Module 11: Inflation and the first instant · Lesson 11.7 · about 17 min
The multiverse: science or speculation?
Inflation may never stop everywhere, making endless bubble universes, perhaps each with different laws. What eternal inflation and the string landscape say, the anthropic argument, and the fierce debate over testing it.
Module 11 · about 17 minutes. Read first: 11.2 Cosmic inflation and 9.3 The vacuum energy puzzle. This lesson is different from the others. Much of it is speculation: ideas that follow from theories but have no direct evidence. Each claim is labelled with how solid it is. By the end you will be able to explain what scientists mean by a multiverse, why some take it seriously, and why others think it is not science at all.
1. Four meanings of “multiverse”
The word is used for very different ideas. From most solid to least:
- Space beyond our horizon (very likely). The observable universe is a sphere 46 billion light-years in radius (Lesson 5.9). Space is flat to 0.2%, so it is probably much larger, perhaps infinite, with more galaxies under the same laws. Few dispute this, though some do not call it a multiverse.
- Bubble universes from eternal inflation (speculative, follows from many inflation models). Regions where inflation ended at different places and times, beyond any horizon we could ever reach.
- Different laws in different bubbles (speculative, needs the string landscape). Different particle masses, forces or dark energy in each.
- Quantum many-worlds (a separate idea about quantum measurement, not cosmology). Not covered here.
2. Eternal inflation
In the slow roll of Lesson 11.2, the inflaton creeps downhill. But it also gets quantum kicks, about H/2π each Hubble time (Lesson 11.3), up or down at random. High on the hill the slope is so gentle that the kicks beat the roll. In the regions kicked uphill, inflation goes faster and lasts longer, and since each Hubble time multiplies an inflating region’s volume by e³ ≈ 20, those regions soon outweigh everything else. Alexander Vilenkin (1983) and Andrei Linde (1986) showed that, in many models, inflation ends in more and more places, but never everywhere. Each place where it ends becomes a hot Big Bang, a bubble; ours would be one.
3. The landscape and the anthropic argument
String theory has extra dimensions curled up in many possible shapes, threaded by fields. Each choice gives a different “vacuum”, a different kind of empty space with its own particle masses, forces and vacuum energy. Estimates made in 2003-04 gave around 10500 such vacua: the string landscape. Combined with eternal inflation, every vacuum could be realised in some bubble.
That links to the worst puzzle of Lesson 9.3: why is dark energy about 10120 times smaller than the natural estimate? In 1987 Steven Weinberg argued that if it were much bigger, the expansion would have sped up before galaxies could form, so no one would be around to measure it. In a landscape where the value differs from bubble to bubble, most bubbles have huge vacuum energies and are empty; observers find themselves only in the rare ones with a small value. This is the anthropic argument. Weinberg used it to predict a small, non-zero value a decade before supernovae found one.
4. Can it be tested?
5. The debate
In 2014 George Ellis and Joe Silk wrote in Nature, “Scientific method: Defend the integrity of physics”, warning that some physicists were proposing to relax the requirement that theories be testable, to accommodate the multiverse and string theory. Supporters reply that science has always accepted unseen consequences of well-tested theories, and that the multiverse is a prediction of theories, not an extra assumption. Paul Steinhardt, one of inflation’s founders, has gone further: he argues that eternal inflation can produce any outcome and so predicts nothing, and favours bouncing alternatives. In 2017 more than 30 physicists, including Guth and Linde, replied that inflation has made and passed real tests. Meanwhile, some string theorists have questioned whether the landscape’s vacua with positive vacuum energy exist at all.
What we can say firmly: our observable universe is very likely a small part of something larger. Whether that larger thing contains other bubbles with other laws is not known and may never be known for certain.
The multiverse in cosmology.
- 1983 Vilenkin: inflation can be eternal; Steinhardt notes the same at the 1982 Nuffield workshop.
- 1986 Linde: self-reproducing chaotic inflation.
- 1987 Weinberg's anthropic bound on the cosmological constant.
- 2000 Raphael Bousso and Joseph Polchinski: string theory can give a vast number of vacuum energies.
- 2003 Kachru, Kallosh, Linde and Trivedi build vacua with positive vacuum energy in string theory; Leonard Susskind names "the anthropic landscape".
- 2011 First search for bubble collisions in the CMB.
- 2014 Ellis and Silk's essay starts a wide debate.
“Scientists have discovered parallel universes.” No observation has detected another universe. What exists are theories with good support (inflation) or deep mathematical appeal (string theory), some versions of which imply a multiverse. When you read a headline about other universes, ask which of the four meanings it uses and what, if anything, was measured.
- “Multiverse” has several meanings: space beyond our horizon (very likely), bubbles from eternal inflation and different laws in each (speculative), quantum many-worlds (separate idea).
- In many inflation models quantum kicks keep some regions inflating; their volume grows faster than bubbles form, so inflation never ends everywhere (Vilenkin 1983, Linde 1986).
- The string landscape may hold around 10500 vacua; with eternal inflation, observers would find themselves only where dark energy is small enough for galaxies (Weinberg 1987).
- Possible tests: bubble-collision marks in the CMB (none found), a closed curvature would rule out simple versions; the measure problem blocks firm predictions.
- Ellis and Silk (2014) and others argue it is not testable science; supporters call it a consequence of tested theories. The question is open.
Check yourself
Sources
- A. Vilenkin (1983), Birth of inflationary universes, Physical Review D 27, 2848 https://doi.org/10.1103/PhysRevD.27.2848
- A. D. Linde (1986), Eternally existing self-reproducing chaotic inflationary universe, Physics Letters B 175, 395 https://doi.org/10.1016/0370-2693(86)90611-8
- S. Weinberg (1987), Anthropic bound on the cosmological constant, Physical Review Letters 59, 2607 https://doi.org/10.1103/PhysRevLett.59.2607
- R. Bousso and J. Polchinski (2000), Quantization of four-form fluxes and dynamical neutralization of the cosmological constant, Journal of High Energy Physics 2000(06), 006 https://doi.org/10.1088/1126-6708/2000/06/006
- S. Kachru, R. Kallosh, A. Linde and S. P. Trivedi (2003), De Sitter vacua in string theory, Physical Review D 68, 046005 https://doi.org/10.1103/PhysRevD.68.046005
- S. M. Feeney, M. C. Johnson, D. J. Mortlock and H. V. Peiris (2011), First observational tests of eternal inflation, Physical Review Letters 107, 071301 https://doi.org/10.1103/PhysRevLett.107.071301
- G. Ellis and J. Silk (2014), Scientific method: Defend the integrity of physics, Nature 516, 321 https://doi.org/10.1038/516321a
- M. Kleban and M. Schillo (2012), Spatial curvature falsifies eternal inflation, Journal of Cosmology and Astroparticle Physics 2012(06), 029 https://doi.org/10.1088/1475-7516/2012/06/029
Your ticks stay in this browser. Save them as a file on the course home.
Updated 2026-11-22. Numbers are rounded for teaching; measured values come from the sources listed in each lesson. Simulations run in your browser; nothing you type is sent anywhere.