1: 

2: 

3: \documentclass[12pt,epsf]{article}

4: %\usepackage{epsfig}

5: \usepackage{cite}

6: \setlength{\topmargin}{-.3in} \setlength{\oddsidemargin}{.0in}

7: %\setlength{\evensidemargin}{-.21in}

8: \setlength{\textheight}{8.5in} \setlength{\textwidth}{6.35in}

9: \setlength{\footnotesep}{\baselinestretch\baselineskip}

10: \newlength{\abstractwidth}

11: \setlength{\abstractwidth}{\textwidth}

12: \addtolength{\abstractwidth}{-6pc}

13: \usepackage[dvips]{graphicx}

14: \usepackage{subfigure}

15: \usepackage{amssymb}

16: 

17: 

18: 

19: 

20: \flushbottom \thispagestyle{empty} \pagestyle{plain}

21: 

22: 

23: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

24: %%%%%%%%%%%%%%%%%%%%%%

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

26: \renewcommand{\thanks}[1]{\footnote{#1}} % Use this for footnotes

27:  \newcommand{\starttext}{

28: \setcounter{footnote}{0}

29: \renewcommand{\thefootnote}{\arabic{footnote}}}

30: \renewcommand{\theequation}{\thesection.\arabic{equation}}

31: \newcommand{\be}{\begin{equation}}

32: \newcommand{\bea}{\begin{eqnarray}}

33: \newcommand{\eea}{\end{eqnarray}}

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

35: \newcommand{\ee}{\end{equation}}

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

37: \newcommand{\N}{{\cal N}}

38: \newcommand{\<}{\langle}

39: \newcommand{\oc}{{\cal O}_C}

40: \renewcommand{\a}{\alpha}

41: \renewcommand{\b}{\beta}

42: \newcommand{\half}{{1\over 2}}

43: \newcommand{\comment}[1]{}

44: \renewcommand{\>}{\rangle}

45: \def\ba{\begin{eqnarray}}

46: \def\ea{\end{eqnarray}}

47: \newcommand{\PSbox}[3]{\mbox{\rule{0in}{#3}\special{psfile=#1}\hspace{#2}}}

48: 

49: \def\corrxy{\langle \phi(x) \phi(y) \rangle}

50: \def\corrf{\langle \phi(\tau_1) \phi(\tau_2) \rangle}

51: \def\t0{$t=0$}

52: \def\14{{1\over4}}

53: \def\12{{1 \over 2}}

54: \def\eq{&=&}

55: \def\tro{\tilde{\rho}}

56: \def\d{\partial}

57: \def\dt{\partial_{\tau}}

58: \def\ds{\partial_{\sigma}}

59: \def\h3{h^{3\over 2}}

60: \def\R{\bar{R}}

61: \def\qft{quantum field theory}

62: \def\>{\rangle}

63: \def\<{\langle}

64: \def\sc {Schwarzschild}

65: \def\ls{\sqrt{\alpha'}}

66: \def\des{de Sitter Space}

67: \def\f{\Phi}

68: \def\cc{cosmological constant}

69: \def\st{string theory}

70: \def\sms{supermoduli--space}

71: \def\lmb{\lambda}

72: \def\lo{\lambda_0}

73: \def\sb{supersymmetry breaking}

74: \def\sbs{supersymmetry breaking scale}

75: \def\bdg{Banks Dine Gorbatov}

76: \def\0cc{$\Lambda = 0$}

77: \def\pcc{$\Lambda > 0$}

78: \def\sds{Schwarzschild-de Sitter}

79: \def\rd{\dot r}

80: \def\Om{   \Omega_{D-1}^2   }

81: \def\om{   \Omega_{D-2}^2   }

82: \def\pt{\tilde \phi}

83: \def\eps{{\ensuremath{{\bf \epsilon}}}}

84: \def\tb{{\ensuremath{\bar T}}}

85: \def\wb{{\ensuremath{\bar W}}}

86: \def\epsb{{\ensuremath{{\bar \epsilon}}}}

87: \def\adst{{\ensuremath{{AdS_3}}}}

88: \def\eone{{\ensuremath{\eps_1}}}

89: \def\etwo{{\ensuremath{\eps_2}}}

90: \def\etwob{{\ensuremath{\bar \eps_2}}}

91: \def\eoneb{{\ensuremath{\bar \eps_1}}}

92: \def\r{{\ensuremath{\xi}}}

93: \def\fa{{\ensuremath{f_a}}}

94: \def\d{{\ensuremath{\zeta}}}

95: \def\tl{{\ensuremath{t_\Lambda}}}

96: \def\I{{ Im(K_T \dot T)}} 

97: \def\tensor{{\otimes}}

98: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

99: %%%%%%%%%%

100: 

101: \begin{document}

102: \renewcommand{\theequation}{\thesection.\arabic{equation}}

103: \begin{titlepage}

104: \bigskip

105: %\rightline{hep-th/yymmnnn}

106: 

107: \bigskip\bigskip\bigskip\bigskip

108: 

109: \centerline{\Large \bf {Anthropic Explanation of the Dark Matter Abundance}}

110: 

111: 

112: \bigskip\bigskip

113: \bigskip\bigskip

114: 

115: \centerline{\bf Ben Freivogel}

116: \medskip

117: \centerline{\small Department of Physics and Center for Theoretical

118: Physics} 

119: \centerline{\small 

120: University of California, Berkeley, California 94720, U.S.A.}

121: \medskip

122: \centerline{\small Lawrence Berkeley National Laboratory, 

123: Berkeley, California 94720, U.S.A.}

124: \medskip

125: \centerline{\small freivogel@berkeley.edu}

126: 

127:  \begin{abstract} I use Bousso's causal diamond measure to make a

128:  statistical prediction for the dark matter abundance, assuming an

129:  axion with a large decay constant $f_a \gg 10^{12}$ GeV.  Using a

130:  crude approximation for observer formation, the prediction agrees

131:  well with observation: 30\% of observers form in regions with less

132:  dark matter than we observe, while 70\% of observers form in regions

133:  with more dark matter.  Large values of the dark matter ratio are

134:  disfavored by an elementary effect: increasing the amount of dark

135:  matter while holding fixed the baryon to photon ratio decreases the

136:  number of baryons inside one horizon volume.  Thus the prediction is

137:  rather insensitive to assumptions about observer formation in

138:  universes with much more dark matter than our own.  The key

139:  assumption is that the number of observers {\it per baryon} is

140:  roughly independent of the dark matter ratio for ratios near the

141:  observed value.  \end{abstract}

142: 

143: \end{titlepage}

144: \starttext \baselineskip=18pt \setcounter{footnote}{0}

145: 

146: 

147: 

148: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

149: %%%%%%%%%%%%%%%%%%%%%%
 

150: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

151: %%%%%%%%%%%%%%%%%%%%%%
 

152: 

153: 

154: \setcounter{equation}{0}

155: 

156: \section{Introduction}

157: Our current understanding of string theory suggests a vast landscape, 

158: with eternal inflation leading to an infinite number

159: of pocket universes containing different low energy physics \cite{bp,

160:   kklt, lenny}. In this

161: context, a prediction requires a combination of landscape

162: statistics, the dynamics of eternal

163: inflation, and anthropic considerations. Some quantities may still

164: have a conventionally natural explanation, while

165: for others anthropic arguments are crucial. The most famous

166: example of a quantity whose observed value is well explained by

167:  anthropics

168: is the cosmological constant \cite{cc}.

169: 

170: It is natural to wonder whether the dark matter abundance has an

171: anthropic explanation. For example, models with a large axion decay

172: constant, $\fa \gg 10^{12}$ GeV, are attractive from the particle

173: physics point of view \cite{pq} and arise naturally in string

174: theory. However, in these theories the dark matter abundance is

175: naturally much larger than what we observe \cite{axbound}. Many

176: authors\cite{axanth, simeon,

177:   tarw, htw}\footnote{I apologize

178:   for missing references throughout this note. Please let me know if your work should

179:   be cited.}, beginning with Linde, 

180: have examined the question of whether the dark matter

181: abundance is typical

182: once anthropic weighting is taken into account. These authors have

183: reached a variety of conclusions. Due to the

184: difficulty in simulating universes different from our own and our

185: ignorance about the conditions necessary for life, it is

186: unclear whether a dark matter abundance of even 100 times the observed value

187: prevents observers from forming. Furthermore, even the most stringent

188: assumptions about the requirements for life \cite{tarw} lead to the

189: conclusion that the observed dark matter abundance is unusually small.

190: 

191: It is not possible to make a prediction in the landscape without

192: regulating the infinities. Eternal inflation produces an infinite

193: number of pocket universes of every type, and each pocket universe is

194: spatially infinite. One attractive recipe for regulating the

195: infinities 

196:  is the causal diamond measure of Bousso \cite{cd}, which explains the

197:  observed cosmological constant well \cite{harnik, by}. 

198: 

199: Here I focus on the simplest possible scenario: I fix all

200: parameters to their observed values and fix the axion decay constant

201: to a large value, $f_a \gg 10^{12}$ GeV. I assume that axions make up

202: all of the dark matter. The only parameter which is allowed to vary is

203: the initial axion misalignment angle $\theta_i$. In terms of

204: observable quantities, this means that the baryon to photon ratio is held

205: fixed, while the ratio of dark matter to ordinary matter is allowed to

206: vary.

207: In this simple context, the causal diamond measure just tells us to

208: count the number of observations within the backward lightcone of one

209: geodesic. In other words, we can get a finite answer from spatially

210: infinite universes by restricting our attention to one horizon

211: volume. (The ``horizon'' throughout this paper refers to everything within

212: the backward lightcone of one geodesic, and never to the apparent horizon.)

213: 

214:  In this note, I show that dark matter abundances much larger than

215:  what we observe are disfavored in the causal diamond measure by an

216:  elementary effect: increasing the amount of dark matter while holding

217:  fixed the baryon to photon ratio decreases the

218:  number of baryons inside our causal patch. Quantitatively, if \d\ is

219:  the ratio of dark matter to ordinary matter, the number of baryons

220:  $N_b$ in the causal patch is \be N_b \propto {1 \over 1 + \d} ~.  \ee

221: 

222: This may be surprising, because it seems that we are not subtracting

223: any baryons when we increase the amount of dark matter. To see that

224: the above formula is correct

225: at least at one time, consider the time \tl\ when the

226: cosmological constant begins to dominate the energy density. The

227: horizon volume at this time is set by the cosmological constant,

228: independent of the dark matter ratio \d. By definition, at the time

229: \tl\ the total matter density is equal to the vacuum energy density,

230: $\rho_\Lambda (\tl) = \rho_m (\tl)$, so the total amount of matter

231: inside the horizon at the time \tl\ is independent of \d. Therefore,

232: the number of baryons inside the horizon is $N_b (\tl) \propto 1/(1 +

233: \d)$. Incidentally, the time \tl\ has only a tiny dependence on \d.

234: 

235:  A similar effect is likely to be present in measures other than the

236:  causal diamond measure. For example, the scale factor measure

237:  \cite{dgsv} in the homogeneous approximation effectively weights

238:  vacua by the physical density of observers \cite{bfy}. The density of

239:  baryons at fixed time is proportional to $1/(1 + \d)$, so the same

240:  weighting factor is recovered.

241: 

242: While this effect is independent of the details of the dark matter, 

243: the prediction for the dark matter abundance depends on the 

244: prior distribution for the dark matter

245: density.

246: The amount of dark matter is determined by the initial axion

247: misalignment angle

248:  $\theta_i$, and symmetry guarantees that $\theta_i$ is randomly

249: distributed. 

250: For ``unnaturally'' small dark matter ratios, $\theta_i$ is near its

251: minimum, and $\d \propto \theta_i^2$. This leads to a 

252: prior distribution for the ratio of dark matter to ordinary

253: matter

254: \be

255: {d P \over d \d}  \propto {1 \over \sqrt{\d}}~\ \ {\rm (prior)}.

256: \ee

257: 

258: In principle, after computing the prior probability distribution  for

259: \d\ and the number of

260: baryons as a function of \d, one should compute the number of observations per baryon

261: as a function of \d. Here I follow Tegmark,

262: Aguirre, Rees, and Wilczek \cite{tarw} and assume that the number

263: of observations per baryon is roughly constant for dark matter ratios

264: in the range $2.5 < \d < 100$, while using the

265: approximation that outside of this range there are almost no observations per

266: baryon.  I find that 30\% of observers see less dark matter than we

267: do, while 70\% see more. Therefore, our observations are quite typical.

268: The

269: distribution is rather broad: in our approximation, 68\% of observers

270: see a dark matter ratio $\d \lesssim 15$, while 95\% of observers see

271: $\d \lesssim 65$

272: 

273: The results are rather robust, which is good because the conclusions

274: of \cite{tarw} are controversial. For example, Hellerman and Walcher

275: \cite{simeon} conclude that, to the best of our current understanding,

276:  observers may form up to $\d \approx

277: 10^5$. I show that the statistical prediction here is only mildly

278: sensitive to assumptions about observer formation at large $\d$,

279: unlike in the conventional anthropic analysis: assuming that observers

280: form efficiently all the way up to $\d = \infty$ only has the effect

281: of changing the number of observers who see less dark matter

282: than we do from 30\% to 25\%.

283: 

284: While the prediction for the dark matter abundance agrees well with

285: observation, it may not be absolutely persuasive to antianthropicists

286: because the probability

287: distribution is rather broad and the dark matter ratio has already

288: been measured. Confirmation of these ideas could come from

289: observation, because axionic dark matter has distinctive observational

290: signals in the form of isocurvature perturbations \cite{htw, kapnel}.

291: 

292: In the next section, I review the prior distribution for the dark

293: matter abundance in high-scale axion models. In section 3, I compute

294: the number of baryons inside the horizon as a function of \d\ and

295: describe the approximation for the number of observers per baryon. In

296: section 4, I combine these results to get a statistical prediction

297: for the dark matter abundance. In section 5, I mention several

298: interesting future directions.

299: 

300: \section{Prior Distribution}

301: We focus on a small part of the landscape. We fix the axion decay

302: constant to a large value, $f_a \gg 10^{12}$ GeV, and assume that the

303: energy scale of inflation is significantly smaller than $f_a$. We 

304: consider the set of vacua exactly like

305: ours, but with varying initial axion misalignment angle $\theta_i$. To

306: be concrete, we could imagine that slow roll inflation begins by a

307: tunneling event from a metastable false vacuum, 

308: and that the shift symmetry of the axion is

309: unbroken in the entire regime of interest. In other words, the axion

310: is basically a spectator in the tunneling event. Then every time a

311: bubble forms the axion is roughly homogeneous throughout the interior

312: of the bubble, with a random initial misalignment angle.

313:  

314: Once the universe cools sufficiently, the potential for the axion

315: becomes important. The axion field is approximately homogeneous in one

316: horizon volume and acts like dark matter \cite{axbound}. The amount of dark matter

317: relative to baryonic matter is determined by the misalignment angle.

318: In models with a large axion decay constant, the

319: natural value for the dark matter abundance is much larger than what

320: we observe, so values near the observed value correspond to the axion

321: being rather close to the minimum of its potential. In this regime,

322: the potential is approximately quadratic, so the dark matter abundance

323: is

324: \begin{equation}

325: \d \propto \theta_i^2~.

326: \end{equation}

327: Therefore, the prior probability distribution for $\sqrt{\d}$ is flat

328: between zero and some large value where our quadratic approximation

329: breaks down.

330: Changing variables, the prior distribution for the dark matter ratio

331: $\d$ in the scenario with a large axion decay constant is

332: \be

333: { dP \over d {\d} }\propto {1 \over \sqrt{\d}} \ \ \ {\rm (prior).}

334: \ee

335: 

336: \section{Anthropic Considerations}

337: We now compute the effect of increasing the dark matter abundance on

338: the number of observers. The analysis in this section is independent

339: of the details of the dark matter. However, inspired by axion

340: models, 

341: we hold fixed the

342: baryon to photon ratio, and simply increase the amount of dark

343: matter. This means that at decoupling, the density of baryons is

344: independent of the dark matter abundance, while the total matter

345: density at decoupling increases. Since we are assuming a flat

346: universe, the Hubble parameter at decoupling will

347: increase as the dark matter abundance increases.

348: 

349: We first compute the number of baryons inside the horizon as a

350: function of time, then discuss the number of observations per

351: baryon. 

352: 

353: \subsection{Number of baryons inside the horizon}

354: We are simply counting the number of baryons in the backward lightcone

355: of future infinity as a function of time.\footnote{Smaller causal diamonds have been used in

356: \cite{harnik}, but the fundamental definition of the causal diamond measure is to

357: count everything within a large causal diamond which extends beyond

358: our own bubble. This is equivalent, within our bubble, to counting

359: everything in the backward lightcone of future infinity.}

360: The number of baryons inside the horizon can be computed in the

361: homogeneous approximation. Assuming a flat universe, the metric is

362: \be

363: ds^2 = - dt^2 + a^2(t) (dx^2 + dy^2 + dz^2)

364: \ee

365: We begin at cosmological constant domination and work backwards.

366:  By

367: definition, $t = t_\Lambda$ when $\rho_m = \rho_\Lambda$. Therefore,

368: \be

369: \rho_b (t_\Lambda) = \rho_\Lambda / (1 + \d)~.

370: \ee

371: The horizon volume at the time $t_\Lambda$ is set by the de Sitter

372: radius, independent of the dark matter abundance. Therefore the baryon

373: mass inside the horizon at the time $t_\Lambda$ is

374: \be

375: M_b (t_\Lambda) \propto 1/(1 + \d)~.

376: \ee

377: More generally, since observations can be made at any time, we need

378: the number of baryons inside the horizon as a function of time. 

379: Note that both the Hubble parameter at vacuum domination and the

380: total matter energy density $\rho_m$ at vacuum domination are independent of the dark

381: matter ratio $\d$ in the approximation that radiation makes up a

382: negligible fraction of the energy density at \tl. 

383: Therefore, the scale factor as a function of time

384: from vacuum domination, $a(t - \tl)$, and the total matter density,

385: $\rho_m(t - \tl)$, are independent of the ratio \d. Since the scale

386: factor is unaffected by \d, the volume inside the horizon, $V(t -

387: \tl)$ is also independent of \d. On the other hand,

388: the energy density in radiation, $\rho_\gamma$, does depend on \d. 

389: 

390: Therefore, as long as radiation is a negligible fraction of the energy

391: density, the total mass inside the horizon, $M(t - \tl)$, is {\it

392:   independent} of the dark matter ratio \d. So the number of baryons

393: inside the horizon, for all times with negligible energy density in

394: radiation, is given by

395: \be

396: N_b (t- \tl) = {1 \over 1 + \d} N_0 (t - \tl)

397: \ee

398: where $N_0 (t - \tl)$ is a universal function which does not depend on

399: \d.

400: 

401: Of course the dark matter ratio does affect the early universe. The

402: baryon to photon ratio is fixed, so at decoupling the ratio $\rho_b/

403: \rho_\gamma$ 

404: is independent of $\d$. So the total

405: matter density at decoupling depends on \d, 

406: \be

407: \rho_m (t_{DC}) \propto 1 + \d

408: \ee

409: Therefore matter-radiation equality happens at a higher temperature as

410: \d\ is increased. Quantitatively, the matter-radiation ratio redshifts

411: as

412: \be

413: {\rho_m \over \rho_\gamma} \sim a

414: \ee

415: and the temperature redshifts as $T \sim 1/a$,

416: so

417: \be

418: T_{eq} \propto 1 + \d ~.

419: \ee

420: The matter density is also higher at equality,

421: \be

422: \rho_m (t_{eq}) \propto T_{eq}^4 \propto (1 + \d)^4

423: \ee

424: 

425: To summarize, the bottom line is that \d\ affects the early universe, changing the

426: time and temperature at matter-radiation equality. But for all values

427: of \d\ the universe matches onto a universal late time behavior (in

428: the homogeneous approximation) as

429: soon as the energy density in radiation is much less than than the

430: energy density in matter. During this era, $a(t- \tl)$ and $\rho_m (t

431: - \tl)$ are both universal functions which are independent of \d, so

432: the total mass inside the horizon $M(t - \tl)$ is also independent of

433: \d. So the number of baryons at any time with negligible energy

434: density in radiation is a universal function times the factor $1/(1 + \d)$.

435: 

436: %% After

437: %% $t_\Lambda$, $M_b$ drops exponentially to zero as baryons are expelled

438: %% from the causal diamond. Before $t_\Lambda$, the universe is matter

439: %% dominated. During this period, the scale factor satsifies

440: %% \be

441: %% a(t) = a(t_\Lambda) \left(t \over t_\Lambda \right)^{2/3}

442: %% \ee

443: %% We need to know the volume inside the horizon as a function of

444: %% time. The horizon is a null surface, so it satisfies

445: %% \be

446: %% a dr = - dt

447: %% \ee

448: %% Plugging in the form of the scale factor, we have during matter domination

449: %% \be

450: %% r(t) = r(t_\Lambda) + 3 (t_\Lambda - t)^{1/3} t_\Lambda^{2/3} /

451: %% a(t_\Lambda)

452: %% \ee

453: %% The number of baryons inside the horizon is proportional to the

454: %% comoving volume inside the horizon, which is proportional to $r^3$. So

455: %% \be

456: %% M_b (t) = M_b(\tl) \left[1 + (1 - t/\tl)^{1/3}\right]^3

457: %% \ee

458: 

459: \subsection{The number of observations per baryon}

460: Now in principle, we should count the number of observations inside

461: the causal patch by

462: multiplying the number of baryons by the number of observations per

463: baryon per unit time:

464: \be

465: N_{obs} = \int dt N_b (t) f(t)

466: \ee

467: where $f(t)$ is the number of observations per baryon per unit

468: time. One estimate for the function $f(t)$ is that it is

469: proportional to the collapsed baryon fraction. We will not try to

470: compute $f$ in detail here, because it is quite difficult and much

471: work has already been done. 

472: 

473: To proceed, we make a crude approximation. First we need a lower bound

474: on the amount of dark matter. As the amount

475: of dark matter is decreased, the size of the biggest nonlinear

476: structures decreases. According to Tegmark, Aguirre, Rees, and Wilczek

477: \cite{tarw}, density perturbations at scales similar to our galaxy

478: become nonlinear only if

479: \be

480: Q \gtrsim 10^{-5} \left( 1 + \d_0 \over 1 + \d \right)^{4/3}

481: \ee

482: where $\d_0 \approx 5$ is the dark matter ratio in our universe and

483: $Q \approx 2 \times 10^{-5}$ is the magnitude of the density perturbations. This

484: criterion gives

485: \be

486: \d_{\rm min} \approx 2.5 ~.

487: \ee

488: It is interesting to note that while the above bound depends on $Q$ and $\Lambda$,

489: \cite{tarw} give other bounds on the dark matter ratio which are more

490: general. 

491: For example, struture formation is seriously

492: impeded by Silk damping when $\d \lesssim 1$, independent of both $Q$

493: and $\Lambda$, so a lower bound in

494: the neighborhood of $\d \approx 1$ is somewhat robust.

495: 

496: What is the upper bound at which the amount of dark matter starts to

497: seriously impair observer formation? This is a controversial subject,

498: and in any case the most stringent upper bound in the literature is

499: that of \cite{tarw} who find $\d \lesssim 10^2$. On the other hand,

500: Hellerman and Walcher \cite{simeon} find no compelling evidence for a bound stronger

501: than $\d \lesssim 10^5$. One additional consideration in the causal

502: diamond measure is that the number of baryons inside the causal

503: diamond becomes very small after the cosmological constant dominates

504: the vacuum energy, so if for large \d\ observer formation is delayed

505: past the time \tl, the probability of observing such large \d\ will be suppressed.

506: 

507: Here, we will use the results of \cite{tarw} and approximate the

508: number of observers per baryon as a constant in the range

509: \be

510: 2.5 < \d < 100

511: \ee

512: while assuming that there are approximately zero observers per baryon

513: outside this range. We will also

514: demonstrate that the prediction is robust against changing the

515: assumptions.

516: Clearly, there is room for improvement in these considerations.

517: 

518: \section{Prediction for the Dark Matter Abundance}

519: Now we can combine the known prior distribution for axion dark matter

520: with the anthropic counting of observers to generate the probability

521: distribution for the observed dark matter abundance. To get the final

522: probability distribution for \d, we multiply the prior

523: distribution

524: \be

525: {d P \over d \d} \propto {1 \over \sqrt{\d}} \ \ \ {\rm (prior)}

526: \ee

527: by the number of baryons inside the causal patch,

528: \be

529: N_b \propto {1 \over 1 + \d}

530: \ee

531: and the number of observations per baryon,

532: \bea

533: {N_{obs} \over N_b} &\approx& {\rm constant,\ for}\ \ 2.5 < \d < 100

534: \nonumber \\

535: {N_{obs} \over N_b} &\approx& 0 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\rm otherwise.}

536: \eea

537: Thus the final probability distribution

538: for the dark matter abundance

539: \bea

540: {d P \over d \d} &\propto&

541:  {1 \over \sqrt{\d} (1 + \d)} \ \ \ \ 2.5 < \d < 100 \nonumber \\

542: {d P \over d \d} &\approx& 0 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\rm otherwise}

543: \label{dist}

544: \eea

545: This probability distribution is pictured in figure 1, along with the

546: ``conventional'' anthropic probability distribution, which instead

547: counts the number of observers per baryon.

548: \begin{figure}[!htb]

549: \center

550: \includegraphics [scale=1]

551: {prob} \caption{The probability distribution for the dark matter ratio \d\

552: is plotted as a function of $\sqrt{\d}$. The conventional anthropic analysis

553: gives approximately a flat distribution in the range $2.5 \lesssim \d

554: \lesssim 100$ (horizontal line), while the causal diamond measure

555: gives an additional factor of $1/(1 + \d)$ (curved line). The vertical

556: line shows the observed value $\d = 5$. In the conventional analysis

557: only about 8\% of observers form in regions with less dark matter than

558: we observe, even with the most favorable assumptions about observer

559: formation.  In

560: the causal diamond measure the observed value is fairly typical:

561: 30\% of observers form in regions with less dark matter than we observe. 

562: The causal diamond

563: prediction is rather insensitive to assumptions about observer

564: formation at large \d.}

565: \label{fig-pot}

566: \end{figure}

567: 

568: To address the typicality of our observations, it makes sense to

569: compute how much of the probability distribution is on each side of

570: the observed value. Integrating (\ref{dist}), we find that 30\% of

571: the probability is at smaller \d, while 70\% is at larger \d, so our

572: observations are quite typical. The probability distribution is fairly

573: broad: 68\% of observers form in regions with $\d \lesssim 15$, and

574: 95\% of observers form in regions with $\d \lesssim 65$.

575: 

576: Since observer formation at large \d\

577: is poorly understood, let's see what happens to our distribution if we

578: asssume that the number of observations per baryon is constant all the

579: way up to $\d = \infty$. With this assumption, 25\% of observers form

580: in regions with less dark matter than us, so our observations are

581: still fairly typical. Since the prediction is not significantly

582: affected, it is not urgent to understand observer formation for very

583: large dark matter ratios $\d > 100$.

584: 

585: To probe the sensitivity of our results further, assume for a moment

586: that the number of observers per baryon is constant for $1 < \d <

587: 100$. With this assumption, 53\% of observers form in regions with

588: less dark matter than we have. While it is probably not realistic to

589: think that structure formation in our universe can proceed effectively

590: down to $\d \approx 1$, as we mentioned in the previous section once

591: $Q$ and $\Lambda$ are allowed to vary the lower bound $\d_{min}

592: \approx 2.5$ is not valid in general, while the lower bound $\d_{min}

593: \approx 1$ is valid \cite{tarw}.

594: 

595: \section{Future Directions}

596: It will be interesting to extend these considerations to a larger

597: landscape in which more parameters are allowed to vary. It is

598: particularly interesting to ask what happens once the cosmological

599: constant and the density contrast $Q$ are allowed to vary along with

600: the dark matter ratio.

601: A more precise prediction within the small landscape studied here would

602: require an improved analysis of observer formation for a range of dark

603: matter ratios $0 < \d \lesssim 1000$, although as we explained large

604: values of \d\ have only a small effect on the prediction.

605: 

606: In supersymmetric field theory and string theory models, one has to

607: worry about cosmological problems arising from the Saxion and other

608: moduli \cite{saxion}. It is important to verify that the positive

609: results found here survive in a more detailed model.  More generally,

610: it would be interesting to repeat the statistical arguments here in

611: the presence of other sources of dark matter in addition to the axion.

612: 

613: An additional interesting direction is the search for observational

614: confirmation of these ideas. As discussed recently by \cite{thomas,

615:   htw, kapnel}

616: axionic dark matter

617: has observational signatures in the form of isocurvature

618: perturbations; these signals are enhanced when the axion misalignment

619: angle is ``unnaturally'' small due to anthropic selection \cite{kapnel}. 

620: These perturbations can be produced during inflation if

621: the inflation scale is high enough. They can also be produced prior to

622: inflation and survive observationally if inflation does not go on for

623: too long.  In fact, Kaplan and Nelson \cite{kapnel} conclude that for

624: some parameter choices, isocurvature perturbations from the

625: preinflationary era will be observable if there are fewer than about 74

626: efoldings of slow roll inflation. (This is in conventions where the

627: bound coming from the observed flatness of the universe corresponds to

628: 60 efoldings.) Combined with arguments \cite{fkms}

629: that inflation is quite likely to last for fewer than 74 efoldings, 

630: this is a very exciting

631: possibility which deserves further study.

632: 				

633: \subsubsection*{Acknowledgements}

634: I would like to thank Raphael Bousso, Simeon Hellerman, Ann Nelson, 

635: Stefan Leichenauer, Leonard Susskind, and especially Jens Niemeyer for

636: helpful discussions.  This work was supported by the Berkeley Center

637: for Theoretical Physics, by NSF CAREER grant 0349351,

638: and by DOE grant DE-AC02-05CH11231.

639: 

640: \begin{thebibliography}{999}

641: 

642: %\cite{Bousso:2000xa}

643: \bibitem{bp}

644:   R.~Bousso and J.~Polchinski,

645:   ``Quantization of four-form fluxes and dynamical neutralization of the

646:   cosmological constant,''

647:   JHEP {\bf 0006}, 006 (2000)

648:   [arXiv:hep-th/0004134].

649:   %%CITATION = JHEPA,0006,006;%%

650: 

651: %\cite{Kachru:2003aw}

652: \bibitem{kklt}

653:   S.~Kachru, R.~Kallosh, A.~Linde and S.~P.~Trivedi,

654:   ``De Sitter vacua in string theory,''

655:   Phys.\ Rev.\  D {\bf 68}, 046005 (2003)

656:   [arXiv:hep-th/0301240].

657:   %%CITATION = PHRVA,D68,046005;%%

658: 

659: %\cite{Susskind:2003kw}

660: \bibitem{lenny}

661:   L.~Susskind,

662:   ``The anthropic landscape of string theory,''

663:   arXiv:hep-th/0302219.

664:   %%CITATION = HEP-TH/0302219;%%

665: 

666: %cc stuff 

667: %\cite{Weinberg:1987dv}

668: \bibitem{cc}

669:   S.~Weinberg,

670:   ``Anthropic Bound on the Cosmological Constant,''

671:   Phys.\ Rev.\ Lett.\  {\bf 59}, 2607 (1987).

672:   %%CITATION = PRLTA,59,2607;%%

673: 

674: %\cite{Vilenkin:1994ua}

675: %\bibitem{Vilenkin:1994ua}

676:   A.~Vilenkin,

677:   ``Predictions from quantum cosmology,''

678:   Phys.\ Rev.\ Lett.\  {\bf 74}, 846 (1995)

679:   [arXiv:gr-qc/9406010].

680:   %%CITATION = PRLTA,74,846;%%

681: 

682: %\cite{Weinberg:1996xe}

683: %\bibitem{Weinberg:1996xe}

684:   S.~Weinberg,

685:   ``Theories of the cosmological constant,''

686:   arXiv:astro-ph/9610044.

687:   %%CITATION = ASTRO-PH/9610044;%%

688: 

689: %\cite{Martel:1997vi}

690: %\bibitem{Martel:1997vi}

691:   H.~Martel, P.~R.~Shapiro and S.~Weinberg,

692:   ``Likely Values of the Cosmological Constant,''

693:   Astrophys.\ J.\  {\bf 492}, 29 (1998)

694:   [arXiv:astro-ph/9701099].

695:   %%CITATION = ASJOA,492,29;%%

696: 

697: 

698: %\cite{Peccei:1977hh}

699: \bibitem{pq}

700:   R.~D.~Peccei and H.~R.~Quinn,

701:   ``CP Conservation In The Presence Of Instantons,''

702:   Phys.\ Rev.\ Lett.\  {\bf 38}, 1440 (1977).

703:   %%CITATION = PRLTA,38,1440;%%

704: 

705: %\cite{Preskill:1982cy}

706: \bibitem{axbound}

707:   J.~Preskill, M.~B.~Wise and F.~Wilczek,

708:   ``Cosmology of the invisible axion,''

709:   Phys.\ Lett.\  B {\bf 120}, 127 (1983).

710:   %%CITATION = PHLTA,B120,127;%%

711: 

712: %\cite{Abbott:1982af}

713: %\bibitem{Abbott:1982af}

714:   L.~F.~Abbott and P.~Sikivie,

715:   ``A cosmological bound on the invisible axion,''

716:   Phys.\ Lett.\  B {\bf 120}, 133 (1983).

717:   %%CITATION = PHLTA,B120,133;%%

718: 

719: %\cite{Dine:1982ah}

720: %\bibitem{Dine:1982ah}

721:   M.~Dine and W.~Fischler,

722:   ``The not-so-harmless axion,''

723:   Phys.\ Lett.\  B {\bf 120}, 137 (1983).

724:   %%CITATION = PHLTA,B120,137;%%

725: 

726: %axion anthropics

727: 

728: %\cite{Linde:1987bx}

729: \bibitem{axanth}

730:   A.~D.~Linde,

731:   ``Inflation And Axion Cosmology,''

732:   Phys.\ Lett.\  B {\bf 201}, 437 (1988).

733:   %%CITATION = PHLTA,B201,437;%%

734: 

735: %\cite{Turner:1990uz}

736: %\bibitem{Turner:1990uz}

737:   M.~S.~Turner and F.~Wilczek,

738:   ``Inflationary axion cosmology,''

739:   Phys.\ Rev.\ Lett.\  {\bf 66}, 5 (1991).

740:   %%CITATION = PRLTA,66,5;%%

741: 

742: %\cite{Linde:1991km}

743: %\bibitem{Linde:1991km}

744:   A.~D.~Linde,

745:   ``Axions in inflationary cosmology,''

746:   Phys.\ Lett.\  B {\bf 259}, 38 (1991).

747:   %%CITATION = PHLTA,B259,38;%%

748: 

749: %\cite{Garriga:2002tq}

750: %\bibitem{Garriga:2002tq}

751:   J.~Garriga and A.~Vilenkin,

752:   ``Testable anthropic predictions for dark energy,''

753:   Phys.\ Rev.\  D {\bf 67}, 043503 (2003)

754:   [arXiv:astro-ph/0210358].

755:   %%CITATION = PHRVA,D67,043503;%%

756: 

757: %\cite{Garriga:2003hj}

758: %\bibitem{Garriga:2003hj}

759:   J.~Garriga, A.~Linde and A.~Vilenkin,

760:   ``Dark energy equation of state and anthropic selection,''

761:   Phys.\ Rev.\  D {\bf 69}, 063521 (2004)

762:   [arXiv:hep-th/0310034].

763:   %%CITATION = PHRVA,D69,063521;%%

764: 

765: %\cite{Wilczek:2004cr}

766: %\bibitem{Wilczek:2004cr}

767:   F.~Wilczek,

768:   ``A model of anthropic reasoning, addressing the dark to ordinary matter

769:   coincidence,''

770:   arXiv:hep-ph/0408167.

771:   %%CITATION = HEP-PH/0408167;%%

772: 

773: %\cite{Hellerman:2005yi}

774: \bibitem{simeon}

775:   S.~Hellerman and J.~Walcher,

776:   ``Dark matter and the anthropic principle,''

777:   Phys.\ Rev.\  D {\bf 72}, 123520 (2005)

778:   [arXiv:hep-th/0508161].

779:   %%CITATION = PHRVA,D72,123520;%%

780: 

781: %\cite{tarw}

782: \bibitem{tarw}

783:   M.~Tegmark, A.~Aguirre, M.~Rees and F.~Wilczek,

784:   ``Dimensionless constants, cosmology and other dark matters,''

785:   Phys.\ Rev.\  D {\bf 73}, 023505 (2006)

786:   [arXiv:astro-ph/0511774].

787:   %%CITATION = PHRVA,D73,023505;%%

788: 

789: %\cite{htw}

790: \bibitem{htw}

791:   M.~P.~Hertzberg, M.~Tegmark and F.~Wilczek,

792:   ``Axion Cosmology and the Energy Scale of Inflation,''

793:   arXiv:0807.1726 [astro-ph].

794:   %%CITATION = ARXIV:0807.1726;%%

795: 

796: %% %\CiteChoi:1985je}

797: %% \bibitem{Choi:1985je}

798: %%   K.~Choi and J.~E.~Kim,

799: %%   ``Harmful Axions In Superstring Models,''

800: %%   Phys.\ Lett.\  B {\bf 154}, 393 (1985)

801: %%   [Erratum-ibid.\  {\bf 156B}, 452 (1985)].

802: %%   %%CITATION = PHLTA,B154,393;%%

803: 

804: %\cite{Bousso:2006ev}

805: \bibitem{cd}

806:   R.~Bousso,

807:   ``Holographic probabilities in eternal inflation,''

808:   Phys.\ Rev.\ Lett.\  {\bf 97}, 191302 (2006)

809:   [arXiv:hep-th/0605263].

810:   %%CITATION = PRLTA,97,191302;%%

811: 

812: 

813: \bibitem{harnik}

814:   R.~Bousso, R.~Harnik, G.~D.~Kribs and G.~Perez,

815:   ``Predicting the Cosmological Constant from the Causal Entropic Principle,''

816:   Phys.\ Rev.\  D {\bf 76}, 043513 (2007)

817:   [arXiv:hep-th/0702115].

818:   %%CITATION = PHRVA,D76,043513;%%

819: 

820: J.~M.~Cline, A.~R.~Frey and G.~Holder,

821:   ``Predictions of the causal entropic principle for environmental conditions

822:   of the universe,''

823:   Phys.\ Rev.\  D {\bf 77}, 063520 (2008)

824:   [arXiv:0709.4443 [hep-th]].

825:   %%CITATION = PHRVA,D77,063520;%%

826: 

827: 

828: %\cite{Bousso:2007er}

829: \bibitem{by}

830:   R.~Bousso and I.~S.~Yang,

831:   ``Landscape Predictions from Cosmological Vacuum Selection,''

832:   Phys.\ Rev.\  D {\bf 75}, 123520 (2007)

833:   [arXiv:hep-th/0703206].

834:   %%CITATION = PHRVA,D75,123520;%%

835: 

836: %\cite{DeSimone:2008bq}

837: \bibitem{dgsv}

838:   A.~De Simone, A.~H.~Guth, M.~P.~Salem and A.~Vilenkin,

839:   ``Predicting the cosmological constant with the scale-factor cutoff

840:   %measure,''

841:   arXiv:0805.2173 [hep-th].

842:   %%CITATION = ARXIV:0805.2173;%%

843: 

844: % \bibitem{Linde:1993nz}

845:   A.~D.~Linde and A.~Mezhlumian,

846:   ``Stationary universe,''

847:   Phys.\ Lett.\  B {\bf 307}, 25 (1993)

848:   [arXiv:gr-qc/9304015].

849:   %%CITATION = PHLTA,B307,25;%%

850: 

851: A.~D.~Linde, D.~A.~Linde and A.~Mezhlumian,

852:   ``From the Big Bang theory to the theory of a stationary universe,''

853:   Phys.\ Rev.\  D {\bf 49}, 1783 (1994)

854:   [arXiv:gr-qc/9306035].

855:   %%CITATION = PHRVA,D49,1783;%%

856: 

857: 

858: %\cite{Bousso:2008hz}

859: \bibitem{bfy}

860:   R.~Bousso, B.~Freivogel and I.~S.~Yang,

861:   ``Properties of the scale factor measure,''

862:   arXiv:0808.3770 [hep-th].

863:   %%CITATION = ARXIV:0808.3770;%%

864: 

865: 

866: %\cite{Banks:1996ea}

867: \bibitem{saxion}

868:   T.~Banks and M.~Dine,

869:   ``The cosmology of string theoretic axions,''

870:   Nucl.\ Phys.\  B {\bf 505}, 445 (1997)

871:   [arXiv:hep-th/9608197].

872:   %%CITATION = NUPHA,B505,445;%%

873: 

874: %\cite{Banks:2002sd}

875: %\bibitem{Banks:2002sd}

876:   T.~Banks, M.~Dine and M.~Graesser,

877:   ``Supersymmetry, axions and cosmology,''

878:   Phys.\ Rev.\  D {\bf 68}, 075011 (2003)

879:   [arXiv:hep-ph/0210256].

880:   %%CITATION = PHRVA,D68,075011;%%

881: 

882: 

883: %\cite{thomas}

884: \bibitem{thomas}

885:   P.~Fox, A.~Pierce and S.~D.~Thomas,

886:   ``Probing a QCD string axion with precision cosmological measurements,''

887:   arXiv:hep-th/0409059.

888:   %%CITATION = HEP-TH/0409059;%%

889: 

890: %\cite{Kaplan:2008ss}

891: \bibitem{kapnel}

892:   D.~B.~Kaplan and A.~E.~Nelson,

893:   ``Inflationary Axion Cosmology Beyond Our Horizon,''

894:   arXiv:0809.1206 [astro-ph].

895:   %%CITATION = ARXIV:0809.1206;%%

896: 

897: %\cite{Freivogel:2005vv}

898: \bibitem{fkms}

899:   B.~Freivogel, M.~Kleban, M.~Rodriguez Martinez and L.~Susskind,

900:   ``Observational consequences of a landscape,''

901:   JHEP {\bf 0603}, 039 (2006)

902:   [arXiv:hep-th/0505232].

903:   %%CITATION = JHEPA,0603,039;%%

904: 

905: 

906: 

907: \end{thebibliography}

908: 

909: \end{document}

910: 

911: 

912: