Results 1 to 15 of 15

Thread: The Beginning of the Universe

  1. #1
    Join Date
    Nov 2019
    Posts
    1

    The Beginning of the Universe

    Are inflaton fields and inflation a good enough explanation of why our universe formed? Or is there a better theory? And why?

  2. #2
    Join Date
    Oct 2005
    Posts
    26,823
    Neither inflaton fields nor inflation explains everything about why our universe formed. They help explain how the history may have played out, but they don't explain the beginning of the formation process. This is because the inflaton field cannot do its thing, and inflation cannot occur, without some initial expansion to allow it to undergo the transitions it undergoes. So that initial expansion is not explained by inflation, so the initial formation is also not explained.

  3. #3
    Join Date
    Sep 2003
    Posts
    12,887
    Quote Originally Posted by chiguchi20 View Post
    Are inflaton fields and inflation a good enough explanation of why our universe formed? Or is there a better theory? And why?
    In addition to what Ken said, there are no scientific theories on "why" or, better, "how" our universe formed. Science requires objective-based reasoning to form a theory. An observable universe allows theories; "theories" about an unobservable universe aren't scientific ones. That doesn't mean metaphysics is pointless and may seem reasonable at some point, but a scientific theory is something special and too valuable to allow corruption.
    We know time flies, we just can't see its wings.

  4. #4
    Join Date
    Feb 2010
    Posts
    771
    More: inflation has quite a few challenges, in terms of consistency (though not with observations, at least not yet).

    Why? Check out this 2017 blogpost by Sabine Hossenfelder (she who owns the BackReAction blog): Is the inflationary universe a scientific theory? Not anymore.

  5. #5
    Join Date
    Oct 2005
    Posts
    26,823
    Quote Originally Posted by Jean Tate View Post
    Check out this 2017 blogpost by Sabine Hossenfelder (she who owns the BackReAction blog): Is the inflationary universe a scientific theory? Not anymore.
    She's quite scathing: "inflationary model building left behind reasonable scientific speculation long ago." Her basic point is that a good scientific theory needs two attributes, and making testable predictions is only one of them. It also needs to give us some reason to expect its predictions to be right. The way I've put this elsewhere is that falsifiability (there has to be some experiment you can imagine coming out in such a way that you could not rationalize away, you'd have to reject the theory-- this is what disqualifies personal belief systems as science theories) by itself is not enough, a good theory also has to make predictions that would seem unlikely to hold unless the theory was valid. It's this latter point that inflationary theories fail on, because there are so many versions of them that can predict so many things, no matter what you observe you can use hindsight to select whichever brand of the initial set of theories is the one that happened to work.

    So I agree with her there, but I would point out that such a scattershot approach to developing inflationary theories can become a scientific theory once there is an observation that allows you to select which version you need. At that point, you can used the culled-down theory to make new predictions, and those new predictions would satisfy both of the above criteria. The situation is similar to the parameters used in post-inflationary cosmology, which have many of the same attributes Hossenfelder criticizes about inflation-- you have not only a set of unconstrained parameters (the Hubble constant, the fraction of dark energy, the fraction of dark energy that is cold, etc.) which can be used to predict almost anything one might get in some new observation, you even have the possibility of adding additional parameters (the dark energy fraction, the dark energy equation of state, etc.) if you observe things (the brightnesses of type Ia supernovae) that can't be fit to any of the existing parameter sets. So that state of affairs would seem to fall victim to her same set of criticisms, yet that would imply that not just inflation, but the entire field of theoretical cosmology, is not a scientific pursuit!

    So I think she goes too far there-- what will produce a scientific theory of post-inflationary cosmology is a set of observations that give a consistent result for all these parameters, followed by an additional new observation that produces a consistent result capable of checking that theory. It is precisely the "Hubble tension" that says we aren't quite there yet, though just a short while ago we thought we were-- so perhaps we are close. We may not be as close in the case of inflation, but it's not the same as saying the pursuit to generate and cull inflationary theories is unscientific. I would simply say that any field of science requires a balance of theory and observation, and sometimes you get a wealth of unexplained data with no way to understand it, and sometimes you get a wealth of potential ways to understand data you don't even have yet, and both of those are not the situation you want to see. That inflation is in the latter camp (as is string theory) does not make it bad science, but it does make it badly unbalanced science, so the solution is to try to get the data that could balance it, or wait until said data becomes available (which is the main solution she seems to be proposing).

  6. #6
    Join Date
    Feb 2010
    Posts
    771
    Quote Originally Posted by Ken G View Post
    She's quite scathing: "inflationary model building left behind reasonable scientific speculation long ago." Her basic point is that a good scientific theory needs two attributes, and making testable predictions is only one of them. It also needs to give us some reason to expect its predictions to be right. The way I've put this elsewhere is that falsifiability (there has to be some experiment you can imagine coming out in such a way that you could not rationalize away, you'd have to reject the theory-- this is what disqualifies personal belief systems as science theories) by itself is not enough, a good theory also has to make predictions that would seem unlikely to hold unless the theory was valid. It's this latter point that inflationary theories fail on, because there are so many versions of them that can predict so many things, no matter what you observe you can use hindsight to select whichever brand of the initial set of theories is the one that happened to work.

    So I agree with her there, but I would point out that such a scattershot approach to developing inflationary theories can become a scientific theory once there is an observation that allows you to select which version you need. At that point, you can used the culled-down theory to make new predictions, and those new predictions would satisfy both of the above criteria. The situation is similar to the parameters used in post-inflationary cosmology, which have many of the same attributes Hossenfelder criticizes about inflation-- you have not only a set of unconstrained parameters (the Hubble constant, the fraction of dark energy, the fraction of dark energy that is cold, etc.) which can be used to predict almost anything one might get in some new observation, you even have the possibility of adding additional parameters (the dark energy fraction, the dark energy equation of state, etc.) if you observe things (the brightnesses of type Ia supernovae) that can't be fit to any of the existing parameter sets. So that state of affairs would seem to fall victim to her same set of criticisms, yet that would imply that not just inflation, but the entire field of theoretical cosmology, is not a scientific pursuit!

    So I think she goes too far there-- what will produce a scientific theory of post-inflationary cosmology is a set of observations that give a consistent result for all these parameters, followed by an additional new observation that produces a consistent result capable of checking that theory. It is precisely the "Hubble tension" that says we aren't quite there yet, though just a short while ago we thought we were-- so perhaps we are close. We may not be as close in the case of inflation, but it's not the same as saying the pursuit to generate and cull inflationary theories is unscientific. I would simply say that any field of science requires a balance of theory and observation, and sometimes you get a wealth of unexplained data with no way to understand it, and sometimes you get a wealth of potential ways to understand data you don't even have yet, and both of those are not the situation you want to see. That inflation is in the latter camp (as is string theory) does not make it bad science, but it does make it badly unbalanced science, so the solution is to try to get the data that could balance it, or wait until said data becomes available (which is the main solution she seems to be proposing).
    Yes, Sabine often takes what seem like extreme positions. She certainly has no qualms about possibly offending a lot of people. Her "community" is theoretical physics, focussing on fundamental or foundational theories; her own work includes quantum gravity, dark matter, the sociology of physics (and much of science), and lately the foundations of quantum stuff.

    As I am, at heart, an astronomer, I like to see what she has to say about astrophysics and cosmology, and look at it through a "now what could be/has been observed?" eye.

    You, and some other readers, may find some of her other blogposts interesting. As she moderates comments on her posts, many of those comments are well worth reading carefully too (it can sometimes be difficult to say just why you think some of those comments are quite wrong, even though you feel quite certain that they are).

  7. #7
    Join Date
    Sep 2003
    Posts
    12,887
    Quote Originally Posted by Ken G View Post
    … We may not be as close in the case of inflation, but it's not the same as saying the pursuit to generate and cull inflationary theories is unscientific. I would simply say that any field of science requires a balance of theory and observation, and sometimes you get a wealth of unexplained data with no way to understand it, and sometimes you get a wealth of potential ways to understand data you don't even have yet, and both of those are not the situation you want to see. That inflation is in the latter camp (as is string theory) does not make it bad science, but it does make it badly unbalanced science, so the solution is to try to get the data that could balance it, or wait until said data becomes available (which is the main solution she seems to be proposing).
    [my bold] Nicely put! It seems she is arguing that there is an imbalance in acceptance or credibility of inflation and, IMO, she is being a bit too hyperbolic on labeling inflation a non-science.

    I agree with her ending statement, "There’s no warning sign you when you cross the border between science and blabla-land." [I prefer Sillyville but "blabla" infers word salads, so that's a plus for a label, and there are warning signs, just ignored too often by many.] I just don't see having a large number of models as a way to move them all to "blabla-land". If so, if x represents the number of models, what value, x, determines when that border is crossed? As long as the models are objective-based, make predictions, and are falsifiable (even in principle), aren't they a service to science?
    We know time flies, we just can't see its wings.

  8. #8
    Join Date
    Feb 2010
    Posts
    771
    Quote Originally Posted by George View Post
    [my bold] Nicely put! It seems she is arguing that there is an imbalance in acceptance or credibility of inflation and, IMO, she is being a bit too hyperbolic on labeling inflation a non-science.

    I agree with her ending statement, "There’s no warning sign you when you cross the border between science and blabla-land." [I prefer Sillyville but "blabla" infers word salads, so that's a plus for a label, and there are warning signs, just ignored too often by many.] I just don't see having a large number of models as a way to move them all to "blabla-land". If so, if x represents the number of models, what value, x, determines when that border is crossed? As long as the models are objective-based, make predictions, and are falsifiable (even in principle), aren't they a service to science?
    If you read some other blogposts by her, you may guess where she's coming from on this.

    In HEP (high energy physics a.k.a. particle physics), in the early days of the LHC, x was in the thousands, maybe tens of thousands. Most based on zero actual hard data; most based on "beauty" dressed in supersymmetry clothes. As the LHC data came in and there was no sign of any SUSY particles, the many hundred/thousand theoretical physicists responded by tweaking their models, and re-tweaking them, etc.

    By your criterion - "are objective-based, make predictions, and are falsifiable (even in principle)" - these many-time-retweaked models should be a service to science, right?

    To what extent is "inflation" like this? There's very little in the way of robust observational results, but a great many tweaked models.

    Maybe it's time to stop going down a rabbit hole and start thinking outside the box?

  9. #9
    Join Date
    Oct 2005
    Posts
    26,823
    Quote Originally Posted by Jean Tate View Post
    Yes, Sabine often takes what seem like extreme positions. She certainly has no qualms about possibly offending a lot of people. Her "community" is theoretical physics, focussing on fundamental or foundational theories; her own work includes quantum gravity, dark matter, the sociology of physics (and much of science), and lately the foundations of quantum stuff.

    As I am, at heart, an astronomer, I like to see what she has to say about astrophysics and cosmology, and look at it through a "now what could be/has been observed?" eye.

    You, and some other readers, may find some of her other blogposts interesting. As she moderates comments on her posts, many of those comments are well worth reading carefully too (it can sometimes be difficult to say just why you think some of those comments are quite wrong, even though you feel quite certain that they are).
    I do respect her insights, as she can serve as a kind of "interpreter" of some of the more difficult aspects of the abstract mathematics involved. The level of the comments is often pretty high also, if you apply the necessary filters!

  10. #10
    Join Date
    Oct 2005
    Posts
    26,823
    Quote Originally Posted by George View Post
    I just don't see having a large number of models as a way to move them all to "blabla-land". If so, if x represents the number of models, what value, x, determines when that border is crossed? As long as the models are objective-based, make predictions, and are falsifiable (even in principle), aren't they a service to science?
    Yes I think there's something more going on there than just number of models, I think her main point is that if you have that many models, you can't really have that many new ideas, so something is wrong or at least highly inefficient. If I throw a pair of distinguishable 6-sided dice there are 36 different possible combinations of what can come up, but I don't need 36 different theories to explain each outcome and then wait to see which one got the right answer, I need one theory that says the outcome cannot be predicted, and a second theory that says it can and here's how, and that's really all I would need when it comes time to do observational testing. Since the "dice of inflation" only got rolled once in the history of our universe, different theories for every possible outcome are not really that useful.

  11. #11
    Join Date
    Sep 2003
    Posts
    12,887
    Quote Originally Posted by Jean Tate View Post
    By your criterion - "are objective-based, make predictions, and are falsifiable (even in principle)" - these many-time-retweaked models should be a service to science, right?

    To what extent is "inflation" like this? There's very little in the way of robust observational results, but a great many tweaked models.
    The ones that qualify as being scientific models deserve to exist, and tweaking is likely. GR got a fair amount of tweaking even by Einstein. I assume this is a given so I was surprised that the article would state, "Inflation is useful because it relates existing observations to an underlying mathematical model, yes. But, we don't yet have enough data to make reliable predictions from it." The article's point of excessive "quantity" is argued with somewhat weak arguments, IMO, attempting to discredit "quality". [I agree with the view that SciAm should be free to note criticisms. The letter to the mag. reminded of the "Hundred scientists against Einstein" one, and apparently that was primarily those too-stuck with Newton.]

    The article comes across to me, Average Joe, like complaining that the forest now has too many trees. Of course, this view is only fair if the "trees" are real objects (legitimate scientific models). If we are instead saying there is too much "air" in the forest, then a claim that all or most are superfluous would appear valid. I still see trees more than air, am I wrong? I sense the articles sees them as something in-between where a certain number of pseudo-trees is okay but not thousands. [The claim that there are "literally infinitely many models one can think up" is more humorous than argumentative and hyperbole even I wouldn't use. Well... I did quote my father once when he said, with a grin, that "Texas was so big, it was bigger than the whole United States!" That was his way to get others to move onto a different topic and actually away from potential hubris. ]

    Maybe it's time to stop going down a rabbit hole and start thinking outside the box?
    I get that point, but I would have preferred to have seen some sort of improved qualifiers to actually cul the herd. If people are just publishing to avoid perishing, then brakes need to be applied, but yet have some sort of way to get the more robust ones from being sucked down that rabbit hole.
    Last edited by George; 2019-Nov-17 at 08:34 PM.
    We know time flies, we just can't see its wings.

  12. #12
    Join Date
    Sep 2003
    Posts
    12,887
    Quote Originally Posted by Ken G View Post
    Yes I think there's something more going on there than just number of models, I think her main point is that if you have that many models, you can't really have that many new ideas, so something is wrong or at least highly inefficient. If I throw a pair of distinguishable 6-sided dice there are 36 different possible combinations of what can come up, but I don't need 36 different theories to explain each outcome and then wait to see which one got the right answer, I need one theory that says the outcome cannot be predicted, and a second theory that says it can and here's how, and that's really all I would need when it comes time to do observational testing. Since the "dice of inflation" only got rolled once in the history of our universe, different theories for every possible outcome are not really that useful.
    Right, and no surprise given your great respect for the SM, you are implying what I think should have been more explicit in the article -- qualifiers for the models themselves. When quantity creates a quagmire then quality needs greater consideration.
    We know time flies, we just can't see its wings.

  13. #13
    Join Date
    Oct 2005
    Posts
    26,823
    Yes, I think her article could be summed up that inflation theories exhibit too much quantity and not enough quality.

  14. #14
    Join Date
    Sep 2003
    Posts
    12,887
    Quote Originally Posted by Ken G View Post
    Yes, I think her article could be summed up that inflation theories exhibit too much quantity and not enough quality.
    Agreed.

    I stumbled upon something that may offer a hint of help -- Escape from Model Land. It's geared for climatology but some of the ideas may be helpful as some sort of model-making guide, perhaps:

    "You may be living in model-land if you…
    1) try to optimise anything regarding the future;
    2) believe that decision-relevant probabilities can be extracted from models;
    3) believe that there are precise parameter values to be found;
    4) refuse to believe in anything that has not been seen in the model;
    5) think that learning more will reduce the uncertainty in a forecast;
    6) explicitly or implicitly set the Probability of a Big Surprise to zero; that there is nothing your model cannot simulate;
    7) want “one model to rule them all”.

    You may be near the exit if you…
    1) are now concerned about your forecast systems Probability of a Big Surprise;
    2) are uncomfortable to realise that given 100x the resources, your team would build a better model rather than refine the probability estimates of your current model;
    3) are careful to distinguish x from its model-land namesake xmodel;
    4) have forsaken the quest for a model that tells you everything, to search instead for models that tell you something about your Quantity of Interest;
    5) are intrigued by the idea of finding your model’s Relevant Dominant Uncertainty – the aspect of the model you should improve first. "
    We know time flies, we just can't see its wings.

  15. #15
    Join Date
    Oct 2005
    Posts
    26,823
    Yet to...
    3) are careful to distinguish x from its model-land namesake xmodel;
    ... I would say "x is a model too, just a more vague one, in the sense of saying less but using fewer required assumptions to say it." So the caution to the modeler is not to avoid taking seriously that xmodel could be the same thing as x, it is to recognize that one cannot pass from a more vague model x to a more specific model xmodel without paying some kind of price for using all those required assumptions and quantitative idealizations.
    And to...
    4) have forsaken the quest for a model that tells you everything, to search instead for models that tell you something about your Quantity of Interest;
    ... I would put it that the key is to match the number of independent questions you would like to answer to the number of independent observable constraints that your model is trying to account for.

    And to...
    5) are intrigued by the idea of finding your model’s Relevant Dominant Uncertainty – the aspect of the model you should improve first. "[/I][/QUOTE]
    ... I would put it that the values of all the observable and fit parameters should be normalized to some kind of "sensitive interval" for each parameter, i.e., how much you would need to change that parameter in order to have an important impact on whatever else you are doing.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •