A Gigaplex of Parallel Lives
Identity across the branches
How many of you are there?
In the Quantum Branching Universe (QBU), the question becomes a modeling question: which decoherent histories satisfy a chosen criterion for being you, and what quantum Measure do they carry? The answer is not a unique headcount. Branches are emergent, their number depends on coarse-graining, and sufficiently fine descriptions need not yield a stable count at all. The companion questions — which continuations count, how much weight they carry, and what makes any of them you — matter more than a census.
Counting Yourself
Any count of “versions of you” depends on what you use to pick them out. The pointing device is the Pattern Identifier (PI) — a precise, reproducible pattern used to select the timelines that contain it — and the count changes drastically with the grain of the pattern.
Identify yourself at the genotype level — your exact genetic sequence — and the selection is precise but not uniquely identifying. Identical twins can share a genotype, and a sequence could in principle be recreated independently. A Strong PI therefore needs provenance as well as sequence: this genome descended from this fertilization event, for example. Even then, genotype is too brittle as a criterion of ongoing identity. Exact sequences diverge under mutation, so the Measure of exact-genotype continuations can fall even while the person recognizably persists.
Identify yourself at the phenotype level — the broader biological and cognitive pattern, the recognizable person — and the picture inverts. Phenotype-level identity is robust by construction: developmental redundancy, functional buffering, and evolutionary constraint mean that enormous swathes of genetic and environmental variation converge on the same recognizable individual. The pattern that is you in practice — the one your friends match against, the one your memories cohere around — persists across branches that differ in millions of microscopic details. The genotype anchors your ancestry; the phenotype carries your survival. When I ask how many of me there are, I am asking at the phenotype level.
The Gigaplex
The title’s Gigaplex is a scale metaphor, not a literal branch count. If we impose a deliberately simple binary coarse-graining — one modeled split per second across an 80-year life, about \(2.52 \times 10^9\) seconds — that toy model contains on the order of
\[ 10^{1{,}000{,}000{,}000} \]
modeled paths before applying any criterion for which still contain a recognizable phenotype-level you. Ten raised to the billion: a Gigaplex in the toy model.
No intuition survives contact with that number. It dwarfs the count of atoms in the observable universe (~\(10^{80}\)). Many choices, accidents, and conversations have decoherent alternatives with no everyday analogue in scale. But quantum theory does not imply that every narratively imaginable variation occurs, nor does the toy count estimate the number of physically meaningful branches.
The Gigaplex is therefore not a license for “everything happens somewhere” reasoning. The physically relevant question is which alternatives have nonzero amplitude under a specified model, and especially how much Measure they carry. Miracles are not guaranteed by a large path space.
Count Is Not Measure
Here is the mistake the Gigaplex invites, and the distinction that blocks it.
The count of modeled paths can grow exponentially. But Measure — squared-amplitude weight relative to a specified state, event, and coarse-graining — is normalized: the weights of mutually exclusive exhaustive alternatives sum to 1. Refining a low-weight alternative into many sub-alternatives does not increase their collective weight.
So the weighted composition of a branching model can look nothing like an unweighted census. Counting branches is especially unreliable because branch number is representation-dependent. Decision and prediction must use the quantum measure supplied by the model, not an arbitrary enumeration of histories. The analogous danger in self-locating belief — counting observers without justifying the sampling measure — is the subject of You’re Not a Random Sample and its quantum sequel You’re Not a Random Branch.
Nowhere does the swap do more damage than in the most seductive argument the branching picture has ever produced.
Quantum Immortality, Priced in Measure
The argument goes like this. Suppose every hazard leaves some nonzero event sector in which you survive — the aneurysm misses, the mutation repairs, the trigger jams. Since those records are real on the Everettian model, the argument says there is always a continuing you and that subjectively you should expect to live forever. This is quantum immortality, and in its usual form it swaps mere nonzero existence for Measure. The premise is not guaranteed: a survival sector must be physically available from the actual state, and even when it is, the rational question is what total weight it carries under a defensible identity and hazard model.
Pricing survival requires an explicit empirical model. The following scenarios are illustrative, not estimates supplied by quantum mechanics:
No advanced technology. Observed human longevity gives no basis for expecting indefinite survival by microscopic luck alone. Everettian branching does not change that evidence or license a numerical survival probability without a biological hazard model.
Moderate medical advancement. Breakthroughs might change the hazard curve, but their probability and effect must be estimated from medicine and forecasting, not inferred from the existence of branches.
Radical life extension. As a sensitivity calculation, suppose interventions reduce annual mortality to one in a million, and suppose — without claiming evidence for the figure — a 1% chance that such technology arrives in time. Conditional on a constant independent annual hazard, surviving the next ten thousand years would have probability
\[ (0.999999)^{10000} \approx e^{-0.01} \approx 0.99 \]
so the Measure of timelines in which you achieve extreme longevity comes out at approximately
\[ 10^{-2} \times 0.99 \approx 10^{-2} \]
The roughly one-percent result merely restates the assumptions; it is not a forecast. The useful lesson is methodological: survival expectations must come from a defensible hazard model and evidence about interventions. No amount of branch-counting can substitute for either. Survival remains an engineering and decision problem, not a metaphysical gift.
Identity as Logical Resonance
So far, “a version of you” has meant a forward continuation — a branch descending from your birth, carrying your body through time. The assumption deserves scrutiny, because the PI machinery does not require it.
One influential computationalist position holds that personal identity depends on realizing the relevant mental organization rather than preserving the same matter. On that view, a sufficiently faithful future or simulated realization could count as the same person.
Translate that into this book’s terms and it becomes an identity proposal, not a theorem of quantum mechanics. A Pattern Identifier can represent the informational and causal constraints a theory of identity treats as sufficient. Whether those constraints are sufficient — whether a copy, reconstruction, or parallel realization is numerically the same person — is a further philosophical claim. Pattern continuity may explain much of ordinary persistence without showing that causal or bodily continuity is irrelevant. I call the proposed relation logical resonance: identity understood as recurrent organization rather than a unique temporal thread. The later discussion of Chaos extends that proposal beyond persons.
The QBU vocabulary can keep this from collapsing into “anything vaguely like me is me.” A Strong PI includes enough provenance that its matches share the stipulated causal origin; a Weak PI such as a name or rough description can select unrelated people. But this formal distinction organizes an identity theory rather than proving one. A theory must still say which memories, dispositions, embodiment, and causal history matter, and why.
Two consequences follow, one for each half of this chapter.
First, on the pattern view, identity would be distributed across qualifying continuations rather than confined to a single thread. Quantum Measure could then weight expectations about those continuations. It would not, by itself, decide which realizations qualify.
Second, if reinstantiation counts as survival, preserving the defining information could become another survival strategy. That conclusion is conditional on the pattern theory; reconstructibility alone does not establish continued subjectivity. Whether a reconstruction preserves what matters — and whether a simulation has experience — belongs to the next volume. Physics constrains which systems can be realized and how quantum weights behave. It does not settle the criterion of personal identity.