Hacker News Reader

> it's clear that observation and model are being conflated

How so?

> [cosmology] isn't a science the way that, for example, physics is.

Why not?

> I'm sure many will disagree with this factually more accurate description, because of the emotional role their ideation of science plays in their lives

Perhaps i will once i understand what you're going on about.

I agree with you here. A phrase that always echoes in my mind when discussing cosmology is

"All models are wrong - some models are useful"

The farther you venture from the verified useful section of a model (by which I mean - the farther you are from model predications that have been validated with observational evidence), the less you should trust it - ALL MODELS ARE WRONG!

And for most sciences - this isn't a huge deal - we can do lots of observational work easily right now. For cosmology... well - our observational data on the history of the universe it just astoundingly, mind-bogglingly, miniscule in comparison to the events we're interested in.

Note that the most distant astronomical object we've observed has a redshift of 11, corresponding to a light travel distance of 13.4Gly. This means that if our models are correct and the universe is indeed about 13.8 billion years old, then we've seen 97% of the universe's history. Including the microwave background in that pushes the number to 99.997%.

....no?

We have a sampling of data from 99.997% of the timeline — but the sample itself is extremely sparse.

You have an observational slice of about 4000 years at the most generous, 400 years more leniently (telescope invention was ~1600s), 40 years for modern tools (OAO-2 in the late 1960s).

You might as well be looking at a single frame of a movie and telling me you know the whole plotline because it happens to have the same pixels on screen for the whole shebang.

But that specific frame has all the previous frames imprinted and increasingly faint in some sort of cosmological motion blur.

> the most distant astronomical object we've observed has a redshift of 11

And even though that seems like a large distance, it is a very small redshift compared to the total time the universe has existed. The cosmic microwave radiation background, for comparison, was emitted at a redshift of about 1100. So a redshift of 11 only covers 1 percent of the expansion of the universe since the CMBR was emitted.

> we've seen 97% of the universe's history

No, we haven't. See above.

> Including the microwave background in that pushes the number to 99.997%.

No, it doesn't, because even though we can see the CMBR, we can't see anything useful in between its redshift and the redshift of 11--not because there's nothing there, but because what's there is too distorted and faint to see. (The only reason we can see the CMBR is that it's black body radiation at a temperature we can independently predict from our knowledge of the physics of recombination, so we can tailor extremely sensitive instruments to looking for its precise signature.) So there is a lot of universe that we haven't seen.

No, we haven't. See above.

No we have. See above: While redshift goes all the way up to infinity, in terms of cosmological time, there's still only 0.4 billion years between z=11 and z=∞, or about 3% of the age of the universe.

To quote you above:

> if our models are correct

What is this whole discussion about?

You're arguing a stance that assumes the model is correct in a discussion about the risks of assuming the model is correct.

But trust me - our model is not correct.

> there's still only 0.4 billion years

In terms of time, yes. I was thinking in terms of observable spacetime volume. Even though there's only about 0.4 billion years before z = 11, the scale factor increased by a huge factor during that time (a factor of 100 from CMBR emission at z = 1100 to z = 11). That's a lot of spacetime volume from which we have no useful observations.

At what point does assumptions built upon assumptions approach being completely incorrect & useless?

The second you need to make a reliable predication based on it.

The good news (or bad depending on your take) is that we're nowhere close to being an advanced enough civilization to have any real stake in making predications about the future state of the universe other than "it will look basically the same as it does today", which we've conveniently used to position ourselves very accurately on the surface of the earth since the sextant and star charts.

No, you are absolutely wrong here. But before I explain why I need to ask a question: are you by chance a young-earth creationist? Because your rhetoric here is very similar to that employed by YECs. (The reason I ask is that my explanation is going to be different if you are not a YEC than if you are.)

I'm engaging in dialectic, not rhetoric. I don't believe rhetoric has a place on this site, at least not from me.

Since this is a site dedicated to intellectual curiosity, why don't you please present me with both?

It is important to note that you did not answer the question.

Only if you think dialectic can't cope with ambiguity.

> why don't you please present me with both?

If you are not a YEC, then the first thing I'm going to do is ask you why you think that cosmology and physics are different?

If you are a YEC then I don't need to ask you that question because I already know the answer. (It's because you have accepted the truth of the Bible as a foundational assumption. From this, the idea that modern cosmology is not science is a logically valid conclusion.) Instead, the first thing I'm going to do at this point is to ask you why you are trying to conceal it.

Also, if you are a YEC, I would be interested to know if you were surprised that I was able to correctly guess this.

If I were not a YEC, then I'd think cosmology and physics are different because physics is a laboratory science where depending on our budget we can do anything from roll balls down inclines to produce quarks while with cosmology we have absolutely no control over conditions and can only make believe we do by sampling in a way that necessarily is corrupted by our biases. And the amount of data to sample from is plenty ample for unconscious bias to have a significant effect.

If I weren't a YEC, I would focus more on issues with radiocarbon dating and the assumption and not the observation that carbon isotope ratios are constant over the history of the Earth and ignore cosmology entirely. In fact I'm not clear on what YEC has to do with cosmology at all, it strikes me as being outside the scope of the arguments I've seen on the subject.

> physics is a laboratory science

That would come as news to most theoretical physicists, who rarely set foot in a laboratory.

> we can ... produce quarks

How do you know we can produce quarks?

Also, where does astronomy fit into your taxonomy? Was Newton doing science or phenomenology when he came up with the inverse square law? Plate tectonics? What about (drum roll, please) biology?

> That would come as news to most theoretical physicists, who rarely set foot in a laboratory.

I agree that it's fair to describe cosmologists as a species of theoretical physicist, given the long standing connection between the observation of heavily bodies and the birth of the modern physical sciences. Still the relation is rather remote.

> How do you know we can produce quarks?

I personally don't. I'm relying on hearsay from a buddy with a PhD in particle physics who worked at Los Alamos. I don't think he's a liar so I'll take him at his word about what's possible in particle physics labs.

> Also, where does astronomy fit into your taxonomy? Was Newton doing science or phenomenology when he came up with the inverse square law? Plate tectonics? What about (drum roll, please) biology?

Astronomy is a phenomenology on account of nobody has a lab big enough to create stars in, or run any other astronomical scale experiments. Newton was doing natural philosophy, which in his case had elements of both what we now call science and phenomenology. Plate tectonics would strictly speaking be a phenomenology. Biology is a bloody wet mess that's mostly phenomenology, but there are disciplines in it which are mature science to the point of being engineering, like breeding domesticated plant and animal species.

> Astronomy is a phenomenology

OK, again this would come as a surprise to most astronomers. Are you aware that we have sent spacecraft to other planets, and that humans have walked on the moon?

I think you might want to review the site guidelines[1]: "Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith."

[1] https://news.ycombinator.com/newsguidelines.html

Except that you have given me evidence that you are not dealing in good faith. You have chosen to be cagey about your beliefs, so I am making as few assumptions about them as possible. I don't know if you're a YEC, a flat-earther, or a lunar landing denialist. People exist who subscribe to all of these positions, so unless you tell me otherwise, I'm perfectly justified in considering the possibility that you might be any of these things.

I agree that it's fair to describe cosmologists as a species of theoretical physicist, given the long standing connection between the observation of heavily bodies and the birth of the modern physical sciences. Still the relation is rather remote.

This is such a weird statement to make.

Let's not forget how we got where we are:

There are two forces active on macroscopic scales, electromagnetism and gravity. At the turn of the previous century, there were some inconsistencies in the way how electromagnetism fit into the rest of physics, which were resolved by the theory of Special Relativity. That shifted the problem to gravity, leading to the theory of General Relativity.

Friedmann then calculated what solutions General Relativity permitted under the assumption of spatial symmetry, the Friedmann models, which form the foundation of the cosmological standard model.

Most of the things I described so far happened before we even had confirmation that other galaxies existed (Friedmann published his 2nd paper in 1924, while Hubble resolved the 'Greate Debate' in 1923 by discovering Cepheids in the 'Andromeda Nebula', nowadays 'Andromeda Galaxy'). That shifted the focus on trying to figure out the model parameters that made Friedmann's model fit reality, and we've been at it ever since.

I'm sorry to say your understanding of physics is pretty wrong.

For both, we control almost nothing, we observe. To that end we build tools to observe. We want to observe to reject hypotheses or to get new insight on how nature behaves. In particle physics these are detectors, like the Super Kamiokande [1], which is just sitting there waiting for neutrinos to arrive from space. In the case of cosmology these are telescopes, radio telescopes and the like, waiting for photons, gravitational wave chirps to arrive from space.

[1] https://en.wikipedia.org/wiki/Super-Kamiokande

You don’t need to apologize for your discomfort at my choosing to understand things in a way I personally find sensible. This has never bothered me because in the areas where it pragmatically matters I’ve achieved success and in the areas where it’s academic I’ve at least somewhat satisfied my curiosity.

Collecting a catalog of phenomenological observations is the defining characteristic of a phenomenology. One may feign hypotheses as one likes to fit that data, but that activity, valuable though it may be, is qualitatively distinct from the proper science of predicting the phenomena before they are observed. Of course phenomenologies can provide the necessary insight for predictive science. Like all human endeavors boundaries can blur.

Sadly much so called science is just computer aided trunk wiggling[1].

[1] https://quotefancy.com/quote/1342664/John-von-Neumann-With-f...

I think observation and model almost necessarily have to be conflated. What we really see is light hitting our retina, or if you want to go one layer further, nerve impulses in our visual cortex. Everything else is model.

So if we accept that everything is model, then what you are really advocating is using verbose descriptions in terms of a fairly established model rather than succinct descriptions in a more speculative model. Phrasing it like this makes it clear that there is a balance to be struck between pros and cons. Maybe we are erring on the speculative side, but you will have to make that case.

Thank you for helping clarify what I'm saying. You've definitely got it right. To elaborate, consider the difference between calculating the distance to a star in the Milky Way using parallax where the model is basically just assume your eyes work plus trigonometry and calculating the distance to other galaxies where far more assumptions are required.

Edit: response from cygx is completely missing the point about control vs selection.

Which is no different in other branches of physics: Have you ever seen a quark?

If you have a model that makes testable predictions, it's science. Cosmology is not only science, but physics, making your original comment contrasting the two kind of nonsensical...

response from cygx is completely missing the point about control vs selection

Because your comment I responded to was arguing something different - namely, the number of required assumptions.

Due to confinement, the existence of quarks rests on quite a lot of them.

> It would better be described as a phenomenology

Not really, no, because current cosmological models are not derived solely by looking at cosmological phenomena and coming up with phenomenological equations that describe them. (That is not to say that past cosmological models didn't do that; only that our current ones, roughly since the early to mid-20th century, don't.) They are derived by starting from laws of physics that already work in other domains, and seeing what those laws say about the universe as a whole.

> to be absolutely clear what the assumptions of the model are, even the most trusted ones

Cosmologists are clear about that.

> maybe this is just an issue in the popular press?

I think it is, since what I see in actual textbooks [1] and peer-reviewed papers is not at all like what you are describing.

[1] A good reasonably current textbook is Liddle's Introduction to Modern Cosmology.

Thanks for the recommendation, I ordered that text.

Great touch. Actually I think we have model based science quite many decades now, but those model need to be anchored on testable and refutable something. (Phenomena is loaded word as I wonder it assumes model free and equipment free a bit). Model and refutable is key here.

We still have the qm to deal with. Just hope our model is even understandable. But at least that model is refutable.

As for correctness, I suspect we might never see 100% ever though. That is partly we are just have a century of major re-think about the fossilization of good science. And there is just happen our maths and mind can understand quite a bit of the universe.

There is always something to be found is good. And there is no guarantees we will know it all. That is good too. As it keeps the most important of science and philosophy alive:

Curiosity.

Wasn't there a time, not so long ago in fact (pre-1920's), where almost all your beloved physics models, beside Netwonian gravities, were wrong as well?

So Astronomy, as a "hard science", is around 100 years behind...what's the big deal with that?

newton was a bit wrong and we are still a bit wrong. where wrong means our models of reality are imperfect.

Even a lot of mathematical assumptions were wrong before Godel and Tarski.