[SOURCE-FIDELITY NOTICE: This content was extracted from a PDF. Two-dimensional equation layout may be lost, including fraction bars, numerator/denominator grouping, superscripts, roots, matrices, and stacked limits. Treat arithmetic or algebraic discrepancies as inconclusive unless they can be verified against an equation-preserving source.]
Typicality and Unitary Refinement: A Conditional
Representation Theorem for the Born Measure
From Homogeneous Branch Abundance to Projective Event Measure
Michael Tolbert
Version 3.1 | September 2026
Abstract. The Everett or many-worlds interpretation retains unitary quantum dynamics and
denies stochastic collapse, yet must explain why branch-relative observations exhibit Born-rule
statistics when every decohered outcome occurs. This paper separates the construction of
a non-arbitrary measure on branch events from the justification for using that measure as
rational credence. For a finite projective decomposition of a pure state, rational squared norms
admit admissible homogeneous equal-norm ancillary refinements. Within that refinement class,
permutation symmetry identifies primitive-equivalent descendants, disjoint cardinal composition
uniquely represents their finite abundance by normalized cardinality, and common-refinement
invariance removes denominator dependence. The rational construction extends uniquely to
arbitrary real squared norms either by continuity or, under an explicit event-mass monotonicity
axiom, by monotone squeezing. The resulting event measure is
T(P | ψ) =∥Pψ∥
2
=⟨ψ|P|ψ⟩.
The argument does not derive probability from unitary dynamics alone. It relocates finite
additivity from an assumed probability measure on projectors to explicit cardinal compo-
sition on homogeneous refinement spaces. A separate credence representation theorem is
proved on a refined finite record algebra using explicit additivity, coarse-graining consistency,
refinement-implementation invariance, and regularity. The principal contribution is a condi-
tional representation theorem that distinguishes symmetry, abundance, event measure, and
credence while making the substantive assumptions and their domains explicit. Keywords:
Everett interpretation; many-worlds interpretation; Born rule; typicality; decoherence; pointer
states; quantum probability.
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
1 Introduction and Scope
The Everett interpretation preserves universal unitary evolution. Measurement therefore does
not select one outcome and eliminate the others; instead, decoherence produces approximately
autonomous records corresponding to multiple outcomes. The familiar Born rule nevertheless
assigns a normalized weight to each record event. A satisfactory account must explain both why a
particular quantitative measure applies and why branch-relative observers should use that measure
in expectation or action.
These are distinct problems. The measure problem asks which functional represents the size
of a branch event. The credence problem asks why such size should constrain rational belief or
decision. Conflating them allows a physical symmetry argument to inherit unannounced probabilistic
premises, or allows a decision-theoretic argument to appear to construct physical measure solely
from preference. The present paper treats the measure problem first and introduces the epistemic
bridge only afterward.
The central result is conditional. Unitary quantum mechanics supplies Hilbert-space geometry,
permits ancillary refinements, and preserves squared norm. Decoherence constrains a coarse algebra
of stable record events. Neither fact by itself selects a probability functional. The additional
structural premise is that equal-norm descendants related by the symmetries of a homogeneous
refinement count as primitive-equivalent units of branch abundance. Disjoint collections of such
units compose cardinally. The paper proves that this conception uniquely induces the Born measure
on finite projective events within the class of admissible homogeneous equal-norm refinements.
The result should not be described as a derivation from the Schrödinger equation alone. It is a
representation theorem: given pointer-defined coarse events, admissible homogeneous equal-norm
refinement, primitive equivalence, disjoint cardinal composition, common-refinement invariance, and
a regular real extension, the Born form is unique within that refinement class. Separate self-locating
or decision-theoretic assumptions are required to identify this measure with single-case credence.
2 Why Familiar Ingredients Do Not Establish Uniqueness
2.1 Geometry is not yet probability
LetHbe a finite-dimensional complex Hilbert space,|ψ⟩∈Hnormalized, and{P
i
}
n
i=1
a projective
decomposition:
P
i
P
j
= δ
ij
P
i
,
n
X
i=1
P
i
= I.
The geometric size of outcome i is
x
i
:=∥P
i
ψ∥
2
, x
i
≥ 0,
X
i
x
i
= 1.
For orthogonal P and Q, Pythagoras gives
∥(P + Q)ψ∥
2
=∥Pψ∥
2
+∥Qψ∥
2
.
This makes squared norm a natural candidate for a measure, but naturalness is not uniqueness.
Unitarity preserves the norm; it does not establish that every acceptable branch-size functional
must be the norm squared.
1
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
Indeed, for any continuous f : [0, 1]→ [0,∞) with f (0) = 0 and f (x) > 0 for x > 0, the rule
W
f
i
f (x
i
)
P
j
f (x
j
)
is normalized, continuous, phase invariant, and permutation covariant. The power familyf(x) =x
α
includes infinitely many non-Born alternatives. Thus normalization, continuity, and symmetry do
not suffice.
2.2 Trace additivity and the logical order
If one definesμ(P) =Tr(ρP), then positivity, normalization, and finite additivity on orthogonal
projectors follow immediately. For a pure state ρ =|ψ⟩⟨ψ|,
μ(P ) = Tr(ρP ) =⟨ψ|P|ψ⟩ =∥Pψ∥
2
.
The trace rule therefore contains the complete Born structure. It does not explain why the relevant
branch measure must be linear in the first place. Gleason’s theorem reverses the direction by
characterizing additive measures on projectors, but additivity remains an initial measure assumption.
The present construction asks whether that structure can be grounded in a prior, non-stochastic
notion of finite abundance.
2.3 Decoherence supplies events, not weights
A measurement interaction correlates system alternatives with environmental records:
X
i
c
i
|s
i
⟩|A
0
⟩|E
0
⟩−→
X
i
c
i
|s
i
⟩|A
i
⟩|E
i
⟩.
Tracing over the environment suppresses off-diagonal terms when⟨E
j
| E
i
⟩is small. Decoherence
thereby identifies approximately stable record sectors and explains why coherent alternatives can
function as exclusive macroscopic events. It does not, without further argument, interpret the
diagonal coefficients as probabilities.
The exact theorems below concern a finite projective idealization. Their physical application is
relative to a pointer-admissible coarse algebra selected to an operational tolerance. This restriction
prevents equal-norm refinement from arbitrarily choosing the macroscopic event basis. Refinement
occurs within already identified record sectors.
3 Ancillary Refinement and the Additivity Gap
3.1 Algebra-relative equivalence
A refinement adds orthogonal microscopic labels while preserving a parent event at the coarse level.
If a record ancilla R is introduced in a normalized state |φ⟩, then
Tr
R
(PρP ⊗|φ⟩⟨φ|) = PρP.
Consequently, system observables have the same restricted expectation data before and after
refinement. The global states are not identical. They agree only on the coarse algebra that ignores
2
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
the ancillary label. Refinement invariance is therefore an operational noncontextuality principle,
not a claim that two globally different vectors are the same physical state.
Bare refinement invariance remains too weak to yield additivity. Let
W
h
(P | ψ) = h
∥Pψ∥
2
,
withh(0) = 0,h(1) = 1. If a parent of squared normxis split into two orthogonal descendants of
squared norm x/2, then for h(x) = x
2
,
W
h
(P
1
- P
2
| ψ) = x
2
, W
h
(P
1
| ψ) + W
h
(P
2
| ψ) =
x
2
2
.
The parent and refined union preserve the same coarse value, although the descendant values do
not add. Operational equivalence therefore cannot silently be replaced by summation.
3.2 Symmetry establishes equality, not aggregation
In an equal-amplitude state withMorthogonal descendants, permutation symmetry forbids privi-
leging any label. It follows that singleton descendants have equal primitive status. It does not follow
that each has numerical value 1/M. A normalized, permutation-invariant set functional can assign
value one to every nonempty subset. The inference from equal status to relative weight requires a
compositional account of finite abundance.
This is the additivity gap. The argument below does not conceal it. Disjoint composition is
stated explicitly on the representative space, and the resulting cardinality theorem is presented as
conditional on that premise.
4 Finite Homogeneous Refinement Spaces
4.1 Equal-norm refinements
Definition 4.1 (Rational branch geometry). The squared-norm vectorx= (x
1
,...,x
n
) is rational
when there exist nonnegative integers m
i
and a positive integer M such that
x
i
=
m
i
M
,
X
i
m
i
= M.
Definition 4.2 (Equal-norm refinement). An equal-normM-refinement maps each nonzero nor-
malized branch vector|ψ
i
⟩=P
i
|ψ⟩/
√
x
i
intom
i
mutually orthogonal descendants, such that every
descendant in the refined total state has amplitude modulus 1/
√
M.
Definition 4.3 (Admissible homogeneous refinement). An ancillary refinement of a pointer-defined
coarse event is admissible and homogeneous when its nonzero descendants are mutually orthogonal,
have equal squared norm, and are related by state-preserving label permutations within the refined
event algebra. Only descendants satisfying these conditions are treated as primitive-equivalent
representatives for cardinal abundance.
Lemma 4.4 (Equal norm is necessary for homogeneous permutation equivalence). Let
|Ψ⟩ =
M
X
a=1
c
a
|a⟩
3
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
be an orthogonal descendant decomposition with everyc
a
̸= 0. Suppose that for every transposition of
descendant labelsaandb, there is a corresponding unitaryU
ab
such thatU
ab
|a⟩=|b⟩,U
ab
|b⟩=|a⟩,
the remaining descendant rays are preserved, and
U
ab
|Ψ⟩ = e
iθ
ab
|Ψ⟩.
Then |c
1
| =··· =|c
M
|.
Proof. Comparing the coefficients of the exchanged orthogonal vectors gives
c
b
= e
iθ
ab
c
a
, c
a
= e
iθ
ab
c
b
,
and hence|c
a
|=|c
b
|. Because the transposition condition holds for every pair, all coefficient moduli
are equal. Thus unequal-norm descendants cannot form one homogeneous permutation-equivalence
class of the kind used in the cardinality theorem.
Theorem 4.5 (Existence of admissible equal-norm refinement). If x
i
=∥P
i
ψ∥
2
= m
i
/M, then an
ancillary isometry, extendable to a unitary on a sufficiently large Hilbert space, represents outcomei
by exactly m
i
orthogonal descendants, each of squared norm 1/M.
Proof.
Choose ancillary vectors|i,k⟩,k= 1,...,m
i
, orthonormal across all pairs (i,k). On the
span of the vectors |ψ
i
⟩|0⟩ define
V
|ψ
i
⟩|0⟩
1
√
m
i
m
i
X
k=1
|ψ
i
⟩|i,k⟩.
The inputs are orthonormal because the projectors have orthogonal ranges, and the outputs are
orthonormal because the ancillary labels are orthogonal. ThusVis an isometry on the specified
subspace. If necessary, enlarge the ancillary space by an orthogonal null sector. Complete the
input vectors to an orthonormal basis of the enlarged input space and complete their images to an
orthonormal basis of an output space of equal dimension. Mapping the completed input basis to
the completed output basis extends V to a unitary. Since
|ψ⟩ =
X
i
r
m
i
M
|ψ
i
⟩,
we obtain
V (|ψ⟩|0⟩) =
1
√
M
X
i
m
i
X
k=1
|ψ
i
⟩|i,k⟩.
There are M nonzero descendants, all with squared norm 1/M, and m
i
lie under outcome i.
The theorem establishes unitary representability only. No typicality or probability has yet been
assigned.
4.2 Why cardinality is the unique finite abundance concept
Let Ω
M
be an M-element homogeneous representative set. An event is a subset A⊆ Ω
M
.
Definition 4.6 (Homogeneous finite-abundance scheme). A family{a
M
}assignsa
M
(A)∈[0,1] to
each A⊆ Ω
M
and satisfies:
4
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
- a
M
(∅) = 0 and a
M
(Ω
M
) = 1;
- a
M
(πA) = a
M
(A) for every permutation π of Ω
M
;
- if A∩ B = ∅, then a
M
(A∪ B) = a
M
(A) + a
M
(B).
The three conditions above form the minimal axiom set used in the cardinality representation.
Disjoint cardinal composition (A3) is not inferred from symmetry. It states the mereological content
of the selected abundance concept: nonoverlapping collections contain the sum of the primitive
units in each collection. If this axiom is rejected, normalized cardinality is not uniquely forced.
Theorem 4.7 (Cardinality representation). Every homogeneous finite-abundance scheme is uniquely
represented by
a
M
(A) =
|A|
M
.
Proof.By permutation invariance, every singleton has one common values
M
. Because Ω
M
is the
disjoint union of itsMsingleton subsets, repeated use of (A3) gives 1 =Ms
M
, hences
M
= 1/M. If
A contains m elements, it is the disjoint union of m singletons, so a
M
(A) = m/M =|A|/M.
Corollary 4.8 (Uniform-refinement invariance). Every homogeneous finite-abundance scheme
automatically satisfies
a
LM
(A
(L)
) = a
M
(A)
under L-fold uniform refinement.
Proof.
By Theorem 4.7,a
M
(A) =|A|/M. Uniform refinement gives|A
(L)
|=L|A|and|Ω
LM
|=LM,
so
a
LM
(A
(L)
) =
L|A|
LM
=
|A|
M
= a
M
(A).
This theorem is rigorous but deliberately conditional. It does not derive disjoint composition
from unitary dynamics. It proves that normalized cardinality is uniquely fixed once finite abundance
is required to compose over nonoverlapping collections of primitive-equivalent representatives.
4.3 Diagnostic countermodels and the role of the assumptions
The assumptions above perform distinct tasks. The following constructions are diagnostic rather than
alternative physical theories; they exhibit recognizable ambiguities when individual requirements
are removed.
Proposition 4.9 (Functional separation of the abundance assumptions). On finite homogeneous
representative spaces:
1.without permutation invariance, normalized additive assignments may distinguish descendant
labels;
without disjoint cardinal composition, nonlinear functions of normalized cardinality may
satisfy normalization, permutation invariance, and uniform-refinement invariance while failing
additivity;
5
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
3.without uniform- and common-refinement invariance, assignments may depend on artificial
descendant multiplicity or rational presentation;
4.
without regularity, agreement with normalized cardinality on rational weights does not determine
irrational weights.
Proof. Let Ω
M
={1,...,M}. For the first claim, choose unequal positive weights w
j
and set
a
(w)
M
(A) =
P
j∈A
w
j
P
M
j=1
w
j
.
This is normalized and additive but not permutation invariant. For the second, set
a
(2)
M
(A) =
|A|
M
2
.
This is normalized, permutation invariant, and unchanged by uniform refinement, but it is not
additive on disjoint nonempty events. For sensitivity to uniform refinement, definea
size
M
(∅) = 0,
a
size
M
(Ω
M
) = 1, anda
size
M
(A) =|A|/(M+ 1) for proper nonemptyA; this changes underL-fold
splitting. For common-refinement sensitivity, define, for a rational presentation m/M,
b
ε
(m,M ) =
0,m = 0,
1,m = M,
m
M
- ε
m(M − m)
M
2
gcd(m,M )− 1
gcd(m,M )
, 0 < m < M,
where 0< ε≤1. The assignment preserves endpoints and remains in [0,1], agrees withm/Mon
reduced presentations, but generally changes under a common scaling (m,M)7→(Lm,LM). Finally,
the function
f (x) =
(
x,x∈ Q,
1− x, x /∈ Q
agrees with the rational construction but is neither continuous nor monotone. Thus each condition
removes a distinct ambiguity.
5 The Rational Born Measure
For an eventS ⊆ {1,...,n}, letP
S
=
P
i∈S
P
i
and letA
S
⊆Ω
M
contain the descendants whose
parent outcome lies in S.
Theorem 5.1 (Rational typicality). For rational squared normsx
i
=m
i
/M, the cardinal abundance
of S is
T
M
(S) =
|A
S
|
M
=
X
i∈S
m
i
M
=
X
i∈S
x
i
=∥P
S
ψ∥
2
.
Proof.By 4.5, outcomeicontributesm
i
representatives. Therefore|A
S
|=
P
i∈S
m
i
. Division by
Myields the cardinal ratio, and orthogonality gives the equality with projected squared norm.
Corollary 5.2 (Finite additivity). If S and R are disjoint parent events, then
T
M
(S∪ R) = T
M
(S) + T
M
(R).
6
|A|
M
.
Theorem 5.4 (Common-refinement independence). Any two rational equal-norm refinements
representing the same squared-norm vector assign the same typicality to every parent event.
Proof.Suppose the denominators areMandN. Refine both toL=lcm(M,N). In either
representation, outcomeithen hasLx
i
descendants. By 5.3, the common refinement preserves each
original ratio, so the original assignments agree.
The common-refinement theorem discards absolute branch number. Typicality belongs to an
equivalence class of homogeneous representations, not to a fundamental census of microscopic worlds.
6 Unique Extension to Real Squared Norms
Let ∆
n−1
be the standard simplex and ∆
Q
n−1
its dense rational subset. On rational points, Theo-
rems 5.1 and 5.4 define the well-posed functional
T
Q
(S,x) =
X
i∈S
x
i
.
Theorem 6.1 (Unique continuous extension). For every eventS, T
Q
(S,·) has a unique continuous
extension to ∆
n−1
, namely
T(S,x) =
X
i∈S
x
i
.
Proof. Choose rational x
(r)
→ x. Finite summation is continuous, so
T
Q
(S,x
(r)
) =
X
i∈S
x
(r)
i
−→
X
i∈S
x
i
.
The limit depends only onx. Any other continuous extension agreeing on the dense rational subset
must have the same limit at every x.
7
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
6.1 Event-mass monotonicity and factorization
For a fixed event S ⊆{1,...,n}, define the event-mass preorder on the simplex by
x⪯
S
y ⇐⇒
X
i∈S
x
i
≤
X
i∈S
y
i
.
Definition 6.2 (Event-Mass Monotonicity). An extension T(S,·) satisfies Event-Mass Monotonicity
when, for every event S and all x,y ∈ ∆
n−1
,
x⪯
S
y=⇒T(S,x)≤ T(S,y).
Thus increasing the total branch mass assigned to an event cannot decrease its typicality. This is
stronger than coordinatewise monotonicity and is stated explicitly because it orders simplex points
by the total mass of the event under evaluation.
Proposition 6.3 (Event-mass factorization). LetSbe a nonempty proper event. If T(S,·) satisfies
Event-Mass Monotonicity, then there exists a monotone function f
S
: [0, 1]→ [0, 1] such that
T(S,x) = f
S
X
i∈S
x
i
!
.
Proof.
Ifxandyassign equal total mass toS, then bothx ⪯
S
yandy ⪯
S
x. Event-Mass
Monotonicity gives T(S,x)≤T(S,y) and T(S,y)≤T(S,x), hence equality. Therefore T(S,x)
depends only ont=
P
i∈S
x
i
. BecauseSis nonempty and proper, everyt ∈[0,1] is attained by
some simplex point: assign total masstwithinSand total mass 1−twithin its complement. Define
f
S
(t) to be the resulting common value. Ift≤ u, choose simplex points of event massestandu;
Event-Mass Monotonicity gives f
S
(t)≤ f
S
(u).
Theorem 6.4 (Unique Event-Mass Monotone extension). Any normalized extension of T
Q
satisfying
Event-Mass Monotonicity is uniquely given, for every event S, by
T(S,x) =
X
i∈S
x
i
.
Proof.
For the trivial events, normalization gives T(∅,x) = 0 and T(Ω,x) = 1, which equal their
event masses. LetSnow be nonempty and proper. By Proposition 6.3, T(S,x) =f
S
(t) with
t=
P
i∈S
x
i
andf
S
monotone. For each rationalq ∈[0,1], choose a rational simplex point with
event massq. Agreement with T
Q
givesf
S
(q) =q. For arbitraryt, choose rational sequencesq
r
↑ t
and s
r
↓ t. Then
q
r
= f
S
(q
r
)≤ f
S
(t)≤ f
S
(s
r
) = s
r
,
so f
S
(t) = t.
Theorem 6.5 (Born measure representation within admissible homogeneous refinements). Under
admissible homogeneous equal-norm refinement, primitive equivalence, disjoint cardinal composition,
common-refinement invariance, and either continuous extension or Event-Mass Monotonicity, every
finite projective event satisfies
T(P | ψ) =∥Pψ∥
2
=⟨ψ|P|ψ⟩
and T is normalized, nonnegative, and finitely additive on orthogonal events.
8
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
Proof.The rational result follows from Theorem 5.1, and representation independence follows from
Theorem 5.4. The real extension follows independently from Theorem 6.1 or Theorem 6.4. Positivity
and normalization are immediate. IfPQ= 0, orthogonality gives∥(P+Q)ψ∥
2
=∥Pψ∥
2
+∥Qψ∥
2
.
For ρ =|ψ⟩⟨ψ|, the trace rule is recovered at the end:
T(P | ψ) = Tr(ρP ).
The trace is thus an operator representation of the measure already constructed, rather than the
starting point of the argument.
6.2 Why alternative powers fail
Lemma 6.6 (Refinement aggregation equation). Suppose an unnormalized contributiong(x) is
assigned to a parent event of squared normx, an admissible equalL-fold refinement producesL
primitive-equivalent descendants of squared normx/L, and the parent contribution equals the disjoint
aggregate of the descendant contributions. Then
g(x) = Lg(x/L).
Proof.
Primitive equivalence gives each descendant contributiong(x/L), and disjoint cardinal
composition gives the parent contribution as their sum.
For g(x) = x
α
, Lemma 6.6 gives
x
α
= L(x/L)
α
= x
α
L
1−α
.
For anyx >0, choosingL= 2 yields 2
1−α
= 1, henceα= 1. Sincexis already squared amplitude,
this selects amplitude exponent two. This argument is illustrative; Theorem 4.7 already excludes
nonlinear cardinal assignments on homogeneous representative spaces.
7 Events, Composition, and Repeated Records
The relevant event operation is disjoint union in a pointer-defined Boolean algebra. IfP
A
P
B
= 0,
then “AorB” is represented byP
A
+P
B
. Tensor products play a different role. For product states
and product projectors,
T(E× F | ψ⊗ φ) =∥(P
E
⊗ P
F
)(ψ⊗ φ)∥
2
= T(E | ψ)T(F | φ).
Additive union and product composition are therefore compatible: direct sum represents exclusive
disjunction, while tensor product represents conjunction across systems or trials.
ForNindependent copies of a two-outcome state with one-trial typicalitiesxand 1− x, the
frequency-k subspace has typicality
T
N
(k) =
N
k
!
x
k
(1− x)
N−k
.
Its relative-frequency mean is x and variance is x(1− x)/N. Chebyshev’s inequality yields
T
N
k
N
− x
≥ ε
≤
x(1− x)
Nε
2
.
9
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
Thus record sectors with frequencies bounded away from the Born value have vanishing total
typicality asN →∞. This is a concentration result for the derived measure. It does not by itself
show why a particular observer must disregard low-typicality successors.
8 From Measure to Rational Credence
LetBbe the finite Boolean algebra of coarse pointer events with atomic outcome set Ω ={1,...,n}.
For an admissible refinementR, let Ω
R
be its finite descendant set, letB
R
= 2
Ω
R
be the refined
record algebra, and let
π
R
: Ω
R
→ Ω
map each descendant to its parent pointer outcome. For a coarse eventS ⊆Ω, the refined
representative is π
−1
R
(S)⊆ Ω
R
.
Write C
B
(S | ψ) for coarse pre-observation credence and C
R
(A| V
R
ψ) for credence on the refined
algebra. Assume:
(C1) Normalization:
C
R
(∅| V
R
ψ) = 0,C
R
(Ω
R
| V
R
ψ) = 1.
(C2) Refined-algebra additivity: if A,B ∈B
R
are disjoint, then
C
R
(A∪ B | V
R
ψ) = C
R
(A| V
R
ψ) + C
R
(B | V
R
ψ).
(C3)
State-preserving permutation indifference: credence depends on the refined state ray, so
global phases are immaterial. For every admissible permutationπof equal-norm descendants
whose implementing unitary satisfies
U
π
V
R
|ψ⟩ = e
iθ
π
V
R
|ψ⟩,
one has, at the same refined physical state,
C
R
(πA| V
R
ψ) = C
R
(A| V
R
ψ).
(C4) Coarse-graining consistency: for every S ∈B,
C
B
(S | ψ) = C
R
(π
−1
R
(S)| V
R
ψ).
(C5)
Regular extension: coarse credence is continuous in the squared-norm vector, or satisfies
the epistemic analogue of Event-Mass Monotonicity: if x⪯
S
y, then
C
B
(S | x)≤ C
B
(S | y),
together with agreement on rational refinements.
Condition (C4) is the formal coarse/refined consistency requirement needed from environmental
separability and diachronic consistency. Refinement-implementation invariance then follows from
(C4), rather than being imposed independently. Diachronic or decision-theoretic principles may
motivate these equations, but the representation theorem uses the equations explicitly.
10
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
Corollary 8.1 (Refinement-implementation invariance). IfRandR
′
are admissible refinements of
the same coarse experiment and both satisfy coarse-graining consistency (C4), then, for every coarse
event S,
C
R
(π
−1
R
(S)| V
R
ψ) = C
R
′
(π
−1
R
′
(S)| V
R
′
ψ).
Proof. By (C4), each side equals the same coarse credence C
B
(S | ψ).
Theorem 8.2 (Born credence representation). Under (C1)–(C5) and Theorem 4.5,
C
B
(S | ψ) = T(P
S
| ψ) =∥P
S
ψ∥
2
for every event S in the finite pointer algebra.
Proof.
For rational branch geometry, an admissible equal-norm refinement hasMdescendants, with
m
i
under pointer outcomei. By (C3), all singleton descendants have a common credencec. By
(C1), (C2), and the disjoint union of the M singleton records,
1 = C
R
(Ω
R
| V
R
ψ) = Mc,
so c = 1/M. Refined-algebra additivity gives
C
R
(π
−1
R
({i})| V
R
ψ) =
m
i
M
= x
i
.
By (C4),
C
B
({i}| ψ) = x
i
.
Applying (C2) to descendant unions and then (C4) gives the result for every coarse unionS.
Corollary 8.1 removes dependence on the ancillary implementation. Finally, (C5) extends the
rational equality either continuously or by the same factorization-and-squeezing argument used in
Theorem 6.4.
The physical and epistemic theorems have different domains. Disjoint composition in Theorem 4.7
concerns abundance of primitive representatives. Condition (C2) concerns coherent degrees of belief
on the refined record algebra. Condition (C4) explicitly bridges refined and coarse credences and
yields implementation invariance as Corollary 8.1; they are not consequences of unitary dynamics.
Critics may reject permutation indifference, coarse-graining consistency, refinement-implementation
invariance, or the self-locating interpretation that motivates them. The theorem’s role is to display
a non-circular division of labor: the physical construction fixes the quantitative measure, while
explicit epistemic principles determine when that measure guides rational expectation.
Decision-theoretic approaches can provide an alternative bridge. If preferences over quantum acts
satisfy an expected-utility representation and the relevant equivalence and branching-indifference
conditions, the act value takes the form
V =
X
i
T(P
i
| ψ)u(R
i
) =
X
i
∥P
i
ψ∥
2
u(R
i
).
This practical role does not need to be built into the cardinality theorem itself.
11
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
9 Relation to Existing Born-Rule Programs
The present argument occupies an intermediate position among established approaches. Gleason
begins with an additive measure on projectors and derives trace form. Frequency-operator and
typicality programs analyze the concentration of repeated-trial records, commonly using Hilbert
norm as the measure of negligibility. Envariance motivates equal treatment of equal-modulus
alternatives through system-environment symmetries and uses fine graining for rationally related
amplitudes. Deutsch-Wallace decision theory addresses rational preference, while self-locating
approaches address branch-relative credence.
The distinctive claim here is narrower than any complete replacement of those programs. Symme-
try identifies primitive equivalence but does not supply aggregation. Cardinal abundance supplies
aggregation on a finite homogeneous representative space. Common refinement supplies repre-
sentation independence. Regularity supplies the real extension. Epistemic or decision-theoretic
principles supply the probability role. The novelty lies in making these stages explicit and in locating
finite additivity precisely at the transition from homogeneous representative units to their disjoint
collections.
This construction also differs from naive branch counting. Decoherence generally does not provide
a unique microscopic branch census, and arbitrary counting changes under coarse graining. The
present measure is assigned to equivalence classes of equal-norm refinements of fixed pointer-defined
coarse events. Absolute descendant number is discarded; only relative cardinality preserved by
common refinement survives. The counted objects are therefore representation devices anchored to
a coarse event algebra, not asserted to be a fundamental inventory of worlds.
10 Principal Objections and Replies
10.1 Does disjoint composition hide probability additivity?
It contains the same algebraic equation as finite additivity, and the paper does not deny this. The
claim is interpretive and structural: (A3) is imposed on collections of primitive counting units before
stochastic selection, credence, or probability on the original projector algebra is introduced. The
theorem relocates the premise and identifies its content; it does not derive it from dynamics. A
critic who rejects cardinal composition rejects the selected conception of abundance, not a concealed
mathematical step.
10.2 Does equal norm already assume the Born rule?
Equal norm is initially a geometric and symmetry condition. Unitary evolution preserves the Hilbert
norm, and equal-norm orthogonal descendants are related by label permutations. No probability
interpretation is needed to state those facts. The substantive premise is that these symmetry-related
descendants are the appropriate primitive units for abundance. The paper explains why squared
norm acquires measure structure; it does not discover squared norm independently of Hilbert
geometry.
10.3 Why should counterfactual refinements constrain actual measure?
Refinements function as representation tests. If an allowed ancillary interaction adds only microscopic
labels, preserves the parent record on the coarse algebra, and can be erased or ignored without
12
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
altering that coarse experiment, a description-independent event measure should not depend on
whether the labels are displayed. This is a noncontextuality premise. It is physically motivated but
not compelled by unitary dynamics.
10.4 Is normalized cardinality only uniform probability renamed?
The two have the same finite algebra. They need not have the same interpretation. Cardinal
abundance answers what fraction of a homogeneous representative space has a property; probability
additionally concerns chance, credence, or action. The distinction is sufficient to separate the
measure theorem from the credence theorem, but it does not make the latter unnecessary.
10.5 Does decoherence select exact branches?
No. Realistic decoherence yields approximate stability and approximate orthogonality relative to a
system-environment split, time scale, and operational tolerance. The finite theorem idealizes those
records as orthogonal projectors. A complete robustness theorem should bound changes in typicality
when record sectors have small residual overlap.
10.6 How general is the result?
The proved theorem concerns normalized pure states, finite projective decompositions, exact
orthogonality, rational refinements, and a regular extension to real squared norms. Mixed states
require purification or preparation noncontextuality. POVMs require invariance under equivalent
Naimark dilations. Countably infinite decompositions require an additional continuity condition
sufficient for sigma additivity. Relativistic quantum theory requires a covariant event framework.
These are open extensions rather than implicit consequences of the finite result.
11 Discussion: Structural Naturalness and Limits
The Born measure is natural within the stated framework for three connected reasons. First,
Hilbert geometry supplies equal-norm refinement units, and unitary permutations make those units
structurally interchangeable. Second, cardinal composition is the unique finite abundance rule for
disjoint collections of such primitive-equivalent units. Third, common refinement makes parent-event
ratios independent of denominator or descriptive resolution. The result is not obtained by labeling
squared norm a probability; it is obtained by showing that the refinement structure gives squared
norm the unique stable representation of relative abundance.
This naturalness claim is conditional rather than absolute. A critic may deny that counterfactual
refinements constrain objective measure, reject cardinal composition as explanatory, or refuse the
epistemic bridge from typicality to credence. The framework’s contribution is to make the location
and cost of each rejection precise. It replaces an ambiguous appeal to symmetry with a sequence of
independently assessable premises.
The most important unresolved physical issue is approximation. Pointer events are emergent,
while the proof uses exact projectors. The most important mathematical extensions concern mixed
states, POVMs, infinite outcomes, and local quantum field theory. The most important interpretive
issue is whether self-locating or decision-theoretic principles provide a compelling probability role
13
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
for the measure. None of these limitations undermines the finite representation theorem, but each
limits the theorem’s reach.
12 Conclusion
This paper has established a conditional representation theorem for the Born measure within the class
of admissible homogeneous equal-norm refinements. Unitary quantum mechanics permits rational
branch geometries to be represented by such ancillary refinements. Permutation symmetry renders
the descendants primitive-equivalent but does not assign numerical weights. When finite abundance
is required to compose over disjoint collections of those units, normalized cardinality is uniquely
fixed. Uniform and common refinement make the resulting ratios representation independent.
Continuity gives a unique extension to arbitrary real squared norms; alternatively, the explicitly
stated Event-Mass Monotonicity axiom yields event-mass factorization and the same unique extension
by monotone squeezing. The induced measure on coarse projective events is
T(P | ψ) =∥Pψ∥
2
=⟨ψ|P|ψ⟩.
The argument does not eliminate additivity. It relocates finite additivity from an assumed probability
measure on projectors to explicit mereological composition on a homogeneous representative space.
Nor does it derive rational credence from dynamics alone. The separate credence theorem defines
beliefs on a refined finite record algebra and uses explicit coarse-graining consistency, refinement-
implementation invariance, and regularity to identify coarse branch-relative credence with the
derived measure.
The framework’s principal contribution is therefore a disciplined separation of four stages:
symmetry identifies homogeneous equivalent representatives; cardinal composition supplies finite
abundance; refinement invariance and regularity induce the Born measure; and epistemic consistency
conditions supply the probability role. The diagnostic countermodels clarify the distinct work
performed by these assumptions. This does not settle every form of the Everettian probability
problem, but it clarifies the additivity gap and gives a mathematically explicit account of how
Born-form measure can arise within a deterministic unitary theory under a transparent package of
structural assumptions.
References
[Busch(2003)]Busch, P. (2003). Quantum states and generalized observables: A simple proof of
Gleason’s theorem. Physical Review Letters, 91, 120403.
[Deutsch(1999)]
Deutsch, D. (1999). Quantum theory of probability and decisions. Proceedings of
the Royal Society A, 455, 3129–3137.
[Farhi et al.(1989)]Farhi, E., Goldstone, J., and Gutmann, S. (1989). How probability arises in
quantum mechanics. Annals of Physics, 192, 368–382.
[Gleason(1957)]
Gleason, A. M. (1957). Measures on the closed subspaces of a Hilbert space. Journal
of Mathematics and Mechanics, 6, 885–893.
[Greaves(2007)]
Greaves, H. (2007). Probability in the Everett interpretation. Philosophy Compass,
2, 109–128.
14
Typicality, Unitary Refinement, and the Born RuleMichael Tolbert
[Hartle(1968)]Hartle, J. B. (1968). Quantum mechanics of individual systems. American Journal
of Physics, 36, 704–712.
[Joos et al.(2003)]Joos, E., Zeh, H. D., Kiefer, C., Giulini, D., Kupsch, J., and Stamatescu, I.-O.
(2003). Decoherence and the Appearance of a Classical World in Quantum Theory. Springer.
[Price(2010)]Price, H. (2010). Decisions, decisions, decisions: Can Savage salvage Everettian
probability? In S. Saunders et al. (Eds.), Many Worlds? Everett, Quantum Theory, and Reality.
Oxford University Press.
[Saunders(2021)]Saunders, S. (2021). Branch-counting in the Everett interpretation of quantum
mechanics. Proceedings of the Royal Society A, 477, 20210600.
[Schlosshauer and Fine(2005)]
Schlosshauer, M., and Fine, A. (2005). On Zurek’s derivation of the
Born rule. Foundations of Physics, 35, 197–213.
[Sebens and Carroll(2018)]Sebens, C. T., and Carroll, S. M. (2018). Self-locating uncertainty and
the origin of probability in Everettian quantum mechanics. British Journal for the Philosophy
of Science, 69, 25–74.
[Wallace(2012)]Wallace, D. (2012). The Emergent Multiverse: Quantum Theory According to the
Everett Interpretation. Oxford University Press.
[Zurek(1981)]Zurek, W. H. (1981). Pointer basis of quantum apparatus: Into what mixture does
the wave packet collapse? Physical Review D, 24, 1516–1525.
[Zurek(2003)]Zurek, W. H. (2003). Decoherence, einselection, and the quantum origins of the
classical. Reviews of Modern Physics, 75, 715–775.
[Zurek(2005)]Zurek, W. H. (2005). Probabilities from entanglement, Born’s rule from envariance.
Physical Review A, 71, 052105.
15
