Fundamental Question on the Nature of Science

Tim Thompson · Apr 11, 2009

I set the context with the following exchange of opinions from the "Lambda-CDM theory - Woo or not?" thread. However, as we can see, the comments in fact have a wide ranging effect on what we can consider to be "scientific".

Tim Thompson said:
... So, is astronomy an empirical science or not? Are astronomical data empirical or not?

Michael Mozina said:
Observations are not "empirical experiments". Astronomical data as you call it is also open to "subjective human interpretation".

Tim Thompson said:
I disagree. Yes they are empirical experiments. Astronomical observations constitute valid tests of astronomical hypotheses, in exactly the same sense as an experiment in an organic chemistry lab. In both cases there is an hypothesis to test, and the observations serve that purpose. In both cases criteria are established prior to the experiment, whereby either the verification or falsification of the hypothesis can be established. There is no difference between the data gained by astronomical observation to test astronomical hypotheses and the data gained in a controlled laboratory experiment to test a laboratory hypothesis.

Michael Mozina said:
Observations are *not* empirical experiments. There is no "control mechanism", and no "experiment" actually occurring. We are "observing" something, and making "subjective interpretations (different ones)" from this uncontrolled observation. ... But Tim, there is no 'control mechanisms' as it relates to events in deep space. All we can do is observe these events without any hope of "controlling" them.

So what we have here is a contest between "observation" and "controlled observation", which is really what a "controlled experiment" is. Now, Mozina's opinion is explicit enough; he says "Observations are not 'empirical experiments' ". It is not much of a stretch to read this as "observations are not empirical" period.

This impacts a wide range of fields of science where "controlled experiments" are usually impossible. Astronomy is one example, where the full complexity of a major astronomical object (planet, star, molecular cloud, galaxy & etc.) cannot be reproduced in any laboratory, ever. But so are zoology, anthropology, and a host of natural sciences affected by the crucial difference between "observations" and "controlled experiments". Are naturalists in the field, observing chimpanzees in the wild, for instance, doing science? Are they doing something "empirical"? If you observe your favorite chimps long enough to predict their behavior, and then observe them to verify the prediction (i.e., to validate the hypothesis), is that not science because it is not "controlled"? Is that not an "empirical experiment" because it is not controlled?

My own opinion is made clear in the exchange above as well, I hope. I think both controlled and uncontrolled observations & experiments are on equal standing when it comes to considering them as "empirical" or "scientific". obviously they are not the same thing, and a "controlled experiment" might well allow the experimenter to more stringently qualify their results than is done with an uncontrolled observation in the field. But that does not detract from the empirical, and experimental nature of an uncontrolled observation.

This seems to me to be a fundamental issue of importance across the scientific spectrum. So I thought it would be worthwhile to make it a topic of its own and solicit the opinions of other scientists (or anyone else who cares to join in) who might not be reading a thread with "Lambda-CDM" in the title. What do you all think? Are "uncontrolled" observations equally empirical, equally scientific, compared to controlled experiments? Do those of you working in areas where field observations are common consider yourselves to be scientists, and that you are doing science?

Skwinty · Apr 11, 2009

My opinion is:
If you don't treat astronomical observations as empirical, then you learn nothing.
Astronomers would then achieve nothing but " ooh look at the pretty image in my telescope"

athon · Apr 11, 2009

Tim Thompson said:
However, as we can see, the comments in fact have a wide ranging effect on what we can consider to be "scientific".

Before I address your other points, there is no single consensus on what science is and is not. This is called the demarkation problem in HPS, and has served as fodder for all manner of philosophical debates.

This does not mean we cannot exclude anything from being science, but it does mean that a precise categorisation of activities is impossible to do from an objective standpoint.

Now, Mozina's opinion is explicit enough; he says "Observations are not 'empirical experiments' ". It is not much of a stretch to read this as "observations are not empirical" period.

First point: like 'science', 'experiment' has no boldly defined boundaries. That said, it's typically understood to be an activity that involves direct manipulation of some variables in order to better understand how they might be related to others.

Going by that, it's possible to argue that astronomy can indeed be experimental. Any time one intervenes in selecting observations according to a possible relationship, they are controlling variables. That's not to say there aren't fields in astronomy where it's difficult to accomplish this, however, which is what this question is pertaining to.

Are naturalists in the field, observing chimpanzees in the wild, for instance, doing science? Are they doing something "empirical"?

Here is the crux of the matter - you seem to be asking if science is defined by the ability to manipulate variables. I'd say 'absolutely not'. I'd go even further to say that our understanding of the universe would be dramatically impeded if this was the defining feature.

Let's look at it this way - science is not a method. I know you're probably taught that at school or at university, but it's bollocks. It's simply false.

Science is a methodology. It consists of a range of values and shared understanding of historical precedents that aim to evaluate a method in order to decide collectively if it presented us with useful information. As such, it is never a matter of finding an absolute truth or arriving at a solid conclusion, but rather a way of building confidence in an idea having some usefulness in explaining an observation.

Difference practices will give us different levels of confidence in a result. If we can directly manipulate the variables and watch the outcome change, we can be rather confident that those variables are related in some way. However, if we can't, even direct observations can be useful. To state 'it is no longer science' is at best wrong, and at worst can lead us to ignore potentially useful information.

Some observations, when combined, can indeed present the same confidence levels as a controlled experiment, especially if the nature of the variables involved are individually well understood.

What do you all think? Are "uncontrolled" observations equally empirical, equally scientific, compared to controlled experiments? Do those of you working in areas where field observations are common consider yourselves to be scientists, and that you are doing science?

'Empirical' simply means 'observed', so it would be contradictory to say observations weren't empirical in nature. As for asking whether they are 'scientific', again, stating otherwise is irrational and ignorant of the history of science philosophy and practice. As for whether controlled experiments present some objective boundary for something to be considered robust enough to be granted the gold star for science, I'd say that science depends on the ability to discuss every situation on a case-by-case basis.

Athon

Zeuzzz · Apr 11, 2009

Good idea for a thread. Cosmology is, at a stretch, a (sort of) science.

Heres a much read paper by MJ Disney which pretty much sums up the whole issue being brought up in the OP, its on the contrainsts of cosmology which distinguish it from any other science. He comes to the (rather strong) conclusion that it not a science but a priesthood.

http://arxiv.org/PS_cache/astro-ph/pdf/0009/0009020v1.pdf

It is argued that some of the recent claims for cosmology are grossly overblown. Cosmology rests on a very small database: it suffers from many fundamental difficulties as a science (if it is a science at all) whilst observations of distant phenomena are difficult to make and harder to interpret. It is suggested that cosmological inferences should be tentatively made and sceptically received [....]

Table 2
PARTICULAR DIFFICULTIES FOR COSMOLOGY AS A SCIENCE
1. Only one Universe.
2. Universe opaque for 56/60 decades since Planck era.
3. Need to extrapolate physics over huge distances.
4. Need to work with what we can currently detect. [But . . . ]
5. Local background very bright.
6. Distances very hard to determine (standard candles).
7. Observational Selection insidious.
8. Distant galaxies hard to measure and interpret unambiguously.
9. Luminosity Functions unreliable.
10. Geometry, astrophysics and evolution often entangled.
11. Physics of early Universe unknown (and unknowable?)
12. Human time-frame so short compared to cosmic.
13. Origin of inertia.
14. The singularity.

(A) There is only one Universe! At a stroke this removes from our armoury all the statistical tools that have proved indispensable for understanding most of astronomy.
(B) The Universe has been opaque to electromagnetic radiation for all but 4 of the 60 decades of time which stretch between the Plank era (10⁻⁴³ sec) and today (10¹⁷ sec). Since as much interesting physics could have occurred in each logarithmic decade, it seems foolhardy to claim that we will ever know much about the origin of the cosmos, which is lost too far back in the logarithmic mists of Time. Even the Large Hadron Collider will probe the microphysics back only as far as 10⁻¹⁰ secs). [2].
(C) Cosmology requires us to extrapolate what physics we know over huge ranges in space and time, where such extrapolations have rarely, if ever, worked in physics before. Take gravitation for instance.. When we extrapolate the Inverse Square Law. ( - dress it up how you will as G.R.) from the solar system where it was established, out to galaxies and clusters of galaxies, it simply never works. We cover up this scandal by professing to believe in “Dark Matter” - for which as much independent evidence exists as for the Emperor’s New Clothes.
(D) Objects at cosmologically interesting distance are exceedingly faint, small and heavily affected by factors such as redshift-dimming and k-corrections, so it will obviously be very difficult, if not impossible, to extract clear information about geometry, or evolution, or astrophysics - all of which are tangled up together. [..........]

5 THE COSMOLOGIST’S CREDO

The cosmologist, who would also be a scientist, must surely subscribe to at least the following assumptions:

(A) “Speculations are not made which cannot, at least in principle, be compared with observational or experimental data, for tests” [the NON-THEOLOGICAL assumption].
(B) “The portion of the Universe susceptible to observation is representative of the cosmos as a whole”. [The ‘GOOD LUCK’ assumption].
(C) “The Universe was constructed using a significantly lower number of free parameters than the number of clean and independent observations we can make of it”. [The ‘SIMPLICITY’ assumption].
(D) “The Laws of Physics which have significantly controlled the Universe since the beginning are, or can be, known to us from considerations outside cosmology itself i.e. we can somehow know the laws which operated during the 56/60 electromagnetically opaque decades”. [The ‘NONCIRCULARITY’ assumption].
Finally the really wishful cosmologist who believes the final answers are just around the corner must confess to the following extra creed:
(E) “We live in the first human epoch which possesses the technical means to tease out the crucial observations”. (As opposed to Hipparcos and parallax, Helmholz and the age of the Earth,Wegener and palaeomagnetic drift) [The ‘FORTUNATE EPOCH’ assumption.]

I can see very little evidence to support any of the last 4 assumptions while it is dismaying to find that some cosmologists, who would like to think of themselves as scientific, are quite willing to abrogate the first. [.........]

There are eight main sections;
THE OBSERVATIONS WHICH BEAR ON COSMOLOGY
THE SPECIFIC DIFFICULTIES OF COSMOLOGY
PARTICULAR DIFFICULTIES FOR COSMOLOGY AS A SCIENCE
THEORY AND OBSERVATIONS (WHAT WE DON’T KNOW ABOUT GALAXIES)
THE COSMOLOGIST’S CREDO
THE PATHOLOGIES OF COSMOLOGY
SOME HISTORICAL MISTAKES IN COSMOLOGY
COSMOLOGY IN PERSPECTIVE

I highly recommend reading it, its a good read.

He ends on a rather blunt note.

COSMOLOGY IN PERSPECTIVE

Of course we would all love to know of the fate of the Universe, just as we’d love to know if God exists. If we expect science to provide the answers though, we may have to be very patient - and literally wait for eternity. Alas professional cosmologists cannot afford to wait that long. For that reason the word ‘cosmologist’ should be expunged from the scientific dictionary and returned to the priesthood where it properly belongs.

I’m not suggesting that cosmology itself should be abandoned. Mostly by accident it has made some fascinating, if faltering progress over the centuries. And if we are patient and build our instruments to explore the Universe in all the crevices of parameter space, new clues will surely come to hand, as they have in the past, largely by accident. But we should not spend too many of our astronomical resources in trying to answer grandiose questions which may, in all probability, be unanswerable. For instance we must not build the Next Generation Space Telescope as if it was solely a cosmological machine. We should only do that if we are confident of converging on “the truth”. If we build it to look through many windows we may yet find the surprising clues which lead us off on a new path along the way. Above all we must not overclaim for this fascinating subject which, it can be argued, is not a proper science at all. Rutherford for instance said “Don’t let me hear anyone use the word ‘Universe’ in my department”.

Shouldn’t we, scientists, be saying something like this to the general public:

“It is not likely that we primates gazing through bits of glass for a century or two will dissemble the architecture and history of infinity. But if we don’t try we won’t get anywhere. Therefore we professionals do the best we can to fit the odd clues we have into some kind of plausible story. That is how science works, and that is the spirit in which our cosmological speculations should be treated. Don’t be impressed by our complex machines or our arcane mathematics. They have been used to build plausible cosmic stories before - which we had to discard afterwards in the face of improving vidence. The likelihood must be that such revisions will have to occur again and again and again.”

I apologise for such a highly opinionated attack, but it does appear to me that the pendulum has swung much too far the other way. Surely the ‘burden of proof’ ought to rest squarely on the proponents of what will always be a fascinating but suspect subject.

I think he makes his point quite well. Cosmology is different from any other science in that it offers no control experiment and is based on a very long list of ambiguities and assumptions, so it should never be considered as definitive as other sciences. Thus by default alternative explanations that seek to explain cosmological data should be given more leeway and more chance before being dismissed out of hand. Some of the members of the aforementioned cosmological preisthood may be claiming that

sol said:
Cosmology is in its golden age - tons of extremely accurate data, rapid narrowing of the cosmological parameter space... for the first time in history there really is a precise standard model of cosmology, and it fits observations very well.

But the history of science tells us something quite different. We would be fools to think we have it worked all out and are merely tweaking the odd parameter here and there, especially in an area such as cosmology, where more observational constraints and problems with interpretting data lie than any other area of science. A new set of data may come along with makes the parameters we've spent years tweaking completely irrelivant anyway.

So is cosmology a science? Yes, I wont go as far as Disney or his counterparts, it just suffers terribly restrictive contraints which make any defintive models highly dubious, and means that alternatives should be considered (and funded) on equal grounds before jumping to presumptuos conclusions.

Some words from the esteemeed cosmologist Richard Lieu may go well to end here:

http://arxiv.org/PS_cache/arxiv/pdf/0705/0705.2462v1.pdf

Astronomy can never be a hard core physics discipline, because the Universe offers no control experiment, i.e. with no independent checks it is bound to be highly ambiguous and degenerate. Thus e.g. while superluminal motion can be explained by Special Relativity. data on the former can never on their own be used to establish the latter. This is why traditionally astrophysicists have been content with (and proud of) their ability to use known physical laws and processes established in the laboratory to explain celestial phenomena. Cosmology is not even astrophysics: all the principal assumptions in this eld are unveried (or unveriable) in the laboratory, and researchers are quite comfortable with inventing unknowns to explain the unknown. How then could, after fty years of failed attempt in nding dark matter, the elds of dark matter and now dark energy have become such lofty priorities in astronomy funding, to the detriment of all other branches of astronomy? I demonstrate in this article that while some of is based upon truth, at least just as much of CDM cosmology has been propped by a paralyzing amount of propaganda which suppress counter evidence and sub- due competing models. The recent WMAP3 paper of Spergel et al (2007) will be used as case in point on selective citation. I also show that when all evidence are taken into account, two of the competing models that abolish dark energy and/or dark matter do not trail behind CDM by much. Given all of the above, I believe astronomy is no longer heading towards a healthy future, unless funding agencies re-think their master plans by backing away from such high a emphasis on groping in the dark.

Dancing David · Apr 11, 2009

Zeuzzz said:
I think he makes his point quite well. Cosmology is different from any other science in that it offers no control experiment and is based on a very long list of ambiguities and assumptions, so it should never be considered as definitive as other sciences.

Nope, there are controls present and available. You don't like that, because your model just don't exist or don't work and you just want your ideas to be right.

Present the
Model,
Show the
Predictions and
How they match the
Data,

Your Steady State/Not the Big bang theory has been trashed on a number of observations.

You don't like that so you resort to silly philosophy.

Thus by default alternative explanations that seek to explain cosmological data should be given more leeway and more chance before being dismissed out of hand.

pretend they are dismissed out of hand Zeuzzz, that is all it is, you pret3ending.

They are ruled out by the data, Complaining doesn't change that.

The abundance of metallics in the Universe says that it is not any older than suggested by the BBT.

You don't like that so you use some silly philosophical point to avoid the
Fact
that your model (it is eternal and not the BBT) doesn't fit the
data.

Some of the members of the aforementioned cosmological preisthood may be claiming that
But the history of science tells us something quite different. We would be fools to think we have it worked all out and are merely tweaking the odd parameter here and there, especially in an area such as cosmology, where more observational constraints and problems with interpretting data lie than any other area of science.

Uh, huh.

Sure, your models fail to match the data Zeuzzz, not because of observational constraints.

But because they don't match the data.

Sure make up excuses for why those PC theories/It's Not the BBE , can't make predictions that match the data.

You have yet to present one that does match the data.

OOOOPS.

A new set of data may come along with makes the parameters we've spent years tweaking completely irrelivant anyway.

Yeah, so, ever heard of science?

that is the way it works.

You can't just sit there and spin moonbeams and say, well maybe data will come along that proves my moon beam theory.

So is cosmology a science? Yes, I wont go as far as Disney or his counterparts, it just suffers terribly restrictive contraints which make any defintive models highly dubious,

this just shows you have some bad idea about what science is Zeuzzz, try reading about Gell-Mann and Feynman.

Science is about approximate models , it is never definitive!

You have a model that approximates reality.

Nothing is ever final, it is always a model.

and means that alternatives should be considered (and funded) on equal grounds before jumping to presumptuos conclusions.

Oh right, show us a moon beam theory that doesn’t fail to meet the data.

Your explanation of why it isn't cosmological redshift, don't match the data.
Your theories of why there was no BBE, don't match the data. Your theories of electric stars, don't match the data.

You suffer from thinking that moon beam theories should be funded.

Yeah right Zeuzzz, I challenge you to present one model, one set of predictions and one set of data that work.

I asked you to do that two or three years ago.

The problem is NOT that these ideas are pooh-poohed out of hand. they are looked at and they don't work.

So stop pretending try putting some meat to your moon beam theories.

Some words from the esteemeed cosmologist Richard Lieu may go well to end here:

http://arxiv.org/PS_cache/arxiv/pdf/0705/0705.2462v1.pdf

This is the [problem Zeuzzz,

All thought, ideas and science are just models of approximate behavior.

The merit of an idea, thought and theory is in its ability to model the behavior of reality, make predictions and match observation.

Hiding behind these cute ideas of ‘observational constraints’ does not mean that a theory a priori has merit.

Show us a theory that
1. Has a model
2. Makes predictions
3. Matches observations

You keep pretending that you have done these things, when you haven’t.

edd · Apr 11, 2009

Zeuzzz said:
Some words from the esteemeed cosmologist Richard Lieu may go well to end here:

http://arxiv.org/PS_cache/arxiv/pdf/0705/0705.2462v1.pdf

The paper makes interesting points but is rather overblown. The objections to the CMB section is mildly concerning, from the point of view that someone would raise many of these as serious objections. Take:
" It comes from a redshift of 1,000. How do you know the redshift of the CMB? We do not have any characteristic emission or absorption line - there is not even a straw to clutch."
When I was doing my PhD I asked precisely this question, and got a very reasonable answer with noone trying to overstate the case. It certainly wasn't a "how dare you?" question either. We were positively encouraged to ask questions like this, and questions like how do we know the CMB is a background source?

And, at the end of the day there's lots of pointing at cosmology not being "laboratory verifiable" but without offering alternative methods of explaining the observations we have.

There's essentially a lot of criticism of LCDM and no reasonable alternative offered.

I'm also rather amused that much of the criticism one might level against LCDM requires a deeper understanding of what one might consider "messy" astrophysics, and more funding going towards more traditional areas of astronomy research than theoretical studies or CMB experiments.

Curiously, I'll also note that one of the "alternative models" has as a coauthor someone who did encourage exactly the questions I mentioned above, and who I highly respect for that and his deeply intelligent approach to such matters. It's rather peculiar to me to find his work somehow framed as repressed material struggling to reach the surface against the oppression of a standard LCDM model. Alternatives do get heavily discussed and we're all ready to jump on better models and theories that do a better job of explaining things than LCDM, which is certainly not a sufficiently explanatory theory for most people's tastes, at least on the fundamental level that there's no explanation as to what the L is yet. Not entirely dissimilar feelings of surprise come at the discussion of other names on that list that I've met, who I'd consider pretty mainstream - we're really not all as thoroughly sold on LCDM being the definite answer as you might like to have everyone believe.

Finally, I'd note that if there's a problem with the amount of money going into cosmology then this is really something to be raised with the funding agencies rather than something to be blamed on scientists successfully applying for that funding. How you divide finite funds across different areas of science (whether within a field or between two very different fields like biology and astronomy) is an endlessly complicated question, and one that needs a higher level discussion than merely discussing any current unanswered questions in cosmology.

That was rather rambly, but my experiences of the field seem to be rather different than those of Lieu's in accommodating more traditional astrophysics as a partner with cosmology rather than a competitor, and with alternatives to the current best model continuing to be studied - I'm a great believer that science has to be a competition between models rather than simply a case of saying "Oh look, that fits, it's good enough".

Zeuzzz · Apr 11, 2009

Dancing David said:
Nope, there are controls present and available. You don't like that, because your model just don't exist or don't work and you just want your ideas to be right.

Present the
Model,
Show the
Predictions and
How they match the
Data,

Your Steady State/Not the Big bang theory has been trashed on a number of observations.

You don't like that so you resort to silly philosophy.

pretend they are dismissed out of hand Zeuzzz, that is all it is, you pret3ending.

They are ruled out by the data, Complaining doesn't change that.

The abundance of metallics in the Universe says that it is not any older than suggested by the BBT.

You don't like that so you use some silly philosophical point to avoid the
Fact
that your model (it is eternal and not the BBT) doesn't fit the
data.

Uh, huh.

Sure, your models fail to match the data Zeuzzz, not because of observational constraints.

But because they don't match the data.

Sure make up excuses for why those PC theories/It's Not the BBE , can't make predictions that match the data.

You have yet to present one that does match the data.

OOOOPS.

Yeah, so, ever heard of science?

that is the way it works.

You can't just sit there and spin moonbeams and say, well maybe data will come along that proves my moon beam theory.

this just shows you have some bad idea about what science is Zeuzzz, try reading about Gell-Mann and Feynman.

Science is about approximate models , it is never definitive!

You have a model that approximates reality.

Nothing is ever final, it is always a model.

Oh right, show us a moon beam theory that doesn’t fail to meet the data.

Your explanation of why it isn't cosmological redshift, don't match the data.
Your theories of why there was no BBE, don't match the data. Your theories of electric stars, don't match the data.

You suffer from thinking that moon beam theories should be funded.

Yeah right Zeuzzz, I challenge you to present one model, one set of predictions and one set of data that work.

I asked you to do that two or three years ago.

The problem is NOT that these ideas are pooh-poohed out of hand. they are looked at and they don't work.

So stop pretending try putting some meat to your moon beam theories.

This is the [problem Zeuzzz,

All thought, ideas and science are just models of approximate behavior.

The merit of an idea, thought and theory is in its ability to model the behavior of reality, make predictions and match observation.

Hiding behind these cute ideas of ‘observational constraints’ does not mean that a theory a priori has merit.

Show us a theory that
1. Has a model
2. Makes predictions
3. Matches observations

You keep pretending that you have done these things, when you haven’t.

David, you appear to be lost. This thread is about the constraints on cosmology and astronomy as a science. Thats EVERY cosmology. Whether BBT or ZHRYFV theory. Here is the thread used for discussing alternative theories to the Big Bang and attacking me: Plasma Cosmology - Woo or not . And my recent critisism about funding (or lack thereof) of alternatives can be seen here in that thread (since this appears to be what set you off): http://www.internationalskeptics.co...ight=telescope+allocation+comitee#post4591550

Please migrate your time to that thread instead instead of derailing this one. Infact repost your above reply there, and I'll pick it apart, piece by piece.

This thread is about the constraints on cosmology as a science, which I responded to by linking to a paper which addresses exactly these differences which separate it from most traditional science, with the section "THE OBSERVATIONS WHICH BEAR ON COSMOLOGY" being the most prudent. Thus answering the specific question in the OP in how cosmology differs from other areas of science, and what constraints it suffers. These constraints apply to any and every cosmological model. I also linked to Lieu's paper which is again directly related to the OP, showing the lack of controls in astronomy and how it differs from other sciences. The conclusion of which was that definitive statements about the state of the universe should always be avoided, as it turns out we very rarely have it all solved and sorted out, and nothing can be directly proved in situ, so all tenable options should be considered. I did not bring up anything that you are talking about. Alternative cosmologies or Plasma cosmology have nothing to do with this thread, unless Tim wants to spend ages debating them here too (which I'm sure he doesn't).

If you want to discuss the Disney paper, or the Leiu paper, or the constraints and problems with cosmology (any cosmology) as a science, then go ahead.

Meanwhile, a couple of other papers that I think would be a good addition to this threads contents are:

http://arxiv.org/PS_cache/astro-ph/pdf/0310/0310214v2.pdf

Mart´ın L´opez-Corredoira
Astronomisches Institut der Universit¨at Basel
Venusstrasse 7. CH-4102-Binningen (Switzerland)
February 2, 2008

ABSTRACT

I will review some results of observational cosmology which critically cast doubt upon the foundations of the standard cosmology: 1) The redshifts of the galaxies are due to the expansion of the Universe; 2) The cosmic microwave background radiation and its anisotropies come from the high energy primordial Universe; 3) The abundance pattern of the light elements is to be explained in terms of the primordial nucleosynthesis; 4) The formation and evolution of galaxies can only be explained in terms of gravitation in the cold dark matter theory of an expanding Universe. The review does not pretend to argue against this standard scenario in favour of an alternative theory, but to claim that cosmology is still a very young science and should leave the door wide open to other positions.

That bit in bold was the point I was trying to make, David. This is nothing to do with alternatives, just the considerable contraints cosmology suffers as a scientific discipline compared to most sciences.

This is what makes cosmology so different to other areas of science, it can never be definitive about things that happened so long ago and so far away and with such limited data (well, a lot of data, but with an extremely limited scope).

Heres another good one, with a slight philosophical overtone. If you're going to try to unravel all of the secrets and origin of the entire universe a bit of philosphy is unavoidable:

http://arxiv.org/ftp/arxiv/papers/0709/0709.3191.pdf

SOME EPISTEMIC QUESTIONS OF COSMOLOGY
PETAR V. GRUJIĆ
Institute of Physics, PO Box 57, 11080 Belgrade, Serbia

ABSTRACT. We discuss a number of fundamental aspects of modern cosmological concepts, from the phenomenological, observational, theoretical and epistemic points of view. We argue that the modern cosmology, despite a great advent, in particular in the observational sector, is yet to solve important problems, posed already by the classical times. In particular the stress is put on discerning the scientific features of modern cosmological paradigms from the more speculative ones, with the latter immersed in some aspects deeply into mythological world picture. We finally discuss the principal paradigms, which are present in the modern cosmological studies and evaluate their epistemic merits.

KEY WORDS: cosmology, epistemology, methodology, mythology, philosophy of science

DeiRenDopa · Apr 11, 2009

Zeuzzz said:
Good idea for a thread. Cosmology is, at a stretch, a (sort of) science.

Heres a much read paper by MJ Disney which pretty much sums up the whole issue being brought up in the OP, its on the contrainsts of cosmology which distinguish it from any other science. He comes to the (rather strong) conclusion that it not a science but a priesthood.

http://arxiv.org/PS_cache/astro-ph/pdf/0009/0009020v1.pdf

There are eight main sections;
THE OBSERVATIONS WHICH BEAR ON COSMOLOGY
THE SPECIFIC DIFFICULTIES OF COSMOLOGY
PARTICULAR DIFFICULTIES FOR COSMOLOGY AS A SCIENCE
THEORY AND OBSERVATIONS (WHAT WE DON’T KNOW ABOUT GALAXIES)
THE COSMOLOGIST’S CREDO
THE PATHOLOGIES OF COSMOLOGY
SOME HISTORICAL MISTAKES IN COSMOLOGY
COSMOLOGY IN PERSPECTIVE

I highly recommend reading it, its a good read.

He ends on a rather blunt note.

I think he makes his point quite well. Cosmology is different from any other science in that it offers no control experiment and is based on a very long list of ambiguities and assumptions, so it should never be considered as definitive as other sciences. Thus by default alternative explanations that seek to explain cosmological data should be given more leeway and more chance before being dismissed out of hand. Some of the members of the aforementioned cosmological preisthood may be claiming that
But the history of science tells us something quite different. We would be fools to think we have it worked all out and are merely tweaking the odd parameter here and there, especially in an area such as cosmology, where more observational constraints and problems with interpretting data lie than any other area of science. A new set of data may come along with makes the parameters we've spent years tweaking completely irrelivant anyway.

So is cosmology a science? Yes, I wont go as far as Disney or his counterparts, it just suffers terribly restrictive contraints which make any defintive models highly dubious, and means that alternatives should be considered (and funded) on equal grounds before jumping to presumptuos conclusions.

Some words from the esteemeed cosmologist Richard Lieu may go well to end here:

http://arxiv.org/PS_cache/arxiv/pdf/0705/0705.2462v1.pdf

You're very fond of this, and another, Disney article, aren't you Z?

I mean, you brought it up here, here (a whole thread devoted to it!), and here (another thread!) ... and maybe elsewhere too.

Two responses, in the last thread, are interesting:

sol invictus (extract) said:
Unfortunately your opinion is utterly worthless, because you don't understand even basic physics, let alone cosmology, you're a proven liar, and you seem to subscribe to every bit of vaguely scientific sounding woo that comes your way so long as it's wrong.

And moreover you did a terrible job (if it was you) of cherrypicking those references - most of them are further evidence for the standard picture, and the rest are mostly fringe nonsense.

(source)

and:

Reality Check (extract) said:
This paper by MJ Disney has already been dicussed in this forum, e.g. starting with this post in Big Bang - Woo or Not. It is a list of difficulties with cosmology in general (not only Big Bang cosmology). the list is a bit mixed - some points are obvious, some well known, others have been overtaken by time (8 years is a short time in astomony these days!).

(source)

Perhaps there's a key aspect, pertinent to this thread, that has not already been covered?

On a side note, I see that you frequently try to bolster your points by an appeal to authority (e.g. "the esteemeed cosmologist", despite the fact that you can get quite, er, vehement if anyone else tries to even hint at the fact that plenty of others - authorities - take a dim view indeed of your pet ideas. How does the English idiom go, the one about sauce, the goose, and the gander?

Zeuzzz · Apr 11, 2009

DRD, pay special attention to what I said above to david, as you are guilty of doing the exact same thing. Its kind of funny really, when I post valid material about the subject at hand that you cant debunk or refute, you just start attacking me, my 'posting habits'. Note the severe lack of scientific material in your above post.

Saying to a person just about to (hopefully) achieve his degree in physics in the coming months that "you don't understand even basic physics, let alone cosmology, and you're a proven liar" isn't very nice.

And yes, I am fond of that paper, as it makes some extremely pertinent points about how ambiguous some aspects of cosmology are compared to other areas of science, showing that people like you that get so wound up about them need not do so, as nothing in cosmology is set in stone and it suffers numerous problems as a scientific discipline that make it an amazingly unique scientific enterprise in comparison to most other sciences.

Sorry Tim. I tried to answer your question with numerous papers I think address your question directly, but since I seem to have posted highly relevent material answering your questions people seem to have got annoyed at me. I resign from this thread. And from the forum. Now I've gone maybe this thread can be filled with the science underlying the publications I posted and not Ad Homs, accusations and attacks.

Tim Thompson · Apr 11, 2009

Cosmology & Science & etc.

Zeuzzz said:
This thread is about the constraints on cosmology and astronomy as a science.

Certainly astronomy and cosmology are the examples I gave, if only because I am familiar with those disciplines. But I did intend a rather broader scope. As I said in my opening post, "This impacts a wide range of fields of science where "controlled experiments" are usually impossible.", and "Are naturalists in the field, observing chimpanzees in the wild, for instance, doing science?". I think that the hard core demand that science is characterized only by "controlled experiments", and that mere "observations" are not scientific is not reasonable.

Zeuzzz said:
Heres a much read paper by MJ Disney which pretty much sums up the whole issue being brought up in the OP, its on the contrainsts of cosmology which distinguish it from any other science. He comes to the (rather strong) conclusion that it not a science but a priesthood.

Indeed he does, but you post only Disney's paper without reference to the other side of the argument. Consider, for instance, Turner (2001) who says in his paper: "As an extreme example, consider Disney (2000), who defines cosmology as a quest to understand the entire space-time history of the universe (in response to Disney see Peebles 2000). Not surprisingly, he concludes that the achievements of cosmologists have been minimal and cosmology may not be a science at all. While more than two decades ago the relativist Ellis (1975) educated us all on the impossibility of Disney's goal - we are absolutely limited in our knowledge of the universe by our past light cone - that has not prevented significant progress toward understanding how the basic features of our portion of the universe came about as well as their implications for the universe as a whole." So if one is to believe Turner, Disney simply re-defined cosmology to fit some idea of his own, and then argued against that instead of arguing against cosmology as it is practiced by cosmologists.

But Peebles (2000) responds more directly and clearly profoundly disagrees with Disney. See section 4 of Peeble's paper, "Is Cosmology a Science?", which comes to detailed conclusions that are exactly the opposite of Disney. Peebles tells us that "... our commonly accepted cosmology did grow by the introduction of hypotheses to fit phenomena.", and this is exactly the point I had tried to make about astronomical observations. They fit the pattern we expect of testing hypotheses; whether or not the test is a "controlled experiment" seems secondary to the question of whether or not the activity constitutes "science" in a reasonable way.

And see section 5, "A Next Generation of Cosmological Tests", where Peebles says: "The rules of evidence in science have evolved to admit quite indirect approaches. The community agrees that the many laboratory tests of quantum mechanics fully validate it as a real and magnificently successful physical science, even though no one has ever seen a state vector in nature." Here we have an example of the controlled experiment leading to an explanation which relies on the assumption of the existence of something (i.e., a state vector) which is not seen in nature. Does that invalidate the controlled experiments because it relies on quantities not previously known in nature (i.e., state vectors)? Are state vectors simply mathematical inventions required to shore up a failed theory (quantum mechanics), or are they just a way for us to understand what we see? What's the essential difference between "inventing" state vectors to make sense of quantum mechanics, and "inventing" dark energy to make sense of cosmology? Aren't they both the same thing? Do we accept quantum mechanics, but then reject cosmology, even though we are doing essentially the same thing in both cases?

No one argues that cosmology is not restricted in what it can achieve. But is any science not restricted? ("But any active physical science is similarly incomplete: each has a well tested center around which is the exciting confusion of ongoing research." - Peebles 2000, page 6).

athon said:
... there is no single consensus on what science is and is not. This is called the demarkation problem in HPS, and has served as fodder for all manner of philosophical debates.

Indeed so. The demarcation of science from non-science, and the demarcation between the "analytic" and "synthetic" approaches are dealt with by Hugh Gauch in his book Scientific Method in Practice (Cambridge University Press, 2003). I think the criticism of the difference between field observations and controlled laboratory experiments are based for the most part on an improperly restricted understanding of "science", in an attempt to force science into a simple recipe.

athon said:
Here is the crux of the matter - you seem to be asking if science is defined by the ability to manipulate variables. I'd say 'absolutely not'. I'd go even further to say that our understanding of the universe would be dramatically impeded if this was the defining feature. ... Some observations, when combined, can indeed present the same confidence levels as a controlled experiment, especially if the nature of the variables involved are individually well understood.

Indeed I think this is the crux. The apparent claim presented by Mozina in my opening post, that uncontrolled observations are too heavily affected by "subjective" interpretation to be scientifically valid cannot be true. If we can say that an essential part of science is the validation of hypotheses, then any activity which does this in a reasonable way must be "scientific" in some broad sense at least.

For me, it's not just about astronomy and cosmology being "restricted" in some way, but whether or not we can say that they are simply not science at all. And by extension, if we demand controlled laboratory experiments in all cases, then a vast collection of natural sciences lose the "science" altogether. I don't think that works.

GeeMack · Apr 11, 2009

Tim Thompson said:
Certainly astronomy and cosmology are the examples I gave, if only because I am familiar with those disciplines. But I did intend a rather broader scope. As I said in my opening post, "This impacts a wide range of fields of science where "controlled experiments" are usually impossible.", and "Are naturalists in the field, observing chimpanzees in the wild, for instance, doing science?". I think that the hard core demand that science is characterized only by "controlled experiments", and that mere "observations" are not scientific is not reasonable.

Good point, Tim. We can force a little humidity to condense into "rain" in laboratory chamber, and making a whirly, tornado-looking wind in a tube isn't so tough. But controlled experiments of this type can only give us the most rudimentary understanding of weather phenomena. And very often those look-alike lab projects are darn near useless for explaining various phenomena. That swirly vortex in a tube is only minimally like a real tornado. The glowing electrons around Birkeland's terrella were, for the most part, quite unrelated to actual activity on the Sun.

We have a pretty sound scientific understanding of solar physics, even without building a scale fusion fueled ball in a weightless vacuum chamber in a lab. We'll certainly continue to improve our knowledge of solar physics, and yet very likely never construct such a model. If we waited until we could create scaleable tornados and thunderstorms in a variable controlled lab, we'd be waiting a very long time before we consider meteorology real science.

Perpetual Student · Apr 11, 2009

Is the theory of evolution empirical? Probably not, according to Mozina. We can only do some limited stuff in a lab to demonstrate natural selection (with flies, bacteria, etc.). Modern genetics gives us some experimental help, but the big picture of evolution requires observations in diverse areas like geology, embryology taxonomy, etc.
Is anyone seriously going to claim that evolution is not a foundational theory in biology? (Creationists aside!)

athon · Apr 12, 2009

Perpetual Student said:
Is the theory of evolution empirical? Probably not, according to Mozina. We can only do some limited stuff in a lab to demonstrate natural selection (with flies, bacteria, etc.). Modern genetics gives us some experimental help, but the big picture of evolution requires observations in diverse areas like geology, embryology taxonomy, etc.
Is anyone seriously going to claim that evolution is not a foundational theory in biology? (Creationists aside!)

Well, theories themselves can't technically be empirical, although they can rely on empirical data. Minor technicality, but 'empirical' simply refers to the nature of the information gathered that goes into the formation of a theory.

In any case, the last attempt to restrict science to only include empiricism was logical positivism. There's a good reason why this didn't survive the 20th century - if we stuck to it as Ernst Mach had outlined it, nobody should have had any confidence in general relativity bearing fruit.

Athon

Tubbythin · Apr 12, 2009

Tim Thompson said:
So what we have here is a contest between "observation" and "controlled observation", which is really what a "controlled experiment" is. Now, Mozina's opinion is explicit enough; he says "Observations are not 'empirical experiments' ". It is not much of a stretch to read this as "observations are not empirical" period.

I think one can become very educated by listening to MM. Why? Because just about everything he says is the exact opposite of reality. So by paying attention and then remembering to revert everything he says one learns quite alot.
I find it particularly bizzare his rant about things not being empricial. When I was an undergrad things that were "empirical" were usually looked down upon. That is to say an "empirical law" was a law which we had no idea of the reasoning behind, only that we could observe it. We could explain the how but not the why... which is deeply unsatisfying for a scientist.
Of course his claims about empiricism are ridiculous, what he's advocating, though he is apparently completely incapable of understanding this, is an abandonment of most of the last half a millenium of scientific progress.
To be honest, I'm not that sure how an astronomical "observation" is all that much less scientific than... say... an LHC "experiment". With the former you select suitable detectors and look at the Universe at a a suitable energy (of EM radiation) then filter out all the extraneous crap. With the latter, you select suitable detectors and look at a little region of our beam pipe with particles colliding at a suitable energy then filter out all the extranaeous crap. Ad there's a lot of it too - gigabytes a second. How is one more scientific than the other?

Tubbythin · Apr 12, 2009

PARTICULAR DIFFICULTIES FOR COSMOLOGY AS A SCIENCE
1. Only one Universe.

Granted but... humans have only evolved once. So does that make the study of human evoltion not science?

2. Universe opaque for 56/60 decades since Planck era.

Its a pretty bizzarre way of putting things... bearing in mind the Universe has only been opaque for ~ 300,000/13.7 billion of its lifetime. Not to mention the fact that we can study earlier times through, eg the abundance of the light elements and extrapolate through eg studies of quark gluon plasmas.

3. Need to extrapolate physics over huge distances.

The only real problm I can see is that we need to assume the laws of physics are independent of position and time. Since nobody has ever shown otherwise it seems a reasonable assumption for the time being.

4. Need to work with what we can currently detect. [But . . . ]

Well duh. The same is true for... erm... everything.

5. Local background very bright.

Physics is often about finding needles in haystacks Nuclear and particle physics invariably involves trying to find a tiny signal in a backgorund of boring, well known stuff.

6. Distances very hard to determine (standard candles).

Perhaps. But its at least reassuring that our cosmic distance ladder is consistent - where the realm ofany two candles overlaps, the distances agree in general.

7. Observational Selection insidious.

I have no idea what that is meant to mean.

8. Distant galaxies hard to measure and interpret unambiguously.

Vague, bordering on meaningless sentence.

9. Luminosity Functions unreliable.

Perhaps.

10. Geometry, astrophysics and evolution often entangled.

Evolution of what exactly?

11. Physics of early Universe unknown (and unknowable?)

We've reached temperatures corresponding to some small fraction of a second after the Big Bang. Cosmologists are often quick to point out thelimitations of extapolating prior to this.

12. Human time-frame so short compared to cosmic.

So? Human time fram is very long compared to the atomic. Doesn't mean we cann't do atomic phyics. Human time frame is pretty short compared to geological time too. Is geology unscientific?

13. Origin of inertia.

Inertia of what?

14. The singularity.

See 11.

Dancing David · Apr 12, 2009

Zeuzzz said:
David, you appear to be lost. This thread is about the constraints on cosmology and astronomy as a science. Thats EVERY cosmology. Whether BBT or ZHRYFV theory. Here is the thread used for discussing alternative theories to the Big Bang and attacking me: Plasma Cosmology - Woo or not . And my recent critisism about funding (or lack thereof) of alternatives can be seen here in that thread (since this appears to be what set you off): http://www.internationalskeptics.co...ight=telescope+allocation+comitee#post4591550

Please migrate your time to that thread instead instead of derailing this one. Infact repost your above reply there, and I'll pick it apart, piece by piece.

This thread is about the constraints on cosmology as a science, which I responded to by linking to a paper which addresses exactly these differences which separate it from most traditional science, with the section "THE OBSERVATIONS WHICH BEAR ON COSMOLOGY" being the most prudent. Thus answering the specific question in the OP in how cosmology differs from other areas of science, and what constraints it suffers. These constraints apply to any and every cosmological model. I also linked to Lieu's paper which is again directly related to the OP, showing the lack of controls in astronomy and how it differs from other sciences. The conclusion of which was that definitive statements about the state of the universe should always be avoided, as it turns out we very rarely have it all solved and sorted out, and nothing can be directly proved in situ, so all tenable options should be considered. I did not bring up anything that you are talking about. Alternative cosmologies or Plasma cosmology have nothing to do with this thread, unless Tim wants to spend ages debating them here too (which I'm sure he doesn't).

If you want to discuss the Disney paper, or the Leiu paper, or the constraints and problems with cosmology (any cosmology) as a science, then go ahead.

Meanwhile, a couple of other papers that I think would be a good addition to this threads contents are:

http://arxiv.org/PS_cache/astro-ph/pdf/0310/0310214v2.pdf

That bit in bold was the point I was trying to make, David. This is nothing to do with alternatives, just the considerable contraints cosmology suffers as a scientific discipline compared to most sciences.

This is what makes cosmology so different to other areas of science, it can never be definitive about things that happened so long ago and so far away and with such limited data (well, a lot of data, but with an extremely limited scope).

Heres another good one, with a slight philosophical overtone. If you're going to try to unravel all of the secrets and origin of the entire universe a bit of philosphy is unavoidable:

http://arxiv.org/ftp/arxiv/papers/0709/0709.3191.pdf

Thanks Zeuzzz, I will respond later as my brain is reaching it's limit for the nonce.

Dancing David · Apr 12, 2009

Zeuzzz said:
Good idea for a thread. Cosmology is, at a stretch, a (sort of) science.

Heres a much read paper by MJ Disney which pretty much sums up the whole issue being brought up in the OP, its on the contrainsts of cosmology which distinguish it from any other science. He comes to the (rather strong) conclusion that it not a science but a priesthood.

http://arxiv.org/PS_cache/astro-ph/pdf/0009/0009020v1.pdf

There are eight main sections;
THE OBSERVATIONS WHICH BEAR ON COSMOLOGY
THE SPECIFIC DIFFICULTIES OF COSMOLOGY
PARTICULAR DIFFICULTIES FOR COSMOLOGY AS A SCIENCE
THEORY AND OBSERVATIONS (WHAT WE DON’T KNOW ABOUT GALAXIES)
THE COSMOLOGIST’S CREDO
THE PATHOLOGIES OF COSMOLOGY
SOME HISTORICAL MISTAKES IN COSMOLOGY
COSMOLOGY IN PERSPECTIVE

I highly recommend reading it, its a good read.

He ends on a rather blunt note.

I think he makes his point quite well. Cosmology is different from any other science in that it offers no control experiment and is based on a very long list of ambiguities and assumptions, so it should never be considered as definitive as other sciences. Thus by default alternative explanations that seek to explain cosmological data should be given more leeway and more chance before being dismissed out of hand. Some of the members of the aforementioned cosmological preisthood may be claiming that
But the history of science tells us something quite different. We would be fools to think we have it worked all out and are merely tweaking the odd parameter here and there, especially in an area such as cosmology, where more observational constraints and problems with interpretting data lie than any other area of science. A new set of data may come along with makes the parameters we've spent years tweaking completely irrelivant anyway.

So is cosmology a science? Yes, I wont go as far as Disney or his counterparts, it just suffers terribly restrictive contraints which make any defintive models highly dubious, and means that alternatives should be considered (and funded) on equal grounds before jumping to presumptuos conclusions.

Some words from the esteemeed cosmologist Richard Lieu may go well to end here:

http://arxiv.org/PS_cache/arxiv/pdf/0705/0705.2462v1.pdf

I know you are a post and tun person Zeuzzz, but the part in bold is exactly what I was talking about.

Moonbeam theries do not get priority just because you like moonbeams.

The rest of it is calling you on the fact that you pretend to have a coherent set of theories and ideas that you never subtatiate.

Skwinty · Apr 12, 2009

Dancing David said:
I know you are a post and tun person Zeuzzz, but the part in bold is exactly what I was talking about.

Moonbeam theries do not get priority just because you like moonbeams.

The rest of it is calling you on the fact that you pretend to have a coherent set of theories and ideas that you never subtatiate.

Dancing David
Get with the program
:h2:

The horse has bolted
http://www.internationalskeptics.com/forums/showthread.php?t=140029

Tubbythin · Apr 13, 2009

Skwinty said:
Dancing David
Get with the program
The horse has bolted
http://www.internationalskeptics.com/forums/showthread.php?t=140029

That's dissapointing.

sol invictus · Apr 13, 2009

Tubbythin said:
That's dissapointing.

I don't find it so. This thread is a perfect example - Tim asked an interesting and reasonable question, and Zeuzzz smothered it in a mountain of boring woo.

As for the OP:

first, the suggestion that observations are not empirical is false by definition of "empirical".

second, the idea that there is a significant difference between, say, astrophysics and any other kind of science is easily disposed of. Take particle physics as an example. I've certainly never heard anyone suggest that isn't science - in fact by some measures it's the most successful in the history of the human race.

And yet in particle physics, there are no controls in the usual sense. Why? Because it tests the fundamental laws of nature, and we only live in one universe with one set of laws. You cannot smash two protons together in our world, and then again in another world with simpler laws as a control. The closest thing to that - and the way it's actually done - is to run a simulation of the collision on a computer, using the set of currently known laws, and compare that to the observed result. A significant difference is a discovery.

Astrophysics is methodologically indistinguishable. Colliders can run for longer and record more collisioons. Telescopes can collect more light. Larger colliders can be built. Larger and more sensitive telescopes can be built. We can build lepton or photon or hadron colliders. We can build visible light or radio or cosmic ray telescopes. Etc., etc.

In the end, the only real difference is that some quantities in astrophysics are fundamentally limited in the quantity of data accessible to us. The large scale power spectrum of the cosmic microwave background is a prime example - we already have all the information about that we ever will. But on the other hand all branches of science face such limitations in practice, and like them astrophysics can overcome these difficulties with new ideas and measurement techniques that open new opportunities for observation.

Dancing David · Apr 13, 2009

Skwinty said:
Dancing David
Get with the program
The horse has bolted
http://www.internationalskeptics.com/forums/showthread.php?t=140029

I believe he will be back, I recall similar stances before.

I saw the thread in FM.

Tubbythin · Apr 13, 2009

sol invictus said:
I don't find it so. This thread is a perfect example - Tim asked an interesting and reasonable question, and Zeuzzz smothered it in a mountain of boring woo.

Granted. But the number of times someone somewhere comes up with something fascintating in response to one of Z's posts makes me think his presence might be missed. Also... I had some unaswered questons.

As for the OP:

first, the suggestion that observations are not empirical is false by definition of "empirical".

second, the idea that there is a significant difference between, say, astrophysics and any other kind of science is easily disposed of. Take particle physics as an example. I've certainly never heard anyone suggest that isn't science - in fact by some measures it's the most successful in the history of the human race.

And yet in particle physics, there are no controls in the usual sense. Why? Because it tests the fundamental laws of nature, and we only live in one universe with one set of laws. You cannot smash two protons together in our world, and then again in another world with simpler laws as a control. The closest thing to that - and the way it's actually done - is to run a simulation of the collision on a computer, using the set of currently known laws, and compare that to the observed result. A significant difference is a discovery.

Astrophysics is methodologically indistinguishable. Colliders can run for longer and record more collisioons. Telescopes can collect more light. Larger colliders can be built. Larger and more sensitive telescopes can be built. We can build lepton or photon or hadron colliders. We can build visible light or radio or cosmic ray telescopes. Etc., etc.

In the end, the only real difference is that some quantities in astrophysics are fundamentally limited in the quantity of data accessible to us. The large scale power spectrum of the cosmic microwave background is a prime example - we already have all the information about that we ever will. But on the other hand all branches of science face such limitations in practice, and like them astrophysics can overcome these difficulties with new ideas and measurement techniques that open new opportunities for observation.

That's pretty much what I was trying to say. But rather more eloquently put.

DeiRenDopa · Apr 15, 2009

I'm rather late to this thread, but I hope people are still reading it ...

Here's a somewhat different perspective on 'observations', one that I raised with MM in that thread; I'd be interested in what others think.

There is a sense in which (almost?) every modern astronomical 'observation' is a 'controlled experiment'; namely, the 'data' which ends up being presented in a published paper is the result of a great deal of 'data reduction' and 'analysis', and nearly all of this is 'controlled experiment'.

The design and construction of the detectors, instruments, systems etc; their testing and calibration; the conversion of detector 'readings' to digital form, their transmission to remote computers (esp in the case of space-based instruments); the specialist software packages used to 'reduce' the data; ... all are vital to the end result (the astronomical observation), and all rest the collective application of a great deal of physics.

Another different perspective on 'observations': if cosmology is the (scientific) study of the observable universe, then its scope is quite strongly constrained by the physics of the time in which it is conducted.

Take Olbers' paradox, based on one of the oldest observations (that the night sky is dark) ... the extent to which hypotheses concerning the resolution of this paradox can be tested depends heavily on how strongly we can constrain 'dark' (what level of light can we detect, how confident we can be about the sources of light we do detect in the night sky, etc). When physics developed sufficiently that light could be understood to be electromagnetic radiation, then Olbers' paradox could be generalised to all wavebands (and, much later, the fact that the night sky is not at all dark in the microwave band demonstrated).

Today, the observable universe is (almost) empty of neutrinos, from direct observation, but a neutrino analog of the CMB should be there (per contemporary cosmological models). Perhaps, even in the lifetimes of some people reading this post in 2009, that neutrino background may become directly observable ... and cosmology become very different than what it is today ...*

* and if you don't buy this conclusion wrt neutrinos, think of what direct detection and characterisation of dark matter (particles) would do, or dark energy, or ...

Dancing David · Apr 15, 2009

I think people, some people, have an objection to not being able to have independant and dependant variables in a 'controlled' experiement.

But observations are valid as observations, regardless of wether they are in a "controlled' experiment or not.

The approximate model that matches the observations is the best fit to reality. It will change with time.

temporalillusion · Apr 15, 2009

Dancing David said:
I think people, some people, have an objection to not being able to have independant and dependant variables in a 'controlled' experiement.

But observations are valid as observations, regardless of wether they are in a "controlled' experiment or not.

The approximate model that matches the observations is the best fit to reality. It will change with time.

This was MM's beef I think when we went over this in the LCDM thread.

My contention was that you can still get your controls, you just have to be more clever about your observations.

sol invictus · Apr 15, 2009

DeiRenDopa said:
There is a sense in which (almost?) every modern astronomical 'observation' is a 'controlled experiment'; namely, the 'data' which ends up being presented in a published paper is the result of a great deal of 'data reduction' and 'analysis', and nearly all of this is 'controlled experiment'.

The design and construction of the detectors, instruments, systems etc; their testing and calibration; the conversion of detector 'readings' to digital form, their transmission to remote computers (esp in the case of space-based instruments); the specialist software packages used to 'reduce' the data; ... all are vital to the end result (the astronomical observation), and all rest the collective application of a great deal of physics.

In what sense does any of that constitute a control?

I'm used to thinking of controls as another experiment or set of experiments similar to the one you're really interested in, but differing in the aspect you'd like to test. For example, to test the efficacy of drug X in treating condition Y you conduct two otherwise identical trials with a placebo and the drug, and compare the results.

In astrophysics one might want to test whether there is a significant amount of dark matter in (say) the Milky Way. The most direct analog of a control would be to observe a galaxy which is identical to the Milky Way in every way except that it lacks dark matter, and compare that to what we observe in our Milky Way. Since that's obviously impossible, astros use simulations or analytic solutions to theories to predict what the Milky Way would look like if it had no DM, note the discrepancy with observations, and conclude that it must have DM. But that's not an experimental control - it's more like a theoretical one.

Everything I just said - and everything you said in the paragraph I quoted - carries over completely to particle physics, by the way. So if from this one were to conclude that astrophysics is not science, one would be forced to conclude the same about particle physics - an odd conclusion indeed.

DeiRenDopa · Apr 15, 2009

sol invictus said:
[...]

In the end, the only real difference is that some quantities in astrophysics are fundamentally limited in the quantity of data accessible to us. The large scale power spectrum of the cosmic microwave background is a prime example - we already have all the information about that we ever will. But on the other hand all branches of science face such limitations in practice, and like them astrophysics can overcome these difficulties with new ideas and measurement techniques that open new opportunities for observation.

(bold added)

Would you mind expanding on this a bit please?

Do you mean the low-l (and m) modes in the angular power spectrum?

DeiRenDopa · Apr 15, 2009

sol invictus said:
DeiRenDopa said:

There is a sense in which (almost?) every modern astronomical 'observation' is a 'controlled experiment'; namely, the 'data' which ends up being presented in a published paper is the result of a great deal of 'data reduction' and 'analysis', and nearly all of this is 'controlled experiment'.

The design and construction of the detectors, instruments, systems etc; their testing and calibration; the conversion of detector 'readings' to digital form, their transmission to remote computers (esp in the case of space-based instruments); the specialist software packages used to 'reduce' the data; ... all are vital to the end result (the astronomical observation), and all rest the collective application of a great deal of physics.

Click to expand...

In what sense does any of that constitute a control?

I'm used to thinking of controls as another experiment or set of experiments similar to the one you're really interested in, but differing in the aspect you'd like to test. For example, to test the efficacy of drug X in treating condition Y you conduct two otherwise identical trials with a placebo and the drug, and compare the results.

In astrophysics one might want to test whether there is a significant amount of dark matter in (say) the Milky Way. The most direct analog of a control would be to observe a galaxy which is identical to the Milky Way in every way except that it lacks dark matter, and compare that to what we observe in our Milky Way. Since that's obviously impossible, astros use simulations or analytic solutions to theories to predict what the Milky Way would look like if it had no DM, note the discrepancy with observations, and conclude that it must have DM. But that's not an experimental control - it's more like a theoretical one.

Everything I just said - and everything you said in the paragraph I quoted - carries over completely to particle physics, by the way. So if from this one were to conclude that astrophysics is not science, one would be forced to conclude the same about particle physics - an odd conclusion indeed.

In that for the 'processed' data which comes out at the end - that which is found in catalogues and papers* - to have the meaning given to it ('Our CO(2-1) image [of the] multiple protostellar system L1551 IRS5', to take an example from a very recent arXiv preprint) everything that went into the observation, directly or indirectly, must work as expected; the engineering and applied physics must be right, if you will ... so every observation is a controlled experiment of that embedded physics.

There is nothing particularly profound about this, from the perspective of working scientists (not just astronomers), who are (I imagine) for the most part quite comfortable that the instrumentation all, and always, works as it is supposed to. However, from the perspective of HPS, or when trying to understand just how deep the misconceptions or misunderstanding are in an alternative/crackpot perspective, it is. And, of course, it's no different than the two us communicating via keystrokes, a PC, the internet, etc; these too are controlled experiments of the relevant physics (albeit poorly designed, or stated ones, and rather scientifically boring ones too).

* very rarely is the 'raw data' published, though much of it is available to anyone; 'raw data' however is rather a misnomer.

rlr · Apr 16, 2009

Observation is empirical. That's kind of what the word "empirical" means. Originating in, or based on, observation.

There is no rule in science that experiment is necessary. The point of an experiment is to get something to happen that you want to study. If it's happening anyway, then there's no reason not to study the thing in situ. The point of experimental controls is largely to rule out human mistakes, not natural confounds. When the thing being observed cannot possibly be open to human tampering -- such as a supernova -- that issue does not arise.

It seems that there is a great deal of misunderstanding over what science actually is or that there is a monolithic "scientific method" that defines what is or is not science. There can be scientific and unscientific methods, but there is no Scientific Method, singular, capitalized, authoritative.

sol invictus · Apr 16, 2009

DeiRenDopa said:
(bold added)

Would you mind expanding on this a bit please?

Do you mean the low-l (and m) modes in the angular power spectrum?

Yes, that's what I meant. If you're interested in deducing the total (i.e. scalar plus tensor) power spectrum of primordial density perturbations using the CMB power spectrum, you already have all the information there is to have (at least at low l).

DeiRenDopa said:
In that for the 'processed' data which comes out at the end - that which is found in catalogues and papers* - to have the meaning given to it ('Our CO(2-1) image [of the] multiple protostellar system L1551 IRS5', to take an example from a very recent arXiv preprint) everything that went into the observation, directly or indirectly, must work as expected; the engineering and applied physics must be right, if you will ... so every observation is a controlled experiment of that embedded physics.

Ah, OK. But that's not a control of the physics being investigated - it's a control of lots of other physics, none of which is very relevant. That is, it's hard to imagine anyone doubting that the existence of dark matter in independent of whether or not your telescope works, or whether its internet connection is functioning as expected. And anyway controls in the sense you refer to there are present in all experiments - but that's certainly not what people usually mean when they talk about "controlled experiments".

So... do you agree that scientific tests of fundamental questions about the universe cannot be controlled, at least not in the ordinary sense?

DeiRenDopa · Apr 16, 2009

sol invictus said:
[...]

DeiRenDopa said:

In that for the 'processed' data which comes out at the end - that which is found in catalogues and papers* - to have the meaning given to it ('Our CO(2-1) image [of the] multiple protostellar system L1551 IRS5', to take an example from a very recent arXiv preprint) everything that went into the observation, directly or indirectly, must work as expected; the engineering and applied physics must be right, if you will ... so every observation is a controlled experiment of that embedded physics.

Click to expand...

Ah, OK. But that's not a control of the physics being investigated - it's a control of lots of other physics, none of which is very relevant. That is, it's hard to imagine anyone doubting that the existence of dark matter in independent of whether or not your telescope works, or whether its internet connection is functioning as expected. And anyway controls in the sense you refer to there are present in all experiments - but that's certainly not what people usually mean when they talk about "controlled experiments".

Indeed.

The point I'm trying to make is that the implications of the idea in the OP (MM's one, not TT's) are even more far-reaching than those TimT explicitly mentions.

So... do you agree that scientific tests of fundamental questions about the universe cannot be controlled, at least not in the ordinary sense?

I find the distinction that MM seems to be trying to make hard to grasp.

As I see it, you formulate one or more hypotheses, and you go do some tests. Some of the tests may be ones you do in a lab, some in the field (in situ), and some remotely. In all three types of test, you have some control ... but there's nothing fundamentally different between the types due simply to the degree of control you have.

So yes, at least some scientific tests of fundamental questions cannot be controlled (I am wary of absolutes - perhaps there is a scientific test of fundamental questions about the universe which can be controlled, in the ordinary sense, even if neither you nor I can think of it ...).

I'll come back to the CMB later (thanks for your clarification).

DeiRenDopa · Apr 19, 2009

low-l CMB

(continued, as promised)

sol invictus said:
DeiRenDopa said:

Would you mind expanding on this a bit please?

Do you mean the low-l (and m) modes in the angular power spectrum?

Click to expand...

Yes, that's what I meant. If you're interested in deducing the total (i.e. scalar plus tensor) power spectrum of primordial density perturbations using the CMB power spectrum, you already have all the information there is to have (at least at low l).

[...]

I'm interested in exploring this in some more detail ...

... but first, can we agree on what "low l" is?

COBE characterised the monopole extremely well, and I think there's very little scope for nailing it down better ... at least locally; the CMB temperature has been estimated at some distant (time and space) places, via really neat spectroscopy, but with very large error bars, so there is a great deal of scope for improvement there.

The dipole was also characterised well by COBE, and much better by WMAP. However, as Hinshaw et al. and Nolta et al. make clear, it's only nailed down to ~one part per thousand, so there is some scope for improvement.

When it comes to the (other) low-l multipoles up to l = 10, this WMAP5 Nolta et al. figure is telling:

IOW, the lowest l datapoint(s) have the largest error bars (greatest estimated uncertainties).

Furthermore, WMAP is the single source of all significant data in this range ... COBE established the existence of the angular power spectrum, but not its magnitude (other than the dipole), and Earth-based (including balloon) observations are essentially completely insensitive to such large-scale features.

Further, it is at these scales that most of the known foregrounds (dust, galactic synchrotron radiation, free-free radiation^) contribute the greatest hard-to-model power.

Finally, it is also at these scales that the to-date unmodelled foregrounds are most likely to be important^^.

Planck is scheduled to launch next month, and so within five years we will have a completely independent set of inputs which will be analysed to estimate the CMB low-l angular power spectrum.

In conclusion then, at least between the dipole and l ~10, the information we have to date comes from only one mission, and we can reasonably expect that the uncertainties on it, from WMAP, can be substantially reduced.

I'm interested to know if the above is consistent with how you view the current state of play wrt observations of the CMB.

^ unresolved point sources are not particularly important at these scales
^^ of course, if we knew what the unmodelled foregrounds were, we could attempt to model them! FWIW, there are only two that might be potentially important, that I've not seen mentioned in the literature; namely, the heliosheath/heliopause and the Local (ISM) Bubble

edd · Apr 19, 2009

DRD - the problem is not having only one satellite observing it, or having a satellite that has technical limitations on observing the power spectrum at low-l.

The problem is having only one CMB to measure, and at low-l there are not enough "realisations" of the structure on that scale that are observable.

That's why sol invictus says we have all the information we ever will. We'd need entirely separate things to observe, not things to observe with in order to improve our knowledge.

MattusMaximus · Apr 20, 2009

athon said:
Well, theories themselves can't technically be empirical, although they can rely on empirical data. Minor technicality, but 'empirical' simply refers to the nature of the information gathered that goes into the formation of a theory.

In any case, the last attempt to restrict science to only include empiricism was logical positivism. There's a good reason why this didn't survive the 20th century - if we stuck to it as Ernst Mach had outlined it, nobody should have had any confidence in general relativity bearing fruit.

Athon

Good point, Athon. This points out, among other things, the importance of inductive reasoning in science. In case anyone's interested, Steve Novella of the Skeptologists has a good blog post summarizing inductive reasoning in science here.

Fundamental Question on the Nature of Science

Muse

Philosopher

Unregistered

Banned

Penultimate Amazing

Master Poster

Banned

Master Poster

Banned

Muse

Banned

Illuminator

Unregistered

Illuminator

Illuminator

Penultimate Amazing

Penultimate Amazing

Philosopher

Illuminator

Philosopher

Penultimate Amazing

Illuminator

Master Poster

Penultimate Amazing

Technical Admin

Philosopher

Master Poster

Master Poster

Scholar

Philosopher

Master Poster

Master Poster

Master Poster

Intellectual Gladiator