Susskind and other usual suspects try hard to convince the world that they are the ones who finally understood 'Many Worlds' and that such is a success of string theory and all that. A media spectacle is going on right now, see here at the New Scientist’s “Ultimate Guide to the Multiverse”, Brian Greene chipping in with the Multiverse episode of the “Fabric of the Cosmos” series on PBS, and many others. Some, such as Lee Smolin and Peter Woit claim to criticize, but apparently were also successfully made to believe that Many Worlds is part of Tegmark/Susskind personality cult and string theory hype.
“Agnostics expect a Multiverse” clarified an important distinction: Many Worlds and Multiverses are not the same. Different forms of Multiverses may or may not belong to possible consistent descriptions of for instance cosmology, but “Many Worlds” correctly understood are true, much like one plus one equals two is true since two is defined that way. Let's explain “Many Worlds”, especially what they are (1), that they do not equal quantum physics (2), and why such is tautologically true, including why those who demand that the concept needs empirical verification do not understand scientific verification properly (3). (UPDATE: Two more recent versions of the general argument for MW/M can be found in the article "Plenitude Principle as necessary for Fundamental Physics and much more".)
(UPDATE: This article is not supporting nonsense such as Tegmark’s four level Multiverse that claims to include Many Worlds as the third level which “adds nothing new” (his words) to the second level – an obvious and fatal problem with his scheme!)
1) “Many Worlds”, What Are They?
“World” in the Many Worlds context means a world of observers (interacting systems, shared observations), not a cosmos or “universe” inside a certain model of the physical world. This is not how Immanuel Kant used the word “world”. Nevertheless, if “Many Worlds” are mentioned especially in the context of modern physics, the worlds are those that are observed by the observers “in” those worlds. These worlds must not be confused with “universe” or “cosmos”. The phrase “Many Worlds” came historically along with so called “Everett branches”, but we do not need to concern us with quantum physics (see below).
Different universes being described inside one description constitute a “Multiverse”. If existent, it is included in Many Worlds. Many Worlds is not included in every Multiverse model, but all possible Multiverses are intrinsic in Many Worlds, because Many Worlds is about all that is self-consistently possible. Susskind, Tegmark, Greene etc. like to confuse this (I am not sure how far Bousso commits this mistake – he seems to be further along, so is Rovelli).
A Many Worlds description is one that includes other possible worlds, also those that are not actualized for you right now, counter factual ones, for example the possible worlds where you wear different clothes while reading this right now.
A truly fundamental description aims to take account of the whole of totality, of all that is self-consistently possible. If you refused Many Worlds in such a description, you would assume that in the very foundation of totality, the very most fundamental laws of nature, intrinsic to the absolute rock bottom core of physics in its most profound and general symmetries, there is something inside that ensures that you, yes you personally with name and address, wear those particular clothes today!
In the most fundamental description, all self-consistent possibilities are equivalent. Even if you assume non-quantum physics and a universe where all future was already determined by initial conditions, the ultimate description treats all self-consistently possible initial conditions equivalently. If gods blew holy dust only into one set of initial conditions, the most fundamental description will still include all the worlds where they fancied different sets. All self-consistently possible gods want you to wear those smelly socks today? Is that a hypothesis worth considering?
2) Many Worlds is Not Quantum Physics
Many Worlds are not quantum physics. Quantum physics is not the foundation of Many Worlds. Many Worlds is the foundation of quantum physics! In general, physics is not the source of self-consistency; self-consistency is at the core of mathematics and thus possible physics.
Quantum physics merely describes the way in which worlds are entangled with each other. Such mutual interference implies that your actualized world cannot be fully described without including other worlds. Entanglement allowed experimental evidence about the other worlds, for example via Elitzur-Vaidman bomb testing type experiments.
Elitzur-Vaidman bomb testing is evidence for entanglement with counter factual worlds.
HOWEVER!!! It is important to keep in mind that interference is not empirical evidence needed for Many Worlds! It is support for quantum Many Worlds. This evidence supports for example the so called Many World Interpretations (MWI) of quantum physics. The models under the MWI label, like this simple model that resolves the Einstein-Podolsky-Rosen paradox, often have probabilities that cannot be normalized. They often suffer from a frequentist probability concept and the so called “measure problem”, which is the arguably most important problem in cosmology and quantum physics today.
Many Worlds as such do not gain much from this evidence, because they are in the fundamental description no matter what. Without quantum mechanics, other self-consistent worlds are just classical parallel worlds that never interfere, but they are still Many Worlds.
3) Do Many Worlds Exist? Do we Need to Verify their Existence?
Now to the aspect that few are able to grasp. Lee Smolin has big problems with this – in fact, I have yet to see a physicist who fully gets it. Maybe Bousso/Rovelli come closest, but I am not sure how far.
It is fashionable to quote Popper and demand that something that cannot be falsified should not enter science. Lets first forget Popper, who is about as insightful as Hilarious Putnam. Such writers confuse with nonsensical terminology and demand others to admit “meaning” to stuff that cannot be verified. Nonsensical terminology can come up with any desired amount of distinctions that make no verifiable difference. People like Wittgenstein and Ayer understood much better what verification is all about: It starts with terminology!
The problem with the question “Do not-actualized counter-factual worlds exist?” is that it is not asking for the “existence” of anything, but instead for the definition of “exist”. Academics turn this into sophistry about “ontological commitment toward modal reality”. If you want progress however, you must consider the following:
If “exist” includes other possibilities, you cannot make the distinction between the modal categories “necessary”, “exist”, and “possible” anymore. All necessary does exist. All that is existent is possible. If you turn back on this terminology via “everything possible actually exists” (“Principle of Plenitude”), everything becomes necessary and the whole terminology crashes!
Vegetable and fruit are distinct because we want to make distinctions. Only through such distinctions can we even express that tomatoes are perhaps wrongly classified. Only after such distinctions can we have the “holistic revelation” that fruit and vegetables are basically the same.
So you cannot start with “Please verify existence of other possible worlds!” You must start with a definition of “exist” that provides a verifiable distinction (in Ayer’s words “significant meaning”) between “possible” and “exist”. If you have not, your aches about that everything possible exists are self-inflicted.
Once more in other words: Your world is actualized for you (now). Other worlds are actualized relative for the observers who observe those other worlds, the other possible “yous now”. Now you ask, “But do those worlds actually exist?” This means something like “Are those worlds really actualized?” It asks “Are they actually actual?” (actually actual - see the problem yet???) It asks for something like an absolute actualization that is different from relative actualization for observers. It claims a distinction that makes no verifiable difference! If you cannot tell me how I am supposed to distinguish “absolute actualized” from “relative actualized”, I will not waste time trying to prove that other worlds are actually absolutely actualized rather than “merely possible”.
Now you may think that this is unfair because the distinction I want you to verify first is exactly the distinction that you wanted me to verify, but here it is that I exclaim: Precisely! That is why Many Worlds are tautologically true even although current Many Worlds Interpretations of quantum mechanics have the measure problem and often assume unobservable universal states (the wave function of the universe). I know better than most that MWI are not the last word, but Many Worlds are still true.
Those who demand to verify Many Worlds think themselves better scientists who care about verification, but like with so many issues, say freedom and democracy, those who shove it down other’s throats seldom understand half of it. Verification starts well before one can ask questions.
Do Many Worlds exist? It depends on how you like to use “exist”. Any truly fundamental description of physics must include Many Worlds, period. We can discuss whether “world” is too misleading and should be replaced by “Mind”, as in Many Minds, which is a good point. But “I don’t like Many Worlds and they should get back to what can be verified by experiment” regresses to a naïve, pre-philosophical, unenlightened scientism.--------------------------------------------
Remark to philosophers: Not mentioning modal realism may seem a harsh omission, however, physics audiences do not know modal realism much and it may mislead, because modal realism is about modal logic (terminology), yet a wider audience may misinterpret the name to mean that such philosophy is all about whether or not stuff is "really real".