## Saturday, January 27, 2018 ... /////

### Features of scientific theories don't have to be separately testable

Two days ago, Ethan Siegel wrote a text about the multiverse that is considered an antidote to Hossenfelder's rant by readers of both authors:

Yes, The Multiverse Is Real, But It Won't Fix Physics
The title summarizes the two basic claims in his text. The multiverse is real. But it is useless for physics. A weird combination, indeed.

He has described lots of insights about the Universe at the longest length scales, about inflation and its predictions, and so on. If you divided the article to very small pieces and reviewed them separately, you could think that they were written by someone who really understands the stuff.

But if one reads everything at once, a completely different picture emerges: A picture of Ethan Siegel as a schoolkid that has memorized lots of sentences but doesn't understand what they mean at all. His misunderstanding may be seen from any "added value". Whenever he tries to use his "knowledge" to produce a new proposition, the new proposition ends up ludicrous because he doesn't really understand the "knowledge".

So there are comments "almost" explaining that some kind of a multiverse is unavoidable because the Universe is observed to be extremely flat which makes it virtually certain that it continues beyond the most distant places that are observable. The full Universe is surely much greater than the observable Universe, at least by some orders of magnitude – but it's much more likely that it's larger by many, many orders of magnitude.

Great. But at lots of points, you figure out that Ethan Siegel doesn't actually understand inflation – or doesn't "believe" it's right. At some moment, we read:
However, the standard predictions that come out of this [inflation] are difficult to do science with. They include:
• That different regions where inflation ends should never collide or interact.
• That the fundamental constants and laws in different regions should be the same as they are here.
• And that unless inflation was truly eternal to the past, there isn't enough "space" to contain all the parallel Universes that the many-worlds interpretation of quantum physics would require.
First, I must say that the last point is complete nonsense. It's totally stupid to deliberately conflate the multiverse with the "many worlds interpretation of quantum mechanics", a philosophy attempting to circumvent quantum mechanics. Even if you thought that there's a reason to consider the "many worlds interpretation" seriously, it would be dumb beyond imagination to look for "enough space" to include all the "parallel worlds" of the many worlds interpretation. Those are parallel worlds exactly because they shouldn't be thought of as co-existing in one world. It's complete nonsense to add their volumes. It's completely wrong to turn them into geometric neighbors, to look for "enough space" somewhere for all of them. They're "parallel" because they're not commensurable. If they could be neighbors, we would have to call them "series worlds", not "parallel worlds", in analogy with circuits.

OK but the first two points are indeed some of the generic predictions of inflation or eternal inflation or some other models. Siegel writes that "they are difficult to do science with". What does it exactly mean and what are the intended consequences of this strange claim?

The predictions or implications of a scientific theory – in this case, inflationary cosmology – are whatever they are predicted to be. Scientists may either rule them out if it is possible; or if they cannot rule them out, they must consider them a viable possibility that must be taken into account. There is no way to eliminate possibilities in science without actually ruling out the theory.

Inflation says that the regions where inflation ends don't collide or interact. In what sense does it make sense "harder" or – as he clearly wants to say in between the lines – in what sense does it make the theory less scientific? It's spectacularly obvious that any such statement is just a reflection of the writer's (in this case Siegel's) bias, his prejudices, his inability or lack of will to consider viable options in a fair and balanced way, impartially.

What else than bias and irrational prejudices could be behind this vague mudslinging, emotional and ill-defined efforts to make one possible answer worse than the other?

When logical arguments based on the most persuasive scientific theory of the relevant questions implies that some regions don't collide, they probably don't collide. Why "should" they collide? Why would you construct rhetorical exercises claiming that the theory "should" say that "they should collide" in order for the theory to be scientific or more scientific? Theories may predict whatever they conclude and the conclusions "regions collide" and "regions don't collide" are obviously equally legitimate and must be treated equally.

The exact same comments apply to all the other predictions that deluded writers like Hossenfelder or Siegel try to demonize. Eternal inflation predicts the low-energy effective field theory – the spectrum of particles, their masses, interactions, and coupling constants – to be different in some sufficiently different regions of the multiverse while some other theories may predict that the low-energy physics has to be the same everywhere.

Now, it was a nontrivial theorists' leap that they realized that something like that was possible at all. Most people before them couldn't even think about that possible new paradigm. But once it's out there in the realm of hypotheses, it's obvious that this scenario is (at least so far) compatible with everything we know about Nature. We can't eliminate this hypothesis now which means that we must be open-minded to it and investigate it. You may invent demagogic slogans that this possibility is unscientific or less scientific or "harder for science" but that ony proves that you are prejudiced. And if something makes the scientific progress harder or impossible in the long run, it's sufficiently strong prejudices.

There's nothing wrong or unscientific or "less scientific" or "harder to do science with" about a theory that predicts different low-energy conditions in different regions of the world. Christopher Columbus and other explorers who were sailing over the Atlantic Ocean said that the world on the other side may be similar to ours – to Europe – in some ways. But it may be different in others. The people over there may have a slightly different skin color. They used to eat different crops than we did – including potatoes and maize which we Europeans found useful. They have converged to a very different exchange rate to convert ethanol to gold – which made it good for Europeans who wanted to get gold. And so on.

Did potatoes or maize or Native Americans' obsession with alcohol make America's existence less scientific? What the fudge? Everyone who says such things is obviously a prejudiced demagogue. "The life elsewhere is similar to ours" and "the life elsewhere is different from ours" are two competing, mutually exclusive statements that must be treated as two equally viable alternatives unless you have actual evidence that one is favored.

Siegel's last paragraph says:
Just don't expect it to solve your most burning questions about the Universe. For that, you need physics you can put to an experimental or observable test. Until that day arrives, the consequences of a Multiverse will likely remain in the realm of science fiction: where they presently belong. It's okay to speculate, but if you insist on attributing a physics problem's solution to an untestable feature of the Universe, you're essentially giving up on physics. We all know that the mysteries of the Universe are hard, but that's no reason to not even try to find a solution. The Multiverse is real, but provides the answer to absolutely nothing.
This paragraph is so fundamentally wrong that I can't imagine in what sense Siegel might understand what the scientific research may mean in any nontrivial context.

If the multiverse framework is correct (or sufficiently justified by direct or indirect evidence) and implies some general consequences – i.e. predictions – than these are obviously scientific predictions, not predictions of "science fiction". Why some people choose these labels – e.g. "science fiction" – for something that follows from a concept that they admit to be "real"? If it's real, its consequences are also real, aren't they? If they're real, how they can be "science fiction"? What the fudge?

In the last sentence, Siegel says that "the multiverse provides the answer to absolutely nothing". He clearly contradicts himself because the previous sentences he wrote discussed whether you should consider a "physics problem's solution" to be science or science fiction. So there clearly exists a physics problem's solution provided by the multiverse – which contradicts his later statement that the concept provides "answer to absolutely nothing". A physics problem's solution clearly is an answer to something, namely to a physics problem. Couldn't he see this flagrant contradiction? Does he care?

But it's the following sentence that is most fanatically anti-scientific:
It's okay to speculate, but if you insist on attributing a physics problem's solution to an untestable feature of the Universe, you're essentially giving up on physics.
Only fudged-up antiscientific morons use the term "speculation" for whole categories of physics research done by some of the smartest people on Earth – who are some 40 IQ points smarter than these critics. Lots of the research of inflation and other things (surely not all of it) is stellar science par excellence that has rightfully earned some of the greatest scientific prizes and some smaller prizes, too. But aside from this obnoxious and utterly wrong "physics is speculation" slur for modern physics which has become nearly an obligatory mantra for almost all "politically correct" writers in recent years, Siegel makes a statement that "you essentially give up on physics once a solution to the problem is attributed to an unobservable feature of the Universe". Oh, really?

This sentence may be viewed as just another example of the screams "it is no physics" that inkspillers have bombarded the readers of junk popular books with in a recent decade or so. But note that Siegel is more specific – and the sentence apparently tries to be more precise than just an emotional scream. It seems that he wants the details of his sentence to matter. He wants to demonize all unobservable features of the Universe. Once they are used in any way – once your conclusions depend on them – you're "giving up on physics".

What the fudge? I don't think that someone who is familiar with modern physics at all – who understands basically anything from physics of the recent century or so – could write such a complete nonsense. Physical theories are full of features that are separately untestable. Quarks' color, gluons are unobservable. Virtual particles are unobservable or at least unobserved by definition. The Earth has a core that is unobservable. Of course, dark matter isn't seen because it's dark. Think about pretty much any concept or feature in modern physics – or modern science (geology and evolution of species would give you tons of new examples) – and you will find out that it is separately unobservable or untestable.

A huge fraction of the concepts and features of scientific theories – and, at a sufficiently deep level, a majority of them – is only observable or testable in combination with other things. You may only observe implications of virtual particles which are implications of not just the virtual particles; you need to use the assumptions about the interference of probability amplitudes, forces acting on articles, insights about some detectors work and what they imply, and so on.

The correct point is that only full-blown complete theories (or, more precisely, models, whenever "models" and "theories" are distinguished) may be required to be testable or observable. Their individual pieces or features may be invisible, untestable, and/or unobservable, and they very often are. Siegel, Hossenfelder, and too many others have joined the cult demanding that every piece and concept in science must be testable, observable, and visible to mediocre people such as themselves. But by articulating this totally ludicrous demand, they only prove that they don't understand any modern science whatsoever. If they understood at least something about modern science, they would know that this requirement about "testable pieces" is totally wrong and they would stop writing this breathtaking stupidity.

It's desirable for a scientist to separate statements as much as possible and to test assumptions as separately as they can – for them not to conflate unequivalent questions. But most of the time, they just cannot. At the end, they may only test predictions of complete enough theories. They may compare two theories that differ as little as possible, e.g. in one feature, if they want to test this feature. But it's not always possible to create a "modified theory" where only one feature is replaced. Theories in real science are coherent and their various features require (or at least support) the existence of other features. That's not bad news. It's good news for science: This coherence, cooperation of the features inside the theory, and the strict or partial dependence of features on other features is what allows concise enough theories to explain and predict such vast and diverse sets of phenomena.

In most cases, we can only test the "final products" but the final products depend on numerous features, theoretical constructs, and assumptions that haven't been tested separately and, in many cases, they even cannot be tested separately in principle. Everyone who has a psychological problem with this very general fact has a problem with modern science as a whole.