“I see now that the circumstances of one's birth are irrelevant. It is what you do with the gift of life that determines who you are.” -Mewtwo, Pokemon (via Takeshi Shudo)
After a week of commenting technical difficulties here on Scienceblogs, Starts With A Bang!'s Comments of the Week series is back with a vengeance! I'm so stoked that it's October, because Treknology, comes out in just two weeks! (And yes, if you want an autographed, signed copy shipped from me directly, there will be an opportunity for all of you.) Star Trek: Discovery is out, and we'll be having reviews every Monday after an episode airs, and so you may have noticed this means the end of Mostly Mute Monday for a while. But don't fret; I've started "Five For Fridays," where we'll be doing a new series on five facts, questions, examples, or some other scientific "thing" each Friday going forward. That, and of course the new Starts With A Bang podcast is live, on the James Webb Space Telescope!
- If matter is made of point particles, why does everything have a size? (for Ask Ethan),
- 5 Things the world needs from Star Trek: Discovery (beginning our Monday ST:DIS reviews),
- New space telescope, 40 times the power of Hubble, to unlock astronomy's future,
- How much fuel does it take to power the world?,
- The Big Bang wasn't the beginning, after all,
- 5 Questions you were too embarrassed to ask about the expanding Universe (for Five For Fridays),
- How can worlds that never get above freezing have liquid water? (for Ask Ethan),
- Star Trek: Discovery, analysis and recap, Season 1, Episodes 1-2 (ST:DIS review),
- It from Bit: is the Universe a cellular automaton? (by Paul Halpern),
- The four ways the Earth will actually end,
- LIGO-VIRGO Detects The First Three-Detector Gravitational Wave,
- Is the inflationary Universe a scientific theory? Not anymore (says Sabine Hossenfelder), and
- Five surprising facts about the first galaxies in the Universe (for Five For Fridays).
So no more delays; it's onto our comments of the week!
Trees are seen blown over in a parking lot as hurricane Irma moves through the area of Pembroke Pines, Florida on September 10, 2017. Making landfall as a Category 4 storm, the 2017 season, featuring both Harvey and Irma, is the first in recorded history where two Category 4 (or higher) storms have made landfall in the same year. Image credit: Saul Loeb/AFP/Getty Images.
From Denier on the impact of hurricanes: "Hurricanes are going to happen no matter what we do, but Irma is perhaps the perfect case in point on the impact of economics. The Category 4 eye wall rolled right across Key West and they’re mostly fine. It came through in the early morning and by that night there were even a couple of bars on the island that opened. There is money in Key West and the structures are well built. There are keys that don’t have Key West’s wealth. On those Keys there are mobile home parks and unrenovated houses built before the 1986+ building codes were enacted. They didn’t fare so well. It wasn’t uncommon to see a newer looking home appear as if nothing happened sitting across the street from a scene of utter devastation. Down in the Caribbean there is even less wealth and many of those islands look like they were hit by an atomic bomb."
We have now had three category five hurricanes this Atlantic hurricane season, all of which did extraordinary damage to United States territories: Harvey, Irma, and Maria. The season isn't over yet, either, but hopefully the worst of the damage is. There is really no amount of building that can save you in the worst-case scenario. Portions of Puerto Rico were prepared for 27 feet of flooding; those parts received 80 feet from the onslaught of Maria.
No, you can't look at one particular event and say, "this was the work of climate change." But you can look at what happened and say, "what can we do to repair the damage, to aid the affected, to rebuild in a more resilient fashion, to reduce the potential damage in the future, and to learn the lessons from the havoc that has been wreaked." That is my big hope for what can come out of all of this, but I have little faith that hope will come to fruition in the near future in this country.
The energy levels and electron wavefunctions that correspond to different states within a hydrogen atom, although the configurations are extremely similar for all atoms. The energy levels are quantized in multiples of Planck's constant, but even the lowest energy, ground state has two possible configurations depended on the relative electron/proton spin. Image credit: PoorLeno of Wikimedia Commons.
From Another Commenter on why matter takes up space: "The Pauli Exclusion Principle goes a long way towards explaining why matter occupies space."
This is true in one particular sense: it explains why atoms are the sizes that they are, and why multiple atoms, bound together, remain the sizes that they are. By preventing two electrons (a great example of a fundamental fermion) from occupying the same quantum state, the Pauli Exclusion Principle prohibits atoms from "shrinking" together or overlapping too much.
But the differing forces, both nuclear and electromagnetic (and to a lesser extent, gravitational), are responsible for why individual protons and neutrons, or single atoms themselves, have the sizes that they do. Yes, Pauli is an important component, but even without it, the building blocks of matter-as-we-know-it would still occupy the same volume they're observed to occupy.
From macroscopic scales down to subatomic ones, the sizes of the fundamental particles play only a small role in determining the sizes of composite structures. Image credit: Magdalena Kowalska / CERN / ISOLDE team.
From Kasim Muflahi on whether point particles would necessarily be black holes: "I agree with the implication of the question i.e. it implies that electrons, quarks etc. shouldn’t be described as zero-volume points because they have mass; and mass is quantised so that it can’t exist in a zero volume point. If it did, it’d be a black hole."
That is not necessarily true. Quarks and electrons are fundamental as far as we can measure, but there is no rule (as you incorrectly posit) that prevents these particles from being as small as the Planck scale, which is some 10^-35 meters. Our observations can constrain them down to scales of around 10^-18 or 10^-19 meters; colliders show that if they do have a physical size, it is smaller than that. We can also infer an interaction cross-section, but that is not equivalent to a physical size according to the rules of quantum mechanics.
Could they be point particles? According to the quantum rules of the Universe, as best as we understand them, yes they could. Your intuition is no substitute for the actual physics.
Ethan Siegel's upcoming new book, Treknology: The Science of Star Trek from Tricorders to Warp Drive. Image credit: Quarto / Voyageur Press, CBS / Paramount, and E. Siegel.
From Steve Blackband on my upcoming book, Treknology: "Looking forward to your Trek book. You know you will have a vociferous audience. Ive only been to one Trek conference and all i can say is that these guys are crazy!
No doubt you will do better than that awful and childish Shatner book on Trek tech. However I will be most interested in how you compare with Lawrence Krauss, of whom I am a big fan."
We're all a little crazy; I take that in a good way!
The book, Treknology, is starting to get its first reviews and so far they're very positive. I've also been doing a whole slew of interviews and podcasts about it, and there's a lot of buzz, as you'd expect, around all things Star Trek right now. But I am curious how you feel this new book compares with Krauss' now-classic The Physics of Star Trek, especially since I was a senior in high school when it came out (IIRC) and I read it. Of course, a lot has happened in the past 20+ years, and many of the technologies featured in Star Trek, including TNG, DS9, and Voyager, were simply undeveloped back in the 1990s, but are well on their way now!
It's a fantastic illustration of how science doesn't end, but progresses, and so much becomes possible in terms of how humanity can benefit when it does.
An artist's conception (2015) of what the James Webb Space Telescope will look like when complete and successfully deployed. Note the five-layer sunshield protecting the telescope from the heat of the Sun. Image credit: Northrop Grumman.
From Patrick Sweetman on the upcoming NASA flagship missions: "I suppose these things take a long time to get off the ground, but we haven’t even hoisted the James Webb Telescope yet."
These things take more than "a long time" to get off the ground. NASA, with the way its budget currently works, gets approximately one flagship mission per decade for astrophysics. In the 1990s, that was Hubble. In the 2000s, we didn't get one, owing to the legacy of "faster, better, cheaper," which gave us two (faster and cheaper) out of the three (it wasn't better). In the 2010s, we're getting James Webb; in the 2020s, it'll be WFIRST. There are a number of candidates for the 2030s, and LUVOIR is one of the finalists and perhaps the most ambitious, exciting, and expensive one.
But I've been sarcastically looking at so much of what's been proposed recently, rolling my eyes and thinking to myself, "why don't you dream a little smaller, if that's even possible." LUVOIR may be the first mission I've seen come down the pipeline, with the exception of Big Bang Observer (which would be a quartet of LISAs at different points around Earth's orbit, which is being floated for the 2050s at the earliest), that actually seems like an ambition worthy of humanity's dreams. I like it.
Also, even though the launch date got bumped, don't be down on Webb. The "five year life" is like how Opportunity (still roving, by the way) was supposed to be a 90 day mission. They've got enough onboard coolant for the mid-IR instrument to last a decade, and even past that point, the near-IR instruments on Webb could propel it into a second decade. Since L2 servicing for LUVOIR will be ideal (if not mandatory), there's no reason why the Webb wouldn't make a great testbed for it. The rewards of a refueled and serviced JWST could be astounding!
Reactor nuclear experimental RA-6 (Republica Argentina 6), en marcha. As long as there's the right nuclear fuel present, along with control rods and the proper type of water inside, energy can be generated with only 1/100,000th the fuel of conventional, fossil-fuel reactors. Image credit: Centro Atomico Bariloche, via Pieck Darío.
From John, quoting me and responding on nuclear energy: "“… Is it only our fears of nuclear disaster that prevents us from using our current technology to better the world for humanity for generations to come?’
I fear that is true. If only it were not so!"
It's easy to point to disasters like Chernobyl, Three Mile Island, and Fukushima, and showcase the highly publicized failures that demonstrate the dangers of nuclear power used irresponsibly, while the potential of using reactor fuel to generate nuclear weapons plays on some of humanity's greatest fears.
But fear is the great mind-killer when it comes to policy, and reason is the only solution. There are scientific solutions to nuclear energy without the possibility of meltdowns, without the waste problems, and without the nuclear weapons danger. If we cared about our world enough to make it so, we could switch away from fossil fuels and onto nuclear power within a decade. Alas, fear has carried the day up until the present, with far less than 10% of the world's energy coming from nuclear. This has the potential to change... if we can all agree. Again, I'm not optimistic about that anytime soon, but the world is changing, and that's a "crisitunity" if there ever was one.
Even though inflation may end in more than 50% of any of the regions at any given time (denoted by red X’s), enough regions continue to expand forever that inflation continues for an eternity, with no two Universes ever colliding. Image credit: E. Siegel.
From Denier on the beginning of the Universe: "Do they have a theory on why it didn’t happen earlier? Why didn’t the Big Bang happen at the beginning? Why wait? What was it about expanding space that didn’t allow a Bang then later did allow a Bang?"
What we can say about inflation is that, by its nature, it wipes out any information (as far as our observable Universe is concerned) that pre-existed before the final 10^-33 seconds (or so) of inflation. It's only those tiny, last moments that leave any sort of information imprint on our observable Universe at all. There are many models that are viable of what happened prior to those final moments of inflation, including:
- that inflation was eternal to the past,
- that there was a singularity in the past, and only a small region was inflating, but that inflating region took over in short order,
- that inflation was a consequence of our Universe "rejuvenating" from a prior state,
and many others. Different regions of space will see inflation end at different times, but they are forever lost to us; we can only access what's physically, causally connected to us, and all we see is all we get.
From Jim Paige on what flat space actually means: "Ethan, I know that the universe is “flat,” but when I think of that description I picture something like a very thin pancake or sheet of paper.
Since we live in a universe with 3 “travel” dimensions & time, combining to form space-time, that seems very different than “flat” to me.
I haven’t been able to get a handle on the explanation of what a flat universe really means. Could you explain the answer to me?"
I'm going to take you down a dimension, because if you want to visualize the full three dimensional space, you'd need to have experience in four dimensions to be outside of it. So let's instead think of a sheet of paper as "flat," which works just fine for two dimensions. If you took a sheet of paper shaped like a sphere, that would be "positive curvature," while if you had a sheet of paper that was shaped like a saddle, there'd be "negative curvature." The think you can ask is what happens to parallel lines, which you can ask in any number of dimensions that's two or more.
- If you have positive curvature, parallel lines will eventually meet, which is why lines of longitude all meet at the poles.
- If you have zero curvature (or perfect spatial flatness), the parallel lines will never meet, always remaining equidistant.
- If you have negative curvature, parallel lines diverge, getting farther apart the farther away you move.
We have used this technique and light from the CMB, the Big Bang's leftover glow, to measure our spatial curvature. It's 0, to a precision of ~10^-2, the best we've ever measured. If we can measure down to about 10^-5 or 10^-6, we should be able to get down to the actual curvature predicted by inflation. Interesting!
The warrior that Burnham kills is given the traditional Klingon death ritual... and then predictably used as a political tool to start a war. Image credit: Jan Thijs/CBS © 2017 CBS Interactive.
From Sinisa Lazarek on the start of Star Trek: Discovery: "the first two episodes were more of an intro into this world (although they don’t show anything of either the klingon world and state of afairs or federation, only brief hints), then they are “get to know the crew” episodes. Sort of like game of thrones but on steroids.. ok, here are the characters, by the end of the 2nd episode most of them will die.. But in the sneak peak after 2ns episode you get to learn that the whole show will more or less revolve around Burnham and the war with klingons."
Well, that's certainly what the start of the show is about, but I'm not entirely sure that we're truly in for "War Trek" as I've feared. The Federation is flawed; the Klingon empire clearly has those who disagree with T'Kuvma. (Don't forget that when it came to the initial warrior killed by Burnham, that warrior's brother would not give into T'Kuvma's demagoguery.) After all, even though they call him "T'Kuvma the Unforgettable," he's never mentioned by name in any other Star Trek series. Clearly, he's been forgotten.
And that alone should be enough to give hope; if interstellar species can learn from their failures to create a more perfect future, perhaps we can, too.
Conway's Game of Life is a popular and very simple algorithm for encoding the evolution of a system, leading to complex but stable/quasi-stable patterns. Image credit: MrJavaFrank / YouTube.
From Frank on why the Universe must be a cellular automaton: "IMHO universe/reality must be a Cellular Automata Quantum Computer operating at Planck scale."
Be very, very careful when you attempt to apply your login and intuition to how the Universe ought to behave. The "rules" that govern the Universe are neither intuitive nor necessarily logical to us; all we can do is ask nature "what are you doing" and listen and try to make sense of it. When we add ourselves into the equations, that's when we most easily are led astray.
You did post an interesting set of thoughts, though; I don't necessarily agree with them, but I don't necessarily disagree fully, either.
A collision between two large, rocky bodies in space can be catastrophic for one or both of them. This has happened to Earth before, and will no doubt happen again. But the end of the Earth? That's happening even if something like this never does. Image credit: NASA / JPL-Caltech.
From Bennett Smith on this blog: "This is a general comment to readers, not a comment on this article in particular. I want to say that Dr. Siegel’s articles are simple enough for me as a layman to understand, but complex enough to be meaningful and challenging. People who use the comments section to post attacks on Dr. Siegel are jerks and should be ashamed of themselves. If they so adamantly disagree with Dr. Siegel, they should create their own blogs. But it’s much easier to disparage than it is to create. I for one hope that Ethan continues his blog for years to come, because I enjoy them and look forwarding to reading them."
Well, wow. I very rarely get a comment this kind and generous directed towards me. It made me feel very good, so thank you for saying, Bennett. People will do what they do for their own internal reasons, and I will likely never know what those reasons are, fully. But this out-of-nowhere kindness means a lot to me, and so thank you.
From Naked Bunny with a Whip on the ultimate locking: "Earth’s rotation won’t be tidally locked to the sun before it becomes a white dwarf, will it? Actually, can it ever be tidally locked to the sun with the moon orbiting it?"
The Earth will be more strongly locked to the Moon than to the Sun, and so the Earth-Moon lock wins. When the Moon spirals away sufficiently from the Earth, the Earth will co-orbit the Moon with a period of 47 days. As our Sun loses mass (after it becomes a white dwarf), our orbit will be pushed out, will take approximately 2-3 years, and the tidal forces on our world will be only about 20% of what they are today due to the Sun. The Moon will cause a permanent deformation in the world.
Interestingly, if we were at the right distance, we could have a perfect locking, where the Moon would always be located at the L2 Lagrange point, but alas, nature didn't give us that setup.
The noise (top), the strain (middle), and the reconstructed signal (bottom) in all three detectors. Image credit: The LIGO and VIRGO scientific collaborations.
From Sinisa Lazarek on the significance of the latest gravitational wave detection: "Looking at the picture in the Forbes article (where all three detectors/signals are shown), Livingston signal does look like an actual signal. Hanford looks so/so, but Virgo looks just like noise. Why are they so different? On the other hand, why does a waveform look different in all three detectors if it;s the same signal?"
Well, three things:
- The top row shows the signal-to-noise ratio. Yes, in Livingston, it's off the charts, peaking at 14. But a SNR greater than 1 you can do something with. At Hanford, it got up to 7, which is robust. At Virgo, it "only" got up to 4.5 (which is still good), an incredible feat considering that Virgo is only about at a third the operating sensitivity of either LIGO detector.
- They are all so different because the gravitational wave has a specific planar orientation as it passes through Earth, and each detector occupies a different two-dimensional plane because the Earth is round! So Livingston is more favorable configured for this particular wave than either Hanford or Virgo (in Italy).
- And if you look at the bottom row, you can clearly visually see the goodness-of-fit in all three detectors; it isn't "just noise" even to your naked eye.
So... pretty incredible.
Also, there's a candidate for the snarkiest comment of the week in here:
Also, as Michael Kelsey notes, Virgo has only 3 km arms, while each LIGO detector has 4 km arms, which makes Virgo less sensitive in principle.
To those who are doubters, skeptics, trolls, etc., however you choose to self-define, as long as you obey the rules of conduct on this blog, you are welcome. But that does not entitle you to a response from me. Remember that.
The quantum fluctuations that occur during inflation do indeed get stretched across the Universe, but the larger feature of inflation is that the Universe gets stretched flat, removing any pre-existing curvature. Image credit: E. Siegel / Beyond The Galaxy.
And finally, from CFT on an actual quality comment on Sabine's article deriding inflation: "I think Sabine Hossenfelder says it precisely and elegantly:
“It is this abundance of useless models that gives rise to the criticism that inflation is not a scientific theory. And on that account, the criticism is justified. It’s not good scientific practice. It is a practice that, to say it bluntly, has become commonplace because it results in papers, not because it advances science.”"
There are a great many successes that inflation has had, and I think Sabine is being grossly unfair to cosmic inflation by defending Steinhardt et al.'s perspective as thoroughly as she does. I think she is dismissive of a great amount of scientifically robust predictions that inflation has given us that have been borne out by observation, and I think I will have no choice but to write a follow-up piece for later this week.
However, I think Sabine was right about the creation of useless model after useless model, which is a hallmark of "not even science" anymore. It was part of -- interestingly enough -- why I wrote that String Theory was not even a scientific theory two years ago, and her defense of string theory as science is completely inconsistent with her criteria for inflation. But you do not have to agree with me 100% of the time, and Sabine is just as much a physicist (if not more!) than I am, and is entitled to her opinion and I am proud to represent that on my platform, even if I don't agree.
But there's much more exciting stuff to come, and with that said, have a great start of October and I hope you're looking ahead to more science and even more fun as Halloween approaches!
- Log in to post comments
IMHO what should be accepted as the size of a quantum particle must be its Compton wavelength. (See below.)
I had also said I think BHs are must be made of Planck particles. I want to clarify that what I think maybe happening in BHs is, when they form, particles (neutrons?) get compressed, their Compton wavelength gets smaller and smaller, until their wavelength/size drops to Schwarzschild radius, then they turn into Planck particles.
""A Planck particle, named after physicist Max Planck, is a hypothetical particle defined as a tiny black hole whose Compton wavelength is equal to its Schwarzschild radius.""
Or maybe original particles get disintegrated into multiple Planck particles, or maybe original particles get destroyed and new Planck particles form from the available energy.
Also if above is really correct about how BHs form, that implies when physicists collide quantum particles with higher and higher energies, and see upper limit for true size of each particle drops smaller and smaller, what must be really happening is, as particles collide they are in effect getting compressed, so their Compton wavelength drops according to collision energy (but physicists interpret that as meaning Compton wavelength and (true) size of a particle are unrelated concepts).
"I’m going to take you down a dimension, because if you want to visualize the full three dimensional space, you’d need to have experience in four dimensions to be outside of it."
What?? Not so. Anyone can directly observe 3-D space by just looking around here in the real world. What is an "experience in four dimensions" anyway? Space (volume) is completely described by three axes. You write nonsense.
" If you have positive curvature, parallel lines will eventually meet, which is why lines of longitude all meet at the poles."
"Parallel lines remain the same distance apart over their entire length. No matter how far you extend them, they will never meet."
In geometry, parallel lines are lines in a plane which do not meet; that is, two lines in a plane that do not intersect or touch each other at any point are said to be parallel."
" If you have zero curvature (or perfect spatial flatness), the parallel lines will never meet, always remaining equidistant."
(One out of three statements correct.)
" If you have negative curvature, parallel lines diverge, getting farther apart the farther away you move."
As with convergence, if they diverge they are not, by definition, parallel.
You really should read Kelley Ross's paper on the Ontology and Cosmology of Non-Euclidean Geometry and quit spouting the standard mainstream confusion based on imaginary geometry with no referent in the real world.
Thanks for the explanation about the Earth/Moon/Sun rotation. I'm always happy to see it when my intuition is roughly correct, though the reality is always more complex (and better quantified).
The reason I asked is that Larry Niven's short story "One Face" popped into my head as I was finishing your article. It didn't seem like the resolution of that story would be physically possible.
There is an adage that goes:
“It is difficult to get a man to understand something, when his salary depends on his not understanding it.”
― Upton Sinclair
In reference to Ethan's comment about Sabine's inconsistency with Super stings, I've noticed that almost everyone that reacts strongly to 'letting them go' is invested heavily in them, usually professionally for many years. I can easily understand why this would make a person not want to acknowledge something isn't working out when they have so many years of their life invested in an idea. That said, when your theory can not falsified (it's true no matter what) it isn't science anymore, it's belief. There is no way around this, brutal as it may seem.
@Frank #1: You wrote, "IMHO what should be accepted as the size of a quantum particle must be its Compton wavelength." Sure, and that _is_ what we call "size" in the very low energy limit. But if you restrict yourself to zero energy, then you miss an awful lot of actual, real-world physics.
So how do you measure the "size" of a quantum entity? Well, you could try using a "ruler" or "calipers": Constrain it to a finite region, and when you can't squeeze any more (like a micrometer on a ball bearing), then you know the size. But I'm sure you know, Frank, as well as I do what happens if you do that: constraining position bumps up the momentum (motion), and doing so enough can change the properties of the entity you're constraining (excitation energies, internal states, even stuff like pair production).
So instead, in the real world, we measure the size of quantum entities by scattering: throw something at what you want to measure and see how it bounces. This works extremely well for charged entities. It's how Rutherford figured out that the gigantic (10^-10 m) atom has a really tiny (10^-14 m) hard little core in the middle, with pretty much empty space around it.
How did he figure that out? And how did he get a size? By looking at the _pattern_ (angular distribution) of scattering for a charged particle (alpha, +2) off a charged sphere (atom or nucleus). For EM, we can calculate what that distribution should be (it's an undergraduate homework problem): for a perfect point charge, the distribution is proportional to 1/sin^4. For a spherical charge, that distribution is modified with a cut-off set by the radius of the sphere.
So we can shoot charged projectiles in a beam at charged targets, measure the angular distribution, and extract the radius of the target. If you do that with electron-electron or electron-proton scattering, you don't really get an interesting result. The e-e repulsion is so strong that you don't really get to probe small radii. But at much higher energies, the beam has enough kinetic energy to get into the target field much closer to the center, and you can start to look for deviations from 1/sin^4.
For electrons on protons at medium energies (a few GeV electron beam on hydrogen), we see deviations from 1/sin^4 that are consistent with a hard sphere of radius about 10^-15 m (1 fm). At even higher energies, we start to probe _inside_ that sphere, and interact with the proton's constituents in such a way as to produce new secondary particles, not just scattering the electron.
But for electron-electron scattering, even at the very highest energies (135 GeV for each electron in a collision) probing as close as we can get to the charges shows _NO_DEVIATION_ from 1/sin^4 angular scattering. The beam energy allows us to translate that result into a maximum possible size (if the size were larger, then we'd see deviations).
That upper limit for the sphere an electron's charge must be enclosed in is currently no larger than 10^-18 m (the same is true for the size of the constituents in a proton). Maybe the electron's charge really _is_ that size. Maybe it's really down at the Planck scale, or maybe it's a true mathematical point (which is what we use in the Standard Model, because it makes the maths easier :-).
We don't know. All we know is that, at the highest precision we can measure, the electric field of an electron _looks_like_ it is a perfect 1/r^2 field coming from a point at the center.
" but I’m not entirely sure that we’re truly in for “War Trek” as I’ve feared."
After watching E03 just now.. I think we're in for Dark Trek.
Again I agree.
This was right to the point : "...when your theory can not falsified (it’s true no matter what) it isn’t science anymore, it’s belief."
It seems that most theoretical physics ( with an emphasis on math) and cosmology 'these days' falls into that category.
Evidence?... What evidence? Playing with toy models in a closed room... mostly it seems.
I somewhat disagree. All this model-building is, in one sense, an attempt to come up with different hypotheses in the hopes that one of them will turn out to be useful, predictive, etc. As with many things in science, you don't know how useful the result will actually be until you get it. Now the vast majority, if not all, of these models will end up being useless. With our perfect hindsight, we see those useless things and think 'what a waste' - but with are extremely limited foresight, we have no idea which future model will be in the useless category.
Sub-fields of science often go into a 'hypothesis exploration' phase like this experimental testing is hard to come by, for two reasons. One, because it's a comparatively cheap and easy activity to do while you're waiting for those terribly difficult, lengthy, and expensive experiments to be developed and performed. And two, because one possible payoff of a new good hypothesis might be cheaper, faster, easier ways to test amongst the candidates.
In short, sometimes in science you have a good idea of which frogs are secret princes and which ones aren't. Those are the good times. But other times you have no idea which - if any - frogs are princes. In those bad times, progress may boil down to doing a lot of kissing. Inflation theory right now is going through a 'kiss a lot of frogs and hope for a prince' phase.
Let's hear it for Paul Feyerabend!