Holes in Big Bang

[sol invictus]

Sol...

Unless you have some other references I'm going to stick to my statement that to my knowledge all inflation 'predictions' are actually "postdicted' from earlier observations. That certainly seems to be the case with both of the early papers you cited.

What? Why?

It's true that it was known in the early 1980s that a roughly flat primordial spectrum over some range of scales would probably be consistent with observations, but that's it (and that's why those guys were excited about their results). But inflation predicted something much more specific, only the first rough version of which is in those 1982 papers. By 1996 paper there were many details (like the acoustic peaks), none of which were observed until years later.

As it relates to the DE/EM issue, I'm still doing some background reading and I have not reached a final conclusions just yet.

OK.

 

[Michael Mozina]

What? Why?

Did you see my earlier post on the two earliest papers you cited? You haven't responded to it yet, unless of course I missed it.

It's true that it was known in the early 1980s that a roughly flat primordial spectrum over some range of scales would probably be consistent with observations, but that's it (and that's why those guys were excited about their results).

But the authors are actively "curve fitting' inflation theory to known observations in those earlier papers.

By 1996 paper there were many details (like the acoustic peaks), none of which were observed until years later.

I guess I'll have to go through the later paper then with a bit more focus on what (if anything) is actually a 'prediction' vs. what is simply a curve fit to a known observation. It is however quite clear that some of these features were already "observed" by the time the first papers were written and in those earlier papers, inflation theory is being revised to take these previous observations into account. That is postdiction, not prediction.

I suppose it's 'possible' to "predict" some some features based purely on mathematical models, but typically, most "predictions' are based upon something that is learned from active experimentation, not simply a mathematical model. GR theory shows that there are exceptions to that rule however.

 

[Perpetual Student]

What? Why?

It's true that it was known in the early 1980s that a roughly flat primordial spectrum over some range of scales would probably be consistent with observations, but that's it (and that's why those guys were excited about their results). But inflation predicted something much more specific, only the first rough version of which is in those 1982 papers. By 1996 paper there were many details (like the acoustic peaks), none of which were observed until years later.

This appears to be a gray area. Initially, inflation was needed to account for observed characteristics of the universe like homogeneity, etc. Later, it was realized that a totally homogeneous era could not lead to the structure currently observed --- some inhomogeneity was needed. As I recall, quantum fluctuations were then presented as the likely cause of the primordial lumpiness leading to galaxy formation, etc. Then, a search for inhomogeneities in the CMB ensued -- and were found.
Would there not have to be quantum fluctuations anyway, for galaxies to form, with or without inflation? Are there not other possible explanations for the CMB and its irregularities? It seems that MM's view that aspects of inflation theory are "postdicted" might have some merit.

 

[sol invictus]

Did you see my earlier post on the two earliest papers you cited? You haven't responded to it yet, unless of course I missed it.

Oops, missed it. What they're saying is that they know that a flat spectrum with amplitude around 10^-5 would roughly fit data. But it's very rough, because at the time no precise data was available - and didn't become available until 20 years later.

Typical inflation models, as I've told you, have 1 or 2 parameters, 1 or 2 numbers that must be put in. As with any scientific theory, you use 1 or 2 numbers from data to fix those (in this case, 10^-5 as the overall amplitude). Everything else you get out is a prediction (or postdiction, as the case may be).

Here, the output is an entire spectrum. It's observable out to angular momentum moment of a few thousand, which means the output is a few million numbers which form a very specific and characteristic pattern. If any one of those had turned out to be significantly off, inflation would have been fasified. But none of them were - there are a few mild and potentially interesting anomalies, but only a few (again, out of millions).

 

[sol invictus]

This appears to be a gray area. Initially, inflation was needed to account for observed characteristics of the universe like homogeneity, etc. Later, it was realized that a totally homogeneous era could not lead to the structure currently observed --- some inhomogeneity was needed.

Actually that was realized immediately, not later.

As I recall, quantum fluctuations were then presented as the likely cause of the primordial lumpiness leading to galaxy formation, etc. Then, a search for inhomogeneities in the CMB ensued -- and were found.

Right - but it wasn't just that inhomogeneities were found. A few million data points were collected by satellite, and they matched the predictions extremely well. This was one of the most precise and impressive achievements in the entire field of astrophysics, by the way. It catapulted cosmology into the status of "precision science".

Would there not have to be quantum fluctuations anyway, for galaxies to form, with or without inflation?

No. Quantum fluctuations can only seed structure if there is an event horizon, and that happens only when the expansion is accelerating (which means some variety of inflation).

Are there not other possible explanations for the CMB and its irregularities? It seems that MM's view that aspects of inflation theory are "postdicted" might have some merit.

I don't see any connection between those two sentences.

Sure, there are other possible explanations. None of them work very well at all, though - they're all much more complicated. As for postdicted, I'm befuddled how you can regard a precise prediction for the shape of the spectrum made at least 7 years before the data was available as a "post"-diction.

 

[sol invictus]

I suppose it's 'possible' to "predict" some some features based purely on mathematical models, but typically, most "predictions' are based upon something that is learned from active experimentation, not simply a mathematical model.

That's not true at all. In all types of physics one uses some experimental input plus certain basic principles and experience to build a mathematical model, and then makes predictions with the model that are then tested. But of course it's the model that makes the predictions. You can't make predictions with experimental results, except perhaps about an absolutely identical experiment - and that wouldn't be interesting. The only way to make predictions is with a model, and in physics the models are mathematical.

The top quark is an excellent example, by the way.

 

[Perpetual Student]

Actually that was realized immediately, not later.

OK

Right - but it wasn't just that inhomogeneities were found. A few million data points were collected by satellite, and they matched the predictions extremely well. This was one of the most precise and impressive achievements in the entire field of astrophysics, by the way. It catapulted cosmology into the status of "precision science".

Thanks for that clarification and additional insight.

No. Quantum fluctuations can only seed structure if there is an event horizon, and that happens only when the expansion is accelerating (which means some variety of inflation).

Unfortunately, I don't understand QM well enough to follow that but I must accept that as a valid point, unless someone else can refute it.

I don't see any connection between those two sentences.

Sorry, I guess there wasn't any connection. See my comments below.

Sure, there are other possible explanations. None of them work very well at all, though - they're all much more complicated. As for postdicted, I'm befuddled how you can regard a precise prediction for the shape of the spectrum made at least 7 years before the data was available as a "post"-diction.

Let me try this rather fuzzy explanation. Over the years, as I followed the development of inflation theory in periodicals like Scientific American, there appeared to be some ad hoc tag-ons to inflation theory in the wake of new astronomical observations. My feeling is based on the historical context as things unfolded at the time, but I cannot recall enough details to demonstrate my point. I simply have no way of now reconstructing the sequence of observations and theoretical modifications of inflation to document that there were times when new developments appeared to be "postdicted."
In any case, I do continue to accept inflation theory, but not with any sense of certainty.

 

[DeiRenDopa]

I suppose it's 'possible' to "predict" some some features based purely on mathematical models, but typically, most "predictions' are based upon something that is learned from active experimentation, not simply a mathematical model.

That's not true at all. In all types of physics one uses some experimental input plus certain basic principles and experience to build a mathematical model, and then makes predictions with the model that are then tested. But of course it's the model that makes the predictions. You can't make predictions with experimental results, except perhaps about an absolutely identical experiment - and that wouldn't be interesting. The only way to make predictions is with a model, and in physics the models are mathematical.

The top quark is an excellent example, by the way.

Here's a quite different kind of example ...

... the Hulse-Taylor pulsar, and gravitational wave radiation.

With GR, and a model (a pair of 'point' masses in orbit around their mutual centre of mass) one can make certain predictions (rate of decay of the orbits); plug some numbers in (masses, distance between them) and an eminently testable hypothesis falls out (the observed pulses from a binary pulsar will behave {like this} when plotted against time).

No experimental results to be found anywhere ... (oh, and Hulse and Taylor got an all-expenses paid trip to Stockholm).

 

[sol invictus]

Let me try this rather fuzzy explanation. Over the years, as I followed the development of inflation theory in periodicals like Scientific American, there appeared to be some ad hoc tag-ons to inflation theory in the wake of new astronomical observations.

Well, that's not an entirely unfair criticism. As I mentioned earlier, there are many different inflation models, and there's quite a lot of room to adjust the pre/post-dictions - particularly if you're willing to add ingredients and make the model more complicated. But there are certain features and predictions they all share, and those were well-understood long before the relevant observations were made.

It's very much like quantum field theory. There are infinitely many QFTs, and only one that describes the standard model of particle physics. They all have certain features in common, and physicists were fairly certain particle physics was described by a QFT long before they knew which one it was.

In fact it took quite a long time and lots of experimental input before the correct theory (SU(3)xSU(2)xU(1) with the right matter content) could be identified. It has 25 or so parameters which must be fixed with data. Once that's done, however, it makes lots and lots of predictions, a huge number of which have since been verified (and a few falsified, and the model adjusted).

Similarly there are lots of inflation theories, particularly if you allow them to be as complex as the standard model. We don't currently have the data to nail down which one is correct, but we do have enough to be quite confident that at least one is.

And remember - nature couldn't care less whether we first discover the correct theory and then predict the data, or first find the data and then discover the correct theory. A theory is either correct or not; pre- versus post-diction is a human distinction.

 

[Zeuzzz]

I think everyone needs to chill. There no point getting personal. This thread has been full of emotive undertones (more-so at the beginning). Fact is that there are a lot of alternative theories out there, and its not a matter of one being the *truth* and all the others crackpot theories made up by *creationists* or [insert stereotype here]. When you put faith in a theory being truth and get religously attatched to it by thinking its beyond reproach from all other theories, your no longer thinking scientifically, but religously.

Theres some good material in this thread. Lets not forget people, multiple theories can be correct at the same time, no matter how different they are. Finding out which theory matches the evidence and data is what should be being done, something called the scientific method. Of which there have been glimmers of in this thread.

I'm not gonna get fully involved either way. Infact I dont spend much time online at all at the moment, real life issues are pressing.

Argue on, and keep it in good faith peeps.

Isn't science wonderful?

 

[Michael Mozina]

Oops, missed it. What they're saying is that they know that a flat spectrum with amplitude around 10^-5 would roughly fit data. But it's very rough, because at the time no precise data was available - and didn't become available until 20 years later.

Typical inflation models, as I've told you, have 1 or 2 parameters, 1 or 2 numbers that must be put in. As with any scientific theory, you use 1 or 2 numbers from data to fix those (in this case, 10^-5 as the overall amplitude). Everything else you get out is a prediction (or postdiction, as the case may be).

Here, the output is an entire spectrum. It's observable out to angular momentum moment of a few thousand, which means the output is a few million numbers which form a very specific and characteristic pattern. If any one of those had turned out to be significantly off, inflation would have been fasified. But none of them were - there are a few mild and potentially interesting anomalies, but only a few (again, out of millions).

My problem with that logic is this: If I already have a number of key "rough" observations from several different wavelengths in the spectrum and I "assume/figure out" that there is a discernible pattern in that data, I can then create a formula to fit that basic pattern.

Unless there really isn't a pattern, my formula, rough as it may be, will still be likely to apply quite well to the later (more refined) measurements. The pattern itself is still "postdicted" from the rough (earlier) observations and the mathematical pattern was still worked out from observation, not from actual "prediction". Unless the postdicted pattern that I come up with is simply wrong, it's going to apply pretty well to later and better measurements. The only thing that is likely to be modified a bit by later, more accurate measurements are the 'variables' but the mathematical pattern was still a postdicted fit.

 

[Michael Mozina]

Here's a quite different kind of example ...

Oh my goodness. I already conceded in my original post (the part you can't see because you're ignoring me and he cut out that part of my post) that there were exceptions to that rule. You're preaching to the choir.

 

[Perpetual Student]

And remember - nature couldn't care less whether we first discover the correct theory and then predict the data, or first find the data and then discover the correct theory. A theory is either correct or not; pre- versus post-diction is a human distinction.

Thanks. Perhaps that's the ultimate key. We have a theory; we have matching observations. If no other theory fits the observations (and until one does) we go with what we have.

 

[Michael Mozina]

And remember - nature couldn't care less whether we first discover the correct theory and then predict the data, or first find the data and then discover the correct theory. A theory is either correct or not; pre- versus post-diction is a human distinction.

I'd be fine with that concept if we could actually demonstrate that inflation is real in a standard test with control mechanisms. As it stands, we're postdicting a fit with an "imaginary" entity that presumably no longer exists and can *never* be verified or falsified in any conventional test with a control mechanism. As long as we're willing to keep modifying inflation theory to fit any and all new observations, it becomes a form of 'dogma' that defies any kind of falsification processes. At that point there is no longer any empirical distinction between science and religion.

 

[sol invictus]

My problem with that logic is this: If I already have a number of key "rough" observations from several different wavelengths in the spectrum and I "assume/figure out" that there is a discernible pattern in that data, I can then create a formula to fit that basic pattern.

In some cases, perhaps. But that couldn't have happened here. Look at the spectrum.

All that was known of that in 1996 (I think, perhaps someone else can cofirm) is the extreme left-hand part of it, stopping well before the first peak at perhaps l~50 (that's the numbers on the horizontal axis) and with large error bars. You tell me - given that much, could you fill in the rest, including relative heights of peaks, their spacing, etc.? Obviously not - and if you look at those papers, you'll see that that isn't at all what happened. Instead, the spectrum was predicted from the model.

I As long as we're willing to keep modifying inflation theory to fit any and all new observations, it becomes a form of 'dogma' that defies any kind of falsification processes. At that point there is no longer any empirical distinction between science and religion.

That's utter nonsense: the progress of science is nothing other than modifying theories to fit new observations. But I'm tired of going around in circles on that, so I'm not going to respond further.

 

[LibraryLady]

I am the third moderator to step in here. I have sent 28 posts to AAH, all of them off topic. Anyone derailing from here on in risks infraction. Behave.

Replying to this modbox in thread will be off topic  Posted By: LibraryLady

 

[DeiRenDopa]

What? Why?

It's true that it was known in the early 1980s that a roughly flat primordial spectrum over some range of scales would probably be consistent with observations, but that's it (and that's why those guys were excited about their results). But inflation predicted something much more specific, only the first rough version of which is in those 1982 papers. By 1996 paper there were many details (like the acoustic peaks), none of which were observed until years later.

This appears to be a gray area. Initially, inflation was needed to account for observed characteristics of the universe like homogeneity, etc. Later, it was realized that a totally homogeneous era could not lead to the structure currently observed --- some inhomogeneity was needed. As I recall, quantum fluctuations were then presented as the likely cause of the primordial lumpiness leading to galaxy formation, etc. Then, a search for inhomogeneities in the CMB ensued -- and were found.
Would there not have to be quantum fluctuations anyway, for galaxies to form, with or without inflation? Are there not other possible explanations for the CMB and its irregularities? It seems that MM's view that aspects of inflation theory are "postdicted" might have some merit.

I'd like to introduce an example from astronomy that shines a bright light on the "predicted/post-dicted" distinction; I think the only reasonable conclusion one can draw from this example is that such a distinction is pretty close to meaningless, and that if you do wish to try to keep it, you would need to pay extraordinary attention to detail.

Consider "dark matter".

Zwicky introduced the term, in the 1930s, to account for an apparent inconsistency in an analysis he did of the data he obtained on a rich galaxy cluster (Coma). At the time, Zwicky had no reason to introduce 'non-baryonic' as a modifier, nor 'cold'; indeed, it is unlikely either *could* have been applied at the time!

Not long afterwards (or possibly before), Jan Oort (yes, of Oort cloud fame) published evidence for the existence of dark matter in the local region of the Milky Way (i.e. within a few hundred pc of sol); several decades later re-analysis together with far more extensive data showed his conclusion was in error.

In the late 1960s, Rubin (and colleagues) published studies of the rotation curves of some nearby normal spiral galaxies, concluding that these galaxies are embedded in a halo of dark matter.

It took another ~25+ years for compact massive halo objects (MACHOs) to be ruled out as the primary component of this inferred massive halo; sometime around then dark matter began to be seriously considered as being non-baryonic (and cold).

Sometime after the first decent x-ray band data on rich galactic clusters became available, the existence of an essentially thermal, hot, diffuse IGM was confirmed; the estimated mass of such was, and still is, considerably greater than the total estimated mass of all the galaxies in such clusters, dark matter halos included. However, the estimated total mass of such clusters exceeded (and still exceeds) that of the IGM by a factor of ~5.

Along the way, and completely independently, cosmological research was converging on an estimate of the average mass density of the universe being ~one-fifth of the critical density; of this mass, several lines of independent evidence strongly suggested that only ~one-fifth was baryonic.

(the actual history is much, much, much more intricate than I have outlined above).

So, to cut to the chase: non-baryonic cold dark matter ("CDM"), as a theory, is extraordinarily successful ... it accounts for millions of independent observations, across the full range of the electromagnetic spectrum, and of an extraordinary range of objects (from dwarf galaxies to normal galaxies to giant galaxies to galaxy groups to galaxy clusters to the universe as a whole). In the multi-decade history of the study of CDM (to be anachronistic for several decades), there have been several 'crises', many curiosities and anomalies, a great deal of refinement and revision, hundreds and hundreds of predictions and post-dictions, etc, etc, etc, etc.

And I haven't even introduced an independent line of research: indications from particle physics of the existence of an entire class of particles hithertofore unseen, the properties of which could well match those of CDM (should CDM be composed of particles).

I have read MM's posts on this topic (there are dozens, if not hundreds), and like si I find that his characterisation of astronomy is grotesque, and his views on 'controlled experiments' etc riddled with misconceptions, internal inconsistencies, etc.

 

[Perpetual Student]

DeiRenDopa:

Thank you for the above historical summary of dark matter. I understand that you necessarily omitted countless details; nevertheless, it was very helpful.
The sad reality is that it is difficult for a layman to access all the context you just provided.
As I mentioned above, I have been following developments in cosmology for many years now (about 50), but obviously (and sadly) the significance of many details and interlinking of concepts have escaped my grasp.

 

[DeiRenDopa]

PS, for a relatively brief, non-technical overview, I recommend "In Search of Dark Matter", by Ken Freeman and Geoff McNamara (2006, Springer/Praxis; ISBN: 0-387-27616-5). Freeman, who must be close to retirement by now, is a professional astronomer who has been working on DM for just about his whole (professional) life, developed at least one of the observational tools used to test various DM hypotheses (i.e. PNe in the outskirts of galaxies), and has authored several hundred papers (not all as sole, or even lead, author of course!). There are other, popular-level, books on the topic, but this is the best that I've read.

 

[Michael Mozina]

DeiRenDopa:

Thank you for the above historical summary of dark matter. I understand that you necessarily omitted countless details; nevertheless, it was very helpful.
The sad reality is that it is difficult for a layman to access all the context you just provided.
As I mentioned above, I have been following developments in cosmology for many years now (about 50), but obviously (and sadly) the significance of many details and interlinking of concepts have escaped my grasp.

FYI, there are new signs that the 'mass estimation techniques' of standard theories will require some revision because they fail to account for some of even the most visible material that is located inside various galaxies:

http://www.sciencedaily.com/releases/2009/08/090819145846.htm

The effects are particularly important in parts of the universe where stars are spread out over a larger volume -- the rural Africa of the cosmos. There could be about four times as many stars in these regions than previously estimated.

"Especially in these galaxies that seem small and piddling, there can be a lot more mass in lower mass stars than we had previously expected from what we could see from the brightest, youngest stars," Meurer said. "But we can now reduce these errors using satellites like the Galaxy Evolution Explorer."

Emphasis mine.