1: 

2: 

3: 

4: 

5: \documentclass[aps,prc,twocolumn]{revtex4}

6: \usepackage{graphicx,amssymb}

7: 

8: \begin{document}

9: 

10: \title{Local Causality and Completeness:  Bell \emph{vs.} Jarrett}

11: \author{Travis Norsen}

12: \affiliation{Marlboro College \\ Marlboro, VT  05344 \\ norsen@marlboro.edu}

13: 

14: \date{August 14, 2008}

15: 

16: \begin{abstract}

17: J.S. Bell believed that his famous theorem entailed a deep and

18: troubling conflict between the empirically verified predictions of

19: quantum theory and the notion of local causality that is motivated by

20: relativity theory.  Yet many physicists continue to accept, usually on 

21: the reports of textbook writers and other commentators, that Bell's

22: own view was wrong, and that, in fact, the theorem only brings out a

23: conflict with determinism or the hidden-variables program or realism

24: or some other such principle that (unlike local causality), allegedly,

25: nobody should have believed anyway.  (Moreover, typically such beliefs

26: arise without the person in question even being aware that the view

27: they are accepting differs so radically from Bell's own.)

28: Here we try to shed some light

29: on the situation by focusing on the concept of local causality that is

30: the heart of Bell's theorem, and, in particular, by contrasting Bell's

31: own understanding with the analysis of Jon Jarrett which has been the

32: most influential source, in recent decades, for the kinds of claims

33: mentioned previously.  We point out a crucial difference between 

34: Jarrett's and Bell's own understanding of Bell's formulation of 

35: local causality, which turns out to be the basis for the erroneous 

36: claim, made by Jarrett and many others, that Bell misunderstood 

37: the implications of his own theorem. 

38: \end{abstract}

39: 

40: \maketitle

41: 

42: 

43: \section{Introduction}

44: 

45: In 1964, J.S. Bell proved the result now known as Bell's

46: Theorem:  any physical theory of a certain

47: type must make predictions (for a certain class of experiment) which

48: respect a so-called Bell Inequality.  

49: \footnote{John S. Bell, \emph{Speakable and Unspeakable in Quantum

50:     Mechanics}, 2nd ed., Cambridge University Press, 2004.  Subsequent

51:   references to Bell's writings in the text will be given in-line with

52:   the year of the referenced paper and page numbers from the book.}

53: Quantum Mechanics (QM) predicts

54: violations of the inequality, and subsequent

55: experiments establish convincingly (though not without loopholes) that

56: the Quantum Mechanical predictions are correct -- i.e., the

57: experiments establish that 

58: the type of theory Bell showed must respect the Inequality, cannot be

59: empirically viable, i.e., cannot be true.  

60: \footnote{For a review of recent experiments and associated loopholes,

61:   see, e.g., Abner Shimony, ``Bell's Theorem'', \emph{The Stanford

62:   Encyclopedia of Philosophy} (Fall 2006 Edition), Edward N. Zalta

63: (ed.), URL = http://plato.stanford.edu/archives/fall2006/entries/bell-theorem/}

64: But the

65: question (which has given rise to an enormous literature) remains:

66: what type of theory is it, exactly, that Bell's Theorem (combined with

67: the associated experiments) refutes?

68: 

69: Bell's own view, expressed already in the opening lines of his 1964

70: paper and subsequently clarified and defended in virtually all of his

71: later writings, was that ``It is the requirement of locality ...

72: that creates the essential difficulty.''  (Bell, 1964, p. 14)

73: By ``locality'' Bell here means 

74: the prohibition, usually taken to be an implication of

75: special relativity (SR), of super-luminal (faster than light) 

76: causation.  Bell thus 

77: took his own theorem to establish a troubling conflict between (the

78: empirically verified predictions of) QM (i.e.,

79: between \emph{experiment}) and SR:

80: \begin{quote}

81: ``For me then this is the real problem with quantum theory:  the

82: apparently essential conflict between any sharp formulation and

83: fundamental relativity.  That is to say, we have an apparent

84: incompatibility, at the deepest level, between the two fundamental

85: pillars of contemporary theory...'' (Bell, 1984, p. 172)

86: \end{quote}

87: Most practicing physicists, however, (and most philosophers of

88: physics) have disagreed with Bell and continued to believe in the

89: unproblematic consistency of QM and SR.  Where do they think Bell went

90: wrong?  

91: 

92: One can divide reasons for disagreement (with Bell's own interpretation

93: of the significance of his theorem) into two classes.  First, there are

94: those who assert that the derivation of a Bell Inequality relies not

95: just on the premise of locality, but on some additional premises

96: as well.  The usual suspects here include Realism, Hidden Variables,

97: Determinism, and Counter-Factual-Definiteness.  (Note that the items

98: on this list are highly overlapping, and often commentators use them

99: interchangeably.)  The idea is then that, since it is only the

100: \emph{conjunction} of locality with some other premise which is in conflict

101: with experiment, and since locality is so strongly motivated by SR, we

102: should reject the other premise.  Hence the widespread reports that

103: Bell's theorem finally refutes the hidden variables program, the

104: principle of determinism, the philosophical notion of realism, etc.

105: \footnote{See, for example:  N. David Mermin, ``What is quantum

106:   mechanics trying to tell us?'' \emph{AmJPhys}, {\bf{66}}(9),

107:   September 1998, pg 753-767; Marek Zukowski, ``On the paradoxical

108:   book of Bell,'' \emph{Stud. Hist. Phil. Mod. Phys.}, {\bf{36}}

109:   (2005) 566-575; A. Zeilinger, ``The message of the quantum,''

110:   \emph{Nature} {\bf{438}}, 743 (8 December, 2005); Daniel Styer,

111:   \emph{The Strange World of Quantum Mechanics} (page 42), Cambridge,

112:   2000; George Greenstein and Arthur Zajonc, \emph{The Quantum

113:     Challenge} (Second Edition), Jones and Bartlett Publishers,

114:   Sudbury, Massachusetts, 2006; John Townsend, \emph{A Modern Approach

115:   to Quantum Mechanics}, McGraw-Hill, 1992; Herbert Kroemer

116: \emph{Quantum Mechanics}, Prentice Hall, New Jersey, 1994; Richard

117: Liboff, \emph{Introductory Quantum Mechanics} (2nd edition),

118: Addison-Wesley, Reading, Massachusetts, 1992}

119: 

120: Here is how Bell responded to this first class of disagreement:

121: \begin{quote}

122: ``My own first paper on this subject ... starts with a

123: summary of the EPR argument \emph{from locality to} deterministic

124: hidden variables.  But the commentators have almost universally

125: reported that it begins with deterministic hidden variables.''

126: (Bell, 1981, p. 157)

127: \end{quote}

128: Here

129: (a footnote, but also in the main text of the article in question)

130: Bell goes out of his way to stress the overall logical structure of

131: his \emph{two-part} argument:  first, an argument \emph{from} locality and

132: certain predictions of QM (namely, perfect anti-correlation for

133: parallel spin measurements on a pair of spin 1/2 particles in the spin

134: singlet state) \emph{to} the existence of deterministic local hidden

135: variables; and then second, from such variables to the inequality,

136: i.e., to a disagreement with certain other predictions of QM.  This

137: whole first class of disagreement with Bell, then, rests on a simple

138: confusion about Bell's argument. 

139: \footnote{For further discussion, see any of Bell's papers and, e.g.:

140:   Tim Maudlin, \emph{Quantum Non-Locality and Relativity} (Second

141:   Edition), Blackwell, Malden, Massachusetts, 2002;  Travis Norsen,

142:   ``Bell Locality and the Nonlocal Character of Nature,''

143:   \emph{Found. Phys. Lett.}, {\bf{19}}(7), 633-655 (Dec. 2006)}

144: 

145: 

146: The more interesting and more subtle second class of disagreement 

147: includes those who accept that the empirically-violated Bell

148: inequality can be derived from Bell's locality condition alone, but

149: who argue that this locality condition is \emph{too strong}, i.e., 

150: that it smuggles in some extra requirements beyond those

151: minimally necessary to respect SR's prohibition on superluminal

152: causation.  At the head of this class is Jon Jarrett, whose 1983 

153: PhD thesis and subsequent 1984 paper 

154: \footnote{Jon Jarrett, ``On the Physical Significance of the Locality

155:   Conditions in the Bell Arguments,'' \emph{Nous} {\bf{18}} (1984) 569-589}

156: argued that Bell's own local causality condition (which Jarrett calls 

157: ``strong locality'') is logically equivalent to

158: the conjunction of two subsidiary conditions, which Jarrett described

159: respectively as ``locality'' and ``completeness.'' 

160: 

161: Roughly speaking, Jarrett's ``locality'' is the requirement that the

162: outcome of a measurement on one particle be independent of the type of

163: measurement performed (at spacelike separation) on a second particle

164: (which, in the interesting sorts of cases, is described by QM as being

165: entangled with the first particle).  Jarrett's ``completeness'' on the

166: other hand requires the outcome of the first measurement to be

167: independent of the outcome of the second, spacelike separated

168: measurement.  He then argues that a violation of ``locality'' would

169: entail the ability to send superluminal signals (Alice could learn

170: something about the setting of Bob's measurement device by examining

171: the outcome of her own experiment) which would allow Bob to transmit

172: a signal across a spacelike interval, something clearly forbidden

173: by SR.  By contrast, a violation of ``completeness'' indicates no

174: conflict with SR.  Thus, Jarrett

175: argues, in the face of the empirical data conflicting with ``strong

176: locality'' we may reject ``completeness'' and thereby achieve -- contra

177: Bell -- a kind of Peaceful Coexistence between QM and SR.  

178: \footnote{``Peaceful Coexistence'' is Abner Shimony's term:  

179:   ``Metaphysical problems in the foundations of quantum mechanics,''

180:   \emph{International Philosophical Quarterly} 18, 3-17.}

181: 

182: Jarrett's project has been widely hailed and widely discussed.  It was

183: the immediate stimulus for almost everything in the ``Bell

184: literature'' for about a decade after its appearance, and continues to

185: set a broadly influential context for much ongoing work in this area.

186: \footnote{See, for example, M.L.G. Redhead, \emph{Incompleteness,

187:     Nonlocality, and Realism: A Prolegomenon to the Philosophy of

188:     Quantum Mechanics} Oxford, 1987; J. Cushing and E. McMullin, eds.,

189:   \emph{Philosophical Consequences of Quantum Theory}, Notre Dame,

190:   1989 (see especially the contributions of Paul Teller and Don

191:   Howard); Brandon Fogel, ``Formalizing the separability condition in

192:   Bell's theorem,'' \emph{Stud. Hist. Phil. Mod. Phys.}, 38 (2007),

193:   920-937} 

194: 

195: The purpose of the present paper is to critically assess

196: Jarrett's analysis and

197: the conclusions it led him to, by comparing his project side-by-side

198: with Bell's own discussions of the relevant issues.  In particular, I

199: will claim that Jarrett simply missed a crucial aspect (having

200: something to do with ``completeness'') of Bell's

201: formulation of local causality; this turns out to be the heart of the

202: thinking behind Jarrett's (\emph{prima facie} rather puzzling) terminology

203: for his two sub-conditions, as well as his central claim that

204: violations of his ``completeness'' criterion indicate no conflict with

205: special relativistic local causality.  The main conclusion is thus

206: that, contrary to Jarrett and his followers, Bell's own local 

207: causality criterion is in no sense ``too strong.''  And this of course

208: undermines the attempt to establish the Peaceful Coexistence of QM and

209: SR, i.e., it supports Bell's own interpretation of the meaning of his

210: theorem. 

211: 

212: The following two

213: sections present, respectively, Bell's own (final and most careful)

214: formulation of the locality premise, and then Jarrett's analysis.  

215: Section IV includes some comparative discussion, highlighting

216: especially the relation of Jarrett's thinking to the EPR argument.  A

217: brief final section then summarizes and concludes.

218: 

219: Before launching into this, however, it is appropriate to briefly

220: survey earlier criticisms of Jarrett's analysis of Bell's locality

221: concept.  To my knowledge, the fullest critical discussion of

222: Jarrett's project is in Maudlin's book. \footnote{\emph{Op cit., 

223: pp. 93-98}}  

224: Maudlin's main points \emph{vis-a-vis} Jarrett are as follows:  

225: (a) Jarrett's identification of his ``locality'' sub-condition with

226: the prohibition on superluminal signals is wrong; (b) Jarrett's

227: identification of superluminal signaling with superluminal causation

228: is wrong; (c) Jarrett's claim that a violation of his ``completeness''

229: condition does not entail any nonlocal causation is wrong.

230: \footnote{Similar points are made in Jeremy Butterfield, ``Bell's

231:   Theorem: What it Takes'' \emph{Brit. J. Phil. Sci.}, {\bf{43}}

232:   (1992) 41-83.  Butterfield, however, (unlike Maudlin) ends up

233:   accepting both the validity/meaningfulness of Jarrett's analysis and

234:   Jarrett's ultimate conclusion about Peaceful Coexistence, 

235:   even though he disagrees with some of Jarrett's arguments.  Since I

236:   disagree with Butterfield's conclusions, I find him a less

237:   convincing overall critic of Jarrett's project than Maudlin.  My

238:   objection to Butterfield's proposed route to Peaceful Coexistence --

239:   which, like the present discussion of Jarrett, involves the claim 

240:   that the commentator has missed or misunderstood a crucial element

241:   of Bell's concept of local causality -- will be presented elsewhere.}

242: 

243: The arguments for (a) and (b) are made clearly and compellingly by

244: Maudlin, who thus really demolishes Jarrett's erroneous identification

245: of his ``locality'' with the relevant requirements of SR.  But the

246: case for (c) is made only indirectly, essentially by dismissing

247: Jarrett and re-asserting Bell's claim to the contrary.  So, to be a

248: bit more precise about the goal of the present paper, the aim is to

249: fill this gap by exploring in detail how Jarrett's ``completeness''

250: condition relates to Bell's local causality criterion and how Jarrett's

251: misunderstanding of the latter led him to the various erroneous

252: conclusions.  

253: 

254: I should note at the outset, however, that the view to be presented

255: here as Jarrett's is almost certainly a bit misleading as to his 

256: (or those I consider his followers') fully-considered views.  Jarrett

257: does actually say all the things I attribute to him, but my gloss will

258: perhaps minimize the extent to which Jarrett (I would argue,

259: inconsistently) also acknowledges points in conflict with the views I

260: will attribute to him.  It is probably best, therefore, to understand

261: the ``Jarrett'' discussed here as a rhetorically clarifying construct,

262: which may or may not correspond to the views of the actual Jon

263: Jarrett.

264: 

265: 

266: 

267: \section{Bell's Concept of Local Causality}

268: 

269: 

270: 

271: \begin{figure}[t]

272: \begin{center}

273: \includegraphics[width=3.3in,clip]{./fig1.eps}

274: \end{center}

275: \caption{

276: \label{fig1}

277: ``Space-time location of causes and effects of events in region 1.''

278: (Figure and caption are from Bell, 1990, p. 239.)

279: }

280: \end{figure}

281: 

282: 

283: 

284: 

285: Bell's fullest and evidently most-considered discussion of local

286: causality occurs in his last published paper, \emph{La nouvelle

287: cuisine} (1990, 232-248).  We will here essentially follow that discussion,

288: supplementing it occasionally with things from his earlier papers.

289: 

290: Bell first introduces what he calls the ``Principle of local

291: causality'' as follows:  

292: ``The direct causes (and effects) of events are near by, and even the

293: indirect causes (and effects) are no further away than permitted by

294: the velocity of light.''

295: Then, referencing what has been reproduced here as Figure 1, Bell

296: elaborates:  ``Thus, for events in a space-time region 1 ... we would

297: look for causes in the backward light cone, and for effects in the

298: future light cone.  In a region like 2, space-like separated from 1,

299: we would seek neither causes nor effects of events in 1.  Of course

300: this does not mean that events in 1 and 2 might not be correlated...''  

301: (1990, p. 239)

302: 

303: 

304: After remarking that this formulation ``is not yet sufficiently sharp

305: and clean for mathematics,'' Bell then proposes the following 

306: version, referencing what has been reproduced here as Figure 2:

307: \begin{quote}

308: ``A theory will be said to be locally causal if the probabilities

309: attached to values of local beables in a space-time region 1 are

310: unaltered by specification of values of local beables in a space-like

311: separated region 2, when what happens in the backward light cone of 1

312: is already sufficiently specified, for example by a full specification

313: of local beables in a space-time region 3...'' (1990, 239-40)

314: \end{quote}

315: Although Bell doesn't immediately formulate this mathematically (which

316: is curious, since he has just advertised it as a formulation which

317: \emph{is} ``sufficiently sharp and clean for mathematics''), we may do

318: so in a way that is clearly (as evidenced by what comes later in the

319: paper) what he had in mind:

320: \begin{equation}

321: P(b_1 | B_3, b_2) = P(b_1 | B_3).

322: \label{eq:loc}

323: \end{equation}

324: Here $b_i$ refers to some beable (or more precisely, its value) 

325: in region $i$, and $B_i$ refers to a

326: ``full specification'' of beables in region $i$.  This simply asserts 

327: mathematically what Bell states in the caption of his

328: accompanying figure:  ``full specification of [beables] in 3 makes 

329: events in 2 irrelevant for predictions about 1.''  

330: Note that Bell here uses the term (which he had earlier coined)

331: ``beable'' (rhymes with ``agreeable'')

332: to denote whatever is posited, by the candidate theory in question, 

333: to be \emph{physically real}:  

334: \begin{quote}

335: ``The beables of the theory are those elements which might

336: correspond to elements of reality, to things which exist.  Their 

337: existence does not depend on `observation'.  Indeed observation

338: and observers must be made out of beables.'' (1984, p. 174)

339: \end{quote}

340: For further discussion of ``beables'' see Bell's \emph{The Theory of

341:   Local Beables} (1975, pages 52-3), \emph{Beables for quantum field

342:   theory} (1984, pages 174-6), and \emph{La nouvelle cuisine} (1990,

343:   pages 234-5).  

344: 

345: 

346: 

347: \begin{figure}[t]

348: \begin{center}

349: \includegraphics[width=3.3in,clip]{./fig2.eps}

350: \end{center}

351: \caption{

352: \label{fig2}

353: ``Full specification of what happens in 3 makes events in 2 irrelevant

354: for predictions about 1 in a locally causal theory.''  

355: (Figure and caption are from Bell, 1990, p. 240.)

356: }

357: \end{figure}

358: 

359: 

360: 

361: Bell then adds the following clarificatory remarks:

362: \begin{quote}

363: ``It is important that region 3 completely shields off from 1 the

364: overlap of the backward light cones of 1 and 2.  And it is important

365: that events in 3 be specified completely.  Otherwise the traces in

366: region 2 of causes of events in 1 could well supplement whatever else

367: was being used for calculating probabilities about 1.  The hypothesis

368: is that any such information about 2 becomes redundant when 3 is

369: specified completely.''  (1990, p. 240)

370: \end{quote}

371: It will be crucial to understand these remarks, so we shall briefly

372: elaborate.  

373: 

374: First, suppose that the region labeled $3$ in Figure 2

375: sliced across the backwards light cone of 1 at an earlier time, such

376: that it failed to ``completely shield off from 1 the overlap of the

377: backward light cones of 1 and 2.''  See, for example, the region 

378: labeled $3^*$ in Figure 3.  Why then would a violation of

379: Equation \ref{eq:loc} (but with $3 \rightarrow 3^*$) 

380: fail to necessarily indicate the presence of 

381: nonlocal causation?  Suppose we are dealing with a non-deterministic

382: (i.e., irreducibly stochastic, genuinely chancy) theory.  And suppose

383: that some event (a beable we shall call ``X'')  comes into existence

384: in the future of this

385: region $3^*$, such that it lies in the overlapping

386: backwards light cones of $1$ and $2$.  By assumption, the theory in

387: question did not allow the prediction of this beable on the basis of a

388: full specification of beables in $3^*$.  Yet once X comes into existence,

389: it can (in a way that is perfectly consistent with local causality)

390: influence events in both 1 and 2.  And there is therefore the

391: possibility that specification of events from 2 could allow one to infer

392: something about X from which one could in turn infer \emph{more}

393: about goings-on in 1 than one could have inferred originally from just

394: the full specification of beables in $3^*$.  In other words (Bell's): ``the

395: traces in region 2 of causes [such as our X] of events in 1

396: could well supplement whatever else was being used for calculating

397: probabilities about 1.''  This mechanism for producing violations of

398: Equation \ref{eq:loc} without any superluminal causation, however,   

399: will clearly not be available so long as ``region 3 completely shields

400: off from 1 the overlap of the backward light cones of 1 and 2.''   

401: 

402: 

403: \begin{figure}[t]

404: \begin{center}

405: \includegraphics[width=3.3in,clip]{./fig3.eps}

406: \end{center}

407: \caption{

408: \label{fig3}

409: Similar to Figure 2, except that region $3^*$ (unlike region 3 of

410: Figure 2) fails to shield off region 1 from the overlapping backward

411: light cones of regions 1 and 2.  Thus, (following the language of

412: Figure 2's caption) even full specification of  what happens in $3^*$ 

413: \emph{does not necessarily make} events in 2 irrelevant for 

414: predictions about 1 in a locally causal theory.  

415: }

416: \end{figure}

417: 

418: 

419: Bell's other clarification is also crucial.  Suppose that events in

420: region 3 of Figure 2 are \emph{not} specified completely.  We may

421: denote such an incomplete description by $\bar{B}_3$.  Then does a

422: violation of Equation \ref{eq:loc} (but with $B_3 \rightarrow \bar{B}_3$)

423: necessarily imply the existence of any nonlocal causation?  No, for it

424: would then be possible that some event X (again in the overlapping past

425: light cones of 1 and 2) influences both 1 and 2 such that

426: information about 2 could tell us something about X which in turn

427: could tell us something about 1 which we couldn't infer from $\bar{B}_3$.

428: We only need stipulate that the beables in region 3 which ``carry''

429: the causal influence from X to 1 are (among) those omitted by $\bar{B}_3$.

430: But since there is, by definition, no such omission in the \emph{complete}

431: specification $B_3$, this eventuality cannot arise, and a violation of

432: Equation \ref{eq:loc} \emph{must} indicate the existence of some

433: nonlocal causation, i.e., causal influences not respecting Bell's

434: original ``Principle of local causality'' (as displayed in Figure 1).  

435: \footnote{Note that Bell stresses the need for a \emph{complete}

436: specification of beables in the relevant space-time region already

437: in his 1975 paper \emph{The theory of local beables}:  

438: ``Now my intuitive notion of local causality is

439: that events in 2 should not be `causes' of events in 1, and vice

440: versa.  But this does not mean that the two sets of events should be

441: uncorrelated, for they could have common causes in the overlap of

442: their backward light cones.  It is perfectly intelligible then that if

443: [$B_3$] in [our Equation \ref{eq:loc}] 

444: does not contain a complete record of events ... it

445: can be usefully supplemented by information from region 2.  So in

446: general it is expected that [$P(b_1|B_3,b_2) \ne P(b_1|B_3)$.]

447: However, in the particular case that [$B_3$] contains already a

448: \emph{complete} specification of beables ... supplementary information

449: from region 2 could reasonably be expected to be redundant.'' 

450: (1975, p. 54)  Emphasis in original.  

451: This is especially relevant since we will eventually criticize

452: Jarrett for failing to appreciate (in his 1984 paper) this particular 

453: aspect.  So it shouldn't be thought that we are criticizing him for 

454: something Bell only understood and clarified later.  It is worth

455: noting, however, that there are some interesting differences between

456: Bell's 1975 and 1990 formulations of local causality; these will be

457: explored elsewhere, though, since they do not bear on Jarrett's analysis.}

458: 

459: 

460: It is worth noting that we cannot necessarily infer, from a violation

461: of Equation \ref{eq:loc}, that $b_2$ exerts any direct or indirect

462: causal influence on $b_1$.  It might be, for example,

463: that a violation of Equation \ref{eq:loc} is produced by some X-like

464: event lying in the future of region 3, which causally

465: influences both 1 and 2.  But in order to exert a \emph{local} causal

466: influence on 1, such an X would have to lie \emph{outside} the past light

467: cone of 2 -- and vice versa.  What is ensured by a violation of Equation

468: \ref{eq:loc} is thus only that \emph{some} violation of local

469: causality (as sketched in Figure 1) is being posited somewhere by the

470: theory in question.  Whether something in 2 is exerting a causal

471: influence on 1 (or vice versa, or neither) the mere violation of

472: Equation \ref{eq:loc} doesn't permit us to say.  

473: \footnote{This is why, despite being an improvement over Jarrett's

474:   terminology for the two sub-conditions to be discussed in Section

475:   III, Abner Shimony's terminology (``parameter independence'' for

476:   what Jarrett calls ``locality'' and ``outcome independence'' for

477:   what Jarrett calls ``completeness'') is also dubious.  For the

478:   terminology implies that a violation of one of the conditions

479:   entails that the event in question causally \emph{depends} on the distant

480:   ``parameter'' or ``outcome'' respectively.  But this need not be the

481:   case.}

482: 

483: Note that

484: everything in the above discussion refers to some particular candidate

485: physical theory.  For example, there is a tendency for misplaced

486: skepticism to arise from Bell's use of the concept of ``beables'' in 

487: the formulation of local causality.  This term strikes the ears of

488: those influenced by orthodox quantum philosophy as having a

489: metaphysical character and/or possibly committing one (already, in the

490: very definition of what it means for a theory to respect relativistic

491: local causality) to something unorthodox like ``realism'' or ``hidden

492: variables.''  Such concerns, however, are based on the failure to 

493: appreciate that

494: the concept ``beable'' is theory-relative.  ``Beable'' refers not to

495: what \emph{is} physically real, but to what some candidate theory

496: \emph{posits} as being physically real.  Bell writes:

497: ``I use the term

498: `beable' rather than some more committed term like `being' or `beer'

499: to recall the essentially tentative nature of any physical theory.

500: Such a theory is at best a \emph{candidate} for the description of

501: nature.  Terms like `being', `beer', `existent', etc., would seem to

502: me lacking in humility.  In fact `beable' is short for `maybe-able'.''

503: (1984, p. 174)

504: 

505: Similar considerations apply to the notion of ``completeness'' that

506: is, as stressed above, essential to Bell's

507: formulation.   A  complete specification of

508: beables in some spacetime region simply means a specification of

509: everything (relevant) that is posited by the candidate theory in

510: question.  There is no presumption that such a full specification

511: actually correspond to what \emph{really exists} in the relevant

512: spacetime region, i.e., no presumption that the candidate theory in

513: question is \emph{true}.

514: And the same goes for the probabilities in Equation

515: \ref{eq:loc} that Bell's locality criterion is formulated in terms

516: of.  These should be read not as empirical frequencies or subjective

517: measures of expectation, but as the fundamental dynamical

518: probabilities described by the candidate theory in question (which we

519: assume, without loss of generality, to be irreducibly stochastic).

520: \footnote{Determinism is simply a special case in which all

521:   probabilities are either zero or unity.}

522: 

523: Since all the crucial aspects of Bell's formulation of locality are

524: thus meaningful only relative to some candidate theory, it is perhaps 

525: puzzling how Bell thought we could

526: say anything about the locally causal character of Nature.  Wouldn't

527: the locality condition only allow us to assess the local character of

528: candidate theories?  It is important to understand that the

529: answer is essentially (at least initially):  Yes!  Indeed, note that

530: Bell begins the formulation with ``A \emph{theory} will be said to be 

531: locally causal if...'' (emphasis added).   Let us state it openly and

532: explicitly:  Bell's locality criterion is a way of distinguishing

533: local theories from nonlocal ones:  

534: \begin{quote}

535: ``I would insist here on the distinction

536: between analyzing various physical theories, on the one hand, and

537: philosophising about the unique real world on the other hand.  In this

538: matter of causality it is a great inconvenience that the real world is

539: given to us once only.  We cannot know [by looking] what would have 

540: happened if something had been different.  We cannot repeat an

541: experiment changing just one variable; the hands of the clock will

542: have moved, and the moons of Jupiter.   Physical 

543: theories are more amenable in this respect.  We can \emph{calculate}

544: the consequences of changing free elements in a theory, be they only

545: initial conditions, and so can explore the causal 

546: structure of the theory.  I insist that [my concept of local causality]

547: is primarily an analysis of certain kinds of physical theory.''

548: (1977, p. 101)

549: %\cite[pg 101]{bell}

550: \end{quote}

551: How then did Bell think we could end up saying something interesting

552: about Nature?  That is precisely the beauty of Bell's theorem, which

553: shows that no theory respecting the locality condition (no matter what

554: other properties it may or may not have -- e.g., hidden variables or

555: only the non-hidden sort, deterministic or stochastic, particles or

556: fields or both or neither, etc.) can agree with the

557: empirically-verified QM predictions for certain types of experiment.

558: That is (and leaving aside the various experimental loopholes), no 

559: locally causal theory in Bell's sense

560: can agree with experiment, can be empirically viable, can be

561: true.  Which means the true theory (whatever it might be) necessarily

562: \emph{violates} Bell's locality condition.  Nature is not locally

563: causal.  

564: \footnote{This sometimes comes as a shock to adherents of orthodox

565: quantum theory, who are used to thinking of their own theory --

566: especially in its allegedly relativistic variants -- as perfectly

567: consistent with SR.  But the nonlocality of orthodox QM is quite

568: obvious, if one knows where to look.  The key here is that the

569: theory is not defined exclusively by the Schr\"odinger (or

570: equivalent) dynamical equation, but also by some version of a

571: collapse postulate.  And this latter is explicitly nonlocal.

572: Indeed, orthodox (collapse) QM is even \emph{more} nonlocal than

573: certain alternative theories, like Bohmian Mechanics, which are 

574: often maligned precisely for displaying an obvious kind of nonlocality.

575: The simplest type

576: of example which suffices to make this point is the ``Einstein's

577: Boxes'' scenario.  (See Travis Norsen, \emph{AmJPhys} 73(2), Feb. 2005,

578: pages 164-176.)  Bell explains beautifully how

579: this scenario manifests the nonlocal causation inherent in orthodox

580: QM:  ``Suppose ... we have a radioactive nucleus which can emit a

581: single $\alpha$-particle, surrounded at a considerable distance by

582: $\alpha$-particle counters.  So long as it is not specified that

583: some \emph{other} counter registers, there is a chance for a

584: particular counter that \emph{it} registers.  But if it is specified

585: that some other counter does register, even in a region of

586: space-time outside the relevant backward light cone, the chance that

587: the given counter registers is zero.  We simply do not have

588: [Equation \ref{eq:loc}].''  (1975, p. 55)  Of course, one might

589: contemplate an alternative theory in which the relevant complete

590: specification of beables included something else in addition to (or

591: instead of) the QM wave function; such an alternative may or may not

592: exhibit any nonlocal causation.  This illustrates the point made

593: earlier:  what we diagnose as ``nonlocal'' by applying Bell's

594: criterion is some particular candidate theory (as opposed to

595: immediately inferring, from either the empirical predictions of a

596: theory or a given set of empirical observations, that some non-local

597: causation is occuring in Nature.)  The point here is that

598: orthodox QM's account of the ``Einstein's Boxes'' scenario involves

599: non-local causation; whether any nonlocality in fact occurs in Nature

600: when one performs the indicated experiment involving an

601: $\alpha$-particle, however, is a very different question.  If, for

602: example, Bohmian Mechanics (rather than orthodox QM) is true, the

603: answer would be no.}

604: 

605: For ease of future reference and to fix some terminology, it will be 

606: helpful to lay out here a bit more explicitly the type of setup

607: involved in the Bell experiments, and to indicate precisely how one

608: gets from locality as formulated by Bell to the somewhat

609: different-looking mathematical condition (sometimes called

610: ``factorizability'') from which standard derivations of Bell's

611: inequality proceed.  

612: 

613: The setup relevant to Bell's theorem involves a particle source which

614: emits pairs of spin-correlated particles, and two spatially separated

615: devices each of which allows measurement of one of several spin

616: components on the respective incident particle.  (In actual

617: experiments, the particles are typically photons with polarization

618: playing the role of ``spin.'')  Two experimenters, traditionally

619: Alice and Bob, man the two devices.  We use the symbols

620: $\hat{a}$ and $\hat{b}$ to refer, respectively, to the ``settings'' of

621: Alice's and Bob's apparatus (one usually thinks here of an axis in

622: space along which the polarizer or Stern-Gerlach magnetic field is

623: oriented), and $A$ and $B$ to refer to the

624: ``outcomes'' of their respective spin-component measurements.

625: Finally, we will use the symbol $\lambda$ to refer to the ``state of

626: the particle pair.''  The scare-quotes around the various terms here

627: are an advertisement for the following discussion.

628: 

629: First, note that all of the symbols just introduced refer to

630: \emph{beables}.  There is a tendency in the literature for all of

631: these things (the apparatus settings, the outcomes, and the physical

632: state of the particle pair) to remain needlessly abstract.  But all of

633: these things are perfectly concrete, at least relative to some

634: particular candidate theory.  The setting of Alice's

635: apparatus, for example, refers to something like the spatial

636: orientation of a Stern-Gerlach device, or some sort of knob or lever

637: on some more black-box-ish device.  Thus, this ``setting'' ultimately

638: comes down to the spatial configuration of some physically real

639: matter, i.e., it must be reflected somehow in the beables posited by

640: any serious candidate theory. 

641: \footnote{A candidate theory which posited no

642: beables corresponding to such things as knobs and levers should not,

643: and probably could not, be taken seriously.  Bell stresses in his very

644: first discussion of beables that:  ``The beables must include the

645: settings of switches and knobs on experimental equipment ... and the

646: readings of instruments.'' (1975, p. 52)  For elaboration of

647: the sense of the term ``serious'' being used here, see Bell's (1986,

648: pp. 194-5).}

649: 

650: Likewise, the outcome of

651: Alice's experiment is not some ethereal event taking place in Alice's

652: consciousness or some other place about which serious candidate physical

653: theories fail to speak directly.  Rather, the outcome should be

654: thought of as being displayed in the post-measurement position of a

655: pointer (or the arrangement of some ink on a piece of paper, etc.) -- in

656: short, the outcome too is just a convenient way of referring to some

657: physically real and directly observable configuration of matter, and

658: so will necessarily be reflected in the beables posited by any serious

659: candidate theory.  The case with $\lambda$ is a little different,

660: because this is not something to which we have any sort of direct 

661: observational access.  But this only means there is significantly more

662: freedom about what sort of beables $\lambda$ might refer to in various

663: different candidate theories. 

664: 

665: 

666: 

667: \begin{figure}[t]

668: \begin{center}

669: \includegraphics[width=3.3in,clip]{./fig5.eps}

670: \end{center}

671: \caption{

672: \label{fig5}

673: Space-time diagram illustrating the various beables of  relevance for

674: a discussion of factorization.  Separated observers Alice and Bob make

675: spin-component measurements (using apparatus characterized by

676: variables $\hat{a}$ and  $\hat{b}$

677: respectively) of a pair of entangled spin-1/2 particles.   The state

678: of the particles (and/or any other appropriately associated beables) is denoted

679: $\lambda$, and the outcomes of the two measurements are represented by

680: the beables $A$ (in region 1) and $B$ (in region 2).  Note that

681: $\lambda$ and $\hat{a}$ jointly constitute a complete

682: specification of beables in space-time region $3a$, which shields off

683: region 1 from the overlap of the past light cones of 1 and 2.

684: Likewise, $\lambda$ and $\hat{b}$ jointly constitute a complete

685: specification of beables in space-time region $3b$, which shields off

686: region 2 from the overlap of the past light cones of 2 and 1.   Thus

687: the joint specification of $\lambda$ and $\hat{a}$ will -- in a

688: locally causal theory -- make $\hat{b}$ and $B$ redundant for a 

689: theory's predictions about $A$ (and likewise, specification of

690: $\lambda$ and $\hat{b}$ will render $\hat{a}$ and $A$ redundant

691: for predictions about $B$).

692: }

693: \end{figure}

694: 

695: 

696: The basic space-time structure of the setup in question is sketched in

697: Figure 4, and the overall logic is explained in the caption.  The idea 

698: is simply to apply Bell's locality condition to the two measurement

699: outcomes $A$ and $B$ in order to assert that the probability assigned

700: to a given outcome (by a locally causal candidate theory) should be

701: independent of both the setting and outcome of the distant

702: experiment.  That is, in a locally causal theory, we must have

703: \begin{equation}

704: P(A | \hat{a}, \hat{b}, B, \lambda) = P(A | \hat{a}, \lambda)

705: \label{eq:BLA}

706: \end{equation}

707: and

708: \begin{equation}

709: P(B | \hat{a}, \hat{b}, A, \lambda) = P(B | \hat{b}, \lambda). 

710: \label{eq:BLB}

711: \end{equation}

712: It is then trivial to apply the definition of conditional probability

713: to arrive at the conclusion that the joint probability (for outcomes

714: $A$ and $B$) \emph{factorizes}:

715: \begin{equation}

716: P(A,B | \hat{a},\hat{b}, \lambda) = P(A|\hat{a},\lambda) \times

717: P(B|\hat{b},\lambda).

718: \label{eq:factor}

719: \end{equation}

720: Bell writes:  ``Very often such

721: factorizability is taken as the starting point of the analysis.  Here

722: we have preferred to see it not as the \emph{formulation} of `local

723: causality', but as a consequence thereof.'' (1990, p. 243)

724: 

725: It is also important that the apparatus settings $\hat{a}$ and

726: $\hat{b}$ are in some sense ``free.''  This is often discussed in

727: terms of Alice and Bob making literal last-minute free-will choices

728: about how to orient their devices.  What is actually required for the

729: proof is merely the assumption that $\hat{a}$ and $\hat{b}$ are

730: (stochastically)

731: independent of the particle pair state $\lambda$.

732: This comes up in the course of the derivation of Bell's inequality

733: when we write an expression for predicted empirical frequency of a certain

734: joint outcome as a weighted average over the candidate theory's

735: predictions for a given $\lambda$ -- that is,

736: \begin{equation}

737: E(A,B|\hat{a},\hat{b}) = \int d\lambda \, P(A,B|\hat{a},\hat{b},\lambda)

738: \, P(\lambda)

739: \end{equation}

740: where $P(\lambda)$ is the probability that a given particle pair state

741: $\lambda$ is produced by the preparation procedure used at the source.

742: 

743: If the probability distribution $P(\lambda)$ actually depended on

744: $\hat{a}$ or $\hat{b}$ -- as one would expect if the 

745: particle pair exerted some causal influence on the settings (or vice

746: versa!) or if the particle pair and the settings were mutually causally

747: influenced in some non-trivial way by events farther back in the past

748: -- then the above expression for the empirical frequency would be

749: invalid and the derivation of the Bell inequality wouldn't go through.

750: \footnote{See Bell, 1990, pp. 243-4.}

751: 

752: This might seem like cause for concern, especially considering that

753: (as displayed in Figure 4) the past light cones of $\hat{a}$ and

754: $\hat{b}$ overlap with the region containing

755: $\lambda$ -- and $\lambda$ by definition is supposed

756: to contain a \emph{complete} specification of beables in this region.

757: Given all that, one wonders how $\hat{a}$ and $\hat{b}$ could possibly

758: \emph{not} be causally influenced by $\lambda$ (in a locally causal

759: theory).  

760: 

761: Here we see the

762: reason for another aspect of Bell's carefully-phrased formulation of 

763: locality, which requires that beables in (the relevant) ``region 3''

764: (which will of course be two different regions for the two measurements)

765: be ``sufficiently specified, for example by a full specification...''

766: Here there is the implication that a

767: specification could be sufficient \emph{without} being complete.  For

768: example, some candidate theory (and this is actually true of every

769: serious extant candidate theory) might provide a specification of the

770: state of the particle pair which is sufficient in the relevant sense,

771: even though it leaves out some fact (say, the millionth digit of the

772: energy of some relic microwave background photon that happens to

773: fly into the detection region just prior to the measurement) which

774: actually exists in the relevant spacetime region.  Such a fact could

775: then be allowed to determine the setting $\hat{a}$ without introducing

776: even the slightest evidence for the problematic sort of correlation

777: between $\hat{a}$ and $\lambda$.  

778: \footnote{See Bell's discussion in his 1977, pages 100-104.}

779: Indeed, this is just an exaggerated

780: version of what happens in the actual experiments, where

781: carefully-isolated and independent pseudo-random-number generators are

782: used to produce the settings at the two stations.  

783: 

784: Finally, as Shimony, Horne, and Clauser have pointed out, 

785: \begin{quote}

786: ``In any scientific experiment

787: in which two or more variables are supposed to be randomly selected,

788: one can always conjecture that some factor in the overlap of the

789: backward light cones has controlled the presumably random choices.

790: But, we maintain, skepticism of this sort will essentially dismiss all

791: results of scientific experimentation.  Unless we proceed under the

792: assumption that hidden conspiracies of this sort do not occur, we have

793: abandoned in advance the whole enterprise of discovering the laws of

794: nature by experimentation.   [Hence, the extra] supposition needed [to

795: derive Bell's inequality from Bell's locality condition] is

796: no stronger than one needs for experimental reasoning generically, and

797: nevertheless just strong enough to yield the desired inequality.''  

798: \footnote{A. Shimony, M.A. Horne, and J.F. Clauser, ``An Exchange on

799:   Local Beables,'' \emph{Dialectica}, 39 (1985) 86-110}

800: \end{quote}

801: So in the end there is really nothing worth worrying about here, i.e.,

802: nothing which, in the face of the experimental data conflicting with

803: Bell's inequalities, one might reasonably reject as an alternative to

804: rejecting Bell's local causality.  What

805: is important is that Equation \ref{eq:factor} (along with the ``freedom''

806: or ``no conspiracies'' assumption just discussed) entails the Bell

807: inequality.

808: The derivation is standard and will not be repeated here.  

809: \footnote{See, e.g., Bell 1975, pages 55-57.}

810: We will instead simply note

811: that (starting in 1975) Bell almost always referred 

812: to the empirically testable inequality as the ``locality inequality.''  

813: One can hopefully now appreciate why.

814: 

815: One final caveat.  There is a sense in which our

816: verbal description of the meanings of the relevant beables ($\hat{a}$,

817: $\hat{b}$, $A$, $B$, and $\lambda$) is potentially confusing.  For

818: example, there is no sense in which Bell makes some dubious ``extra

819: assumption'' that, e.g., \emph{particles} fly from the ``particle

820: source'' to the measurement devices.   

821: \footnote{Bell, 1981, p. 150}

822: A theory which posits no

823: particle beables, but only (say) waves or a wave function or \emph{whatever}

824: is perfectly fine and the formal locality criterion

825: will apply to it just the same way.  Only the words would need to

826: change.  Relatedly, many papers in the Bell literature raise issues

827: about hidden variables associated not with the particle pair, but with

828: the measurement devices.  Have we excluded such variables, since our

829: $\hat{a}$ refers only to some knob setting on Alice's device and

830: $\lambda$ refers only to the state of the particle pair?  No.  The

831: only important distinction here is that $\hat{a}$ and $\hat{b}$ refer 

832: to beables which are ``free'' (or ``random'') in the above sense,

833: while the variables $\lambda$ are somehow set by past events.

834: 

835: This is no doubt a fuzzy distinction, but nothing important hinges on

836: it.  The point is that any

837: microscopic features of Alice's device (hidden or not)  -- or anything

838: else relevant to the candidate theory's predictions for the probabilities in

839: question -- can be included under the ``setting'' variables ($\hat{a}$,

840: $\hat{b}$) \emph{or} the ``particle state'' variables ($\lambda$),

841: whichever seems more natural.  In short, Bell's formulation of

842: locality is significantly more general than might otherwise be

843: suggested by some of the words used to describe it.

844: 

845: 

846: 

847: 

848: 

849: \section{Jarrett's Analysis}

850: 

851: Jon Jarrett's influential analysis of Bell's locality criterion 

852: begins with Equations \ref{eq:BLA} and \ref{eq:BLB}, which he dubs

853: ``strong locality.''  (Of course, we are innocuously 

854: changing -- and occasionally

855: simplifying -- Jarrett's notation to make it consistent with that

856: introduced above.)  For simplicity, let us focus the discussion 

857: on Equation \ref{eq:BLA}.  Jarrett defines two sub-conditions which,

858: he subsequently proves, are jointly equivalent to this ``strong locality.''

859: 

860: The first sub-condition Jarrett dubs ``locality.''  (It is also

861: sometimes referred to as ``simple locality'' or ``parameter

862: independence'' or ``remote context independence'' in the literature.)

863: \begin{equation}

864: P(A | \hat{a},\hat{b}, \lambda) = P(A| \hat{a}, \lambda).

865: \label{eq:JL}

866: \end{equation}

867: As Jarrett explains, ``Locality requires that the probability for the 

868: outcome

869: [$A$] ... be determined `locally'; i.e., that it depend only on the

870: state $\lambda$ ... of the two-particle system and on the state

871: [$\hat{a}$] of the measuring device.  In particular, that probability

872: must be independent of which (if any) component of spin the distant

873: measuring device is set to measure.''  

874: \footnote{Jon Jarrett, \emph{op cit.}, p. 573}

875: 

876: Jarrett's second sub-condition, which he dubs ``completeness'' (and is

877: also known as ``predictive completeness,'' ``outcome

878: independence,'' ``remote outcome independence'' and ``conditional 

879: outcome independence'') is the following:  

880: \begin{equation}

881: P(A | \hat{a}, \hat{b}, B, \lambda) = P(A | \hat{a}, \hat{b}, \lambda)

882: \label{eq:JC}

883: \end{equation}

884: which ``asserts the stochastic independence of

885: the two outcomes in each pair of spin measurements.''

886: \footnote{\emph{Ibid.}, p. 578}

887: 

888: It is easy to see that, indeed, ``locality'' and ``completeness'' are

889: jointly equivalent to Bell's locality condition as expressed in

890: Equation \ref{eq:BLA}.  First, $P(A|\hat{a},\hat{b},B,\lambda)$ is,

891: under the assumption of ``completeness,'' equal to

892: $P(A|\hat{a},\hat{b}, \lambda)$.  And this, in turn, is equal to $P(A |

893: \hat{a}, \lambda)$ under the assumption of ``locality''.  So

894: ``locality'' and ``completeness'' jointly entail ``strong locality.''

895: And likewise ``strong locality'' clearly entails both ``locality''

896: and ``completeness.''  So the two sub-conditions are, indeed,

897: equivalent to Equation \ref{eq:BLA}.  

898: 

899: The alleged significance of this decomposition, however, emerges

900: only from Jarrett's discussion of the physical interpretation of his

901: two sub-conditions.  

902: 

903: First, Jarrett argues that ``locality'' is equivalent to the

904: prohibition of superluminal signaling, and 

905: hence expresses just what relativity requires of other theories.  As

906: was mentioned in the introduction, both steps of this argument have

907: been found wanting.  First, it is only in combination with some assumptions

908: about the \emph{controllability} of various beables (notably

909: $\lambda$) that Jarrett's ``locality'' is equivalent to the

910: prohibition on superluminal signaling.  There is at least one

911: extant, empirically viable theory (Bohmian Mechanics) which violates

912: Jarrett's ``locality'' condition and yet doesn't permit the

913: possibility of superluminal signaling, precisely because the relevant

914: states cannot (as a matter of principle, as predicted by the

915: theory) be sufficiently controlled.  And second, it is dubious to

916: claim that the prohibition of superluminal \emph{signals} adequately

917: captures relativity's fundamental speed limit.  This would, for

918: example, render Bohmian Mechanics consistent with SR despite its need

919: to postulate a dynamically privileged reference frame -- a ``gross

920: violation of relativistic causality'' according to Bell. (1984, p. 171)

921: But since these

922: problems have been discussed elsewhere in the literature, we leave

923: them aside here and focus instead on Jarrett's physical interpretation

924: of his second sub-condition, ``completeness.''  

925: 

926: Jarrett elaborates the meaning of his ``completeness'' condition with

927: the following example:

928: \begin{quote}

929: ``A simple (and incorrect) model for the Bell-type correlated spin

930: phenomena may serve as a useful illustration.  Suppose, purely for the

931: sake of illustration, that spin is correctly represented as an

932: ordinary classical angular momentum.  Suppose further that when a pair

933: of particles is prepared in the singlet state, the spin vectors for

934: the two particles are aligned exactly anti-parallel to each other.

935: Moreover, given an ensemble of such two-particle systems, suppose that

936: each direction in space is equally likely to be the direction of

937: alignment for an arbitrarily selected member of the ensemble.

938: Finally, if the unit vector [$\hat{a}$] gives the direction along

939: which the

940: axis of the Stern-Gerlach apparatus is aligned ... and if [$\hat{s}$]

941: is the spin vector of the particle ... then the outcome of that

942: measurement is $+1$ if [$\hat{a}\cdot \hat{s} > 0$] and $-1$ if

943: [$\hat{a}\cdot \hat{s} < 0$].  

944: \end{quote}

945: Jarrett then makes what amounts to the following point:  for this

946: model, we clearly have that Alice's particle (from a randomly selected

947: member of the ensemble) is equally likely

948: to be found in the $A = +1$ and $A = -1$ states along any arbitrary

949: direction $\hat{a}$.  Thus, for example,

950: \begin{equation}

951: P (A \! = \! +1 | \hat{a}, \hat{b}, \lambda) = \frac{1}{2}

952: \label{eq:comp1}

953: \end{equation}

954: where $\lambda$ is the singlet state -- evidently meaning, in the context of

955: this example, the state description according to which the particle has

956: some definite but completely unknown spin direction $\hat{s}$.  

957: 

958: On the other hand, it is built into the model that, for $\hat{a} =

959: \hat{b}$, the outcomes of Alice's and Bob's measurements will be

960: perfectly anti-correlated.  Hence, if we additionally specify the

961: outcome of Bob's experiment, the outcome of Alice's is \emph{fixed}.

962: Suppose, for example, that $B = +1$.  Then $A = +1$ is forbidden,

963: i.e., 

964: \begin{equation}

965: P (A \! = \! +1  | \hat{a}, \hat{b}, B, \lambda) = 0. 

966: \label{eq:comp2}

967: \end{equation}

968: Comparing the previous two equations, we see that ``completeness is

969: clearly violated.''  \footnote{\emph{Ibid.}, p. 580}

970: 

971: 

972: Jarrett elaborates: 

973: \begin{quote}

974: ``The probabilities specified for this model are

975: grounded in a blatantly incomplete description of the two-particle

976: state.  In the context of this model, if the theory assigns

977: probabilities only on the basis of the occupancy by the two-particle 

978: system of the singlet state, then conditioning on the outcome [$B$] of a

979: [$\hat{b}$]-component spin measurement on [Bob's particle] may well

980: yield a different probability for the outcome of a spin measurement on

981: [Alice's particle] than would have been given by the corresponding

982: unconditioned probability (i.e., 1/2).  This is so because, if the

983: outcome of the measurement on [Bob's particle] is $+1$ [and if

984: $\hat{a} = \hat{b}$], then it may be

985: inferred that [$\hat{a} \cdot \hat{s}_A < 0 $] (with probability 1),

986: where [$\hat{s}_A$] is the spin of Alice's particle....

987: The outcome of the measurement on [Bob's particle thus] provides

988: information about [Alice's particle] which was not included in the

989: incomplete state description [$\lambda$].''  \footnote{\emph{Ibid.},

990:   p. 580}

991: \end{quote}

992: The important conclusion is this:

993: the fact that the probability assigned to a certain outcome

994: for Alice's experiment 

995: depends, in violation of Equation \ref{eq:JC}, on the

996: outcome of Bob's experiment, does not mean that there is any

997: relativistically-forbidden superluminal \emph{causal influence} (e.g.,

998: from Bob's outcome to Alice's).  

999: That is the whole point of the illustrative example, in

1000: which (by assumption) the outcome of Alice's experiment is determined

1001: exclusively by factors (namely $\hat{a}$ and $\hat{s}_A$) which are

1002: present at her location.  No nonlocal causal influence exists.  

1003: Instead, the violation of Equation \ref{eq:JC} indicates only that we

1004: were dealing with \emph{incomplete state descriptions}, such that

1005: Bob's outcome provides some \emph{information} (usefully supplementing

1006: what was already contained in $\lambda$) which warrants an updating of

1007: probabilities.  

1008: 

1009: 

1010: On the basis of this example, 

1011: Jarrett thus urges the following physical interpretation of

1012: his two sub-conditions:  a violation of ``locality'' would allow the

1013: possibility of sending superluminal signals, and hence indicates

1014: clearly the existence of some relativity-violating superluminal causal

1015: influences; on the other hand, a violation of ``completeness'' does

1016: not indicate the existence of any relativity-violating influences, but

1017: instead suggests only that the state descriptions of the theory in

1018: question are not complete.  It is clear that these physical

1019: interpretations of the two conditions were the basis for Jarrett's

1020: decision to name them ``locality'' and ``completeness'' respectively.  

1021: 

1022: Here is Jarrett's summary of the cash value of this decomposition

1023: \emph{vis-a-vis} Bell's theorem and the associated experiments:  these

1024: together provide very strong evidence ``that

1025: strong locality cannot be satisfied by any empirically adequate

1026: theory.  Since locality is contravened only on pain of a serious

1027: conflict with relativity theory (which is extraordinarily

1028: well-confirmed independently), it is appropriate to assign the blame

1029: to the completeness condition.  ...[O]ne must conclude that certain

1030: phenomena simply cannot be adequately represented by any theory which

1031: ascribes properties to the entities it posits in such a way that no

1032: measurement on the system may yield information which is both

1033: non-redundant (not deducible from the state descriptions) and

1034: predictively relevant for distant measurements.  That `information' is

1035: not (neither explicitly nor implicitly) contained in the `incomplete'

1036: state description.''  \footnote{\emph{Ibid.}, p. 585}

1037: 

1038: 

1039: \section{Comparison}

1040: 

1041: The fundamental origin of the disagreement between Bell and Jarrett

1042: should now be clear:  the two authors do not understand (e.g.)

1043: Equation \ref{eq:BLA} in the same way.  For Bell, the variables

1044: $\lambda$ in this formula (together with $\hat{a}$ as per the previous 

1045: discussion) constitute a \emph{complete} (or perhaps merely

1046: sufficient) specification of beables in some

1047: space-time region that has the same relation to Alice's experiment  

1048: that region 3 (of Figure 1) had to region 1.  Jarrett, by contrast,

1049: is agnostic about the completeness of the description afforded by

1050: $\lambda$.  

1051: 

1052: Strictly speaking,

1053: therefore, Jarrett's decomposition of Equation \ref{eq:BLA}

1054: is not a decomposition of

1055: Bell's locality condition, but, rather, a decomposition of some sort of

1056: no-correlation condition 

1057: \begin{equation}

1058: P(b_1 | \bar{B}_3, b_2) = P(b_1 | \bar{B}_3)

1059: \label{eq:nocorr}

1060: \end{equation}

1061: which is analogous to Equation \ref{eq:loc}, except that, following

1062: the notation of Section II, the variables $\bar{B}_3$ are \emph{not} 

1063: assumed to provide a complete specification of

1064: beables in the relevant spacetime region.  But as we have already

1065: discussed in Section II, and as Bell was perfectly aware,

1066: a violation of Equation \ref{eq:nocorr} does

1067: not necessarily indicate the presence of any nonlocal causation in the

1068: candidate theory in question.  Indeed, it is precisely to close off the

1069: avenue eventually taken by Jarrett (that is, blaming a violation of this

1070: ``no-correlation'' condition on the incompleteness of the specified

1071: beables) that Bell specifically stresses the importance that ``events

1072: in [the relevant region] be specified completely.'' (1990, p. 240)

1073: 

1074: 

1075: This confusion -- this departure from Bell's

1076: actual locality criterion -- is the ultimate basis for Jarrett's

1077: choice of terminology, and also for any initial plausibility of his

1078: project of establishing ``Peaceful Coexistence'' by showing that a

1079: violation of the locality criterion needed for Bell's inequality (in

1080: particular, a violation of his ``completeness'' sub-condition) need

1081: not indicate any conflict with relativistic local causality. 

1082: 

1083: Of course, one could simply apply Jarrett's decomposition strategy to

1084: Bell's actual locality condition.  That is, it is true that, as

1085: Jarrett claimed to have shown, Bell's locality condition can be

1086: decomposed into two Jarrett-like sub-conditions -- namely, our

1087: Equations \ref{eq:JL} and \ref{eq:JC} but now with the requirement

1088: (inherited from Bell's actual locality condition) that $\lambda$ 

1089: provide a complete (or sufficient) specification of relevant beables

1090: (as posited by some candidate theory whose locality is being assessed

1091: by the locality condition).  

1092: But \emph{this} decomposition fails to have any of the physical

1093: implications urged by Jarrett.  In particular, a violation of the

1094: (strengthened) ``completeness'' condition 

1095: \begin{equation}

1096: P(A|\hat{a},\hat{b},B,\lambda) = P(A|\hat{a},\hat{b},\lambda)

1097: \label{eq:BJcomp2}

1098: \end{equation}

1099: (formally equivalent to Jarrett's ``completeness'' sub-condition, but now,

1100: with Bell but contra Jarrett, with the insistence that $\lambda$ and 

1101: $\hat{a}$ jointly provide a complete description of beables in some

1102: spacetime region through the past of $A$ which divides $A$ off

1103: completely from the overlap of the past light cones of $A$ and $B$)

1104: has absolutely nothing to do with the completeness of state descriptions,

1105: but instead indicates the

1106: presence of some nonlocal causation (in violation of the causal

1107: structure outlined in Figure 1) in the candidate theory in question.

1108: 

1109: The most that could be said to distinguish the two sub-conditions is that,

1110: since $\hat{b}$ is (by definition) controllable and $B$ (most likely) 

1111: isn't, a violation of \ref{eq:JL} is (all other things being equal) 

1112: more likely to yield the possibility of superluminal \emph{signaling}

1113: than a violation of \ref{eq:JC}.  But that only matters if we drop

1114: what Bell calls ``fundamental relativity'' and instead read SR

1115: instrumentally, as prohibiting superluminal signalling but allowing in

1116: principle superluminal causation (so long as it can't be harnessed by

1117: humans to transmit messages).   That is, at best, a dubious and 

1118: controversial reading of SR, as already mentioned.  \footnote{See Bell's

1119: discussions of ``controllable'' beables and causation \emph{vs.} 

1120: signaling in his 1975, pp. 60-61; 1984, p. 171; and 1990, pp. 244-246.}

1121: 

1122: It is sometimes suggested 

1123: \footnote{See, e.g., p. 153 of Harvey Brown's half of ``Nonlocality in Quantum

1124:   Mechnics,''  Michael Redhead and Harvey Brown, \emph{Proceedings of

1125:     the Aristotelian Society, Supplementary Volumes}, Vol. 65 (1991),

1126:   pp. 119-159.}

1127: that the relation between Bell and Jarrett

1128: on this point is one of basic agreement, since both held that the

1129: validity of Equation \ref{eq:JC}

1130: has something to do with the completeness of the physical theory in

1131: question.  This is doubly incorrect.  First, the mathematical

1132: condition in question is, for Jarrett, a \emph{formulation} of

1133: completeness; it is, for Bell, (part of) a formulation of local

1134: causality which functions appropriately as such \emph{only if the 

1135: relevant symbols in the formula stand for a complete specification 

1136: of the relevant beables}.  According to Jarrett, Equation \ref{eq:JC} is

1137: supposed to tell us whether completeness holds.  For Bell, on the

1138: other hand, the equation (correctly understood) already presupposes

1139: that completeness holds.  (If it doesn't hold, the condition is useless for

1140: his purposes, and states only that the two outcomes are uncorrelated.)  

1141: And second, where Jarrett (and most subsequent commentators)

1142: regard the ``completeness'' in question as a property of

1143: \emph{theories}, Bell regards his ``completeness'' as a property of a

1144: certain \emph{specification of beables} relative to some candidate

1145: theory.  In this context, the separate question of whether the theory

1146: itself is complete (i.e., whether its posited beables capture

1147: everything that really exists) simply doesn't come up.

1148: 

1149: Note also that, even on his own terms, i.e., even leaving aside his 

1150: departure from Bell's actual

1151: locality criterion, Jarrett's formulation of completeness actually

1152: fails as a criterion for assessing the completeness of the relevant

1153: physical state descriptions.

1154: 

1155: Consider again Equation \ref{eq:JC}

1156: where for the moment we follow Jarrett and interpret $\lambda$ here as

1157: providing some kind of state description, but not necessarily a

1158: complete one.  Jarrett has shown (with the example discussed

1159: in Section III) that a violation of this condition can sometimes be blamed

1160: on the use of incomplete state descriptions, with no implication of

1161: any superluminal causation.  But it is equally easy to display an

1162: example in which violation of Jarrett's condition \emph{cannot} be 

1163: blamed on incomplete state descriptions, but instead indicates the

1164: presence of superluminal causation.

1165: 

1166: Consider a toy model discussed by Maudlin, in which each particle in the

1167: pair is indeterminate (in regard to its spin along any particular

1168: direction) until one of the particles encounters a spin-measurement

1169: apparatus; at this point, this ``first'' particle flips a coin to

1170: decide whether to emerge from the $+1$ or the $-1$ port of the

1171: apparatus, and sends an instantaneous tachyon signal to the other

1172: particle in the pair, instructing it as to how it should behave in

1173: order to give rise to the correct quantum mechanical joint outcomes.

1174: \footnote{See Maudlin, \emph{op cit.}, pg 82.  After presenting this model as

1175:   a simple example of how the observed spin correlations might

1176:   arise, Maudlin uses it to counterexample the claim, made by Don

1177:   Howard and others, that Jarrett's ``completeness'' and ``locality''

1178:   can be mapped, respectively, onto the ``separability'' and

1179:   ``locality'' conditions which emerge from some of Einstein's

1180:   comments about local causality.  As Maudlin points out, the toy

1181:   model is perfectly separable in the sense of Einstein, and yet

1182:   violates what Howard \emph{et al.} would have us take as a mathematical

1183:   formulation of Einstein's ``separability'' (namely, Jarrett's

1184:   ``completeness.'')  Curiously, however, Maudlin does not

1185:   mention this model in his (earlier) discussion of Jarrett.}

1186: Assuming, as before, that $\hat{a} = \hat{b}$, and that Bob's particle

1187: arrives at its detector first

1188: \footnote{Of course, since Alice's and Bob's measurements are, by

1189:   hypothesis, space-like separated, there is no relativistically

1190:   unambigous meaning to ``first.''  But that is really the whole

1191:   point.  This model explicitly involves anti-relativistic

1192:   superluminal causation, so part of the model is that relativity is

1193:   wrong and there exists some dynamically privileged reference frame

1194:   which gives an unambiguous meaning to this ``first'' (and also to

1195:   the ``instantaneous'' in the description of the tachyon signal).}

1196: with the result $B = +1$, 

1197: this model makes the same predictions, Equations \ref{eq:comp1} and

1198: \ref{eq:comp2}, as the local deterministic model discussed by

1199: Jarrett.  And so this model, like Jarrett's, violates Jarrett's

1200: ``completeness'' condition.  And yet clearly with this model there is

1201: no blaming that violation on the use of incomplete state descriptions

1202: -- instead here it is obvious (by construction) that the violation is

1203: due to the presence of superluminal causal influences (in the

1204: particular form of ``tachyon signals'').    

1205: 

1206: This should not be surprising.  We have already argued that, if one

1207: follows Bell in requiring $\lambda$ to constitute a complete state

1208: description, then a violation of Jarrett's ``completeness'' can \emph{only} be

1209: understood as indicating the presence of nonlocal causation.  The

1210: point here is that, even if we follow Jarrett in remaining agnostic

1211: about the completeness of the description afforded by $\lambda$, we

1212: cannot \emph{necessarily} say that a violation of ``completeness'' is

1213: compatible with relativity's prohibition on superluminal causation.

1214: It might be (as shown by Jarrett's model).  But it might not be (as

1215: shown by Maudlin's).  

1216: 

1217: The correct conclusion is therefore as follows:  a violation of

1218: Jarrett's ``completeness'' condition (where we are openly agnostic

1219: about the completeness of the state description $\lambda$) means

1220: \emph{either} that we have relativity-violating nonlocal

1221: causation, \emph{or} that we were dealing with incomplete state 

1222: descriptions.  And \emph{this} dilemma is precisely that posed already in

1223: 1935 by Einstein, Podolsky, and Rosen:  \emph{either} we concede Bohr's claim

1224: that the QM description of states is complete and accept the reality

1225: of the nonlocal causation present in that theory; \emph{or} we reject the

1226: completeness claim and adopt a different (``hidden variables'') theory

1227: which (the authors thought) could restore locality.

1228: \footnote{A. Einstein, B. Podolsky, and N. Rosen, ``Can

1229:   quantum-mechanical description of physical reality be considered 

1230:   complete?'' \emph{Phys. Rev.} 47 (1935), 777-780.  For some more

1231:   recent discussion, see T. Norsen, ``Einstein's Boxes,'' \emph{op

1232:     cit.} and references therein.}

1233: 

1234: 

1235: Jarrett also sees a similarity between his conclusion and that of EPR

1236: (namely that QM is not complete).  But where EPR considered this a

1237: defect to be corrected in some alternative ``hidden variable'' theory

1238: (which they hoped would also restore local causality), Jarrett argues

1239: that incompleteness is not a defect of orthodox QM, but, rather, a

1240: fact of nature:  

1241: \begin{quote}

1242: ``By separating out the relativistic component of the

1243: strong locality condition ... there emerges a clarification of that

1244: class of theories excluded by the Bell arguments:  the class of

1245: theories which satisfy completeness.  Although the term

1246: `incompleteness' may connote a defect (as if, as was the case for the

1247: model discussed [above], all incomplete theories may be

1248: `complet\emph{ed}'), incomplete theories (e.g., quantum mechanics) are

1249: by no means \emph{ipso facto} defective.  On the contrary, when the

1250: result of Bell-type experiments are taken into account, the truly

1251: remarkable implication of Bell's Theorem is that incompleteness, in

1252: some sense, is a genuine feature of the world itself.''

1253: \footnote{\emph{Op cit.}, p. 585}

1254: \end{quote}

1255: This sort of claim is echoed also in Jarrett's later writings. 

1256: \footnote{L.E. Ballentine and Jon P. Jarrett, 

1257: ``Bell's theorem:  Does quantum mechanics contradict relativity?'' 

1258: \emph{Am.J.Phys.} 55(8) (August 1987), 696-701; Jon Jarrett, ``Bell's

1259: Theorem:  A Guide to the Implications'' in J. Cushing and E. McMullin,

1260: eds., \emph{op cit.}}

1261: 

1262: 

1263: Ballentine and Jarrett also give an argument that the ``completeness''

1264: of interest to EPR is a stronger form of Jarrett's ``completeness'' in

1265: the sense that the former entails the latter.  (This occurs in the

1266: context of their agreeing with EPR that QM is incomplete, which they

1267: establish by arguing that, since the correct conclusion from

1268: Bell/experiment is that Jarrett's ``completeness'' condition fails,

1269: then the stronger EPR completeness condition must also fail, just as the

1270: EPR authors claimed.)  But this association is mistaken, since

1271: the argument Ballentine and Jarrett display sneaks in the additional

1272: premise of (EPR's version of) local causality.  \footnote{``But this

1273:   prediction was made without in any way disturbing particle $L$,

1274:   since the $R$ device is at spacelike separation from it...''  This

1275:   mistake was also noted by Andrew Elby, Harvey R. Brown, and Sara

1276:   Foster in ``What Makes a Theory Physically Complete?''

1277:   \emph{Found. Phys.} 23(7),  971-985 (1993).  Note

1278:   also that this error leads Ballentine and Jarrett to remark, in

1279:   passing, that EPR's concept of completeness entails determinism all

1280:   by itself, which is surely a misrepresentation of the worries of

1281:   Einstein and his followers.  See Section 1 of Bell's 1981,

1282:   pp. 139-145, and the important footnote 10 -- already partially

1283:   quoted in the introduction -- on p. 157.}

1284: And this is the premise that does (literally) 

1285: all the work in the EPR argument.  

1286: 

1287: There is thus no apparent sense at all in which EPR's completeness

1288: has anything to do with Jarrett's, except that it was precisely in

1289: the face of QM's violation of Jarrett's ``completeness'' that EPR

1290: argued (correctly) that QM was a non-local theory which, perhaps, could be

1291: replaced by a locally causal alternative theory by adding hidden

1292: variables (or jettisoning the description in terms of wave functions

1293: entirely).  Indeed, at the end of the day, it is pretty clear that 

1294: Jarrett's condition has nothing to do with the completeness of physical

1295: state descriptions.  In concluding (from Bell's Theorem and the

1296: associated experiments) that reality itself is ``in some sense''

1297: incomplete, Jarrett makes clear that he is no longer using the term

1298: with anything like its ordinary meaning -- namely, a description which 

1299: leaves nothing (relevant) out.  By what standard, exactly, could

1300: ``the world itself'' be supposed to have left something out?  

1301: 

1302: It is worth stepping back, therefore, and clarifying what, if 

1303: anything, one can say about ``completeness.''

1304: There are two related senses on the table.  First of all, a theory may

1305: be said to be ``complete'' or ``incomplete'' in relation to external

1306: physical reality.  In this sense, a theory is complete if and only if

1307: it captures or describes everything (relevant) that in fact really

1308: exists.  This is of course just the sense of completeness of interest

1309: to EPR.  Their argument, in essence, was that relativistic local

1310: causality (which they simply took for granted) combined with certain

1311: empirical predictions of QM entailed the existence

1312: of some ``elements of reality'' which had no counterpart in the QM

1313: description of reality.  That is, the QM description left something out;

1314: it was hence

1315: an incomplete theory.  

1316: 

1317: Since this usage of ``completeness'' involves a comparison between

1318: theories and external reality (to which our best access is precisely

1319: through theories!), there is a tendency for it to be regarded as

1320: ``metaphysical''.  

1321: Perhaps this apparently metaphysical flavor is

1322: responsible for Jarrett's suppression of

1323: Bell's requirement that $\lambda$ contain a complete specification of

1324: the relevant beables.  Since there'd be no way to verify whether a given

1325: specification was or wasn't complete in this sense, one might think,

1326: Bell's requirement is meaningless and might as well just be dropped.  

1327: 

1328: But this attitude fails to appreciate one of 

1329: Bell's important advances -- namely, that his formulation of local

1330: causality is a criterion for assessing the locality of \emph{candidate

1331:   theories}.  As already discussed in Section II, Bell's ``complete

1332: specification of beables'' simply does not mean a specification that

1333: captures everything which in fact really exists; rather, it means a

1334: specification which captures everything which is \emph{posited} to

1335: exist \emph{by some candidate theory}.  There is thus nothing the

1336: least bit metaphysical or obscure about Bell's requirement.  For any

1337: unambiguously formulated candidate theory, there should be no question 

1338: about what

1339: is being posited to exist.  And so there will be no ambiguity about

1340: what a complete description of relevant beables should consist of, and

1341: hence no ambiguity about the status -- \emph{vis-a-vis} local

1342: causality  -- of a given well-formulated theory.  

1343: 

1344: There will of course still be

1345: difficult questions about how to decide whether a given candidate

1346: theory is \emph{true}, and hence whether the particular \emph{sort} of

1347: non-local causation contained in it accurately describes some aspect

1348: of Nature.  But the miracle of Bell's argument is that we

1349: need not know which theory is true, in order to know that the true

1350: theory (whatever it turns out to be) will have to exhibit non-local,

1351: super-luminal causation.  There is thus no escaping Bell's conclusion

1352: that some sort of non-local causation (in violation of the structure

1353: displayed in Figure 1) exists in Nature -- in apparent conflict with

1354: what most physicists take to be the requirements of SR.

1355: 

1356: 

1357: \section{Discussion}

1358: 

1359: In the previous section, we interpreted Jarrett as claiming that a

1360: violation of his ``completeness'' condition in no way implied the

1361: presence of non-local causation, but instead only implied that the

1362: state descriptions used in the test had been incomplete.  This is both

1363: fair and unfair -- fair because Jarrett does hang his entire case for

1364: the plausibility of his terminology and his physical interpretation of

1365: the two sub-conditions on precisely this view, but also unfair because

1366: Jarrett also later seems to acknowledge that, ultimately, his

1367: ``completeness'' condition has to be understood very differently.  For

1368: example, he remarks in a footnote that ``completeness, too, has the

1369: character of a `locality' condition.''  \footnote{Jon Jarrett,

1370:   \emph{op cit.}, p. 589}  And the trend in the Bell

1371: literature since Jarrett's paper has certainly been to concede that a

1372: violation of Jarrett's ``completeness'' cannot be quite so trivially

1373: written off (as involving a mere updating of information in the face

1374: of having previously used incomplete state descriptions), but rather

1375: must be understood as indicating \emph{some} sort of non-locality or

1376: ``holism'' or ``non-separability'' or non-causal ``passion at a

1377: distance.''  

1378: 

1379: We do not, therefore, wish to claim that Jarrett (and those who follow

1380: him in thinking his decomposition is in some way or other helpful in

1381: understanding Bell's locality condition and/or in establishing the

1382: peaceful coexistence of SR and QM  \footnote{For a systematic

1383:   review of the recent literature, see Berkovitz, Joseph, "Action at a

1384:   Distance in Quantum Mechanics", The Stanford Encyclopedia of

1385:   Philosophy (Spring 2007 Edition), Edward N. Zalta (ed.), URL =

1386:   http://plato.stanford.edu/archives/spr2007/entries/qm-action-distance/}

1387: ) fully commits precisely the mistake presented (through the stark

1388: contrast to Bell's views) in the previous section.  Rather, we intend

1389: only the weaker claim that Jarrett \emph{et al.} have been led, by

1390: Jarrett's initial analysis, down a path which is \emph{obviously

1391:   untenable} once one clearly understands Bell's own formulation of

1392: local causality (including especially the parallel status -- namely,

1393: both are \emph{beables} -- of ``settings'' and ``outcomes,'' and the 

1394: crucial distinction between superluminal causation and superluminal 

1395: signaling).  

1396: 

1397: 

1398: 

1399: 

1400: 

1401: It turns out to be a rather subtle question whether or not SR

1402: genuinely requires local causality in the sense of Figure 1.  

1403: \footnote{Much of Maudlin's excellent book, \emph{op cit.}, is

1404:   dedicated to exploring just this question.}

1405: But if one grants this (and virtually all physicists and commentators

1406: \emph{do}), then it really is possible to establish an 

1407: \begin{quote}

1408: ``essential

1409: conflict between any sharp formulation [of QM] and fundamental

1410: relativity.  That is to say, we have an apparent incompatibility, at

1411: the deepest level, between the two fundamental pillars of contemporary

1412: theory...'' (Bell, 1984, p. 172)

1413: \end{quote}

1414: The widespread claims to the contrary -- i.e., the claims that instead

1415: Bell's theorem refutes only some already-dubious, dogmatic,

1416: philosophically-motivated program to restore ``determinism'' or

1417: ``classicality'' or ``realism'' (and I mean here both classes of such

1418: claims mentioned in the introduction) -- turn out inevitably to have

1419: their roots in a failure to appreciate some aspect of Bell's own

1420: arguments.  

1421: 

1422: There is, in particular, a tendency for a relatively 

1423: superficial focus on the relatively formal aspects of Bell's

1424: arguments, to lead commentators astray.  For example, how many

1425: commentators have too-quickly breezed through the prosaic first

1426: section of Bell's 1964 paper (p. 14-21) -- where his reliance on the EPR

1427: argument ``\emph{from locality to} deterministic hidden variables''

1428: is made clear -- and simply jumped ahead to section 2's 

1429: Equation 1 (p. 15), hence erroneously inferring (and subsequently

1430: reporting to other physicists and ultimately teaching to students) 

1431: that the derivation ``begins with deterministic hidden variables''?  

1432: (1981, p. 157)  Likewise, we have here explored in detail 

1433: a similar case of too-quickly accepting some formal version of a

1434: premise used in Bell's derivation (such as ``factorizability'') 

1435: while failing to appreciate the rich conceptual context that gives it

1436: the precise meaning Bell intended.  

1437: 

1438: Our final conclusion, therefore, is a plea -- directed at physicists

1439: in general, but commentators on Bell's theorem, textbook writers, and

1440: students in particular -- to simply read (and not just read, but

1441: \emph{read}) Bell's writings.  They are

1442: truly a model of clarity and physical insight, and almost always

1443: convey the essential ideas much more lucidly and tersely than anything

1444: in the secondary Bell literature.  (I have no doubt this applies even to

1445: the current essay!)  

1446: Bell himself, in the preface to the first edition

1447: of his compiled papers (\emph{Speakable and Unspeakable in Quantum

1448:   Mechanics}) suggests that ``even quantum experts might begin with

1449: [chapter] 16, `Bertlmann's socks and the nature of reality', not

1450: skipping the slightly more technical material at the end.''  It is

1451: hard to disagree with that advice, although a strong case could be

1452: made also for Bell's 1990 essay (written after the first edition of

1453: the book, and hence included only in the more recent second edition)

1454: `La nouvelle cuisine,' in which the central importance and meaning of

1455: ``local causality'' is emphasized in lucid detail.  

1456: 

1457: If more physicists would only study Bell's papers instead of relying 

1458: on dubious secondary reports, they would, I think, come to appreciate

1459: that there really is here a serious inconsistency to worry about.  A

1460: much higher-level inconsistency between quantum theory 

1461: and (general) relativity

1462: has been the impetus, in recent decades, for enormous efforts spent

1463: pursuing (what Bell once referred to as) ``presently fashionable

1464: `string theories' of `everything'.''  (1990, p. 235)  How might

1465: a resolution of the more basic inconsistency identified by Bell shed

1466: light on (or radically alter the motivation and context for) attempts 

1467: to quantize gravity?  We can't possibly know until (perhaps long

1468: after) we face up squarely to Bell's important insights.

1469: 

1470: 

1471: 

1472: \end{document}

1473: 

1474: 

1475: