Wednesday, January 31, 2018

There has never been any multiverse mania

But the multiverse became a possible, important sketch of the grand scheme of the world and it remains one as of today

Anti-string crackpot Peter W*it who was notorious a decade ago celebrated the 15th anniversary of a string theory paper in his rant titled
15 Years of Multiverse Mania
The celebrated 2003 paper was the so-called KKLT, an article by Kachru, Kallosh, Linde, and Trivedi that provided us with a reasonably complete construction of a large, googol-like discrete set of string theory vacua (vacuum-like solutions to the fundamental equations of string theory which have different particle spectra and interaction strengths and are candidates to describe the world around us) which has shifted the string theorists' opinions about the uniqueness or inevitability of the "right solution" to string theory.

It's weird for a string theory hater to commemorate 15th anniversaries of string theory preprints. I haven't thought about that anniversary at all. Well, in fact, I think that almost no string theorist has even thought about the two-months-old 20th anniversary of Maldacena's paper that discovered the AdS/CFT correspondence (the paper has some 15,000 citations as of now).

It's been known for decades that string theory generally predicts the precise, discrete values of all a priori adjustable, continuous, dimensionless parameters that affect the low-energy physics. If there's such a seemingly free parameter, you may add an operator changing its value to the world sheet Lagrangian density. (As some semi-complete proofs and anecdotal evidence suggest, this argument is likely to hold non-perturbatively as well but I will only sketch the proof for the deformations of a perturbative string theory.)

But this operator may be multiplied by \(\exp(ik\cdot X(\sigma,\tau))\), an operator on the world sheet, to obtain another marginal operator, namely a vertex operator with a momentum \(k^\mu\). That proves that for every adjustable scalar parameter, string theory actually predicts a scalar field, a whole new dynamical field whose value may change from one point to another. Massless scalar fields almost certainly don't exist because they would lead to new long-range forces that would destroy the equivalence principle (all bodies accelerate at the same acceleration) that has been precisely tested and sits in the foundations of general relativity. Also, such precisely massless scalar field would change with the cosmic time and imply the evolution of the fine-structure constant etc., and observations indicate that those constants are unchanging, too.

For that reason, or perhaps two reasons, the world around us almost certainly contains no precisely massless scalar fields – with a vanishing potential \(V(\phi)=0\) – the so-called "moduli". We should better look for a solution to string theory's equations where a potential is generated for every scalar field which means that the stationary point – the minimum of the potential – is completely determined locally. And indeed, there have been qualitative arguments showing that the potential is indeed generally generated for all the scalar fields, especially after supersymmetry is broken.

It's sort of cool how these deep ideas may be conveyed rather accurately and almost without any equations.

Because of this ability of string theory to pinpoint all the parameters exactly, it was believed in the 1980s that the right solution of string theory that describes the Universe around us is basically unique. But there was always a loophole – that was already known since the late 1980s but wasn't taken too seriously. The set of vacuum-like solutions to string theory may be discrete but it may be large, anyway. Wolfgang Lerche was among those who argued around 1987 that the number could be as high as 101,500.

OK, such papers were ignored for the following 15 years and string theorists overwhelmingly preferred to believe that the "right vacuum of string theory relevant for our four-dimensional physics" was unique or almost unique or at least easy enough to be found from a limited list of candidates. Note that only the discreteness of the set of vacua may actually be justified by a solid string theory argument – which I sketched above. The smallness of the set of string theory vacua couldn't ever be justified in a similar way and that assumption was therefore pure faith.

At some level, string theorists knew almost immediately that the faith had to be wrong. For example, the first realistic vacua – heterotic strings on Calabi-Yau three-folds – were discovered in 1985 and within weeks, it was known that there were some 10,000 topologies of the Calabi-Yau manifolds. So the number of solutions didn't seem to be one even if you require them to be "qualitatively analogous" to the vacua that are needed for real-world particle physics.

For half a century or so, popular books on physics were also full of the anthr*pic principle, the idea that the low-energy laws of physics may vary in a much grander version of the Universe – the multiverse – and our Solar System sort of randomly sits in a corner where the laws of physics are compatible with the intelligent life. That paradigm has been partly orthogonal, partly contradicting to the picture that was apparently emerging from string theory (well, a picture built on string theory calculations plus the faith I have mentioned).

Things began to change in the late 1990s when the positive cosmological constant was discovered by astronomical observations. The cosmological constant still seems to be nonzero but as an energy density, it's tiny, some 10–122 times the Planck density. It means that it's some one-hundred and twenty-two damn orders of magnitude smaller than the natural estimate for what the constant should generically be if it's nonzero – the Planck density.

This relationship has been known for some time, before the cosmological constant was actually measured to be positive, which is why string theorists generally assumed that the cosmological constant in our world was exactly zero. If it were nonzero, it had to be tiny but because of those 122 orders of magnitude, its tiny value would be extremely unnatural. So the assumption of naturalness led the wise folks to assume that the value was exactly zero. Zero may look like an "infinitely unlikely" number but it's not – robust qualitative reasons such as new symmetries might exist that explain why the value is exactly zero. For example, unbroken supersymmetry may guarantee this constant to be zero. Even though supersymmetry has to be broken in the world around us, there could still be some stringy sense in which it's "morally unbroken". No one knew how to make such arguments quantitative and persuasive but it was a possibility that someone else would have succeeded later.

But once the cosmological constant was measured to be nonzero, string theorists reacted a lot. Perhaps, they overreacted. To compensate for their previous, apparently wrong, expectations that the constant should have been zero, they began to spend lots of big shots' man-hours on ideas that assume or imply the idea that the cosmological constant is positive. At least, they wanted the compatibility with the apparently positive cosmological constant because if string theorists really care about something, it's the observations that have actually been made (they don't care about "observations" that haven't been made, that haven't even been described in detail, but that are constantly talked about by šitty demagogues such as Peter W*it).

OK, in 2000, Bousso and Polchinski described their first semi-detailed picture in which string theory can produce a discrete but very large number of candidate vacua, a "discretuum". They noticed that there can be some generalized quantized magnetic fluxes, extra integer-valued labels associated with each topological cycle of the Calabi-Yau, and at some level, these integers have so many possible values that their spectrum behaves almost like a continuum, and this discrete approximation of the continuum, the "discretuum", provides physics with a huge, googol-like number of vacuum-like solutions.

Bousso-Polchinski have slightly over 1,000 citations now. But yes, you can be sure that lots of string theorists were thinking along these lines already in 2000. For example, in Santa Cruz, Michael Dine, my adviser, and I wrote a paper whose main point was a technical criticism of the new "anthr*pic" solution. Aside from saying the obvious thing that certain things (including low-energy gauge groups) are not predicted (which lots of people love to repeat millions of times and take credit for that), we argued that even assuming a tiny cosmological constant, a very heavy Higgs field still seemed more likely, and therefore a generic (incorrect) prediction. Well, indeed, it's assumed today that the anthr*pic-like selection is supposed to make both cosmological constant and the electroweak scale tiny. If you increase the "tasks" that the anthr*pic selection is supposed to solve, our criticism is largely irrelevant. At any rate, I don't think that ours was a groundbreaking paper in any way, just to be sure. We may have wanted to kill the anthr*pic reasoning but I don't think that we did.

KKLT was a much more important paper but I would still view it as a "technical appendix" to Bousso and Polchinski. They specifically talked about type IIB vacua and wanted all the potentials for the scalar fields to be generated by particular effects that may be at least named – and, ideally, calculated as well-established stringy effects through procedures to quantify them. They constructed lots of AdS (anti de Sitter, i.e. negative cosmological constant) vacua with the nonzero potential coming from either gaugino condensates or D-brane instantons, and argued that the addition of D3-branes guarantees the existence of some realistic, de Sitter vacua for each such AdS vacuum. In those, all the moduli are perfectly stabilized and there's a large, googol-like number of such vacua. Because of the large number, some of these vacua have "unnaturally" tiny cosmological constants that can match the observed tiny value.

By now, KKLT have collected some 2,500 citations. It's a lot but it's much less than Maldacena's 15,000 for the AdS/CFT. And I think that among the 2,500 followups, most are still very modest, low-key, and try not to overstate the anthr*pic reasoning. In fact, the word "anthr*pic" itself has often been treated as a slur – which is why people semi-jokingly referred to it as the "A-word". You may notice that even in this blog post, there's no restriction that would prevent us from writing "fucking" but the word anthr*pic ends up in the same category of hardcore šitty taboos and expletives as Peter W*it, to mention an obvious example. You may eat pounds of šit but please never touch W*it or the anthr*pic principle, physicists generally teach their babies.

Peter W*it is therefore full of šit when he says that there's been a multiverse mania. With a possible exception of Lenny Susskind who is both a top physicist and a passionate popular writer defending the anthropic principle (and Susskind's philosophical views on physics are often hybridized with his political opinions in irrational ways), no physicist was really enthusiastic about the multiverse and in fact, most of the successful people have been ashamed of any similarity between their research and the "anthr*pic principle". You won't find dozens of papers that would scream that we have to HODL the multiverse or bring the multiverse to the Moon. I think that you wouldn't even find a single talk by a physicist with a detectable enthusiasm for the concept of the multiverse let alone the anthr*pic reasoning – with a possible counterexample of Lenny's.

For decades, even the multiverse has been treated as an unwanted baby by the physicists – but the competent ones generally don't deny that the multiverse was born as a possibility.

As discussed in hundreds of TRF blog posts that were written years ago, there exist lots of basic yet unequivalent claims about the multiverse. Some of them are backed up by rather convincing scientific evidence and others are not. Most importantly, we must carefully distinguish the
  • the mathematical existence of the landscape, the huge, googol-like set of vacuum-like solutions to the equations of string theory, the underlying theory
  • the existence of the multiverse itself – as realizations of the vacua in the landscape within the "real world"
  • correctness, relevance, or exclusivity of the anthropic principle as an explanation of the low-energy laws of physics
OK, by now, I am probably more than 90% persuaded of the first point, the landscape. The number 90% is somewhat random. I don't know how to calculate it, it reflected some "balance of sentiments" at the moment when I was writing the sentence. I find it plausible that some of the deep criticisms of KKLT and other things are right and the number of the vacuum-like solutions actually isn't large. Relatively well-defined, doable research could settle this first question.

I am more than 80% sure about the existence of the multiverse because it's produced from eternal inflation and I think that inflation is very likely needed and it's also more likely than not to come in an eternal flavor. One needs some less well-defined reasoning to become confident of one answer to this question or another.

Concerning the last point, it's the most "metaphysical one", the number of bad papers promoting the principle that suffer from some rudimentary logical mistakes is the highest one, and my subjective probability that the reasoning is needed (in some correct flavor, not one of the atrociously wrong ones) is some 50%. I just don't know whether the anthr*pic selection – we just live in a random vacuum whose "compatibility with intelligent life" is among the most selective properties of the vacuum – is important or true or relevant in physics. Well, I can imagine that it's right for some questions and not others and a complete picture that will be found in the future will be a hybrid of the anthr*pic principle and conventional "calculate results blindly, paying no attention who could live in a Universe" picture of physics.

I don't know whether the anthr*pic reasoning or the multiverse are correct – i.e. will be proven by truly solid proofs in the future – but I think it's right to say that the multiverse and even the anthr*pic principle are the most explicit known candidate solutions to the relevant puzzles as of today and that's enough for our obligation to consider them as serious contenders.

Lots of partial advances have been made in the recent decades – and from the 2003 KKLT paper – but we still haven't found rock-solid proofs and highly persuasive stories that would change some of the numbers 50%, 80%, 90% to 0% or 100%. Because we don't have game-changing evidence, we simply have to remain open-minded, live with the uncertainty, and work on various possibilities – at least the community must allow itself to think in different directions.

Of course the efforts by aggressive filth and nasty crackpots like Peter W*it shouldn't affect what the people are thinking about. The real world is qualitatively an approximation of that idealized scientific world but it is also a messy place so I think that the witch hunts organized by that anti-physics filth have actually affected the real world physicists' work and what they're willing to publicly say.

Peter W*it and his brain-dead readers probably can't live with the uncertainty so they would love to ban the anthr*pic principle if not the multiverse entirely. But thankfully, they are just irrelevant crackpots and filth and not dictators of the world so the ban hasn't materialized, not even after the 15 years after KKLT that e.g. Peter W*it dedicated to persistent demagogic tirades against physics.

Genuine scientists remain uncertain about the multiverse and they can simply live with the uncertainty, along the lines that Richard Feynman described in the monologue above. Indeed, physics may remain uncertain about the relevance of the multiverse (and/or anthr*pic reasoning) for additional decades, centuries, or millenniums. That's how things sometimes be. Who is existentially or viscerally terrified by the uncertainty simply shouldn't be and cannot be a scientist because the life within this uncertainty is the daily routine of every scientist. If you need all the cutting-edge questions to have settled answers, switch from science to a religious cult – you will have to accept that all the predetermined answers are certain, settled, and wrong.

Note that Feynman's monologue above isn't relevant here just because it talks about the uncertainty and incomplete knowledge as inseparable parts of the scientist's thinking about the real world. He also justifies his disbelief in Biblical and similar stories by saying that they're too provincial. Why should the creator of the world have a special relationship to humans, Earth, the Solar System, and so on? It looks out of proportion. And yes, it's possible that He doesn't even have to have a special attachment to the vacuum around us – with its particular low-energy laws of physics – i.e. to the particular string theory vacuum that seems directly relevant for our physics experiments.

We know that the Earth isn't the only planet – but we don't really know whether the vacuum around us is one of many stringy vacua that "really" exist to a similar extent. Because of this different status of the two questions (other planets, other vacua), the answers may be different, too. But they can be analogous, too. The analogy may work. We don't have really solid evidence or argument that would disprove the relevance of the anthr*pic principle, let alone the multiverse, which is why honest physicists remain open-minded about that possibility and, even if they don't work on the multiverse issues themselves, they surely "allow" their colleagues to do so.

But this open-mindedness doesn't mean that they're enthusiastic about the multiverse or the current status of our understanding of it. Open-mindedness and mania are very different things.

No comments:

Post a Comment