{Tachyonic Handles Lead To Topology Changing Babies}

1124: %\centerline{How Tachyons Handle Babies}

1125: %\centerline{How Babies Handle Tachyons}

1126: %\centerline{Beyond The Zero}

1127: %\centerline{Beyond Type Zero}

1128: %\centerline{Burninating the Countryside}

1129: %\centerline{I Said Consummate V's!}

1130: %\centerline{A Tachyonic Scalpel for Siamese donuts}

1131: %\centerline{Taking a Tachyonic Scalpel to Riemann's Siamese Tori}

1132: %\centerline{Tachyonicology}

1133: %\centerline{Perturbative Topology Change and Baby Universe Formation}\centerline{from Localized Stringy Tachyon

1134: %Condensation}

1135: %\centerline{Topology Change and Disconnected Universes}\medskip\centerline{ from Closed String Tachyons}}}

1136: %\centerline{OK, Panic: Topology Change and Baby Universes}\medskip\centerline{ from Closed String Tachyons}}}

1137: \centerline{Things Fall Apart:}\bigskip\centerline{Topology Change from 

1138: Winding Tachyons}}}

1139: 

1140: \bigskip

1141: \bigskip

1142: \centerline{A. Adams$^1$, X. Liu$^2$, J. McGreevy$^2$, A. Saltman$^2$, E. Silverstein$^2$}

1143: \bigskip

1144: 

1145: \centerline{$^1$ \it Jefferson Physical Laboratory, Harvard University, Cambridge, MA 02138}

1146: \centerline{$^2$ \it SLAC and Department of Physics, Stanford University, Stanford, CA 94305-4060}

1147: \bigskip

1148: \bigskip

1149: \noindent

1150: 

1151: We argue that closed string tachyons drive two spacetime topology

1152: changing transitions -- loss of genus in a Riemann surface and

1153: separation of a Riemann surface into two components. The tachyons

1154: of interest are localized versions of Scherk-Schwarz winding

1155: string tachyons arising on Riemann surfaces in regions of moduli

1156: space where string-scale tubes develop. Spacetime and world-sheet

1157: renormalization group analyses provide strong evidence that the

1158: decay of these tachyons removes a portion of the spacetime,

1159: splitting the tube into two pieces.  We address the fate of the

1160: gauge fields and charges lost in the process, generalize it to

1161: situations with weak flux backgrounds, and use this process to

1162: study the type 0 tachyon, providing further evidence that its

1163: decay drives the theory sub-critical. Finally, we discuss the

1164: time-dependent dynamics of this topology-changing transition and

1165: find that it can occur more efficiently than analogous transitions

1166: on extended supersymmetric moduli spaces, which are limited by

1167: moduli trapping.

1168: 

1169: \bigskip

1170: \Date{January, 2004}

1171: %\draftmode

1172: 

1173: \newsec{Introduction and Setup}

1174: 

1175: One of the most basic aspects of 

1176: %processes in

1177: %issues in

1178: quantum gravity is

1179: the possibility of

1180: dynamical change of spacetime topology.  In classical general

1181: relativity such processes would be singular in the spacetime

1182: metric, but in string theory the singularities may be smoothed out either

1183: classically (via effects having to do with the extent of the

1184: string) or quantum mechanically. Previous examples of the former

1185: include \refs{\classtop,\davetal} and the latter \quantop.  In this note

1186: we show that a natural conjecture for the endpoint of

1187: condensation of a localized

1188: %closed string

1189: tachyon

1190: %condensation

1191: in string theory implies the

1192: existence of two simple topology-changing processes. The

1193: conjecture has substantial supporting evidence that we review and

1194: develop in a similar spirit to \AdamsSV.

1195: 

1196: We compactify type II string theory down to eight dimensions on a

1197: %compact

1198: Riemann surface, which is topologically characterized by

1199: the spin structure and Euler character

1200: %

1201: \eqn\eulerchar{\chi=2-2h,}

1202: %

1203: of each connected component, where the genus $h$ counts the number

1204: of handles of a component. We will start with a single component.

1205: Each handle adds energy density to the surface, as demonstrated by

1206: the eight-dimensional Einstein frame potential energy of a

1207: constant-curvature Riemann surface

1208: %

1209: \eqn\poten{U_{8E} \sim {1\over l_8^8}

1210: %\biggl({g_s^2\overV_\Sigma}\biggr)^{4/3}{1\over g_s^2}

1211: %{ g_s^{2/3} \over V_\Sigma^{4/3} }

1212: \left( { g_s \over V_\Sigma^2 } \right)^{2/3} (2h-2),}

1213: %

1214: with 8-d Planck length $l_8$, string coupling $g_s$, and volume $V_\Sigma$, in string units.

1215: %For a metric whose curvature

1216: %is not uniformly distributed,

1217: %the energy is bigger.

1218: The dependence on the dilaton and volume yields time-dependent

1219: expansion and evolution toward weak coupling; we take the 

1220: initial

1221: volume large 

1222: %and curvature weak 

1223: so that these effects are under control

1224: in the low-energy effective theory.\foot{In \SaltmanJH, two

1225: additional Riemann surface factors in the compactification

1226: manifold and extra brane and flux ingredients were introduced to

1227: perturbatively metastabilize this system in four dimensions,

1228: leading to static solutions for the dilaton, complex structure,

1229: and three volume moduli at large radius and weak coupling, away

1230: from extreme limits of complex structure moduli space. Here we are

1231: interested in a different regime of complex structure moduli

1232: space, accessible from a simpler compactification on a single

1233: Riemann surface down to eight dimensions, and we will necessarily

1234: consider a time-dependent compactification as a result.} The

1235: equations of motion for the metric in the low-energy theory yield

1236: a mild FRW expansion, to which similar comments apply. Starting

1237: from a more general metric of nonconstant curvature on the Riemann

1238: surface, each volume element of negative curvature expands due to

1239: the local energy \poten\ and each element of positive curvature

1240: contracts. Further, the initial value of $g_s$ can be made small,

1241: so that string interactions are negligible.

1242: 

1243: Because of the negative curvature of the compactification for $h

1244: > 1$, this energy density \poten\ is positive and the system will tend to

1245: reduce the genus if there is a dynamical mechanism by which it can

1246: do so.\foot{In the superstring, a world-sheet Witten index valid

1247: classically in the spacetime theory predicts further that there

1248: will be additional discrete components to the target space, for

1249: which we also find independent evidence.  Such effects have also

1250: been seen before for example in \refs{\chicago,\othervacuarefs}.}

1251: %because of the constant term in the Euler character \eulerchar,

1252: Factoring the Riemann surface into multiple components is also

1253: energetically favored. If a Riemann surface of genus $h$ splits

1254: into one of genus $0 < h_1 < h$ and another of genus $h-h_1$, the

1255: Euler character on each of the resulting surfaces is smaller and

1256: hence the potential energy \poten\ is smaller in each decoupled

1257: sector than it used to be in the original connected space.  In all

1258: cases, since we consider small string coupling the energy density

1259: liberated from \poten\ is parameterically smaller than the Planck

1260: energy density.

1261: %net Euler character becomes $2-2h_1+ 2- 2(h-h_1) = 4 - 2 h$, decreasing the energy \poten.

1262: In this note, we will present evidence that both types of transitions are mediated by localized Scherk-Schwarz

1263: winding tachyons.

1264: 

1265: 

1266: \doublefig\before{ The two transition regions:  a) a thin handle;

1267: b) a factorized surface. }{\epsfxsize2.0in\epsfbox{pinching.eps}}

1268: {\epsfxsize2.0in\epsfbox{separating.eps}} We start from a

1269: controlled regime in which curvatures are everywhere weak, but

1270: where the Riemann surface degenerates in some local region to form

1271: a long, thin (sub-string-scale) tube, with antiperiodic boundary

1272: conditions for fermions around the circular direction. As depicted

1273: in \before, this can happen when a handle degenerates or when the

1274: surface nears a factorization limit. In the former case, in order

1275: to ensure antiperiodic boundary conditions, we must choose the

1276: spin structure appropriately.  In the latter case, the fermions

1277: automatically have antiperiodic boundary conditions around the

1278: thin cycle -- this can be seen by thinking of the Riemann surface

1279: as a string world-sheet, in which case this statement is a

1280: consequence of spacetime fermion number conservation.

1281: %Factorization therefore represents a possibly ubiquitous

1282: %instability in compactifications involving higher-genus Riemann

1283: %surfaces (though as emphasized above in the presence of fluxes and

1284: %other ingredients the system is generically metastabilized away

1285: %from such regions \SaltmanJH).

1286: %because it is homologically trivial--shrinking this cycle to zero reveals that the antiperiodic

1287: %boundary conditions reduce to the usual reflection of spacetime fermions under a $2\pi$ rotation.

1288: We can consistently consider such regions while also maintaining small curvature (${\cal R} \alpha^\prime \ll

1289: 1$) everywhere: although a small length scale is developing on the thin tube, it is nearly flat.

1290: 

1291: The regions of complex structure moduli space containing small

1292: tubes with the specified boundary conditions arise naturally from

1293: perturbative dynamics.  Although classically the potential is flat

1294: for the complex structure moduli (in the absence of stabilizing

1295: fluxes \SaltmanJH), the 1-loop contribution to the potential

1296: drives the radius of the tube to smaller values \RohmAQ.

1297: 

1298: The small circle introduces stringy physics -- classical geometric

1299: notions break down. In particular, with the above specifications,

1300: this tube is locally a Scherk-Schwarz (SS) circle \ScherkTA\ times

1301: a line. Therefore, the theory develops tachyonic modes in its

1302: winding string spectrum as the handle becomes thinner than the

1303: string scale. For example, in ten-dimensional type II string

1304: theory, the Scherk-Schwarz circle of radius $L l_s$ can be

1305: obtained from a real line by an orbifold action $(-1)^F$ times a

1306: translation by $L l_s$, where $F$ is the spacetime fermion number.

1307: The world-sheet vacuum energy in the $n$th twisted sector is

1308: $-1/2+n^2L^2$.

1309: %(as can be seen most

1310: %simply by working in the Green-Schwarz formalism)

1311: For sufficiently small $L$, this is negative, yielding tachyonic

1312: modes in the spacetime spectrum.\foot{As $L\to 0$, these tachyonic

1313: winding modes are T-dual to momentum modes of the Type 0 bulk

1314: tachyon, whose condensation we will also examine below using

1315: similar methods.}

1316: 

1317: \doublefig\after{

1318: %The two transition regions:

1319: The centre cannot hold: a) a thin handle capped off; b) a

1320: factorized surface capped off.

1321: }{\epsfxsize2.0in\epsfbox{pinched.eps}}

1322: {\epsfxsize2.0in\epsfbox{separated.eps}} Tachyon condensation

1323: generically reduces the number of degrees of freedom, and in the

1324: case of the Scherk-Schwarz tachyon we will show that this effect

1325: follows from standard results concerning the mass gap of a

1326: corresponding supersymmetric sine-Gordon theory in two dimensions.

1327: Including both dimensions of the tube, we argue in a consistent

1328: approximation scheme that condensing this tachyon causes the

1329: system to lose the region where the circle is smaller than string

1330: scale, breaking the tube there and capping off the remaining

1331: regions (see \after). Our specification of small curvatures

1332: everywhere guarantees that the time evolution induced by the

1333: spacetime curvature term \poten\ is parameterically slower than

1334: that of the tachyon decay process (at least initially).

1335: 

1336: That bulk Scherk-Schwarz tachyon condensation yields a subcritical

1337: theory has been conjectured before in general terms

1338: \refs{\dealwisetal,\earlierSStach, \FabingerJD}.  Here we analyze

1339: this idea in detail in the more controlled setting described

1340: above, in which the tachyon is localized, so that we can apply the

1341: result to obtain the topology-changing processes just described.

1342: The suggestion that Hagedorn tachyons

1343: have an endpoint similar to the one described here 

1344: has been made in \Barbon.

1345: An interesting comment about winding tachyons 

1346: in nonlinear sigma models was made in \friedan.

1347: 

1348: \subsec{Setup and Plan}

1349: 

1350: Classically condensing the tachyon in real time requires adding to

1351: the world-sheet action the on-shell marginal tachyon vertex

1352: operator and solving the resulting path integral in the deep IR.

1353: In type II string theory, the world-sheet has (1,1) local

1354: supersymmetry.  The zero ghost picture tachyon vertex operator is

1355: of the form

1356: %

1357: \eqn\Top{\int d^2 \sigma d\theta^+ d\theta^- ~T(X)}

1358: %

1359: where we work in ${\cal N}=1$ superspace, with spacetime embedding coordinates $x$ extended to $\CN=1$ scalar

1360: multiplets $X=x+\theta^+\psi^-+\theta^-\psi^+ + \theta^+\theta^- F$.

1361: 

1362: We denote the direction around the Scherk-Schwarz circle $\theta$,

1363: its T-dual $\tilde\theta$, and the superspace coordinate

1364: corresponding to the latter $\tilde\Theta$. We also parameterize

1365: the direction lengthwise along the tube by $r$, with corresponding

1366: superspace coordinate $R$.  Finally we denote the target-space

1367: time by $t$, extended to superspace coordinate $X^0$.  The tachyon

1368: vertex operator for our system is

1369: %

1370: \eqn\Tsupvert{T(X)=e^{\kappa X^0}\hat T(R)\cos[w\tilde\Theta],}

1371: %

1372: where $w=nL/l_s$ with $n$ the winding number and $L$ the radius of the tube.  The second reference in

1373: \DaCunhaFM\ studied aspects of this system for other applications, and we will use some of its observations in

1374: our analysis.

1375: 

1376: As we will discuss in more detail below, if we add a tachyon vertex operator with  mild $r$ and $t$ dependence,

1377: %

1378: \eqn\regime{\kappa^2 < k_r^2 \ll w^2}

1379: %

1380: the action generated by \Tsupvert\ is the supersymmetric sine-Gordon model (SSG) for $\tilde\Theta$, corrected

1381: by subleading pieces depending on $\del_{X^0}T$ and $\del_r T$.

1382: 

1383: Classically, the world-sheet theory with the interaction terms

1384: \Tsupvert\ has a mass gap for $\tilde\Theta$ sector in the region

1385: of the tachyon condensate. The physics of the tachyon condensation

1386: process is governed by the quantum theory in the IR limit. In the

1387: regime \regime\ we will see that we can treat the IR quantum

1388: dynamics of the SSG theory as the leading physics of the tachyon

1389: condensation in the $\tilde\Theta$ sector.  As we will review,

1390: there is strong evidence for a mass gap in this system.

1391: 

1392: In addition to its role in the full time-dependent problem induced

1393: by the condensation of $T$ \Tsupvert, the renormalization group

1394: (RG) flow in the world-sheet matter sector (which is induced by

1395: the addition of the $\kappa=0$ vertex operator to the world-sheet

1396: Lagrangian) may well reflect part of the off-shell configuration

1397: space of string theory, since string field theory is built on

1398: off-shell string states and vertices which sample non-conformal

1399: regimes in the world-sheet matter theory. This has been argued in,

1400: for example

1401: \refs{\BanksQS,\VafaUE,\dealwisetal,\AdamsSV,\VafaRA,\otherRG,\chicago};

1402: in all known cases the RG results agree with the pattern found by

1403: probes of the time-dependent process

1404: \refs{\sen,\AdamsSV,\timeprobes}.\foot{One can consider limits in which the

1405: second order equations governing the time-dependent evolution of

1406: the tachyon $T$ become first order, via a large coupling of the

1407: form $\dot\Phi\dot T$ where $\Phi$ is some combination of dilaton

1408: and volume moduli \otherRG. However, this approximation can break

1409: down at later stages in the process.}

1410: 

1411: We thus start in \S2\ by reviewing the RG flow of the sine-Gordon

1412: theory (supersymmetric and otherwise), as well as the RG flow of

1413: this theory weakly coupled to $R$, in the matter theory on the

1414: world-sheet ignoring the effects of $X^0$.  We see that the RG

1415: flow removes the space in the region of the tachyon condensate. In

1416: the appendix, we present a linear sigma model which exhibits the

1417: flow for both dimensions at once; the model has a relevant

1418: operator which changes the vacuum manifold from a connected

1419: hyperboloid to a two-sheeted disconnected one. We also apply this

1420: result to the problem of the condensation of the Type 0 tachyon.

1421: We then apply the RG result to the classical time-dependent

1422: problem \Tsupvert\ in \S3, exhibiting a barrier to penetration of

1423: the world-sheet into the region of tachyon condensate. We also

1424: comment on the efficiency of our topology change process

1425: (including effects of nonzero string coupling), comparing and

1426: contrasting to earlier tachyon and topology change analyses.

1427: 

1428: In \S4\ we discuss the fate of the winding charge when a homology

1429: cycle is destroyed by tachyon condensation. We find that the gauge

1430: field under which the condensing tachyon is charged is Higgsed and

1431: that the dual gauge field appears to confine. In addition, we

1432: comment on the behavior of D-brane charges under the tachyon

1433: condensation, and extend our discussion to Riemann surfaces with

1434: weak flux backgrounds.  In \S5\ we discuss some of the potential

1435: implications of our results and possible directions for further

1436: research.  In appendix A we 

1437: describe a linear sigma model 

1438: which captures the RG.

1439: 

1440: While this paper was in preparation we 

1441: noticed that our results

1442: had been anticipated by the 

1443: prescient early work of Yeats.

1444: In appendix B, we reproduce his argument,

1445: with a translation to more

1446: modern notation.

1447: Note that his analysis applies

1448: under more general circumstances.

1449: 

1450: 

1451: 

1452: 

1453: \newsec{Renormalization Group Structure}

1454: 

1455: In order to study the RG problem in the matter sector, we suppress the $X^0$ dependence in the vertex operator

1456: \Tsupvert. The resulting relevant operator in the region of the tube is (in the zero ghost picture) of the form

1457: %

1458: \eqn\Top{\int d^2 \sigma d^2\theta ~T(X)}

1459: %

1460: where we work in ${\cal N}=1$ superspace, with world-sheet fields

1461: $x$ extended to $\CN=1$ scalar multiplets $X$.  We denote the

1462: direction around the Scherk-Schwarz circle by $\theta$ (not to be

1463: confused with the world-sheet superspace coordinates

1464: $\theta^\pm$!), its T-dual by $\tilde\theta$, and the superfield

1465: corresponding to the latter by $\tilde\Theta$. We also denote the

1466: direction lengthwise along the tube by $r$, with corresponding

1467: superfield $R$. The tachyon vertex operator at zero momentum is

1468: %

1469: \eqn\TsupvertII{T(X)=\hat T(R)\cos[w\tilde\Theta].}

1470: %

1471: \foot{Note that the winding tachyon is actually always complex,

1472: because we could add the winding and antiwinding modes with

1473: different phases

1474: $$ T e^{ i w \tilde \theta} + \bar T e^{- i w \tilde \theta} $$

1475: while preserving reality of the action. This phase in $T = e^{i

1476: \alpha} |T|$ changes the cosine in \TsupvertII\ to $ \cos \left[ w

1477: \tilde \Theta + \alpha\right] $. }Before proceeding, it is worth

1478: commenting further on the supersymmetry structure of the tachyon

1479: vertex.  In the case of \AdamsSV, the lowest-dimensional

1480: zero-momentum tachyon vertex operators were chiral primaries, and

1481: thus preserved ${\cal N}=2$ supersymmetry, leading to useful

1482: ${\cal N}=2$ linear sigma model treatments of the problem

1483: \refs{\VafaRA,\otherRG}.  In our case, the matter sector vertex

1484: operator \TsupvertII\ is not chiral primary except in the limit

1485: $L=0$.  In the appendix, we describe a linear sigma model which

1486: clarifies some physical features, but here we focus mostly

1487: directly on the physical theory on the world-sheet.

1488: 

1489: In the $\theta$ direction, this operator is the interaction Lagrangian of the ${\cal N}=1$ supersymmetric

1490: sine-Gordon theory, while the unperturbed action contains the kinetic term for $\theta$.  In components, the

1491: classical action is

1492: %

1493: \eqn\SSGaction{\int d^2\sigma\left( g_{\tilde\theta

1494: \tilde\theta}(r) \del_a\tilde\theta\del^a\tilde\theta +

1495: g_{rr}(r)\del_a r\del^a r + \hat T(r)

1496: \sin(\tilde\theta)\psi_{\tilde \theta}\tilde\psi_{\tilde\theta} +

1497: \hat T^2(r) \cos^2\tilde\theta\right)}

1498: %

1499: where $\hat T(r)$ 

1500: and $g_{ab}(r)$ vary slowly with $r$ near the point $r_0$ in the middle of the tube where

1501: $g_{\theta\theta}$ reaches its minimum value.

1502: 

1503: Because the system varies only mildly with changes of $r$, let us

1504: start by approximating $g$ and $T$ as independent of $r$, and

1505: study the flow of the ${\cal N}=1$ supersymmetric sine-Gordon

1506: theory in the $\tilde\Theta$ direction. Classically, all modes

1507: (elementary and solitonic) have masses from the tachyon vertex. In

1508: the quantum theory, there is compelling evidence that a mass gap

1509: persists. The model is integrable \FerraraJV, and the well-checked

1510: proposal for its exact factorized S-matrix

1511: \refs{\ShankarCM,\AhnUQ,\HollowoodEX,\BajnokDK}

1512: %based on an ansatz realizing a standard assumption of factorization

1513: %\ZamolodchikovXM,

1514: contains no massless singularities. Physically, this mass gap

1515: results from a generalization of the Kosterlitz-Thouless

1516: transition, in which vortex condensation destroys long range

1517: order. To clarify the important aspects of this phenomenon, we

1518: give a brief review of how it works in the bosonic case. This

1519: analysis can be directly applied to winding tachyons in the

1520: bosonic string if one is willing to tune to zero the omnipresent

1521: bulk tachyon.

1522: 

1523: % insert here bosonic background material.

1524: \subsec{A brief review of vortex-induced confinement}

1525: %\subsec{Gyre condensation and confinement}

1526: 

1527: 

1528: The XY model provides an excellent demonstration

1529: of the aphorism

1530: %[Mandelstam, 't Hooft, Fradkin-Susskind, earlier condensed matter refs,]

1531: that magnetic higgsing is electric confinement. Models in this

1532: universality class are two-dimensional and have a $U(1)$ symmetry,

1533: and therefore include a $U(1)$ Goldstone boson $\theta(z)$, of

1534: some radius $L$, in their low energy description. We will describe

1535: this phenomenon from the point of view of the line of fixed points

1536: parametrized by the radius of the circle in units of the self-dual

1537: radius\foot{In conventional string theory normalization $ S = {1

1538: \over 4 \pi l_s^2 } \int d^2 \sigma \del \theta \bar \del \theta

1539: $.}, $L_c = \sqrt 2 l_s$ (in the realization as a statistical

1540: mechanics problem, this parameter can be traded for the

1541: temperature). Single-valued winding and momentum modes of a boson

1542: at radius $L$ ($\theta \simeq \theta + 2 \pi L$ ) are created by

1543: the operators \eqn\windingmomentum{ \CO_{n,m} \equiv \exp{ i\left(

1544: { n\over L}  \theta + mL  \tilde \theta \right)} , ~~ m,n \in \IZ.

1545: } This operator has conformal dimension $

1546: %\eqn\conformaldimensionofmn{

1547: \Delta_{n,m} = \left({n \over L}\right)^2 +  ( m L ) ^2 .

1548: %}

1549: $

1550: Therefore, when $L < L_c $,

1551: there is a {\it relevant} winding operator $\CO_{0, \pm 1}$, with

1552: (chiral) conformal dimension

1553: $ \Delta_{0,\pm 1} = L^2  < 1 .$

1554: 

1555: The XY model provides an example of a system with `topological order' \KosterlitzXP: in the low-temperature

1556: phase, although there is no local order parameter, correlators fall off as a power law with distance. Above a

1557: critical temperature, 

1558: this order is destroyed by the condensation of {\it disorder} operators, $\CO_{0,1}, \CO_{0,

1559: -1}$ -- in the presence of the vertex operator for a winding mode, the $\theta$ field is constrained to have a

1560: discontinuity. A gas of such insertions destroys the long-range correlations.

1561: %\ifig\xyphasediagram{

1562: %RG trajectories in the xy model near the critical point.

1563: %}{\epsfxsize2.0in\epsfbox{phasediagram.eps}}

1564: 

1565: 

1566: The quickest way to see that the condensation of vortices induces

1567: a nonzero correlation length is to use the fact that this model

1568: can be fermionized \refs{\MandelstamHB, \ColemanBU}. The operator

1569: $\del \theta \bar \del\theta$ which changes the radius of $\theta$

1570: fermionizes to a Thirring four-fermion operator; at a radius other

1571: than the aptly named free-fermion radius, $L = L_{{\rm ff}}=\sqrt

1572: 2 L_c$, the fermions interact via a critical four-fermi term, but

1573: are massless. In this Thirring description, the neutral winding

1574: deformation $T \cos \tilde \theta$ is simply the fermion mass

1575: operator.  Starting close to the limit of a free massive fermion,

1576: this makes it clear that the IR limit that one reaches

1577: by perturbing by the corresponding deformation

1578: by $T$ is a massive theory (a result which also

1579: obtains for any value of the mass). 

1580: 

1581: It is worth noting that in the case of $\IC^q/\IZ_N$ orbifolds 

1582: \AdamsSV, the twisted tachyons at the orbifold singularities are

1583: winding modes when pulled away from the tip.  For large $N$

1584: this allows us to model the twisted tachyon vertex operator to

1585: a good approximation in this region as a SSG winding operator.

1586: Our analysis here of the corresponding mass gap thus provides

1587: a useful corroboration of the decay seen in \AdamsSV\ via 

1588: a more explicit analysis in the physical worldsheet theory.

1589: 

1590: 

1591: 

1592: 

1593: \subsec{Return to the superstring}

1594: 

1595: The claim that the supersymmetric sine-Gordon

1596: model flows to a theory with a mass gap

1597:  can also be

1598: supported

1599: %understood

1600: via a process of elimination as follows.  The operator \Top\ is

1601: relevant at small enough radius, and hence its condensation

1602: reduces the central charge in the $\theta$ direction from 3/2 to

1603: something smaller. The representation content of all unitary

1604: %nontrivial

1605: ${\cal N}=1$ SCFTs with $c<3/2$ have been classified \FQS, and

1606: all explicit examples 

1607: include relevant operators preserving supersymmetry.  As long as some relevant operators survive under the

1608: GSO projection applicable in our theory (and they do survive under the basic type II and type 0 GSO projections

1609: manifest in the Landau-Ginzburg description \KMS), the system has no possible nontrivial infrared stable

1610: endpoint.

1611: 

1612: \subsubsec{Dependence on $r$}

1613: 

1614: Now let us reinstate the $r$-dependence in our problem.  This can

1615: be done by expanding $g_{ab}$ and $T$ in a Fourier series in the

1616: region of the tube (equivalently in a derivative expansion with

1617: respect to $r$).  The zero momentum term yields the flow just

1618: described in the theta direction.  This indeed dominates as we

1619: flow toward the infrared, since including the factor

1620: $:e^{ik_rr}:$, the higher order terms involve higher dimension

1621: operators than the leading term in the expansion in $k_r$.

1622: 

1623: Thus the $\Theta$ direction is removed by the flow.  Once this has

1624: occurred, the remaining theory is subcritical and contains other

1625: bulk tachyons (relevant operators on the world-sheet).  These

1626: reduce the central charge further--in particular, in the $r$

1627: direction the tachyonic modes appearing for sufficiently small

1628: $k_r$ condense, again producing a theory of central charge $c<3/2$

1629: which must flow to a trivial fixed point.  These deformations also

1630: involve supersymmetric sine-Gordon interactions locally.

1631: 

1632: As we move away from the minimum $r_0$, we come to a value of $r$

1633: at which the tachyon gradient ${\del T \over \del r}$ becomes

1634: non-negligible. In this region there the capping-off process we

1635: have inferred from the mass gap must occur. In the appendix, we

1636: present a linear sigma model involving both the $R$ and $\Theta$

1637: directions which exhibits a flow from a connected tube to two

1638: disconnected copies of the complex plane. This confirms the result

1639: argued here directly in the physical theory.

1640: 

1641: Once this process has removed the $R,\Theta$ degrees of freedom in the middle of the tube, the tube has split as

1642: indicated in \before\ and \after.

1643: In one case a handle has been lost (changing the Euler character of the Riemann

1644: surface), and in the other the space has split into two disconnected parts (changing the most basic topological

1645: invariant: the number of connected components).

1646: 

1647: Because of the invariance of the Witten index of the world-sheet

1648: theory in this process, the change in Euler character in the

1649: Riemann surface must be accompanied by the generation of other

1650: isolated vacua.  This is evident from linear sigma model

1651: descriptions of the process, as we discuss in the

1652: appendix \AspinwallRB, and also emerges in our analysis of the

1653: %initial 

1654: time-dependent effects of the tachyon condensation in the

1655: next section.

1656: 

1657: %\break

1658: \subsubsec{Remnants}

1659: 

1660: Our arguments do not rule out the possibility that there is some

1661: nontrivial string scale fixed point, crucially involving both the

1662: $r$ and $\theta$ degrees of freedom, at which the

1663: handle-destruction process can abort. The theory would have to

1664: take some trajectory other than the one above, which led to the

1665: trivial IR fixed point in the region of the tachyon condensate. As

1666: we just discussed, the flow of the sine-Gordon model itself is

1667: well understood, so this possibility can be ruled out if the

1668: $\theta$ direction does not mix with others.

1669: %, and the dimensions decrease sequentially.

1670: However, it is logically possible that there are other

1671: trajectories leading to some nontrivial fixed point for the

1672: $r$-$\theta$ theory. It is only in the case that this fixed point

1673: is IR stable (including projection by the appropriate GSO action)

1674: that this could arise as a stable endpoint of the process for

1675: appropriate initial conditions. An alternative trajectory and

1676: endpoint would be fascinating in its own right, but in the absence

1677: of an example of such a fixed point, we will focus here on the

1678: generic flow to the trivial theory.

1679: 

1680: 

1681: 

1682: \subsec{c-curity Check and the Type 0 Tachyon}

1683: 

1684: The nonlinear sigma model (NLSM) on a {\it compact} Riemann

1685: surface satisfies the hypotheses of Zamolodchikov's $c$-theorem

1686: \refs{\ZamolodchikovGT, \Cappelli}, 

1687: if we define the theory with a cutoff. On the

1688: other hand, we claim the dimension of the target space remains two

1689: after perturbation by the winding tachyon, a relevant operator.

1690: %during the condensation of a winding tachyon.

1691: Does this imply that our conjectured RG flow from the NLSM on a

1692: compact Riemann surface with genus $h$ to one with genus $h-1$

1693: (plus extra states elsewhere) violates the c-theorem?  The answer

1694: is negative because the $c$-function is not simply the

1695: dimensionality of the target space in the presence of curvature.

1696: 

1697: We distinguish two cases, $h > 1$ (the case considered above) and

1698: $h=1$ (closely related to Type 0), and discuss them in turn.

1699: 

1700: 

1701: \subsubsec{$h>1$}

1702: 

1703: While the flow induced by tachyon condensation from genus $h$ to $h-1$ leaves the dimension of spacetime fixed

1704: at 2, it does not leave the $c$-function fixed: since the Riemann surface is not flat, the NLSM pertaining to

1705: the matter sector alone is not conformal. The effective $c$ for a Riemann surface target space, ignoring the

1706: $X^0$ direction, is greater than 2 and a monotonically decreasing function of scale, unless $h-1=1$. In

1707: particular, it is never 2 (or $\hat 2$) when the volume is finite for a higher genus Riemann surface.

1708: 

1709: This is analogous to the familiar case of positively curved target spaces, for which the effective $c$ is less

1710: than the dimensionality of the target space.  For example, in the NS5-brane solution there is an $S^3$ component

1711: stabilized by fluxes whose $\hat c$ is less than 3; there the spacelike linear dilaton makes up this deficit. In

1712: our case, time-dependence of the dilaton plays a similar role\foot{

1713: Unlike our case, however, 

1714: %there is however the important distinction that 

1715: the WZW model is actually conformal, 

1716: so its c-function is just the central charge.}

1717: 

1718: The $c$-function for the NLSM on a Riemann surface 

1719: of genus $h$

1720: and constant curvature $R \sim (2-2h)/V$, with volume

1721: $V = v V_0$

1722: (where $V_0$ is the volume of

1723: some fiducial metric, and $v$ is a dimensionless coupling)

1724: in perturbation theory around the free fixed point

1725: at $v \to \infty$ is

1726: \eqn\cfunction{

1727: c = c_0 + b \alpha' R  + \CO(\alpha'R)^2,

1728: }

1729: where $c_0$ is the central charge at the fixed point,

1730: and $b$ is a positive constant.

1731: This statement applies both for the bosonic and for the 

1732: supersymmetric NLSM,

1733: though hats must be sprinkled appropriately on the RHS in the latter case.

1734: This follows directly from equation (12)

1735: of \ZamolodchikovGT, 

1736: %Zamolodchikov's paper,

1737: which gives the c-function in a perturbation expansion

1738: about a weakly-coupled fixed point.

1739: %is

1740: %$$ c(g)

1741: %\equiv C(g) + 4 \beta^i H_i - 6 \beta^i \beta^j G_{ij} $$ where

1742: %$G_{ij}$ is the Zamolodchikov metric, $C$ is the leading residue

1743: %of the $TT$ OPE, and $H_i$ is the leading residue of OPE of $T$

1744: %with the operator whose beta function is $\beta^i$. The

1745: %contribution to $c(g)$ from $ \beta_V \beta_V G_{VV} $ (where $V$

1746: %is the volume operator of the Riemann surface) is the world-sheet

1747: %origin of the $2h-2$ in the dilaton potential \poten, and indeed

1748: %decreases during the flow.  Note that this result applies for any

1749: %$h$; indeed, we will use this below in discussing $h<1$.

1750: 

1751: There are several possible endstates of such tachyonic decays.  After decay from genus $h$ to $h-1$, fluxes and

1752: branes wrapping surviving cycles might stabilize the moduli of an $h>1$ endstate \SaltmanJH\ in regions of

1753: complex structure moduli space without small Scherk-Schwarz handles, terminating the perturbative decay (though

1754: non-perturbative effects may destabilize these moduli, sowing the seeds of further decay). In the absence of

1755: such stabilizing effects, the end of the process depends on the spin structure on all the cycles; if a torus

1756: remains with periodic boundary conditions, the RG flow can end with a flat metric on the resulting $h=1$

1757: surface, yielding $\hat c=2$. It can also end at infinite volume ($\hat c =2$) via the overall flow toward large

1758: volume at higher genus, if the unstable handles remain large and the system stays far from factorization limits.

1759: 

1760: However, as we discuss next, if there are remaining antiperiodic

1761: boundary conditions for $h=1$, the system can decay further to a

1762: genus 0 surface (a 2-sphere), for which $\hat c < 2$; this then

1763: evolves to a trivial IR fixed point.  Thus, in all cases, the

1764: $c$-theorem is respected.

1765: 

1766: 

1767: 

1768: \subsubsec{The Fate of the Scherk-Schwarz and Type 0 Tachyons: $h=1$}

1769: 

1770: Consider a torus with Scherk-Schwarz boundary conditions on the A cycle in a 

1771: regime in which the A cycle shrinks

1772: below string scale over some portion of the B cycle: condensing the 

1773: tachyonic wound string breaks this handle.

1774: This time, however, the resulting spacetime is topologically a sphere, and 

1775: the corresponding NLSM has a gap,

1776: flowing to $c=0$ in the IR and respecting the $c$-theorem in the strongest 

1777: way possible for a unitary theory.

1778: 

1779: This process is intimately related to the condensation of the Type

1780: 0 tachyon: taking the radius of the A cycle to zero and

1781: T-dualizing gives precisely Type 0, with the tower of winding

1782: tachyons dualizing to the momentum modes of the bulk tachyon on

1783: the non-compact dual circle. The decay of a Scherk-Schwarz torus to

1784: a sphere thus provides a convenient model in which to locally

1785: condense the Type 0 bulk tachyon, strongly suggesting that the

1786: fate of Type 0 under tachyon condensation is indeed a non-critical

1787: string theory, as has been widely conjectured.

1788: 

1789: The fate of the Type 0 tachyon has received much attention, and several 

1790: interesting alternative conjectures

1791: exist besides a flow to a non-critical string theory.  Perhaps the most 

1792: fascinating is the conjecture that Type

1793: 0 is connected to Type II via condensation of a mode of the type 0 theory 

1794: \refs{\CostaNW,\GutperleMB}\foot{The study of

1795: localized tachyon condensation in non-compact twisted circle 

1796: compactifications \refs{\otherRG,\GutperleBP} has also

1797: been interpreted as supporting this conjecture; however, there is an open 

1798: question regarding the order of limits

1799: involved.}. In light of the $c$-theorem, as well as the decrease of the 

1800: energy under tachyon condensation, this

1801: conjecture requires that the world-sheet description breaks down along the 

1802: flow if this connection between type

1803: 0 and type II is to be made by condensation of the type 0 tachyon (in 

1804: \refs{\CostaNW,\GutperleMB} 

1805: two modes, flux and

1806: tachyons, were discussed). Indeed, the authors of 

1807: \refs{\CostaNW,\GutperleMB} 

1808: posed 

1809: this conjecture in the context of

1810: strongly coupled fluxbranes. It would be interesting if indeed some 

1811: mechanism for connecting Scherk-Schwarz to

1812: type II on a circle exists; in our context this would provide a physical 

1813: connection between Riemann surfaces of

1814: different spin structures.

1815: 

1816: A condensate of the zero mode tachyon in flat space does in fact source the 

1817: dilaton (as it must, by our

1818: discussion of the $c$-theorem above). By condensing the bulk tachyon in a 

1819: localized fashion as described above,

1820: however, the effect of tachyon condensation on the string coupling may be 

1821: controlled during the perturbative,

1822: tachyonic decay from torus to sphere (by keeping the volume of the torus 

1823: large in string units), leading to an

1824: NLSM on a sphere with an initially mild time dependent evolution toward 

1825: smaller spheres. While the sphere

1826: eventually experiences a big crunch, so that literal endstate of this system 

1827: is necessarily nonperturbative, it

1828: appears rather far removed from type II on a torus.

1829: 

1830: \ifig\typezip{ Tachyon condensation drives a Scherk-Schwarz torus to a 

1831: sphere.

1832: }{\epsfxsize2.5in\epsfbox{type0.eps}}

1833: 

1834: 

1835: \subsubsec{$c$ and disconnection: $h\to h'\oplus [h-h']$}

1836: 

1837: Decays in factorization limits pose a slightly different question for the 

1838: $c$-theorem.  Considering only the set

1839: of degrees of freedom which localize on one component of the degenerate 

1840: Riemann surface and repeating the above

1841: arguments, $c$ certainly appears to be decreasing.  However, since we end 

1842: with two decoupled perturbative

1843: world-sheet theories, is this the correct measure? Stated differently, what 

1844: is the correct $c$-function on the

1845: endstate string theory with disconnected target space?

1846: 

1847: This problem is similar to the question of defining the c-function in a 

1848: field theory with a double well

1849: potential, working at energy scales below the barrier height.  (As we will 

1850: see in the next section, this is a

1851: good model for our system in fact.) Equivalently, we may consider the 

1852: scaling of the entropy of free fields as a

1853: function of temperature in such a double well configuration.  The c-function 

1854: decreases consistently in the full

1855: quantum field theory.  Of course, this does not have operational 

1856: observational meaning in the causally

1857: disconnected regions, but is a consistency check on the analysis of the 

1858: quantum field theory on the world-sheet.

1859: 

1860: 

1861: 

1862: 

1863: \newsec{Spacetime Physics of the Transition}

1864: 

1865: %In general, explicitly analyzing the time-dependent physics of

1866: %topology-changing processes and tachyon decays is more difficult

1867: % than finding the connection in configuration space

1868: %implied by the RG problem. 

1869: We start this section by analyzing the

1870: effects of the time-dependent perturbation on propagation

1871: of modes through the tube, which builds on the RG results just

1872: reviewed and applies to the process at the level of string

1873: perturbation theory. We then make some comments on the

1874: time-dependent process in the full theory, particularly

1875: emphasizing the role of particle and string production in both

1876: topology-changing and tachyon decay processes.

1877: 

1878: 

1879: 

1880: 

1881: \subsec{Transmission Barrier}

1882: 

1883: In this subsection, we use the nonlinear sigma model to study the

1884: effect of winding tachyon condensation on the propagation of the

1885: string world-sheet into the tube.

1886: 

1887: Consider a tube coordinatized by $\theta$ and $r$, with radius in the $\theta$ direction reaching a sub-string

1888: scale minimum at a point $r=r_0$ (or equivalently the radius of the T-dual variable $\tilde\theta$ reaching a

1889: maximum at $r_0$). We would like to consider the scattering of modes going into the tube from large negative $r$

1890: to see if they get transmitted through the tube after tachyon condensation. If our conjecture about the

1891: topology-changing process is correct, the waves should be completely reflected back to large negative $r$ by the

1892: end of the process.  We will check that the tachyon condensation introduces a barrier to penetration through the

1893: tube.

1894: 

1895: Including the time-dependence, the tachyon vertex operator is

1896: %

1897: \eqn\Ttime{\int d^2\sigma d^2\theta ~T(X)}

1898: %

1899: with $T(X)=\hat T(R) e^{\kappa X^0}\cos[w\tilde\Theta]$.

1900: %(ADD CONSTANTS)

1901: Because of the weak curvature, the $r$-dependence in the

1902: lowest-lying tachyon vertex operator is mild.  It decays to zero

1903: exponentially for values of $r$ outside the region where the

1904: winding mode is tachyonic and rises to a maximum at $r=r_0$, where

1905: the tube radius reaches its minimum. More generally, we can

1906: consider momentum modes in the $r$ direction, producing

1907: oscillation at a scale $k_r$ on top of the mild exponential

1908: falloff of the wavefunction.

1909: 

1910: In components, this leads to a bosonic potential in the

1911: world-sheet matter theory which is of the form

1912: %

1913: \eqn\Ttimebos{U[X]=\del_\mu T\del^\mu T= \biggl\{(-\kappa^2

1914: \cos^2[w\tilde\theta]+w^2\sin^2[w\tilde\theta]) \hat

1915: T(r)^2+{(\del_r\hat T)^2} \cos^2[w\tilde\theta]\biggr\}e^{2\kappa X^0} }

1916: %

1917: The kinetic terms include a mild $r$-dependent metric

1918: %

1919: \eqn\kinT{{\cal L}_{kin}\sim -\del_aX^0\del^a X^0+ g_{\tilde\theta\tilde\theta}

1920: (r)\del_a\tilde\theta\del^a\tilde\theta + g_{rr}\del_ar\del^ar} .

1921: %

1922: 

1923: Upon condensing the tachyon, the world-sheet theory becomes a

1924: nontrivial sigma model on a time-dependent target space, subject

1925: to the constraints of two-dimensional (1,1) supergravity.  The

1926: essential classical effects of supergravity in the world-sheet

1927: theory are as follows.  As discussed in \eg\ \DaCunhaFM, $X^0$,

1928: with its negative kinetic term, can be traded for the conformal

1929: factor of the $2d$ metric; its action arises from a conformal

1930: transformation of the form $\exp{(\gamma X^0)}$ in the

1931: gravitational action on the world-sheet.\foot{Given the time

1932: dependence in the ambient dilaton and volume from the tadpoles in

1933: \poten, the action should be supplemented by a linear dilaton

1934: term, \ie\ a conformal coupling proportional to $\int d^2\sigma

1935: R[e^{ \gamma X^0} \hat g] (2h-2)\Phi'(X^0)$ where $\Phi$ is the

1936: log of the effective string coupling and $\hat g$ is the fiducial

1937: metric from gauge fixing which we take to be flat.} The negative

1938: term in the potential energy \Ttimebos\ arises from supergravity

1939: in this sense--it came from the derivative of the superspace

1940: potential $T(X)$ by $X_0$ -- and is reminiscent of similar

1941: negative terms in the scalar potential arising in supergravity in

1942: higher dimensions.

1943: 

1944: The potential \Ttimebos\ exhibits a barrier to penetration into

1945: the region where the tachyon vertex operator has support. We

1946: consider a regime of parameters

1947: %

1948: \eqn\ineqk{\kappa^2<k_r^2\ll w^2 .}

1949: %

1950: The first inequality ensures that the momentum $k_r$ in the $r$

1951: direction is sufficient to overcompensate the negative $-\kappa^2

1952: \cos^2[w\tilde\theta] $ term for some range of $r$, so that these

1953: terms in combination with the leading term proportional to $w^2

1954: \sin^2[w\tilde\theta]$ provide a classical barrier for all values

1955: of $\tilde\theta$. The second inequality ensures that the leading

1956: interaction term for $\tilde\theta$ is the supersymmetric

1957: sine-Gordon interaction, a fact that will facilitate analysis of

1958: the quantum theory.

1959: 

1960: In particular, although the action exhibits a classical potential barrier for the string in the region where the

1961: tachyon exists, in general one needs to include quantum corrections in order to determine the net effect of the

1962: interaction terms on modes impinging on the tube.  We can use the known results for the RG flow in the model to

1963: analyze this, as follows.

1964: 

1965: First, for modes for which $r$ depends weakly on the world-sheet

1966: coordinates, \ie\ modes with small momentum in the $r$ direction

1967: and no oscillator excitations, this barrier indeed survives in the

1968: quantum theory. This is because the $r$-dependent factors in

1969: \Ttimebos\ behave to good approximation like coupling constants

1970: multiplying $\tilde\Theta$-dependent operators. Since we required

1971: \ineqk\ so that the $w^2 \sin^2[w\tilde\theta]$ term dominates,

1972: the long distance physics of the perturbed sigma model is the long

1973: distance physics of the supersymmetric sine-Gordon theory, which

1974: has a massive spectrum as we reviewed in the last section.  At low

1975: energies in the $r$ sector, the massive $\tilde\Theta$ excitations

1976: will not be generated. Putting the $\tilde\Theta$ sector in its

1977: vacuum then yields a surviving potential for $r$ in the action

1978: \Ttimebos, which blocks low-energy modes from getting through the

1979: tube (ignoring tunneling effects at very weak string coupling).

1980: Thus for low-energy modes, this leading-order contribution in the

1981: tachyon is sufficient to block penetration. Moreover, after

1982: integrating out the $\Theta$ sector the remaining interactions for

1983: $r$ are themselves tachyonic perturbations, whose effects grow in

1984: the infrared.

1985: 

1986: Secondly, one can use the same procedure for more general modes,

1987: as long as the masses in the $\tilde\Theta$ sector exceed the

1988: scale of fluctuation in the $r$ sector.  As one takes the limit of

1989: large tachyon vev in \Ttimebos, the $\tilde\Theta$ masses

1990: increase, as can be seen from the S-matrix results and direct

1991: calculations of soliton masses in

1992: \refs{\ShankarCM,\AhnUQ,\BajnokDK}. Thus as we increase the

1993: tachyon perturbation, the barrier affects more energetic modes. It

1994: is reasonable to expect that a limit exists in which all modes are

1995: blocked.

1996: 

1997: Note also that the $(\del_r T)^2$ term introduces at least one

1998: local minimum in $r$ (and generally more depending on $k_r$) in

1999: the middle of the tube in addition to the potential barrier. This

2000: may reflect the extra vacua predicted by the Witten index.

2001: 

2002: Finally, in order to define the world-sheet path integral, it may

2003: be useful to continue to Euclidean signature in the target space,

2004: setting $X^0\equiv i t_E$.  In doing this, we must consider

2005: $\cosh[\kappa X^0]=\cos[\kappa t_E] $ instead of the pure

2006: exponential, which is reminiscent of spacebrane constructions.

2007: This approach commits us to studying amplitudes in the Euclidean

2008: vacuum, which is a reasonable choice.  From this we obtain

2009: positive kinetic terms and a bosonic potential

2010: %

2011: \eqn\TEuclbos{\eqalign{U_{{\rm bosonic}}= & \hat T(r)^2\bigl(\kappa^2

2012: \sin^2[\kappa

2013: t_E]

2014: \cos^2[w\tilde\theta]+w^2\cos^2[\kappa t_E]\sin^2[w\tilde\theta]\bigr) \cr &+{(\del_r\hat T)^2}

2015: \cos^2[w\tilde\theta]\cos^2[\kappa t_E]\cr }}

2016: %

2017: Like \Ttimebos, this exhibits a potential barrier and a similar

2018: analysis to that above applies.

2019: 

2020: \subsec{Comments on the time-dependent process}

2021: 

2022: The renormalization group results just reviewed in \S2\ and

2023: applied in the last subsection provide strong evidence for

2024: connections between different spacetime topologies in the

2025: configuration space of string theory. Such a connection was

2026: established on the moduli space in Calabi-Yau target spaces in

2027: \refs{\classtop,\quantop}. 

2028: %In all cases, the time-dependent problem

2029: %requires further analysis, though in some cases the qualitative

2030: %features of the transition can be seen from basic energy

2031: %arguments. In general, 

2032: In the quantum theory at nonzero string coupling,

2033: the time-dependent dynamics of topology

2034: change processes and tachyon decay processes are very rich, with

2035: particle, string, and brane production effects playing a crucial

2036: role. Before describing this process in our case, let us briefly

2037: discuss the topology-changing processes studied previously as well

2038: as earlier examples of tachyon decay processes.

2039: 

2040: In highly supersymmetric situations where the moduli space has

2041: been shown to contain configurations of different topology,

2042: particle production effects play an important role in limiting the

2043: extent to which topology-changing processes occur dynamically. For

2044: example, in the conifold transition in ${\cal N}=2$ vacua of type

2045: II string theory \quantop, the low-energy effective description of

2046: the transition is as a transition between the Higgs and Coulomb

2047: branch of an ${\cal N}=2$ supersymmetric field theory \stromcon\

2048: (a realization which resolved a longstanding puzzle about the

2049: behavior of the string amplitudes near this point). Hence one

2050: might attempt to effect the change of topology dynamically by

2051: rolling the scalar fields toward the point where one branch joins

2052: another, for example rolling on the Coulomb branch toward the

2053: point where a hypermultiplet becomes massless. However, because

2054: the mass of the hypermultiplet is decreasing toward zero along

2055: this trajectory, the moduli space approximation breaks down badly,

2056: regardless of how small one takes the initial scalar field

2057: velocity; the nonadiabaticity parameter is of order $\dot

2058: m/m^2\sim \dot\phi/\phi^2$ where $m=\phi$ is the mass of the light

2059: hypermultiplet in terms of the canonically-normalized scalar field

2060: $\phi$. Light hypermultiplets are created, and their energy

2061: density back-reacts on the motion of the Coulomb branch scalar in

2062: a simple, calculable manner.  This results in rapid trapping of

2063: the scalar field at the intersection point between the two

2064: branches \trapping, rather than dynamical evolution toward a large

2065: radius space with different topology. In less symmetric

2066: situations, potential energy and cosmological evolution may

2067: dominate over these kinetic effects and yield dynamical

2068: topology-changing transitions.  In the case of flop transitions,

2069: there are no light states involved in the stringy resolution of

2070: the topology-changing transition, so these may occur dynamically

2071: even in cases with extended supersymmetry, perhaps as in

2072: \GreeneYB.

2073: 

2074: In earlier studies of tachyon decay processes, the time-dependent

2075: problem at nonzero string coupling involved the production of many

2076: excitations, including a gas of localized D-branes as well as

2077: strings and gravity modes \tachtime. In a technical analysis, this

2078: production can be avoided by considering a strict limit of zero

2079: coupling \refs{\sen,\AdamsSV}, but it is expected to occur in the

2080: physical problem, at nonzero coupling.

2081: 

2082: In our case, the dynamical process involves both effects just reviewed.  The beginning of the process consists

2083: of rolling in complex structure moduli space $\tau$ toward a point with a small handle or toward a factorization

2084: limit\foot{Note that as mentioned above there is a 1-loop contribution to the scalar potential driving the

2085: Scherk-Schwarz circle to shrink, so this is a natural process to consider in the absence of metastabilizing

2086: fluxes.}. Again, the masses of other modes depend on $\tau$, and those which become light (such as the winding

2087: modes $T$ around the tube) get produced. This traps $\tau$ in the region where $T$ has mass squared less than or

2088: equal to zero. Once in this region, the tachyon condenses, producing many excitations from the energy liberated

2089: in the process, as in \tachtime.

2090: 

2091: It is worth noting that the back reaction effects of the decay products may be mitigated by a more elaborate

2092: setting, as follows. We could embed our setup in the class of de Sitter models discussed in \SaltmanJH, while

2093: relaxing the flux contributions to allow for the small handle to develop.  If the resulting system after the

2094: tachyon condensation is in a basin of attraction of one of the de Sitter solutions \SaltmanJH, then the late

2095: time de Sitter expansion will dilute the decay products of the tachyon condensation process.

2096: 

2097: 

2098: 

2099: 

2100: \newsec{Charges, D-branes, and Fluxes}

2101: 

2102: In this section, we address the fate of charges in the system, and

2103: generalize our analysis to situations with fluxes turned on.

2104: 

2105: \subsec{Fundamental String Charges}

2106: 

2107: When a handle decays, the conserved charges associated with it are

2108: also lost in the process.    By condensing {\it one} winding mode

2109: to annihilate the handle, we have destroyed {\it two} cycles of

2110: the Riemann surface, namely the $a$-cycle $A$ on which the winding

2111: string is wrapped, and its intersection-dual $b$-cycle $B$. Each

2112: cycle corresponds to a conserved fundamental string winding

2113: charge, whose gauge field in the eight-dimensional effective

2114: theory we will refer to as $F_A$ and $F_B$ respectively.

2115: 

2116: The effective action for these field strengths is (by a simple application of the analysis in \S2\ of

2117: \SaltmanJH, starting from the 3-form field-strength $H$ and decomposing it in a basis of 1-forms on the surface

2118: to reduce to the 2-form field strengths $F$ pertaining to the gauge fields in the $8d$ effective theory)

2119: %

2120: \eqn\couplinggeneral{\int_{8d} \int_\Sigma H \wedge \ast H = \int_{8d} {\cal A}_{jk}(\tau) F^j\wedge_8 \ast_8

2121: F^k. }

2122: %

2123: Here $j,k$ index the quantum numbers on the $a$ and $b$ cycles of

2124: the Riemann surface, and ${\cal A}_{jk}(\tau)$ is the $2h\times

2125: 2h$ matrix

2126: %

2127: \eqn\Amatrix{{\cal A}(\tau) =  i \pmatrix{ 2 \tau (\tau-\bar\tau)^{-1} \bar\tau &

2128:     -(\tau + \bar\tau)(\tau-\bar\tau)^{-1} \cr

2129:                 -(\tau-\bar\tau)^{-1} (\tau + \bar\tau) & 2 (\tau-\bar\tau)^{-1}}}

2130: where $\tau$ is the $h \times h$ period matrix of the Riemann surface. This coupling reduces at genus 1 to

2131: %

2132: \eqn\couplingfunction{\int d^8 x\sqrt{g} {1\over\tau_2^2} | F_A + \tau F_B|^2 }

2133: %

2134: where $A$ and $B$ are the two one-cycles of the surface. The

2135: formula \couplingfunction\ is also a good approximation at higher

2136: genus in situations where the handles not participating in the

2137: process are nearly decoupled from the $a$-cycle $A$ on which the

2138: winding tachyon is wrapped and its dual $b$-cycle $B$. As we

2139: evolve well into the tachyonic regime, $\tau_2$ becomes small, so

2140: $F_A$ is weakly coupled while $F_B$ becomes strongly coupled.

2141: 

2142: The condensing tachyon, since it is a string wound around the

2143: cycle $A$, is charged under the $A$-cycle gauge field. This gauge

2144: field is therefore massed up by the Higgs mechanism. The fate of

2145: the gauge field coming from the $B$-cycle is more mysterious --

2146: the strong coupling suggests that it confines classically

2147: \KogutSN.  This puzzle and its resolution have been seen in the

2148: open string tachyon problem \openconfine as well.

2149: \ifig\breakage{

2150: A pictorial representation of the fate

2151: of the B-cycle winding charge.

2152: If there are $N$ strings on cycle $A$,

2153: with indefinite $N$, then

2154: a string on cycle $B$ can break.

2155: %At large $N$, the tilt required

2156: %to wrap this configuration

2157: %around a circle becomes smaller.

2158: }{\epsfxsize2.0in\epsfbox{joining.eps}}

2159: 

2160: As another probe of the fate of the $B$-cycle gauge group, consider, before the condensation, a state with a

2161: string wound on the $B$-cycle at some point $x$ in the remaining 7 spatial dimensions and a string wound with

2162: the opposite orientation at a point $y$. From the 8-dimensional point of view, this is a charge-anticharge pair.

2163: Since these charges are codimension seven, they cannot affect the condensation process globally. Nevertheless,

2164: as the $A$-cycle tachyon condenses, and the effective $B$-cycle coupling grows stronger, the lines of $F_B$-flux

2165: running from one charge to the other will collapse into a tensionful flux string \KogutSN. These lines of flux

2166: may temporarily prevent the handle from collapsing. However, the charges will experience a strong attraction and

2167: annihilate as soon as possible, allowing the handle to decay subsequently.

2168: 

2169: 

2170: \ifig\confimenentfigure{ A geometric realization of classical confinement. Each point in this picture represents

2171: a circle homotopic to the $A$-cycle. $y$ is the direction in the remaining 7 spatial dimensions in which the

2172: charges are separated. The dashed lines represent $H_{{\rm NS}}$-flux lines between the fundamental string

2173: sources. }{\epsfxsize3.0in\epsfbox{cruderendering.eps}}

2174: 

2175: 

2176: On a general Riemann surface, there are other cycles not directly

2177: participating in the decay process occurring on the $A_1-B_1$

2178: handle.  For those handles with antiperiodic boundary conditions

2179: for fermions on one or both cycles, one might wonder if their

2180: decays are induced by the decay process occurring on the first

2181: handle. In the full time-dependent system, in the absence of

2182: fluxes or other ingredients to metastabilize, the allowed decays

2183: will all occur in time. At the classical level however, one can

2184: say more.  The strings wound around these other cycles do not

2185: couple linearly to the tachyon wrapped around the $A_1$-cycle, so

2186: at the classical level their condensation (and concomitant Higgs

2187: mechanism) will not be induced.  Their gauge couplings are also

2188: not getting strong as in \couplingfunction, so we do not expect

2189: confinement to set in. From the world-sheet point of view, the

2190: global symmetries associated with the winding currents for these

2191: other cycles should be unaffected by the destruction of the

2192: $A_1-B_1$ handle.

2193: 

2194: \subsec{D-branes}

2195: 

2196: In type II theories, D-brane charges also disappear in the tachyon

2197: decay process.  These charges are analagous to the $B$-cycle

2198: fundamental string charge described in the previous subsection:

2199: the gauge groups in question are not Higgsed by the fundamental

2200: winding string condensate on the $A$-cycle, but could be Higgsed

2201: by a condensate of a heavy non-perturbative object later in the

2202: time-dependent process.

2203: 

2204: In previous studies of tachyon condensation processes, D-brane

2205: probes yielded useful independent information \AdamsSV; in the

2206: present context this does not happen because the initial flow

2207: occurs when the geometry is not singular, as it was in \AdamsSV.

2208: 

2209: \subsec{Flux Backgrounds}

2210: 

2211: We can also consider this process in a situation with mild fluxes

2212: on the Riemann surface.  As long as these are sufficiently weak,

2213: the region of complex structure moduli space with small handles

2214: remains accessible.  Suppose, for example, that we have 1-form

2215: flux $F$ on the compact cycle $\gamma_\theta$ of the tube (where

2216: $F$ could come from a $p$-form field strength in higher

2217: dimensions, with one leg on the Riemann surface in question):

2218: %

2219: \eqn\tubeflux{\int_{\gamma_\theta} F = Q}

2220: %

2221: When the tube caps off in the tachyon decay process, the integral

2222: \tubeflux\ still holds in the remaining region, so a source must

2223: appear. In particular, a brane charged under $F$ must appear at

2224: the tip of one cap while the corresponding antibrane appears at

2225: the tip of the other.  Hence, in the presence of flux, this

2226: process produces brane-antibrane pairs as well as change of

2227: topology.

2228: 

2229: 

2230: \newsec{Discussion}

2231: 

2232: We have shown that a tachyon decay process in which a localized

2233: Scherk-Schwarz tachyon decays away the central charge in its

2234: region leads to topology change processes in which handles decay

2235: away and Riemann surfaces break into separate components. Both are

2236: expected from spacetime energetics \poten\ and follow from

2237: world-sheet renormalization group arguments.  The effect is likely

2238: to be very efficient dynamically as it is driven by a tachyonic

2239: mode rather than a massless modulus field.

2240: 

2241: It is interesting to consider the possibility of a dual gauge

2242: theory describing the local physics of the tube, perhaps via some

2243: embedding of the system in AdS/CFT. In such a situation, we would

2244: expect the winding string to be dual to a Wilson loop operator.

2245: Condensation of this Wilson loop would then imply that the cycle

2246: on which the string is wrapped is contractible, since it can be

2247: filled in with a string world-sheet. Considering the winding

2248: circle as Euclidean time, this is the same argument that shows

2249: that a vev for the Polyakov loop implies that the dual geometry

2250: contains a black hole horizon \AharonySX. This agrees with our

2251: 'capping-off' picture. It would be interesting to understand such

2252: a relationship in detail.

2253: 

2254: 

2255: 

2256: 

2257: The change in the number of components of the Riemann surface,

2258: suggesting formation of baby universes perturbatively via stringy

2259: physics, is particularly striking.

2260: %There has been much discussion of this possibility

2261: %of baby universe formation in quantum cosmology, but generally in the context of difficult to define

2262: %manipulations of the Euclidean gravitational path integral.  In our case, given the tachyon decay process we

2263: %described, the effect occurs already classically in string theory.\foot{

2264: %\LindeSK

2265: Previous work on baby universe creation involves

2266: quantum-gravitational tunneling computations

2267: %that are

2268: %difficult to incorporate into a complete

2269: %well-defined framework -- for example, the sign of the action appearing

2270: %in the Euclidean quantum gravity computations remains undetermined

2271: (see \eg\ \LindeSK\ and references therein). The separation of

2272: components we see here from a {\it perturbative} instability is

2273: quite different,

2274: %much more direct, 

2275: and does not depend on subtleties involved in

2276: Euclidean quantum gravity computations.\foot{ At {\it large}

2277: radius for the Scherk-Schwarz tube, the Witten ``bubble of

2278: nothing" \WittenGJ\ may play a similar role. However, this

2279: solution is subdominant to the perturbative dynamics, including

2280: higher order corrections driving the complex moduli toward the

2281: tachyonic regime in situations without stabilizing fluxes.  In

2282: situations with stabilizing fluxes, the decay of fluxes by brane

2283: nucleation appears to be the leading decay mode so it is not clear

2284: if and when the bubble of nothing dominates.} As such, our results

2285: provide simpler motivation for the need to make sense of the

2286: observables in situations where space separates into distinct

2287: components. 

2288: (See for example \wormholerefs, though here the baby

2289: universe is not necessarily Planck sized as in \wormholerefs; the

2290: `dust' vacua may however play this role.).\foot{Of course, other

2291: solutions exist with causally disconnected regions, such as the

2292: connected discretuum of metastable de Sitter solutions in string

2293: theory, whose resolution may be similar.}

2294: 

2295: Given the dramatic nature of this effect, let us

2296: note two conceivable loopholes that could evade this conclusion (neither of

2297: which

2298: appears plausible):  (1)  perhaps the effect takes infinite time according

2299: to the appropriate

2300: observer, or (2) perhaps there is some very attractive endpoint for the RG

2301: flow in the $\tilde\Theta$

2302: sector which is a nontrivial IR stable fixed point rather than the trivial

2303: IR limit

2304: with a mass gap that emerges from

2305: all analyses of the model to date.  Realizing loophole (2) would

2306: require some extremely surprising RG behavior,

2307: %require the proposed exact

2308: %S-matrices for the SSG model to be incorrect,

2309: while (1) would require some mechanism for slowing down the

2310: transfer of energy from potential energy \poten\ to decay products

2311: via the channel we have identified.  Some of our arguments (such

2312: as the world-sheet $c$-theorem) depend on a small string coupling,

2313: something we can tune to obtain control 

2314: as in \refs{\AdamsSV,\sen}.  The

2315: basic spacetime energetics of the process (in particular the

2316: reduction of the 8d potential energy by the condensation of $T$)

2317: appears robust in the presence of interactions, though the

2318: possibility of a `remnant' alternative could in principle arise

2319: via as yet unknown strong coupling effects.

2320: 

2321: 

2322: In order for a truly separated baby universe to form, the process

2323: we described must take place everywhere in the remaining eight

2324: dimensions. It may be more natural in early universe cosmology for

2325: the complex structure moduli to be different in different regions,

2326: and hence only experience a given topology-changing process

2327: locally. However, if it happens the same way within a given

2328: horizon volume, the effect is the same as far as any given

2329: observer is concerned.

2330: 

2331: In either case, the resulting dynamics of the effective 8d effective field theory appears remarkable: the tachyon

2332: perturbation must 

2333: %effectively 

2334: be such that the spectrum decouples into two sets of observables whose interactions

2335: vanish - including the appearance of a new spin-2 state whose longitudinal mode must decouple at the endstate.

2336: Various authors have considered such processes within the context of effective field theory

2337: (see \eg\ \ArkaniHamedSP); it would be fascinating to understand this process in the stringy effective field theory.

2338: 

2339: 

2340: The tachyonic topology-changing dynamics we identified here also

2341: suggest the possibility of 

2342: {\it topological} topological defects.  In

2343: particular, the type II theory in the backgrounds we have focused

2344: on has a complex tachyon, which could yield cosmic strings whose

2345: cores have different topology in the internal dimensions.

2346: 

2347: 

2348: In addition to sharpening the conceptual puzzle of how to treat

2349: causally disconnected regions in gravity, these considerations

2350: point to new scenarios for string cosmology.  For example, the

2351: complex structure modulus, with appropriately tuned potential,

2352: combined with the tachyon, provide an analogue of hybrid

2353: inflation, similar to brane-antibrane inflation but in the closed

2354: string sector.  More generally, it will be interesting to

2355: investigate whether the existence of one-cycles in the early

2356: universe (a generic feature given stringy energy densities at

2357: early times) and their subsequent decay may produce interesting

2358: signatures.  In the context of low-scale gravity scenarios it is

2359: also interesting to ask about signatures of decay (and formation)

2360: of handles.

2361: 

2362: 

2363: One could also consider a configuration in which a sphere bubbles off

2364: of a surface of any genus, starting from a configuration with a locally

2365: pinched neck.  As in the factorization case, this automatically

2366: has antiperiodic boundary conditions for fermions.  However, the starting

2367: pinched configuration required is not a stable endpoint of the Ricci

2368: flow in this case, and hence 

2369: this pinched configuration is separated from 

2370: the constant-curvature configurations by a potential barrier.

2371: %generically this process would require tunneling through a barrier.

2372: It may be relevant in the early universe; high temperature effects could

2373: allow the system to fluctuate over the barrier.

2374: 

2375: 

2376: 

2377: Finally, 

2378: on rather more speculative ground, 

2379: the fact that these

2380: topology-changing processes involve the removal (or appearance, in

2381: principle) of handles 

2382: though tachyon condensation

2383: %when the mass squared of some mode passes through zero 

2384: suggests a host of intriguing generalizations.  For

2385: example, since string theory is not only a theory of

2386: strings, it is natural to wonder whether a $p$-brane wrapping a

2387: sufficiently small 

2388: non-supersymmetric

2389: %non-spin$_c$ 

2390: cycle might go tachyonic 

2391: (as suggested in \openconfine),

2392: and

2393: mediate topology change through the removal of the shrinking

2394: $p+1$-handle.  

2395: This would provide a lovely link between handlebody

2396: theory and string theory.  

2397: %Along similar lines, since the mass

2398: %squared of the tachyon passing through zero is the signal to

2399: %remove a handle, one might wonder whether the tachyon mass (or,

2400: %more generally, the full non-perturbative spectrum) provides a

2401: %morse\foot{Vafa has also noted the possible role of the

2402: %$c$-function as a morse function on the space of conformal field

2403: %theories \VafaUE, and thus on the configuration space of string

2404: %field theory.} function for the target space, with dynamical time

2405: %playing the role of a height variable.  Since renormalization

2406: %group flow in the associated NLSM is simple Ricci flow, this is

2407: %somewhat evocative of 

2408: %the recent proof via Ricci flow of the

2409: %geometrization conjecture.

2410: Mathematicians \ricciflow\ have studied RG 

2411: flows in target spaces of various dimensions

2412: % higher dimensional spaces 

2413: at the level of the leading order in $\alpha'$.  

2414: They have been motivated to

2415: perform surgery manually, `capping off' their spaces 

2416: at a 

2417: %some for-them-arbitary 

2418: cutoff length scale in a way very reminiscent 

2419: of the process we have described.  It would

2420: be very interesting to see if string or brane theoretic tachyons could 

2421: effect this surgery naturally. 

2422: 

2423: 

2424: 

2425: \bigskip

2426: \centerline{\bf{Acknowledgements}}

2427: 

2428: We would like to thank Mohsen~Alishahiha, Nima~Arkani-Hamed, Savas~Dimopoulos, Steve~Giddings, Matthew~Headrick, Albion~Lawrence,

2429: Alex~Maloney, Shahin~Sheikh-Jabbari, Larus~Thorlacius, David Tong and especially Shamit~Kachru and Steve~Shenker

2430: for discussions. A.A. is supported by a Junior Fellowship from the Harvard Society of Fellows. The rest of the

2431: authors are supported in part by the DOE under contract DE-AC03-76SF00515 and by the NSF under contract 9870115.

2432: E.S. thanks the IPM and ISS String School in Iran and the PIMS and PNW String Seminar in Vancouver for

2433: hospitality during parts of this work.

2434: 

2435: \appendix{A}{Linear Sigma Model description}

2436: 

2437: In this appendix, we describe the effects of the winding tachyon using linear sigma models.  That is, we

2438: consider a field theory model whose low-energy limit describes the 

2439: physical problem of interest, following

2440: a method developed in \WittenYC\ in Calabi-Yau compactifications 

2441: and applied 

2442: %beautifully 

2443: to previous

2444: tachyon decay problems \AdamsSV\ in, for example \refs{\VafaRA,\chicago}.

2445: % using methods developed in, for example \HoriKT\MorrisonYH. 

2446: We focus on a noncompact region of the Riemann surface containing a single handle, and

2447: construct a linear model with a low-energy configuration space that flows from a one-sheeted hyperboloid to a

2448: two-sheeted one (see the figure below), exhibiting the local topology-changing process of interest via a description in

2449: which the relevant operator corresponding to the tachyon is realized in a simple way. This method will apply to

2450: the renormalization group flow aspect of the problem, discussed in \S2. It will also make manifest the extra

2451: vacua required by conservation of the Witten index, and will clarify the fate of the GSO projection in the flow.

2452: 

2453: Relative to the case \refs{\AdamsSV,\VafaRA,\chicago}, there is a

2454: complication regarding the level of supersymmetry in the problem.

2455: As we discussed below \TsupvertII, the relevant operator arising

2456: from the tachyon vertex operator does not preserve ${\cal N}=2$

2457: supersymmetry unless the smallest circle in the tube is zero size.

2458: Hence, in order to describe the process starting from a smoother

2459: space, we ultimately consider a model containing soft ${\cal N}=1$

2460: deformations away from a $(2,2)$ supersymmetric gauged linear

2461: sigma model (GLSM).  We analyze this model using its

2462: superrenormalizability and its reduction to $(2,2)$ in an

2463: appropriate limit. The deep IR regime of the model contains

2464: nontrivial corrections to potential energy terms, but we find we

2465: can tune these effects away to sufficient accuracy for our

2466: purposes. It would be interesting to pursue the physics of this

2467: model in more detail than we have space for here.

2468: 

2469: Before moving to our ultimate construction, it will be useful to

2470: keep in mind the following simple ${\cal N}=1$ model which

2471: contains some of the essential features that will be encoded in

2472: our larger model.

2473: %

2474: \doublefig\caps{ The vacuum manifold a) at $\rho < 0$; b) at $\rho

2475: > 0$. }{\epsfxsize2.0in\epsfbox{tube.eps}}

2476: {\epsfxsize2.0in\epsfbox{caps.eps}}

2477: %

2478: The desired geometry looks like the locus

2479: %

2480: \eqn\desired{

2481:  w^2 - |\phi|^2 = \rho  , ~~  w \in \IR, \phi \in \IC. }

2482: %

2483: As $\rho$ goes from negative to positive, the manifold changes

2484: from a one-sheeted hyperboloid to a two-sheeted one, as in 

2485: the figure.

2486: A very simple (1,1) model with this vacuum manifold in the

2487: low-energy limit has real multiplets $w, \sigma$ and a complex

2488: multiplet $\phi$, with potential terms

2489: $$

2490: \int d \theta_+ d \theta_-

2491: %q_+ q_-

2492: ~ \sigma \left(  w^2 - |\phi|^2 - \rho  \right) ;

2493: $$

2494: here $\theta_\pm$ are left- and right-moving coordinates on a $(1,1)$ superspace.  For our purposes we will need

2495: a model (to be elaborated below) for which this geometry arises at low energies, including quantum corrections.

2496: Quantum corrections also determine the direction of flow of $\rho$, and we will find that our model predicts the

2497: appropriate direction, with $\rho$ flowing toward more positive values.

2498: %The width of the throat is $\rho$...

2499: Before moving on, we can use this model (in which many of the NLSM

2500: couplings are hidden in $\rho$) to see that the coupling $\rho$

2501: multiplies a winding operator, as follows.

2502: 

2503: %%I pedandically removed "chemical potential" because we are doing QM not SM--ES %%

2504: 

2505: Consider T-dualizing in the direction of ${\rm Arg}(\phi) =

2506: \theta$, which goes around the hyperboloid. This is accomplished

2507: \refs{\BuscherQJ, \RocekPS} by introducing a gauge field $F = dA$

2508: under whose gauge transformation $\theta$ shifts, and whose theta

2509: angle is an extra field $\tilde \theta$,

2510: \eqn\buscheraction{ S =  \int \left( (\del \theta + A)^2 +

2511:  \tilde \theta F +{1\over e^2}F^2 \right) .}

2512: This dynamical theta

2513: angle $\tilde \theta$ acquires kinetic terms upon integrating out

2514: the gauge fields, and becomes the T-dual coordinate.\foot{

2515: The kinetic term we have included in \buscheraction\ 

2516: does not affect the analysis of 

2517: the T-duality below the high scale $e$.}

2518: With this in

2519: mind, the coupling $\rho$, which in (2,2) models is the

2520: Fayet-Iliopoulos (FI) parameter, turns on the winding tachyon

2521: because it controls the action of a vortex of the gauge field and

2522: scalars. In a vortex configuration, the support of $F$ is

2523: delta-function localized near the centers of the vortices. So, in

2524: a background of $\tilde \theta$, a vortex at a point $z$ on the

2525: world-sheet contributes \eqn\vortexcontribution{ e^{ - S_{{\rm

2526: cl}}(\rho) } e^{ i \tilde \theta (z)}, } where $S_{{\rm

2527: cl}}(\rho)$ is the $\rho$-dependent vortex action.

2528: \vortexcontribution\ is an insertion of the winding vertex

2529: operator with a $\rho$-dependent chemical potential, $T =

2530: T(\rho)$. The sum over such vortices automatically reproduces the

2531: coulomb gas expansion in the winding deformation.

2532: 

2533: 

2534: This is elegantly packaged in the (2,2) {\it gauged} linear sigma

2535: model (GLSM). Here the gauge field is present from the beginning,

2536: and its vortices can be BPS. In the $(2,2)$ case

2537: \refs{\MorrisonYH, \HoriKT}, the T-dual direction couples as a

2538: dynamical theta angle to the GLSM gauge field. The coupling of the

2539: winding operator is complexified to $ e^{ \rho + i \theta_G}$,

2540: where $\theta_G$ is the theta-angle of the GLSM gauge field. This

2541: claim is closely related to the fact that in the Calabi-Yau case

2542: \WittenYC\ the vortex sum reproduces the sum over world-sheet

2543: instantons of the nonlinear sigma model.

2544: 

2545: It is for this reason that we obtain our hyperboloid geometry from

2546: an equation of the form \desired\ following from a $(2,2)$ D-term.

2547: Another possible starting point would be to obtain the initial

2548: hyperboloid from a (2,2) F term via a superpotential $ W = P (\phi

2549: \eta - \mu ) $ where $\mu$ is a constant, and $P, \phi, \eta$ are

2550: chiral fields. The vacuum manifold of this model is the 'deformed'

2551: one-sheeted hyperboloid $\phi \eta = \mu$. However, the tachyon is

2552: still a solitonic winding operator so the linear sigma model

2553: description would not provide a simple description of the flow.

2554: 

2555: %%Morrison-Plesser-Hori-Vafa duality...

2556: 

2557: Indeed, there is a good precedent for

2558: the FI parameter playing the role

2559: of the vev of a localized (winding) tachyon

2560: \VafaRA.

2561: %Develop the analogy between orbifold tachyons

2562: %and winding tachyons before:

2563: In these $\IC^n/\IZ_N$ models, the twisted tachyons arise from

2564: strings which stretch from a point to its image under the

2565: orbifold; if these strings from the $k$th twisted sector are

2566: pulled away from the tip of the cone, they are forced to wind

2567: around it $k$ times. In the limit $N\to \infty$, this $k$ becomes

2568: an integer-valued winding charge. The models we are studying,

2569: therefore, can be considered to promote this `fractional'

2570: localized winding tachyon to an honest winding tachyon. It is

2571: therefore natural that we should find models which realize a

2572: condensation whose behavior is not dissimilar to that of \VafaRA.

2573: 

2574: 

2575: %shrinking of the blowup of a smooth point.

2576: 

2577: 

2578: 

2579: \subsec{Stringy (2,2) model}

2580: 

2581: 

2582: \def\etap{\eta_+}

2583: \def\phip{\phi_+}

2584: %\def\phitwo{\phi_{-2}}

2585: %\def\etatwo{\eta_{-2}}

2586: %\def\etap{\eta_2}

2587: %\def\phip{\eta_1}

2588: \def\phitwo{\phi_{-2}}

2589: \def\etatwo{P_{-2}}

2590: 

2591: With this motivation, we will start by considering the following

2592: (2,2) gauged linear sigma model, following the conventions in

2593: \WittenYC. Consider the GLSM with one $U(1)$ under which the

2594: D-term is

2595: $$ D = |\phip|^2 + |\etap|^2 - 2 |\phitwo|^2 - 2 |\etatwo|^2 - \rho .$$

2596: The subscripts of the fields label their

2597: charges under this $U(1)$.

2598: We will add the superpotential

2599: $$ W = m \etatwo \phip \etap .$$

2600: Here $m$ is a scale which sets the masses of the fluctuations

2601: transverse to the vacuum manifold of the F-terms, and for

2602: convenience can be considered equal to the scale $e$ of the gauge

2603: coupling. The coupling $\rho$ flows logarithmically to $+ \infty$

2604: in the IR \WittenYC\ because the sum of the charges is

2605: $$ Q_{{\rm T}}  \equiv \sum_i Q_i = - 2 .$$

2606: There is an $SU(2)$ symmetry acting on

2607: $ ( \phip, \etap)$ as a doublet.

2608: The F-term equations are

2609: \eqn\ftermsofnubbin{

2610: 0 = F_{\etatwo} = m \phip \etap, ~~~

2611: 0 = F_{\phip} = m \etatwo \etap, ~~~

2612: 0 = F_{\etap} = m \etatwo \phip.

2613: }

2614: 

2615: We now quickly analyze the vacuum manifold in the two

2616: semiclassical regions where $ |\rho| $ is large. This will

2617: determine the target space of the effective nonlinear sigma model

2618: below the scales $m, e$. When $\rho \to - \infty$, either

2619: $\phitwo$ or $\etatwo$ must be nonzero on the vacuum manifold. If

2620: $\etatwo =0$, the remaining F-terms give $\phip \etap = 0$.  This

2621: is an apparently singular hyperboloid

2622: with two branches $\phip\neq 0, \etap\neq 0$. If $\etatwo \neq 0 $, both

2623: $\phip$ and $\etap$ must vanish. If both $\phip$ and $\etap$

2624: vanish, $\etatwo, \phitwo$ are unconstrained and parameterize a

2625: $\IP^1$.

2626: %This is the 'nubbin' prefigured in the title of this subsection.

2627: There is a residual $\IZ_2$ of the gauge group which is unbroken on this branch. The three branches intersect at

2628: $\phip=\etap=\etatwo=0$. Note that for generic $\rho$ the model is not actually singular since the extra branch

2629: is compact. Further, as long as $\theta_G \neq 0$, the model is nonsingular for any $\rho$.  

2630: 

2631: It may be interesting to consider in more detail the singularity

2632: at $\rho=0=\theta_G$.  In Calabi-Yau models \WittenYC, the analogous

2633: singularity reflects the presence of light non-perturbative objects 

2634: in this limit \stromcon.  In our case, it may reflect light

2635: wrapped D-strings, whose condensation as discussed in \S4.2\ would provide

2636: a natural mechanism for confinement of RR charges in the process

2637: along the lines of \openconfine.  

2638: 

2639: When $\rho \to + \infty$, $\phip$ and $\etap$ cannot

2640: simultaneously vanish.  The F-terms therefore set $\etatwo =0$.

2641: Further, the two branches of $\phip \etap =0$ are disconnected. On

2642: the $\phip \neq 0$ branch, the D-term equation is of the form

2643: \eqn\Dtermoncap{ |\phip|^2 - 2 |\phitwo|^2 = \rho . } The $U(1)$

2644: gauge symmetry can be fixed to set $\phip$ real and positive. This

2645: is a single `cap', with a negative curvature related to $\rho$. An

2646: identical analysis applies to the branch with $\etap$ nonzero,

2647: which gives a second `cap' disconnected from the first.

2648: 

2649: In this model, therefore, the physics of the condensed phase is exactly as desired, and depicted in \caps. The

2650: starting point, however, is somewhat obscured by string-scale features of the geometry. In order to clarify that

2651: the starting point is indeed in the universality class of the nonlinear sigma model on a narrow handle, in the

2652: next subsection we will describe a deformation of the model which removes the extra $\IP^1$ in the geometry of

2653: the $(2,2)$ model, and visibly smooths the connecting region between the $\phip$ and $\etap$ throats, but

2654: preserves the RG trajectory of $\rho$.

2655: 

2656: 

2657: 

2658: 

2659: 

2660: \subsec{The (1,1) deformation}

2661: 

2662: %Inspired by the proof that nonzero $\mu$ and

2663: %(2,2)-preserving tachyon were mutually inconsistent,

2664: %we found a (1,1) deformation of the improved nubbin model

2665: %which seems to do what we want.

2666: Recall that the finite-size throat is incompatible with a winding

2667: tachyon preserving $(2,2)$ supersymmetry. Armed with this

2668: information, we seek out a deformation which preserves only

2669: $(1,1)$ supersymmetry.  We will set it up so that the resulting

2670: geometry of the low-energy vacuum manifold is determined by an

2671: equation of exactly the form \desired\ discussed above.

2672: 

2673: Breaking $(2,2)$

2674: to $(1,1)$ means that the chiral multiplets

2675: each split into two real multiplets, and the

2676: superpotential term is now a

2677: full superspace integral of a real function

2678: (which preserves $(2,2)$ if it's the

2679: real part of a holomorphic function of

2680: the formerly-chiral combinations).

2681: This allows many more gauge invariant monomials

2682: because $(1,1)$-preserving operators

2683: can depend on both $\phi$ and $\bar \phi$.

2684: 

2685: We will modify the (2,2) model

2686: of the previous subsection

2687:  by adding the small 'real superpotential' term

2688: %

2689: $$ \delta w = -\mu ( \etatwo \bar \etatwo +

2690: \phip \bar \etap + \bar \phip \etap  ).$$

2691: %

2692: By this we mean that we add to the lagrangian

2693: %

2694: \eqn\breaksNequalstwo{ L = q_+ q_- \left( \delta w \right) }

2695: %

2696: where $ q_\pm \equiv (1 / \sqrt{2}) ( Q_\pm + \bar Q_\pm )$ are

2697: the two real supercharges we are going to preserve. For $(1,1)$

2698: supersymmetry, the superpotential will include everything other

2699: than the kinetic terms. Noting that

2700: $$ \bar Q W = 0 \Longrightarrow

2701: Q_+ Q_- W + \bar Q_+ \bar Q_- \bar W =

2702: q_+ q_- ( W + \bar W), $$

2703: the total real superpotential

2704: (not including D-terms yet) is:

2705: $$ w = m \left( \etatwo \phip \etap + {\overline{ \etatwo \phip \etap }}

2706: \right) - \mu ( \etatwo \bar \etatwo + \phip \bar \etap + \bar

2707: \phip \etap)  .$$ Note that $\mu$ must be real in order for the

2708: action to be real. We are going to use the fact that the real

2709: superpotential can still be differentiated with respect to complex

2710: combinations of fields to figure out the F-term vacuum equations.

2711: These equations are:

2712: %

2713: \eqn\Fetatwo{0= F_{\etatwo} = m \phip \etap - \mu \bar \etatwo }

2714: \eqn\Fphi{0= F_{\phip } = m \etatwo \etap - \mu \bar \etap }

2715: \eqn\Feta{ 0= F_{\etap } = m \etatwo \phip - \mu \bar \phip  }

2716: %

2717: and their complex conjugates. $F_{\phitwo} = 0$ trivially. Adding

2718: \Fphi\ to \Feta\ determines $\etatwo$ in terms of the positively

2719: charged fields (when $\phip, \etap \neq 0$):

2720: %

2721: \eqn\fixetatwo{ \etatwo = { \mu \over m} \left. { \overline{\phip

2722: + \etap} \over \phip +  \etap } \right. }

2723: %

2724: The second factor on the RHS is a phase, so this says that

2725: $$ |\etatwo| = {\mu  \over m}.$$

2726: Dividing \Fphi\ by \Feta\ gives

2727: %

2728: \eqn\relativephase{ {\phip \over \etap} = {\bar \phip \over \bar

2729: \etap} }

2730: %

2731: which implies that this ratio is real.  Call it $x$:

2732: %

2733: \eqn\eliminateeta{ \etap = x \phip .}

2734: %

2735: Plugging these into \Fetatwo\ gives

2736: %

2737: \eqn\fixphi{ x |\phip|^2 = { \mu^2 \over m^2} ,}

2738: %

2739: which tells us that $x > 0$ by the reality of $\mu$ and $m$. Now

2740: the D-term reads:

2741: %

2742: \eqn\Dtermeqn{ \left(x + {1 \over x}\right) { \mu^2 \over m^2} =

2743: \rho + 2 {\mu^2 \over m^2}+ 2 |\phitwo|^2 }

2744: %

2745: Moving the $2 \mu^2/m^2 $ to the LHS gives

2746: %

2747: \eqn\victory{ \left( \sqrt x - {1 \over \sqrt x} \right)^2 {\mu^2

2748: \over m^2} = \rho + 2 |\phitwo|^2 .}

2749: %

2750: Note that the double-valuedness from taking this square root is

2751: not harmful to us since $x$ is always positive on the vacuum

2752: manifold.

2753: %The RHS is positive.

2754: 

2755: To recap:  $\etatwo$ is eliminated by \fixetatwo.

2756: The equation \eliminateeta\ determines $\etap$

2757: in terms of $x, \phip$.  The equation \fixphi\

2758: eliminates $|\phip|$ in terms of $x$.

2759: Because $\phip$ doesn't vanish on this branch of the

2760: moduli space we can use the $U(1)$ gauge symmetry to

2761: fix its phase.

2762: The remaining variables are $x \in \IR_+$ and $\phitwo \in \IC$

2763: related by the eqn \victory.

2764: Solving this equation for the norm of $\phitwo$

2765: gives a circle (the phase of $\phitwo$)

2766: fibered over the $x$ direction.  Its radius is

2767: $$

2768: 2 |\phitwo|^2 =

2769: ( \sqrt x - {1 \over \sqrt x} )^2 { \mu^2\over m^2} - \rho .

2770: $$

2771: This is the desired equation

2772: \desired\ with $ w = \sqrt x - {1 \over \sqrt x} $.

2773: 

2774: In the IR, $\rho \to + \infty$\foot{

2775: In the next subsection, we analyze the RG behavior

2776: of the (1,1) theory and show

2777: that nonzero $\mu$ does not

2778: change the running of $\rho$ from the (2,2) result.},

2779: and this equation

2780: has no solutions for an interval of values of $x$ where

2781: the RHS dips below zero.  There are thus two caps,

2782: as there were for $\mu =0$.

2783: 

2784: In the UV, $\rho \to - \infty$ and the RHS is positive definite.

2785: The radius of the $\phitwo$ circle is big

2786: at the two ends ($ x \to \infty$ and $x \to 0$),

2787: and has a minimum when $ x = 1$.

2788: 

2789: \subsec{Extra Vacua}

2790: 

2791: In the above analysis we have described the part of the vacuum

2792: manifold that for large $|\rho|$ describes a geometrical target

2793: space.  In general, there are other discrete massive vacua of the

2794: system.  Physically, in the spacetime theory, these describe

2795: subcritical-dimension target space components, which we will refer

2796: to as `dust vacua' for want of a better name. From the point of

2797: view of the world-sheet theory, the Witten index contributions

2798: contained in the 10-dimensional geometrical regions jump from 0 in

2799: the hyperboloid to -2 on the pair of caps.  The residual

2800: contribution must be contained in the subcritical dust.

2801: 

2802: In general, we could start with some extra subcritical dust vacua

2803: as well.  Indeed in the above model there is another solution of

2804: the F-term equations when $\phip = \etap = \etatwo = 0 $. The

2805: D-term equation becomes

2806: %

2807: \eqn\evilDterm{ 0 = \rho + 2 |\phitwo|^2 }

2808: %

2809: which in the UV, $\rho < 0$, has solutions. The gauge symmetry

2810: fixes the phase of $\phitwo$, but there is a residual $\IZ_2 $

2811: gauge symmetry, which acts on nothing.

2812: 

2813: In general there can also be vacua in which the scalar $\sigma$ in

2814: the (2,2) gauge multiplet has a vev \WittenYC. We will discuss

2815: these below, with the results that there are no such vacua for

2816: $\rho < 0$, but for the IR limit $\rho

2817: >0$ there are indeed $\sigma$ vacua.

2818: %GIVE SOME ARGUMENT THAT $\chi({\rm evil~vacua}) =0$?.

2819: Altogether, the Witten index for the sum of all components is conserved under the flow.

2820: 

2821: 

2822: \subsec{GSO action on the extra `dust' vacua}

2823: 

2824: The question we want to answer in this subsection is the

2825: following.  We start with a type II string theory, with its chiral

2826: GSO projection.  The vertex operator \TsupvertII\ that we add is

2827: of course invariant under this symmetry, and also does not break

2828: the $Z_2\times Z_2$ global `quantum symmetry' corresponding to

2829: this GSO projection (i.e. it is an NS-NS state).  It is therefore

2830: interesting to ask how the GSO projection acts on the `dust'

2831: components of the target space we obtain after the tachyon

2832: condensation.

2833: 

2834: In particular, the isolated vacua (hereafter called 'dust') lead a priori to disconnected eight-dimensional

2835: universes, which are flat at very weak coupling, and the obvious type II GSO projection in eight dimensions is

2836: not modular invariant, while the type 0 one is. We will see how the discrete R symmetries in our linear model

2837: act in a way which precisely reduces the type II GSO we start with to the type 0 GSO after the flow.  More

2838: precisely, we obtain the eight dimensional target space in the dust components as an intermediate step available

2839: in the RG flow; these subcritical theories are themselves tachyonic (they have relevant operators) and will

2840: ultimately flow further down to 2 or fewer dimensions.

2841: 

2842: This works out as follows. In the linear model realization, the

2843: dust vacua reside on the $\sigma$ branch of the field space, where

2844: as mentioned above $\sigma$ is the scalar in the (2,2) vector

2845: multiplet. (We can set $\mu = 0$ for this discussion, because it

2846: is not important for the $\rho \to + \infty$ physics.) They arise

2847: \WittenYC\ by solving the SUSY-preservation equations arising from

2848: the quantum-corrected twisted chiral superpotential term

2849: \eqn\twistedsp{ \int d\theta _+ d\bar \theta_- ~\tilde W = \int

2850: d\theta_+ d\bar \theta_- ~ \left( t \Sigma + Q_{{\rm T}} \Sigma

2851: \ln \Sigma \right)} where $ t = \rho + i \theta$, $ Q_{{\rm T}} $

2852: is the net charge of the chiral fields, and the second term arises

2853: from the chiral multiplets running in a loop.

2854: %\twistedsp\ is derived in the large-sigma region.

2855: Given that the chiral R-symmetries act by

2856: \eqn\chiralR{ \theta_+ \mapsto e^{i \alpha_+} \theta_+, ~~~

2857: \theta_-  \mapsto e^{i \alpha_-} \theta_- ,}

2858: the presence of the first term implies that

2859: $ \sigma $ transforms like

2860: $$ \sigma \mapsto e^{ i \alpha_+ - i\alpha_-} \sigma .$$

2861: The second term in \twistedsp\ reproduces the R-charge anomaly. Although the $U(1)_{{\rm axial}}$ is anomalous,

2862: a $\IZ_2$ subgroup is preserved because the anomaly is even; this is important because it is by this $\IZ_2$

2863: generator $g$ that the chiral GSO acts on this system.  That is, we have two independent $Z_2$ symmetries which

2864: provide the chiral $\IZ_2\times\IZ_2$ symmetry by which we can orbifold to enforce the GSO projection

2865: appropriate to the type II theory.  (By comparison, the type 0 GSO projection is obtained by the vectorlike

2866: combination of these $\IZ_2$ symmetries.)

2867: 

2868: Now the key point is just that the sigma vacua are at values

2869: of $\sigma$

2870: such that

2871: $$ 0 = {\del \tilde W \over \del \sigma } $$

2872: which are

2873: $$ \sigma_\pm = \pm e^{ |t|/2 }. $$

2874: These two values are permuted by the action of $g$. Therefore, a

2875: fundamental domain for the action of $g$ is obtained by forgetting

2876: about one of the sigma vacua; the remaining theory is one isolated

2877: vacuum, modulo the action of the vectorlike GSO projection, namely

2878: one copy of 8-dimensional type 0.

2879: 

2880: \subsec{Renormalization properties of the (1,1) GLSM}

2881: 

2882: Since the deformation \breaksNequalstwo\ breaks

2883: $N=(2,2)$ supersymmetry,

2884: the usual nonrenormalization theorems

2885: used in \WittenYC\

2886: need not apply.

2887: In order to ascertain whether

2888: our linear model has the desired

2889: flow, we analyze the running of

2890: the $N=(1,1)$ couplings,

2891: and of the possible dangerous terms that are now

2892: allowed by symmetries.

2893: Such dangerous terms include:

2894: \eqn\dangerousterms{

2895: q_+ q_- \left( \tilde m \bar \phitwo \phitwo +

2896:  m_1 \left( \bar \phip \phip + \bar \etap \etap \right) \right).}

2897: The generation of such terms at scales comparable to $e$ or $m$ or $\mu$ would destroy the vacuum manifold.

2898: 

2899: This analysis is facilitated by the

2900: following observation.

2901: Since $\phitwo$ does not appear in

2902: any of the superpotential terms ($N=2$ or otherwise),

2903: if we ignore the coupling to the vectormultiplet

2904: there is a shift symmetry

2905: %

2906: \eqn\shiftsym{ \phitwo \mapsto \phitwo + \alpha }

2907: %

2908: with $\alpha$ an arbitrary constant superfield. This means that

2909: potential terms that involve $\phitwo$ must be accompanied by a

2910: positive number of powers of $e$ (an even number, by charge

2911: conjugation symmetry) as well as $\mu$. Because of this, it is

2912: actually quite difficult to generate the troublesome mass terms

2913: \dangerousterms, particularly when combined with the

2914: superrenormalizible nature of this field theory.

2915: 

2916: 

2917: \itemized \ifig\dangerous{ a) a $\phi^4$ one-loop bubble; b) a

2918: fermion loop; c) a $ \phi^2 \psi^2 $ bubble.

2919: }{\epsfxsize3.0in\epsfbox{dangerous.eps}}

2920: 

2921: \itemaut{First of all, we find that in the theory of interest, the

2922: only three classes of diagrams with two external scalars which can

2923: have UV divergent parts are shown in \dangerous. This is just the

2924: statement of superrenormalizibility of the theory, in detail for

2925: this particular term of interest.}

2926: 

2927: \itemaut{There are no $|\phi|^2 |\psi|^2 $ couplings

2928: in this model, so we never have to worry about a

2929: diagram of type c.}

2930: 

2931: \ifig\manyloops{ Here are the two least-UV-finite diagrams that

2932: renormalize the mass of $\phitwo$. The scalar propagators have an

2933: orientation because the fields are complex and charged, which

2934: isn't indicated. For this figure, $\phi$ means $\phip$. The two

2935: vertices that have been used here are $ \mu \phip^2 \etatwo $ from

2936: the '$\mu$-term' and its conjugate, and $ e^2 |\phitwo|^2

2937: |\phip|^2 $ from the D-term. There are other diagrams where

2938: $\phip$ is replaced by $\etap$ and there is another diagram of the

2939: form of the first where the $\phip$ that hits the $\phitwo$ is

2940: replaced by $\etatwo$ and $\etatwo$ and the other $\phip$ are both

2941: replaced by either $\phip$ or $\etap$.

2942: }{\epsfxsize4.0in\epsfbox{manyloops.eps}}

2943: 

2944: \itemaut{For the $\phitwo$ mass renormalization diagrams, any

2945: putative extra divergence\foot{ By 'extra' we mean in addition to

2946: the divergences arising from the renormalization of $\rho$ or of

2947: the kinetic terms, which {\it do} appear here if you use the

2948: D-term vertices to let the other scalars run in the loop in

2949: diagram a -- but this is $\mu$-independent and therefore not

2950: germane.

2951: %(I barely managed to prevent myself from saying 'relevant').

2952: }

2953: must be accompanied by

2954: two powers of $e$

2955: and

2956: one power of $\mu$ (because $\mu$ is the only thing that breaks $N=2$).

2957: There is no diagram of the form a or b that has this property.

2958: The least non-divergent diagram that we've found with this property

2959: has two loops, and four scalar propagators,

2960: and is show in \manyloops.

2961: }

2962: 

2963: 

2964: \ifig\phimass{

2965: The leading correction to the mass of

2966: $\phip$ comes from this finite amplitude.

2967: The vertices are from the

2968: $ \mu \phip^2 \etatwo $ term and its conjugate.

2969: }{\epsfxsize2.0in\epsfbox{phimass.eps}}

2970: \itemaut{For the $\phip$ and $\etap$ mass renormalization,

2971: and for the renormalization of $\mu$ itself,

2972: we only know that it must depend on $\mu$.

2973: There is still no $\mu$-dependent diagram of the form of diagram a.

2974: For diagram b, the only option is to use the cubic vertices from the

2975: $N=2$ superpotential to make $\phip$ turn into two fermions.

2976: The simplest diagram of this form does indeed diverge

2977: logarithmically in the UV,

2978: but it doesn't depend on $\mu$.  This is just the wavefunction

2979: renormalization of $\phip$ from loops of fermions

2980: (it vanishes at zero external momentum,

2981: and so doesn't renormalize the mass

2982: -- an important fact for the $N=2$ LSM);

2983: it is an innocuous correction to the kahler potential.

2984: We can try to insert $\mu$-dependence by

2985: using the 'mass vertices' from the $N=1$ superpotential

2986: (note that there are {\it no} other terms involving

2987: fermions that come from the $(N=2)$-breaking terms):

2988: $$ L_f = \mu ( \psi^{\etatwo}_+ \bar \psi^{\etatwo}_-

2989: + \psi^{\etap}_+ \bar \psi^{\phip}_-

2990: + \bar \psi^{\etap}_+ \psi^{\phip}_-  )

2991: + \left({ + \leftrightarrow - }\right) .$$

2992: But the powers of $\mu$ that you need to

2993: insert are accompanied by as many extra propagators,

2994: even one of which makes the integral UV finite.

2995: Alternatively, we could just renormalize the fermion propagators

2996: by summing the geometric series, and use these in diagram b.

2997: The result is again that the $\mu$-dependent part

2998: will not diverge.

2999: The leading finite contribution to the $\phip, \etap$ masses

3000: is shown in \phimass.}

3001: 

3002: \itemaut{Similarly, it is simple to show that

3003: $\mu$ itself does not receieve any

3004: UV divergent corrections.  Further,

3005: there is no $\mu$-dependent

3006: UV divergent correction to the FI parameter $\rho$.}

3007: 

3008: From the results 1-4, we can conclude that the beta function for

3009: the dangerous couplings, which we collectively denote as $\tilde

3010: m$, is of the form

3011: %

3012: $$ \beta_{\tilde m^2} = \sum_g {\del \over \del g} \delta \tilde m^2

3013: \beta_g  , $$

3014: %

3015: where $\delta \tilde m^2 $ is the sum of the finite mass-shifts

3016: discussed above. This is because there are no extra divergent

3017: counterterms required to cancel these diagrams -- all of the

3018: scale-dependence comes from the fact that these diagrams, and

3019: hence the counterterms which cancel them, depend on the other

3020: renormalized couplings. We expect that these finite diagrams have

3021: the property that their dependence on the running coupling ${ \del

3022: \delta \tilde m^2 \over \del \rho} $ is small when $|\rho|$ is

3023: large. This is a consequence of the fact that the running coupling

3024: $\rho$ appears, when at all, as a mass term, and that IR

3025: enhancements are at worst logarithmic.

3026: 

3027: The final result of this analysis

3028: is that it is possible to

3029: fend off the dangerous terms \dangerousterms\

3030: in the two large $|\rho|$ phases

3031: by the addition of scale-independent counterterms.

3032: The RG behavior of the

3033: $(1,1)$ linear model

3034: is therefore as described in

3035: subsection A.2.

3036: 

3037: 

3038: 

3039: \subsec{A puzzle about the flow and its resolution}

3040: %sign flip of tachyon beta function}

3041: 

3042: At large negative $\rho$, the vacuum manifold seems to be a

3043: one-sheeted hyperboloid of size $\rho$, independent of $\mu$. If

3044: this vacuum manifold were in fact the target of the NLSM in the

3045: IR, the following puzzle would arise. The string wound around the

3046: waist of the hyperboloid would have a mass of order $l_s \sqrt

3047: \rho \gg l_s$ and should therefore be very irrelevant. On the

3048: other hand, $\rho$, which (as we have shown in \S A.1) turns on

3049: this operator, has a constant, relevant beta function

3050: $$ \beta_\rho = Q_{{\rm T}} = -2 .$$

3051: 

3052: The resolution of this puzzle involves two important subtleties of

3053: the model. First of all, near the center of the throat, the

3054: $\etatwo$ field is light. This is clear since before $\mu$ was

3055: turned on, $\etatwo$ actually became massless at the point

3056: ($\phi_+ = \eta_+ =0$) where the extra $\IP^1$ was attached. This

3057: is an indication of the presence of string-scale features in that

3058: region.

3059: 

3060: More precisely, this opens up a fascinating possible loophole: for

3061: small $\mu$, it does not cost much energy to move the $\etatwo$

3062: field around. By moving it around, we will see that the energy of

3063: the winding mode can be decreased from its apparent value if the

3064: geometry really has small curvatures (of order ${1 \over |\rho|

3065: }$) everywhere.

3066: 

3067: So we consider the energy of the wound string as a probe of how

3068: big the cycle is. We pick a gauge where $\phitwo$ is the field

3069: that winds

3070: $$ \phitwo(\sigma) = e^{ i \sigma } \phitwo^0 ,$$

3071: where $\sigma$ is the world-sheet space coordinate. The kinetic

3072: energy density is something like $ e^2 | \phitwo^0 |^2 $. We add

3073: this to the bosonic potential, with $\phip, \etap$ evaluated on

3074: their vacuum solutions, $ \phip = x \mu, \etap = \mu / x $.

3075: Further, we know that the wound string wants to be at the

3076: narrowest part of the neck, $x=1$.

3077: 

3078: 

3079: 

3080: This gives the following answer for the energy, in string units:

3081: $$

3082: E( \etatwo^0, \phitwo^0 )

3083: = e^2

3084: \left(| \phitwo^0 |^2 + ( 2 {\mu^2 \over m^2} - \rho - 2 |\phitwo^0|^2

3085: - 2 | \etatwo^0|^2 )^2 \right)

3086: + \mu^2 | {\mu\over m} - \etatwo^0 |^2 .

3087: $$

3088: Setting $\etatwo^0 = { \mu \over m}$,

3089: its vacuum value, this gives

3090: $$E(\etatwo = \mu/m) = e^2 \rho. $$

3091: But taking into account the fact that

3092: $\etatwo$ is light (\ie\ assuming $\mu^2 \ll e^2 $)

3093: this actually has a much lower minimum, at

3094: $$E = \mu^2 \rho. $$

3095: %Note that it still grows with $\rho$.

3096: Thus for small $\mu$ (as the (2,2) model becomes a better approximation), the winding mode may become tachyonic,

3097: as required by the direction of flow of $\rho$ combined with its role multiplying the winding operator in the

3098: action.

3099: 

3100: A related point is that from the GLSM instanton expansion, the tachyon vev (the coupling of the tachyonic

3101: winding operator in the 2d action) is actually

3102: $$ T(\rho) = f(\rho, \mu) e^{ \rho } . $$

3103: Here $f(\rho)$ is a ratio of fermionic to bosonic one-loop determinants in the vortex background.  These

3104: precisely cancel in the ${\cal N}=2$ limit, $f_{{\cal N}=2}=1$. This tells us that the beta function for $T$ is

3105: actually

3106: $$

3107: \beta_T = \beta_\rho ~{ \del T \over \del \rho}

3108: = Q_{{\rm T}} ~ ( { f' \over f} + 1 ) e^{ - \rho }.

3109: $$

3110: This can change sign as $\mu$ and $\rho$ vary.  Hence, for {\it large} $\mu$, far from the regime where ${\cal

3111: N}=2$ results apply to a good approximation, the direction of flow may turn around, consistently with the

3112: irrelevance of the winding operator for the large radius connected hyperboloid.

3113: 

3114: 

3115: To summarize, the rewards of the linear sigma model analysis are the following:

3116: %

3117: \itemized

3118: %

3119: \itemaut{ The linear sigma model gives a global picture of the RG

3120: flow, including the B-cycle direction, which corroborates the

3121: analysis of the previous sections. }

3122: %

3123: \itemaut{ We found a {\it gauged} linear sigma model \WittenYC\

3124: which `linearizes' the tachyon, in the sense that the winding

3125: vertex operator becomes directly related to the vortices of the

3126: GLSM. In such models, the condensation of vortices is controlled

3127: by the Fayet-Iliopoulos parameter, whose renormalization

3128: properties are well-understood. }

3129: %

3130: \itemaut{ The GLSM gives independent evidence for the existence of additional vacua, as in \refs{\othervacuarefs,\chicago}.}

3131: %

3132: \itemaut{The GLSM provides a nice picture of the mechanism by

3133: which these lower-dimensional vacua manage to have a consistent

3134: GSO projection.}

3135: 

3136: %%In the last one I took out "type II" because it is potentially

3137: %% unclear to the reader--the result is type 0 effectively...-ES

3138: 

3139: 

3140: 

3141: \vfill\eject

3142: \appendix{B}{Towards a transformative hermeneutics 

3143: of off-shell string theory}

3144: 

3145: \vskip0.7in

3146: 

3147: {\bf The Second Coming} -- W. B. Yeats

3148: 

3149: %\marginpar{elephants}

3150: %Author: C.G. van der Laan, Hunzeweg 57, 9893PB

3151: %   Garnwerd, The Netherlands

3152: %   05941-1525, cgl@risc1.rug.nl.

3153: %Purpose: Verbatim macros via plain TeX, 

3154: %         to be used with AnyTeX.

3155: %                         *numvrb

3156: %User toks variables

3157: \newtoks\thisverbatim

3158: \newtoks\everyverbatim

3159: %

3160: %User customization

3161: \let\preverbatim\medskip

3162: \let\postverbatim\medbreak

3163: %                         *vrblin100

3164: %User `options': \numvrb

3165: \newcount\vrblin

3166: \def\numvrb{\vrblin0

3167:    \everypar{\advance\vrblin1

3168:     \llap{\sevenrm\the\vrblin\quad}}}

3169: \def\nonum{\everypar={}}

3170: %              : \emc

3171: \def\makeescape#1{\catcode`#1=0 }

3172: \def\makeactive#1{\catcode`#1=13 }

3173: {\makeactive\<

3174:  \gdef\emc{\makeactive\<%

3175:    \def<##1>{$\langle##1\rangle$}}}

3176: %                         *vrblin200

3177: %User macro

3178: \def\verbatim{\preverbatim\begingroup

3179:    \tt\setupverbatim

3180:    \the\everyverbatim\relax

3181:    \the\thisverbatim\relax

3182:    \verbatimgobble}

3183: %

3184: \def\endverbatim{\endgroup\postverbatim

3185:    \thisverbatim={}}

3186: %

3187: \def\setupverbatim{\makeactive\`%

3188: %   \let\!=!\makeescape\!%Knuth&Levy  

3189:    \let\*=*\makeescape\*%Knuth&Levy  

3190:    \def\par{\leavevmode\endgraf}%TB381

3191:    \obeylines\uncatcodespecials

3192:    \obeyspaces}

3193: %

3194: {\obeyspaces\global\let =\ 

3195:  \obeylines\gdef\verbatimgobble#1^^M{}%

3196:  \makeactive\` \gdef`{\relax\lq}}%TB381 

3197: %

3198: \def\uncatcodespecials{\def\do##1{%

3199:    \catcode`##1=12 }\dospecials}

3200: %                         *vrblin250

3201: %Minimal | tag for inline verbatim

3202: \def\vrt{{\tt\char`\|}}\makeactive\|

3203: \def|{\bgroup\tt\setupverbatim

3204:       \the\everyverbatim\relax

3205:       \the\thisverbatim\relax

3206:       \def|{\egroup\thisverbatim{}}}

3207: %\endinput %14/2/94 cgl@risc1.rug.nl

3208: %                         *nonum

3209: %Contents

3210: %Newtoks

3211: %   \thisverbatim.................2

3212: %   \everyverbatim................3

3213: %Customing

3214: %   \preverbatim..................6

3215: %   \postverbatim.................7

3216: %Options

3217: %   \numvrb.....................103   

3218: %   \nonum......................106

3219: %   \makeescape.................108

3220: %   \makeactive.................109

3221: %   \emc........................111   

3222: %   <#1>........................112   

3223: %User macro

3224: %   \verbatim...................202

3225: %   \endverbatim................208

3226: %   \setupverbatim..............211

3227: %   \verbatimgobble.............218

3228: %   `...........................219

3229: %   \uncatcodespecials..........221

3230: %Inline verbatim

3231: %   \vrt........................252

3232: %   |...........................253

3233: %History of changes vrb.tex

3234: %March  94   Block comments omitted.

3235: %            Verbatim mode with exclamation 

3236: %            mark as escape char.

3237: %Febr 1994   Release Version 1.

3238: %

3239: %Examples. (Extract the above vrt.tex file

3240: %           and store this as vrt.tex.)

3241: %           

3242: %

3243: 

3244: \def\marginsert#1{

3245: \vskip-0.361in\hskip3.6in\rlap{\footnotefont #1} 

3246: \verbatim}

3247: 

3248: %!endverbatim

3249: %\llap{\footnotefont Am I in the margin?} \vskip-0.33in

3250: %\verbatim 

3251: 

3252: 

3253: \verbatim

3254: Turning and turning in the widening gyre

3255: *endverbatim

3256: \marginsert{%({\it gyre} -- {\bf n.} 

3257: Vortex-induced}

3258: The falcon cannot hear the falconer;

3259: *endverbatim

3260: \marginsert{causal disconnection}

3261: Things fall apart; the centre cannot hold;

3262: *endverbatim

3263: \marginsert{follows from tachyon condensation,}

3264: Mere anarchy is loosed upon the world, 

3265: *endverbatim

3266: \marginsert{creating an explosion of}

3267: The blood-dimmed tide is loosed, and everywhere

3268: *endverbatim

3269: \marginsert{decay products, and opening}

3270: The ceremony of innocence is drowned;

3271: *endverbatim

3272: \marginsert{a can of worms.

3273: %concerns about randomization of coupling constants.

3274: %information loss.}

3275: }

3276: The best lack all convictions, while the worst

3277: Are full of passionate intensity.

3278: *endverbatim

3279: \marginsert{%\joequote}

3280: }

3281: 

3282: 

3283: Surely some revelation is at hand;

3284: Surely the Second Coming is at hand.

3285: The Second Coming! Hardly are those words out

3286: When a vast image out of Spiritus Mundi

3287: Troubles my sight: somewhere in sands of the desert

3288: A shape with lion body and the head of a man,

3289: A gaze blank and pitiless as the sun,

3290: Is moving its slow thighs, while all about it

3291: Reel shadows of the indignant desert birds.

3292: The darkness drops again; but now I know

3293: That twenty centuries of stony sleep

3294: Were vexed to nightmare by a rocking cradle,

3295: And what rough beast, its hour come round at last,

3296: Slouches towards Bethlehem to be born?

3297: *endverbatim

3298: 

3299: 

3300: 

3301: 

3302: 

3303: \listrefs

3304: \end

3305: