How much do you know
π Which multiverse are we talking about?
Eight questions on Tegmark's four levels, eternal inflation, many-worlds and the measure problem β and on which of them anthropic reasoning actually needs.
Lukasz Szramuk Β·
Question 1 of 8
Which Tegmark level contains regions with DIFFERENT physical constants?
Every question, and every answer
The whole quiz, written out β so you can read it without playing, and so a search engine can too.
Which Tegmark level contains regions with DIFFERENT physical constants?
Answer: Level II β bubble universes from eternal inflation
This is the level that does the anthropic work. Level I has the same laws and constants with different initial conditions; Level III has one Hamiltonian and no physical collapse. Only Level II gives you an ensemble of different low-energy constants for a selection effect to select from.
Read the essay βWhat separates bubbles in an inflationary multiverse from branches in many-worlds?
Answer: Bubbles separate by expanding spacetime; branches separate by decoherence
Both multiply descriptions of reality, and that surface similarity is why they get conflated. They do not start from the same mechanism: one is cosmology, with causally separated regions of spacetime that may have different vacuum states; the other is quantum mechanics, with decohered branches of one universal state.
Read the essay βThe string landscape is often quoted at roughly how many vacuum states?
Answer: About 10^500
The figure is a rough count of metastable vacua from compactifying the extra dimensions, and it is the reason the landscape gets paired with anthropic selection: enough distinct values of the constants that a habitable one is not surprising. The number is an estimate of a count, not a measurement.
Read the essay βWhat is the measure problem?
Answer: In an eternally inflating multiverse the fractions depend on the order you count in
Eternal inflation never finishes producing bubbles, so there is no completed set to take fractions of. Different cutoff rules give different answers to 'what fraction of observers see X', and nothing in the theory says which rule is right. Without a measure the multiverse predicts nothing in particular.
Read the essay βWhy does the strongest anthropic argument need a multiverse?
Answer: A selection effect can only select from an ensemble that exists
If the constants could only ever have taken one set of values, there is nothing to select and no puzzle. Weinberg's bound on the cosmological constant is a genuine prediction only if a range of values is realised somewhere; without that it is an observation that we could not have been anywhere else.
Read the essay βWhich multiverse prediction has actually been tested?
Answer: Weinberg's 1987 anthropic bound on the cosmological constant
Weinberg predicted a small but nonzero value from the requirement that galaxies form, before the 1998 supernova results. It is the multiverse's one genuine forecast. Bubble-collision signatures in the CMB have been searched for and nothing significant has been found.
Read the essay βWhat does Tegmark's Level IV claim?
Answer: Every self-consistent mathematical structure exists physically
It is the most radical of the four and the one critics press hardest on, because a structure that contains everything makes nothing more likely than anything else. It is also where the measure problem becomes acute: with no way to weight structures, no prediction follows.
Read the essay βWhat is the strongest objection to using a multiverse to explain fine-tuning?
Answer: Many universes make it likely SOME universe is tuned, not that THIS one is
Ian Hacking called it the inverse gambler's fallacy and Roger White gave it its standard form. The usual counter is that observer selection breaks the analogy β you could not have found yourself in a sterile bubble, whereas a gambler could walk in at any time β and that counter is doing a great deal of work.
Read the essay β