Jump to ContentJump to Main Navigation
The Emergent MultiverseQuantum Theory according to the Everett Interpretation$

David Wallace

Print publication date: 2012

Print ISBN-13: 9780199546961

Published to Oxford Scholarship Online: September 2012

DOI: 10.1093/acprof:oso/9780199546961.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 30 March 2017

(p.429) Appendix A Proof of the Branching-Decoherence Theorem

(p.429) Appendix A Proof of the Branching-Decoherence Theorem

Source:
The Emergent Multiverse
Publisher:
Oxford University Press

(p.429) Appendix A

Proof of the Branching-Decoherence Theorem

Recall that the theorem is

Branching-Decoherence Theorem: If Ƥ = { P ^ j i } is a history space and |ψ〉 is a quantum state, then

  1. (i) If Ƥ has branching structure (relative to |ψ〉) and α is a history then C ^ α | ψ 0 iff α is realised (with respect to |ψ〉.

  2. (ii) If the set Hist of all histories α such that C ^ α | ψ 0 has branching structure (that is, if no two histories in Hist agree on their nth index but not on all previous indices), then Ƥ also has branching structure (relative to |ψ〉), and the realised histories in that branching structure are just the histories in Hist.

  3. (iii) If Ƥ has branching structure (relative to |ψ〉), Ƥ satisfies the decoherence condition.

  4. (iv) If Ƥ satisfies the decoherence condition, it is a coarse-graining of a (decoherent) history space relative which has branching structure relative to |ψ〉.

To develop the proof, it will be helpful to modify Chapter 3’s index notation somewhat. Recall that a history is uniquely specified by a string of indices, one for each time-index. This being the case, I identify each history with that string of indices: a boldface index j denotes a sequence 〈j 0, … j n〉 of indices (one for each time t 0, … t n) such that the projector in the history corresponding to time t k is P ^ j k k . I then generalise the notion of history so that j can stand for any sequence of indices for times t 0, …t m, where mn; ℒ(j) will denote the length of the sequence minus one (so if the sequence is defined for times t 0, … t m then ℒ(j) = m). I write ij (p.430) to indicate that one of i and j is an initial segment of the other (so if ℒ(i) = ℒ(j), then ij entails i = j).

I define history operators for our generalised indices in the obvious way: if ℒ(j) = m, then we have

C ^ j = P ^ j m m P ^ j 0 0 .
(A.1)

Notice that

C ^ j = ( k l P ^ k l m + l ) ( k 1 P ^ k 1 m + 1 ) C ^ j = ( k ) = l + m ; k j C ^ k .
(A.2)

It follows from (A.2) that the decoherence criterion may be rewritten for generalized sequences as

ψ | C ^ i C ^ j | ψ = 0 unless i j .
(A.3)

Finally, I say that a generalized history is realized (for a branching history space) iff it is an initial segment of a realized history.

Part (i) of the theorem may now be proved by induction on the length of histories. For histories of length 1–2, it is a trivial consequence of the definition of a branching history; for longer histories, let j be an arbitrary history with ℒ(j) = m. We have

C ^ j | ψ = P ^ j m m C ^ j | ψ
(A.4)

where j is the initial segment of j with length ℒ(j ) = m − 1. If the theorem holds for sequences with ℒ(k) 〈 m, then it follows that this vanishes unless j is realized. Assuming j is indeed realized, it also follows that

P ^ j m 1 m 1 | ψ = P ^ j m 1 m 1 ( k ) = m 2 C ^ k | ψ = P ^ j m 1 m 1 ( k ) = m 2 , k realised C ^ k | ψ
(A.5)

and therefore, since j is the unique realized history with final element j m−1,

P ^ j m 1 m 1 | ψ = C ^ j | ψ .
(A.6)

Hence C ^ j | ψ = P ^ j m m P ^ j m 1 m 1 | ψ , which, by the definition of branching, vanishes unless j m−1 is the penultimate term on the unique history whose final term is j m.

(p.431) To prove part (ii), notice that if Hist has branching structure, we have

P ^ j m | ψ = ( k ) = m 1 , k m = j C ^ k | ψ = C ^ j | ψ
(A.7)

where j is the unique history in Hist with ℒ(j) = m and j m = j. (If there is no such, then P ^ j m | ψ = 0 . Then

P ^ j m + 1 P ^ j m | ψ = P ^ j m + 1 C ^ j | ψ = C ^ j+ | ψ
(A.8)

where j + is the history with final element j′ and initial history j. Since C ^ j+ | ψ is nonzero for at most one such j + P ^ j m + 1 m + 1 P ^ j m m | ψ is nonzero for at most one j.

Part (iii) is a trivial consequence of (i). To prove part (iv), suppose that our history space { P ^ j i } is decoherent (for given |ψ〉). I then define, for each time index m,

| j, m = C ^ j | ψ / C ^ j | ψ
(A.9)

where j is any sequence with ℒ(j) = m. If P ^ j m is any projector in the mth PVM in the history space, then

P ^ j m | j , m = | j , m ( j m = j ) P ^ j m | j , m = 0 ( otherwise )
(A.10)

Hence the set of operators

O ^ j m = | j, m j, m | , O ^ j m = P ^ j m ( j ) = m ; j m = j | j, m j, m |
(A.11)

is a PVM and { P ^ j m } is a coarse-graining of that PVM.

To see that the history space consisting of the family of these PVMs (for each m) has branching structure, it suffices to note that:

  1. 1. Since |ψ〉 can always be written as a superposition of the | j , m , O ^ j m | ψ = 0.

  2. 2.

    O ^ j m + 1 Q ^ k m + 1 | ψ = k O ^ j m + 1 P ^ k m + 1 Q ^ k m | ψ = j:k j C ^ j | ψ O ^ j m + 1 | j , m + 1
    (A.12)

    so O ^ j m + 1 Q ^ k m | ψ = 0 . (p.432)

  3. 3.

    Q ^ k m+1 Q ^ j m = | k, m + 1 k, m + 1 | j, m j,m|
    (A.13)

    and

    k, m + 1 j, m α ψ | C ^ k C ^ j | ψ
    (A.14)

    so Q ^ k m + 1 Q ^ j m vanishes unless j is an initial segment of k.