Manny's Reviews > The Hidden Reality: Parallel Universes and the Deep Laws of the Cosmos
The Hidden Reality: Parallel Universes and the Deep Laws of the Cosmos
by Brian Greene
by Brian Greene
Manny's review
bookshelves: science, multiverse
May 18, 11
bookshelves: science, multiverse
Recommended to Manny by:
Mary
Read from May 14 to 18, 2011
I've now read three books about the multiverse in rapid succession: the first two were Rees's
Before The Beginning
(1996) and Davies's
The Goldilocks Enigma
(2007). This one came out just a few months ago, so I'm hopefully up to date for the moment.
Well: I'm starting to feel quite familiar with the arguments, but each book has an interestingly different slant. Rees concentrates on presenting the experimental evidence for the existence of other universes  basically, the physical constants of our own universe appear to be tuned exactly right for life to be possible, it's unlikely that this is just chance, so we're probably one of many universes  and explicitly avoids philosophical speculation. Davies recapitulates Rees's arguments, but then goes overboard on philosophy; whatever he says, it's clear that he very much wants there to be a spiritual dimension to the story. Greene is similar to Davies, but his obsession is with string theory. He's spent his life working on it, and he desperately wants it to be part of the final explanation.
Despite Greene's exaggerated love for all things stringy, I liked this book and found it well worth reading. The author presents nine different theories which in one way or another involve multiple versions of the universe. Most of these I had already seen in Davies, but I found Greene's exposition clearer and more detailed. In particular, I thought his chapter on "inflation" was the best account I had read of this mysterious process, which got our universe started by exponentially blowing it up, in a minute fraction of a second, from the size of a proton to a macroscopic scale. People now seem to understand inflation moderately well, and there are mathematical models which can be checked against the detailed maps of the cosmic background radiation which have been constructed over the last 10 years.
I also thought the chapter on the Many Worlds interpretation of quantum mechanics was excellent; I hadn't understood that it's just a historical accident that the Copenhagen interpretation is the mainstream one. As Greene points out, it's actually more conservative to think of quantum processes as continually splitting off new universes, each of which is equally "real". The Copenhagen interpretation, with its notion of the collapse of the wavefunction, is the one which involves handwaving and mystical invocations of the privileged role of the "observer". If you just look at the underlying mathematics, the Schrödinger equation, the Many Worlds interpretation is the straightforward way to translate the numbers into words.
As already noted, Greene wants string theory to be in there, so there are chapters on "braneworlds" (our universe is floating in a higher dimensional space), the "Cosmic Landscape" (there is a multitude of universes reflecting different flavors of string theory) and the Holographic Universe (we are the projection of a lowerdimensional process on a distant boundary surface). Like Davies, he also speculates about the possibility that our universe is a simulation running in some gigantic computer, or that we live in a "Platonic Universe", where mathematical theories exist simply by virtue of being consistent, and there is no other reality.
All of these ideas were fun to read about, but I felt they were somewhat beside the point; I was surprised not to see him focussing on what, at least to me, seems like the obvious account. As he says, the Many Worlds interpretation of quantum mechanics is perfectly respectable, and inflation has accumulated a great deal of credibility as a theory of how the universe got started. Inflation starts at a scale where quantum processes are allimportant. We can actually see the fossils of those quantum processes in the unevenness of the background radiation.
Putting those two things together, we already have a multiverse theory that's mainstream to the point of being conservative. The quantum realities which split off during the very early history of the universe will be substantially different from each other. The question is whether they're going to be different enough, since we want them to differ with respect to things including settings of the physical constants. This part is still speculative. But you need it anyway in order to make most of the other theories work, and there seem to be moderately wellworked out accounts of how it could happen. I am sure people must be exploring in this direction.
It sounds incredible when you write it down, but we appear to be on the edge of constructing a coherent scientific theory of Life, The Universe and Everything. I'm really curious to see what happens next.
Well: I'm starting to feel quite familiar with the arguments, but each book has an interestingly different slant. Rees concentrates on presenting the experimental evidence for the existence of other universes  basically, the physical constants of our own universe appear to be tuned exactly right for life to be possible, it's unlikely that this is just chance, so we're probably one of many universes  and explicitly avoids philosophical speculation. Davies recapitulates Rees's arguments, but then goes overboard on philosophy; whatever he says, it's clear that he very much wants there to be a spiritual dimension to the story. Greene is similar to Davies, but his obsession is with string theory. He's spent his life working on it, and he desperately wants it to be part of the final explanation.
Despite Greene's exaggerated love for all things stringy, I liked this book and found it well worth reading. The author presents nine different theories which in one way or another involve multiple versions of the universe. Most of these I had already seen in Davies, but I found Greene's exposition clearer and more detailed. In particular, I thought his chapter on "inflation" was the best account I had read of this mysterious process, which got our universe started by exponentially blowing it up, in a minute fraction of a second, from the size of a proton to a macroscopic scale. People now seem to understand inflation moderately well, and there are mathematical models which can be checked against the detailed maps of the cosmic background radiation which have been constructed over the last 10 years.
I also thought the chapter on the Many Worlds interpretation of quantum mechanics was excellent; I hadn't understood that it's just a historical accident that the Copenhagen interpretation is the mainstream one. As Greene points out, it's actually more conservative to think of quantum processes as continually splitting off new universes, each of which is equally "real". The Copenhagen interpretation, with its notion of the collapse of the wavefunction, is the one which involves handwaving and mystical invocations of the privileged role of the "observer". If you just look at the underlying mathematics, the Schrödinger equation, the Many Worlds interpretation is the straightforward way to translate the numbers into words.
As already noted, Greene wants string theory to be in there, so there are chapters on "braneworlds" (our universe is floating in a higher dimensional space), the "Cosmic Landscape" (there is a multitude of universes reflecting different flavors of string theory) and the Holographic Universe (we are the projection of a lowerdimensional process on a distant boundary surface). Like Davies, he also speculates about the possibility that our universe is a simulation running in some gigantic computer, or that we live in a "Platonic Universe", where mathematical theories exist simply by virtue of being consistent, and there is no other reality.
All of these ideas were fun to read about, but I felt they were somewhat beside the point; I was surprised not to see him focussing on what, at least to me, seems like the obvious account. As he says, the Many Worlds interpretation of quantum mechanics is perfectly respectable, and inflation has accumulated a great deal of credibility as a theory of how the universe got started. Inflation starts at a scale where quantum processes are allimportant. We can actually see the fossils of those quantum processes in the unevenness of the background radiation.
Putting those two things together, we already have a multiverse theory that's mainstream to the point of being conservative. The quantum realities which split off during the very early history of the universe will be substantially different from each other. The question is whether they're going to be different enough, since we want them to differ with respect to things including settings of the physical constants. This part is still speculative. But you need it anyway in order to make most of the other theories work, and there seem to be moderately wellworked out accounts of how it could happen. I am sure people must be exploring in this direction.
It sounds incredible when you write it down, but we appear to be on the edge of constructing a coherent scientific theory of Life, The Universe and Everything. I'm really curious to see what happens next.
Sign into Goodreads to see if any of your friends have read
The Hidden Reality.
Sign In »
Reading Progress
05/14/2011  page 65 

17.0%  "My third pop cosmology book in as many weeks. Reading them in order of publication, it's interesting to see that, each time I get to that bit, they seem a bit more certain about the validity of "inflation". Though I'd still hesitate to say that anyone's sure." 
05/15/2011  page 110 

29.0%  "I really liked his account of inflation  quite the best one I've read. But now he's explaining why string theory is actually doing very well, despite all the unkind things people say about it..." 
05/15/2011  page 175 

46.0%  "I didn't know that Weinstein had used the Anthropic Principle to make a fairly good estimate of the strength of Dark Energy, ten years before it was discovered. What a clever guy!" 
05/15/2011  page 205 

53.0%  "Everett's dissertation on the Many Worlds interpretation of quantum mechanics was heavily edited because Bohr didn't like it, Everett's advisor Wheeler deferred to Bohr's feelings, and Everett needed a PhD urgently to take up a job in the US nuclear weapons program. An interesting story!" 2 comments 
05/17/2011  page 250 

65.0%  "A good, more or less selfcontained account of the Many Worlds interpretation of quantum mechanics. Though I was surprised not to see explicit discussion of links to cosmology. Maybe that's in the last chapter?" 
05/17/2011  page 275 

72.0%  "The Holographic Principle. I don't see yet what this has to do with the multiverse  looks to me like he just put it in because it's cool and shows off string theory in an advantageous light. But maybe he's got a trick up his sleeve." 
Comments (showing 150 of 114) (114 new)
message 1:
by
Kristen
(new)
May 15, 2011 04:13PM
I heard Greene on NPR the other day discussing two of the theories covered in this book, one of them appeared to be just a new incarnation of eternal return, yet he never used the term and sort of explains it like he just thought it up now. Maybe I only perceived it that way because I have a deep hated from Greene (using Simpson characters in you explanation isn't fun, it just makes me want to run you over with my car).
reply

flag
*
If I'm inclined to read a Brian Greene book (which I am) would you recommend this one instead of Elegant Universe given the theoretical advances over the last decade?
Hi Kristen! Yes, Greene and Davies are both totally into Eternal Return, though confusingly they never quote Nietzche. They calculate how far you need to travel in the multiverse to find the next you  it's a long way.
Given what you said, this book will make you want to do more than just run him over. The metaphor he uses to explain the inflaton field involves countless copies of Cartman sitting on the tops of the spiky mountains you see in the background of South Park. What does that merit?
Ian, I would strongly recommend this over The Elegant Universe. Unless you're also allergic to people using cartoon characters to explain complex mathematics.
Given what you said, this book will make you want to do more than just run him over. The metaphor he uses to explain the inflaton field involves countless copies of Cartman sitting on the tops of the spiky mountains you see in the background of South Park. What does that merit?
Ian, I would strongly recommend this over The Elegant Universe. Unless you're also allergic to people using cartoon characters to explain complex mathematics.
Thanks, Manny. I'll definately check out this book. I have no objection to South Park or any other cartoon characters making a point :)
Right now I'm reading From Eternity to Here by Sean Carroll. He tackles the question of why time only goes one direction, when in all other dimensions we're aware of direciton doesn't matter. It's fascinating and very accessible.
Right now I'm reading From Eternity to Here by Sean Carroll. He tackles the question of why time only goes one direction, when in all other dimensions we're aware of direciton doesn't matter. It's fascinating and very accessible.
notgettingenough wrote: "I really don't see the point of writing this review without any Simpson reference."
Doh! How could I have missed that??
Doh! How could I have missed that??
I mean a person of your goodreads standing should be brave enough to take the challenge even if it means losing a vote. I can see I'm going to have to do it myself.
Trots off to review book.
http://www.goodreads.com/review/show/...
Trots off to review book.
http://www.goodreads.com/review/show/...
Okay, okay. I get the message.
Well, to get an idea of what the braneworld universe looks like, imagine Lisa standing on the top of Springfield's nuclear power station playing her saxophone...
Well, to get an idea of what the braneworld universe looks like, imagine Lisa standing on the top of Springfield's nuclear power station playing her saxophone...
OK  arriving here from that other review, I find that I disagree with quite a bit:
1. The Many Worlds Interpretation is generally considered unproveable, therefore possibly respectable but certainly not science.
2. You tacitly assume one must choose between Many Worlds and Copenhagen. This is false: one has the option of saying, hey, look  if you specify the intitial and final conditions you can calculate the probability of an event! No, mystery, no spontaneous generation of an infinity of infinities of universes, no weird Observer Problem.
3."It's unlikely it's just chance so we are probably one of many universes." Exactly what is scientific about this argument? The assumption is wrong and the conclusion doesn't follow from it.
4. The Eternal Inflation idea multiplies entities infinitely without evidence and so should be considered as merely a hypothesis until a method of observational/experimental testing is developed.
5. String "Theory" is not necessarily selfconsistent, bares no known relation to reality and even if every variation exists we still don't know which one(s) would match the current "low" energy observations.
So the multiverse isn't mainstream as anything more than an idea.
1. The Many Worlds Interpretation is generally considered unproveable, therefore possibly respectable but certainly not science.
2. You tacitly assume one must choose between Many Worlds and Copenhagen. This is false: one has the option of saying, hey, look  if you specify the intitial and final conditions you can calculate the probability of an event! No, mystery, no spontaneous generation of an infinity of infinities of universes, no weird Observer Problem.
3."It's unlikely it's just chance so we are probably one of many universes." Exactly what is scientific about this argument? The assumption is wrong and the conclusion doesn't follow from it.
4. The Eternal Inflation idea multiplies entities infinitely without evidence and so should be considered as merely a hypothesis until a method of observational/experimental testing is developed.
5. String "Theory" is not necessarily selfconsistent, bares no known relation to reality and even if every variation exists we still don't know which one(s) would match the current "low" energy observations.
So the multiverse isn't mainstream as anything more than an idea.
Robert wrote: "OK  arriving here from that other review, I find that I disagree with quite a bit:
1. The Many Worlds Interpretation is generally considered unproveable, therefore possibly respectable but certainly not science.
Well, I agree it's unprovable. As the name suggests, it's basically a way of giving a commonsense meaning to the calculations, isn't it? Greene just points out that it's actually the simplest way to do that.
2. You tacitly assume one must choose between Many Worlds and Copenhagen. This is false: one has the option of saying, hey, look  if you specify the intitial and final conditions you can calculate the probability of an event! No, mystery, no spontaneous generation of an infinity of infinities of universes, no weird Observer Problem.
Agreed, you don't need to adopt any interpretation. But most people want one, don't they? Or why all this heated discussion?
3."It's unlikely it's just chance so we are probably one of many universes." Exactly what is scientific about this argument? The assumption is wrong and the conclusion doesn't follow from it.
Ah, we've done this already. Thinking about it some more, I'm pretty sure you can formalize the probabilistic argument in a Bayesian framework. I feel your objections are too narrowly frequentist.
4. The Eternal Inflation idea multiplies entities infinitely without evidence and so should be considered as merely a hypothesis until a method of observational/experimental testing is developed.
5. String "Theory" is not necessarily selfconsistent, bares no known relation to reality and even if every variation exists we still don't know which one(s) would match the current "low" energy observations.
"
Neither Eternal Inflation nor string theory are in any way relevant to the minimalist multiverse I suggest, which only requires ordinary inflation and the Many Worlds interpretation! So what entities are needlessly being multiplied, unless you think Occam's Razor is a valid argument against MWI? I'm sure there are many philosophers of science who'd disagree with you on that score.
1. The Many Worlds Interpretation is generally considered unproveable, therefore possibly respectable but certainly not science.
Well, I agree it's unprovable. As the name suggests, it's basically a way of giving a commonsense meaning to the calculations, isn't it? Greene just points out that it's actually the simplest way to do that.
2. You tacitly assume one must choose between Many Worlds and Copenhagen. This is false: one has the option of saying, hey, look  if you specify the intitial and final conditions you can calculate the probability of an event! No, mystery, no spontaneous generation of an infinity of infinities of universes, no weird Observer Problem.
Agreed, you don't need to adopt any interpretation. But most people want one, don't they? Or why all this heated discussion?
3."It's unlikely it's just chance so we are probably one of many universes." Exactly what is scientific about this argument? The assumption is wrong and the conclusion doesn't follow from it.
Ah, we've done this already. Thinking about it some more, I'm pretty sure you can formalize the probabilistic argument in a Bayesian framework. I feel your objections are too narrowly frequentist.
4. The Eternal Inflation idea multiplies entities infinitely without evidence and so should be considered as merely a hypothesis until a method of observational/experimental testing is developed.
5. String "Theory" is not necessarily selfconsistent, bares no known relation to reality and even if every variation exists we still don't know which one(s) would match the current "low" energy observations.
"
Neither Eternal Inflation nor string theory are in any way relevant to the minimalist multiverse I suggest, which only requires ordinary inflation and the Many Worlds interpretation! So what entities are needlessly being multiplied, unless you think Occam's Razor is a valid argument against MWI? I'm sure there are many philosophers of science who'd disagree with you on that score.
I dunno what "frequentist" means but I refer you to:
1) Feynman's registration plate.
2) The unproven assumption that "constants" are in fact variables.
Actually, Occam's Razor applies better to MWI vs. Copenhagan vs. What Observer Problem? on the grounds that one cannot expect a scientific resolution to the issue and we therefore need some other way to decide. The one with the least number of entities is WOP followed by Copenhagen followed by WMI.
As regards all the heated debate...people who don't think there is an Observer Problem are less likely to write philosophical treates on the topic...
1) Feynman's registration plate.
2) The unproven assumption that "constants" are in fact variables.
Actually, Occam's Razor applies better to MWI vs. Copenhagan vs. What Observer Problem? on the grounds that one cannot expect a scientific resolution to the issue and we therefore need some other way to decide. The one with the least number of entities is WOP followed by Copenhagen followed by WMI.
As regards all the heated debate...people who don't think there is an Observer Problem are less likely to write philosophical treates on the topic...
When I say "Bayesian" as opposed to "frequentist", I mean that you should work out probabilities given one's state of knowledge, rather than by counting instances. Evidently, the constants are what they are in our universe. But one can meaningfully talk about the probability of a constant's having a value in a certain range given one's state of knowledge.
And if you think Occam's Razor refutes MWI, then doesn't it also refute the view that there is any universe at all? I mean, it's simplest just to posit the existence of sense impressions and nothing else. Who needs reality? But most people don't accept this argument.
And if you think Occam's Razor refutes MWI, then doesn't it also refute the view that there is any universe at all? I mean, it's simplest just to posit the existence of sense impressions and nothing else. Who needs reality? But most people don't accept this argument.
But our state of knowledge is such that the possible range of variation in the constants is that of the observational error! The fundamental problem with this "coincidence" argument is that there is a hidden assumption that there are a range of universes to do statistics on, but you can't do statistics on a sample of one from set of one! And if you assume a set of universes it's hardly surprising that your argument backs there being a set of universes...
The reason Occam's razor is particularly suited to the (nonexistant) Observer Problem is because a scientific approach can't be taken, however a scientific approach to solipsism can be taken, so the Razor goes a bit blunt.
The reason Occam's razor is particularly suited to the (nonexistant) Observer Problem is because a scientific approach can't be taken, however a scientific approach to solipsism can be taken, so the Razor goes a bit blunt.
you can't do statistics on a sample of one from set of one
Well, this is exactly why I said we should use Bayesian methods. A standard example is the statement "global warming is probably occurring right now over the Earth". Evidently, either it is or it isn't  there's only one Earth to look at. But the statement is perfectly reasonable.
You are taking an extreme position on what "probability" means if you just rule out the coincidence argument absolutely. I mean, do you think people like Martin Rees have never heard your counterargument? The question is more how unlikely the coincidence is.
Well, this is exactly why I said we should use Bayesian methods. A standard example is the statement "global warming is probably occurring right now over the Earth". Evidently, either it is or it isn't  there's only one Earth to look at. But the statement is perfectly reasonable.
You are taking an extreme position on what "probability" means if you just rule out the coincidence argument absolutely. I mean, do you think people like Martin Rees have never heard your counterargument? The question is more how unlikely the coincidence is.
This is why I go quietly nuts when as weather forecaster tells me that there is a 10% chance of rain today...On 1 in 10 days exactly like today (of which there will only ever be 1) it will rain! They really mean something different.
In this case they mean, if you could run a zillion Big Bangs, what's the probability of ending up here? This assumes that there is a range of variability in the "constants" but the only evidence of such variability is the observational error and we know that with that range we get...this universe. This is different from the weather where the error in the initial conditions is enough for the model to give rain on 10% of runs.
In this case they mean, if you could run a zillion Big Bangs, what's the probability of ending up here? This assumes that there is a range of variability in the "constants" but the only evidence of such variability is the observational error and we know that with that range we get...this universe. This is different from the weather where the error in the initial conditions is enough for the model to give rain on 10% of runs.
I already mentioned in other review that when "looking time wise far back from where we are right now" is not bringing (yet ? or ever ?) the clear picture of what has happened then  then why not to try additionally to complement this effort by *extrapolation into the future* (may be projecting future could be easier than understanding the past ?) to see where we are going.
Under *we* I mean the entire Mankind, given (or just assuming it as a starting point  to see where it brings us) that its development represents the local *front wave* of the Evolution of the species, which by itself locally represents the most significant result of the cosmic development.
Under *we* I mean the entire Mankind, given (or just assuming it as a starting point  to see where it brings us) that its development represents the local *front wave* of the Evolution of the species, which by itself locally represents the most significant result of the cosmic development.
But Robert, the weather forecasters just mean "given what we know, it would be rational to offer odds of 1:9 that it will rain." What's wrong with that?
Similarly, you can for example put together models of the early universe which do spontaneous symmetry breaking resulting in different settings of the physical constants. For example, it doesn't seem that farfetched that the Higgs field, if it exists, could have settled down at a different level, resulting in different masses for the particles. Then you can ask how probable a given combination of ranges is.
Similarly, you can for example put together models of the early universe which do spontaneous symmetry breaking resulting in different settings of the physical constants. For example, it doesn't seem that farfetched that the Higgs field, if it exists, could have settled down at a different level, resulting in different masses for the particles. Then you can ask how probable a given combination of ranges is.
Alex, he certainly does a lot of extrapolation into the future, as do Rees and Davies. If you're worried about our chances of surviving the next hundred billion years, these are the guys to consult. I got the impression that they've all read their copies of Star Maker until they've fallen apart. I think all three cited it.
I meant to consider extrapolation into the future, which starts from the ongoing observable development (which is the social development in all spheres of human activity), which we see NOW and which could be considered independent of our guesstimated knowledge of the cosmological past.
Alex, I'm not completely sure I understand what you mean, but you might find it interesting to look at these books. Particularly Rees.
Nothing is wrong with what the weather forecasters really mean...
Let me paraphrase you: If something that we don't know exists can take a range of values for which there is no evidence it's possible that the universe would be different. How probable is the one outcome that we know has occurred?
Well, the question is absurd but if you cut out the absurd bits, you are left with: How probable is the one outcome that we know has occurred? Answer: 1.
Let me paraphrase you: If something that we don't know exists can take a range of values for which there is no evidence it's possible that the universe would be different. How probable is the one outcome that we know has occurred?
Well, the question is absurd but if you cut out the absurd bits, you are left with: How probable is the one outcome that we know has occurred? Answer: 1.
But look, Robert, there is abundant evidence that the early universe was ruled by quantum effects. You can see the traces in the cosmic background radiation. Quantum processes aren't deterministic, hence the universe could have turned out some other way.
So even though we only have one universe to look at, it surely makes sense to ask some questions of the form "How likely was it that the universe would have ended up with property X?" The only thing that's unclear is how far you can push it.
So even though we only have one universe to look at, it surely makes sense to ask some questions of the form "How likely was it that the universe would have ended up with property X?" The only thing that's unclear is how far you can push it.
Do they discuss the question why Evolution (in the strict sense of it being the Evolution of the species) stopped when it produced intelligent human being ?
Particularly do they discuss:
Do we (human beings) continue to evolve biologically / anatomically and if not why ?
Why there are no other species on Earth, which are continuing to evolve with the potentiality to reach the state of intelligence ?
Are there any objective laws (to be discovered yet ?), which describe what is going on in this domain ?
Particularly do they discuss:
Do we (human beings) continue to evolve biologically / anatomically and if not why ?
Why there are no other species on Earth, which are continuing to evolve with the potentiality to reach the state of intelligence ?
Are there any objective laws (to be discovered yet ?), which describe what is going on in this domain ?
I'm not aware of any theory that states that the values of the fundamental constants are in fact variable.
Further, you are wrong with the quantum argument: the Universe could not turn out another way because it did not. The probability of an event that has occurred is 1.
Any such argument is really saying, if I had a heap of universes in state 1, what's the chances of one of them changing to State 2. But then the assumption and the conclusion are the same: a multiverse.
Further, you are wrong with the quantum argument: the Universe could not turn out another way because it did not. The probability of an event that has occurred is 1.
Any such argument is really saying, if I had a heap of universes in state 1, what's the chances of one of them changing to State 2. But then the assumption and the conclusion are the same: a multiverse.
There's nothing about evolution, just some wayout stuff about how some kind of life could survive in the very distant future...
I'm not aware of any theory that states that the values of the fundamental constants are in fact variable.
I thought that was exactly what the Higgs field was for  to assign a mass to the elementary particles which depends on the field's strength, hence could be variable?
Further, you are wrong with the quantum argument: the Universe could not turn out another way because it did not. The probability of an event that has occurred is 1.
I think you're falling for some version of the fatalism argument. Look at my review of Fate, Time and Language.
I thought that was exactly what the Higgs field was for  to assign a mass to the elementary particles which depends on the field's strength, hence could be variable?
Further, you are wrong with the quantum argument: the Universe could not turn out another way because it did not. The probability of an event that has occurred is 1.
I think you're falling for some version of the fatalism argument. Look at my review of Fate, Time and Language.
If I am about to roll a sixsided die, we say the probability of rolling a "six" is 1 in 6, or 16.7%. Let's say I then roll the die and a "six" indeed comes up. Do we then say the probability of that six having come up is 100%, or is it still 16.7%?
I don't mean this question rhetorically in an attempt to take sides. I'm actually curious what you guys would say.
I don't mean this question rhetorically in an attempt to take sides. I'm actually curious what you guys would say.
The probability that a six did come up is evidently 100%, assuming that we can trust our eyes.
I could conceivable argue that the probability that a six was going to come up may now be slightly higher than 16.7%, if we want to take account of the fact that rolling a six gives us new evidence which marginally increases the probability that the die is biased towards sixes. But I'm not sure this is coherent. I think it's simplest to say that the probability was 16.7%, and that I may think the probability of throwing a six next time is slightly higher.
I could conceivable argue that the probability that a six was going to come up may now be slightly higher than 16.7%, if we want to take account of the fact that rolling a six gives us new evidence which marginally increases the probability that the die is biased towards sixes. But I'm not sure this is coherent. I think it's simplest to say that the probability was 16.7%, and that I may think the probability of throwing a six next time is slightly higher.
Like I said, a theory for which there is no evidence...
You are mistaken about me falling for some version of the fatalism argument. The probability of an event may be anywhere in the range 0 > 1 before it happens. If it is 0 then it doesn't happen. If it does happen, then the probability switches from whatever it was to 1. But once it happens, it is meaningless to ask what the probability of it occurring was before it happened  the answer is 1. If I roll a die and score a 6, the probability of that happening is 1. Next time 1:6, twice in a row, 1:6^2 but last time? 1:1. This in no way eliminates stochastic processes from occurring or from having occurred.
You are mistaken about me falling for some version of the fatalism argument. The probability of an event may be anywhere in the range 0 > 1 before it happens. If it is 0 then it doesn't happen. If it does happen, then the probability switches from whatever it was to 1. But once it happens, it is meaningless to ask what the probability of it occurring was before it happened  the answer is 1. If I roll a die and score a 6, the probability of that happening is 1. Next time 1:6, twice in a row, 1:6^2 but last time? 1:1. This in no way eliminates stochastic processes from occurring or from having occurred.
Ha! I just came up with the dierolling excample without having seen Ian's comment! What are the chances of that? 100%!
Manny, I'm afraid to have to say this and mean no offense, but you're answer to Ian is wrong. Which about wraps up this debate, as it seems to be the crux of where our opinions diverge.
I'll be interested to hear what you think about Julian Barbour's book!
Also, have you come across this:
http://thenextweb.com/shareables/2011...
I'll be interested to hear what you think about Julian Barbour's book!
Also, have you come across this:
http://thenextweb.com/shareables/2011...
Robert, you are just not using probabilities in the way people do in normal life, when they reflect uncertain beliefs based on partial knowledge.
Here's yet another example. You're a detective who's investigating a serial killer. The guy has killed ten people in various different places. Now a colleague notices that, every time a murder was committed, the Prime Minister was in the immediate vicinity within half an hour of the crime. This is pretty weird, they tell you. It surely can't be just chance. They suggest ideas like checking out the PM's entourage, or seeing if there's anyone who might be trying to frame him.
Do you reply "Well, it happened, so the probability is 100%! There is nothing to explain?" And if you did, how do you think they would react?
Here's yet another example. You're a detective who's investigating a serial killer. The guy has killed ten people in various different places. Now a colleague notices that, every time a murder was committed, the Prime Minister was in the immediate vicinity within half an hour of the crime. This is pretty weird, they tell you. It surely can't be just chance. They suggest ideas like checking out the PM's entourage, or seeing if there's anyone who might be trying to frame him.
Do you reply "Well, it happened, so the probability is 100%! There is nothing to explain?" And if you did, how do you think they would react?
Also, have you come across this:
http://thenextweb.com/shareables/2011/05...
Aaargh! Chess variants! The very work of the devil!
http://thenextweb.com/shareables/2011/05...
Aaargh! Chess variants! The very work of the devil!
You're talking about correlation, not probability, now. If you then say, "ha! But what is the probability of the correlation being random?" I say, when people do that kind of statistics they are in truth saying, if I in the future do x many tests of this type, I would expect this level of correlation y times by chance. If you say, what's the chance of getting this exact correlation with this exact data? 100%!
Just thought I'd save some time...
Oh, and:
http://imgs.xkcd.com/comics/significa...
Just thought I'd save some time...
Oh, and:
http://imgs.xkcd.com/comics/significa...
I know that xkcd cartoon  a very witty presentation of the Bonferroni correction, which we in fact discussed in connection with a recent conference paper we wrote.
I don't quite understand why we're disagreeing here. I use statistical tests all the time, and I assume you do too. I am well aware of the various pitfalls involved, and how things like Bonferroni mean you need to be careful about assigning probabilities.
So if I were the hypothetical detective, and my boss asked me to estimate how likely it was that it could have happened by chance, I'd take that kind of reasoning into account and say it was hard to pin down exactly. But after doing a little figuring, my guess is I'd say that the probability was so low that we needed to follow it up. And so would you, whatever formal language you used to express your conclusion.
So are we just disagreeing about the right form of words, or what?
I don't quite understand why we're disagreeing here. I use statistical tests all the time, and I assume you do too. I am well aware of the various pitfalls involved, and how things like Bonferroni mean you need to be careful about assigning probabilities.
So if I were the hypothetical detective, and my boss asked me to estimate how likely it was that it could have happened by chance, I'd take that kind of reasoning into account and say it was hard to pin down exactly. But after doing a little figuring, my guess is I'd say that the probability was so low that we needed to follow it up. And so would you, whatever formal language you used to express your conclusion.
So are we just disagreeing about the right form of words, or what?
"Quantum mechanics is certainly imposing. But an inner voice tells me that it is not yet the real thing. The theory says a lot, but does not really bring us any closer to the secret of the "old one." I, at any rate, am convinced that He does not throw dice."
"I believe in Spinoza's God, Who reveals Himself in the lawful harmony of the world, not in a God Who concerns Himself with the fate and the doings of mankind."
from "Born Einstein" by Albert Einstein
PS However (IMHO  AP) "the fate and the doings of mankind" could also be the manifestation of the *lawful harmony of the world*
"I believe in Spinoza's God, Who reveals Himself in the lawful harmony of the world, not in a God Who concerns Himself with the fate and the doings of mankind."
from "Born Einstein" by Albert Einstein
PS However (IMHO  AP) "the fate and the doings of mankind" could also be the manifestation of the *lawful harmony of the world*
Alex wrote: ""Quantum mechanics is certainly imposing. But an inner voice tells me that it is not yet the real thing. The theory says a lot, but does not really bring us any closer to the secret of the "old one." I, at any rate, am convinced that He does not throw dice."
"I believe in Spinoza's God, Who reveals Himself in the lawful harmony of the world, not in a God Who concerns Himself with the fate and the doings of mankind."
from "Born Einstein" by Albert Einstein"
Yes, Greene quotes that in connection with the Many Worlds Interpretation, and says that it does indeed respect Einstein's intuition that God doesn't decide things by throwing dice...
"I believe in Spinoza's God, Who reveals Himself in the lawful harmony of the world, not in a God Who concerns Himself with the fate and the doings of mankind."
from "Born Einstein" by Albert Einstein"
Yes, Greene quotes that in connection with the Many Worlds Interpretation, and says that it does indeed respect Einstein's intuition that God doesn't decide things by throwing dice...
Manny  you know me...i find Robert's argument more compelling. There's no proof constants can be variable, we got what we got. So there's no proof that "probability" applies to evaluating the "odds" that they came out in some fashion. I don't think that necessarily rules out multiple universes. We have no way of knowing whether. if they did exist, they could all have the same constants.
Ha! It's funny, I read this book and I was all, wow! Many Worlds Interpretation + Inflation = 42! Though, oddly enough, Greene never actually says that, most likely because it doesn't require string theory.
To me, it's an extraordinary revelation: a more or less mainstream theory of the multiverse, which already comes close to explaining the bizarre settings of the physical constants. But other people look at this, and say, hohum. Is there anything that actually needs to be explained?
As Greene says, you get the same pattern all the way up to the world's most distinguished scientists. Some people think it's a new scientific revolution, others are completely unmoved and don't even see it as science. It's weird.
To me, it's an extraordinary revelation: a more or less mainstream theory of the multiverse, which already comes close to explaining the bizarre settings of the physical constants. But other people look at this, and say, hohum. Is there anything that actually needs to be explained?
As Greene says, you get the same pattern all the way up to the world's most distinguished scientists. Some people think it's a new scientific revolution, others are completely unmoved and don't even see it as science. It's weird.
First up, I only linked to the cartoon because it was funny, not as a veiled criticism of you!
Here's what you do in the serial killer case: You say, assume that the killer will strike again. This is physically possible. Then you say, based on the evidence, what is the probability that in the future the Prime Minister is in proximity when the killing takes place, by chance alone?
The fundamental difference is that we cannot, even in principle, assume that the Big Bang will recur with exactly the same initial conditions, ever, so we can't do statistics on the probability of the same outcome by chance. Further, even if, tomorrow, new evidence transpires to say the universe will actually do a Big Crunch afterall and therefore may be continually oscillating, we don't have evidence that the constants have variability really, hence we cannot do statistics based on a range of values of them beyond that of the observational error.
Now, as I've said before, people are entitled to feel that life is freaky because they can imagine universes with all sorts of variations from observation and then go on to research theories and experiments prompted by that and maybe one day a proper such theory, or compelling experimental evidence will be produced, but I think it's jumping the gun to say we are at that point now. As with many aspects of fundamental physics/cosmology/astronomy all of this will have to be reevaluated in the light of results from CERN regarding the Higgs boson. I guess we all have to try to contain our excitement until then!
Here's what you do in the serial killer case: You say, assume that the killer will strike again. This is physically possible. Then you say, based on the evidence, what is the probability that in the future the Prime Minister is in proximity when the killing takes place, by chance alone?
The fundamental difference is that we cannot, even in principle, assume that the Big Bang will recur with exactly the same initial conditions, ever, so we can't do statistics on the probability of the same outcome by chance. Further, even if, tomorrow, new evidence transpires to say the universe will actually do a Big Crunch afterall and therefore may be continually oscillating, we don't have evidence that the constants have variability really, hence we cannot do statistics based on a range of values of them beyond that of the observational error.
Now, as I've said before, people are entitled to feel that life is freaky because they can imagine universes with all sorts of variations from observation and then go on to research theories and experiments prompted by that and maybe one day a proper such theory, or compelling experimental evidence will be produced, but I think it's jumping the gun to say we are at that point now. As with many aspects of fundamental physics/cosmology/astronomy all of this will have to be reevaluated in the light of results from CERN regarding the Higgs boson. I guess we all have to try to contain our excitement until then!
I understand it's traditional for discussions of this topic to be heated, so under the circumstances I feel we're being remarkably restrained :)
Here's what you do in the serial killer case: You say, assume that the killer will strike again. This is physically possible. Then you say, based on the evidence, what is the probability that in the future the Prime Minister is in proximity when the killing takes place, by chance alone?
But this isn't the question your boss is asking you. He wants to know how likely you think it is that the link to the Prime Minister is a coincidence, based on what you know now. And surely it isn't an unreasonable question? He has to decide whether or not to divert resources into following it up. Making the wrong decision could cost lives.
Here's what you do in the serial killer case: You say, assume that the killer will strike again. This is physically possible. Then you say, based on the evidence, what is the probability that in the future the Prime Minister is in proximity when the killing takes place, by chance alone?
But this isn't the question your boss is asking you. He wants to know how likely you think it is that the link to the Prime Minister is a coincidence, based on what you know now. And surely it isn't an unreasonable question? He has to decide whether or not to divert resources into following it up. Making the wrong decision could cost lives.
I don't understand the distinction you are making. Are you asking me to guess the outcome of the statistical analysis? The statistical analysis is based on the current knowledge and the necessary physical possibility of a genuinely comparable event occurring in the future. The distinction between this and the case under discussion is the twofold impossibility of the event under discussion recurring in the future.
If you are asking me to guess, well, I just reiterate that the status of such guesses is that they are neither science nor statistics, but are acceptable as ways of choosing what to research!
If you are asking me to guess, well, I just reiterate that the status of such guesses is that they are neither science nor statistics, but are acceptable as ways of choosing what to research!
I don't see that the possibility of a similar event occurring in the future is relevant. As I've said above, this is an overly frequentist view. When people say "the probability that anthropogenic global warming is occurring is well over 90%", they are saying something meaningful, though the situation is not repeatable in any interesting way.
To take an example from cosmology that's closer to the one we're discussing, suppose you're doing a study of the background microwave radiation map and think you notice a pattern. Perhaps you find that there are no hotspots in a given range of angular diameters, though there are hotspots with sizes above and below that range. It seems perfectly reasonable to ask how likely it is that the lack of spots in the given range of sizes is a coincidence. If the range is very narrow, then presumably it's not interesting. The broader the range, the stronger your case will be. But you can never repeat the experiment: the map is what it is.
To take an example from cosmology that's closer to the one we're discussing, suppose you're doing a study of the background microwave radiation map and think you notice a pattern. Perhaps you find that there are no hotspots in a given range of angular diameters, though there are hotspots with sizes above and below that range. It seems perfectly reasonable to ask how likely it is that the lack of spots in the given range of sizes is a coincidence. If the range is very narrow, then presumably it's not interesting. The broader the range, the stronger your case will be. But you can never repeat the experiment: the map is what it is.
Your quote about global warming is the same sloppy terminology as the weather forecasters'. If you fully describe the method used, I'm sure the real meaning will become clear, just as it does with regard to ensemble forecasting of the weather.
You can make as many random maps as you like and see what proportion match the "pattern" in the observed one. You can also make as many observational maps as you like and see what range of results you get.
A common and subtle error with statistics can be illustrated with reference to the good old pvalue: You calculate a pvalue for whatever it is and say, "look, the chance of this being random chance is <0.001." WRONG! Actually, it is, "Look, the chance of my doing this type of test again and getting this level of correlation by pure chance is <0.001." The error only rears it's head when you try to ask the probability of something genuinely unrepeatable: What's the probability that Hitler will turn out to be leader of a German political party? <unaswerable.
What's the chance that a randomly selected German will become leader of a political party? <Possibly answerable, if you do it right.
You can make as many random maps as you like and see what proportion match the "pattern" in the observed one. You can also make as many observational maps as you like and see what range of results you get.
A common and subtle error with statistics can be illustrated with reference to the good old pvalue: You calculate a pvalue for whatever it is and say, "look, the chance of this being random chance is <0.001." WRONG! Actually, it is, "Look, the chance of my doing this type of test again and getting this level of correlation by pure chance is <0.001." The error only rears it's head when you try to ask the probability of something genuinely unrepeatable: What's the probability that Hitler will turn out to be leader of a German political party? <unaswerable.
What's the chance that a randomly selected German will become leader of a political party? <Possibly answerable, if you do it right.
You can make as many random maps as you like and see what proportion match the "pattern" in the observed one. You can also make as many observational maps as you like and see what range of results you get.
There's only one observational map  or at any rate, let's assume that so as not to get sidetracked into an argument about observational error. The background radiation was only created once, and when you've mapped it accurately, there you are.
When you talk about "random maps", the randomness has to depend on some underlying model. Why can't we do the same thing with models of the early universe which allow setting of physical constants during spontaneous symmetry breaking? The main difference seems to be that one is more speculative than the other, but they're both speculative. Why is only one admissible?
There's only one observational map  or at any rate, let's assume that so as not to get sidetracked into an argument about observational error. The background radiation was only created once, and when you've mapped it accurately, there you are.
When you talk about "random maps", the randomness has to depend on some underlying model. Why can't we do the same thing with models of the early universe which allow setting of physical constants during spontaneous symmetry breaking? The main difference seems to be that one is more speculative than the other, but they're both speculative. Why is only one admissible?
If you develop a theory that allows variables across some range instead of constants, you can do what you say. Then you would be doing statistics on some imaginary set of early universes for which there is no evidence, but real statistics about imaginary universes don't tell us anything about the real universe. If you could find evidence for the existence of a fixed number or infinite number of universes, you could say, what are the chances that a randomly selected universe has these variables "frozen" at values that allow life? But as we are, with imaginary extra universes, that just leads us back to the premise and the conclusion being the same.
Regarding the CBR, you have to choose random values within the known range of assignable values. This relies on no particular physical model. Or if you want to be more subtle, you could say, I'm going to randomly assign these structures based on the observed range of parameters for these structures. Then you are of course getting into murkier waters and have to very careful about your assumptions. You can generate as many of these random maps as you like and you ask, what is the chance that one of them has the same distribution of structures according to whatever criteria you choose, looks like the observational one. You can really make random maps; you really can't make random universes.
Regarding the CBR, you have to choose random values within the known range of assignable values. This relies on no particular physical model. Or if you want to be more subtle, you could say, I'm going to randomly assign these structures based on the observed range of parameters for these structures. Then you are of course getting into murkier waters and have to very careful about your assumptions. You can generate as many of these random maps as you like and you ask, what is the chance that one of them has the same distribution of structures according to whatever criteria you choose, looks like the observational one. You can really make random maps; you really can't make random universes.
Yes yes yes. So here's the core question. Your argument appears to suggest that, no matter what pattern you found in the CBR, you could just say "only one universe, this is the way it is, nothing to explain". But you know perfectly well that some patterns would have you scratching your head and saying wow, this is very weird, whatever can have caused it? And you'd try and construct models to explain the pattern. No?
I've repeatedly said one can use anything to prod one's research in one direction or another.
But this comes back to something that had me scratching my head back in round one of this discussion: what do you mean by a pattern amongst the constants? If someone can show me a robust simple, or even complicated relationship between the constants, I might well think, is there a deeper explanation of that? If someone says, these constants all exist in a narrow range that allow life as we recognise it, but there is no other genuine scientific support for the idea that these are really not constants, I say, pretty weak! But good luck to you!
But this comes back to something that had me scratching my head back in round one of this discussion: what do you mean by a pattern amongst the constants? If someone can show me a robust simple, or even complicated relationship between the constants, I might well think, is there a deeper explanation of that? If someone says, these constants all exist in a narrow range that allow life as we recognise it, but there is no other genuine scientific support for the idea that these are really not constants, I say, pretty weak! But good luck to you!