Nonsense Pretending: Probability as a Disguise

Jul 31 2008 Published by under Bad Physics, Bad Probability, Debunking Creationism

Once again, you, my readers, have come through with some really high-grade crackpottery. This one was actually sent to me by its author, but I didn't really look at it until several readers sent me the same link because they thought it was my kind of material. With your recommendations, I took a look, and was rewarded. In a moment of hubris, the author titled it A Possible Proof of God's Existence from Multiverse Assumptions.

This article is basically a version of the classic big-numbers probabilistic argument for God. What makes this different is that it doesn't line up a bunch of fake numbers and saying "Presto! Look at that great big probability: that means that it's impossible for the universe/life/everything to exist without God!". Instead, it takes a more scientific looking approach. It dresses the probability argument up using lots of terms and ideas from modern physics, and presents it as "If we knew the values of these variables, we could compute the probability" - with a clear bias towards the idea that the unvalued variables must have values that produced the desired result of this being a created universe.

Aside from being an indirect version of the big-numbers argument, this is also a nice example of what I call obfuscatory mathematics. See, you want to make some argument. You're dead sure that it's right. But it doesn't sound convincing. So you dress it up. Don't just assume your axioms - make up explanations for them in terms of math, so that it sounds all formal and mathy. Then your crappy assumptions will look convincing!

With that said, on to his argument!

He starts out by playing with words to produce a description of a multiverse that will work for his proof. The basic idea of his multiverse is that what we call the universe is only one out of many - that is, instead of what we call the universe being everything that there is, it's just something that appears, to those of us trapped inside of it, as everything. In fact if we could look in from outside, we'd see lots of different "universes".

He starts by imposing an artificial limit on the kinds of multiverses. That is, he says that the many universes in the multiverse all share the same basic rules - the only possible differences between different universes is parameterized physics, and physical states.

The idea of parametrized universes is that the basic laws of physics must be the same in all possible universes, but different universes have different values for some finite set of parameters. We're already running into trouble here. There's no particularly good reason to believe that if there are multiple universes that they have the same basic physical laws, or that they can only vary by a specific, finite number of constants.

To illustrate what I mean by that: in his parameterized universes, you'd have the same basic four forces. But numbers like the relative strengths of the forces, the masses of basic particles, the plank length, etc., could vary. But you couldn't have a universe in which there was no electromagnetic force at all. Why can't there be a universe without electromagnetic forces? Why couldn't there be a universe where there was an additional force beyond the four basics we observe?

The answer is obvious: because if you didn't constrain the multiverse in this way, his proof wouldn't work. So he just handwaves his way past it.

The state differences - which he calls "non-parameterized multiverses" basically comes down to the idea that there can be multiple universes with exactly the same physical laws, where all possible constants (parameters) are exactly the same. The only differences between these universes are what a computers scientist like me calls "state" - the exact set of particles, and their positions, masses, and velocities can be different in different universes.

And once again, he's off the rails of the math. He claims that quantum physics requires a non-parametric multiverse. That's wrong. One interpretation of quantum physics is based on multiple universes, constantly branching. But it's just one possible interpretation of the math - not an actual requirement of it, nor an observed fact.

Based on his argument that quantum physics requires non-parametric multiverses, he concludes that we need to consider hybrids - a multiverse where there are groups of universes, each with their own set of physical constants. Within a group, the constants are fixed; between different groups, the constants can vary, but the basic physical laws cannot.

An important thing to point out here is just how much he's constrained things already. He's trying to form an argument that probability points towards a created universe because of the structure of the universe - but he's arbitrarily constrained that structure quite strictly for absolutely no reason.

And now, he starts to really make mistakes.

Necessarily we exist in a universe in which the origin, evolution and on-going survival of life is possible.

But what can we say about the probabilities of origination and evolution?

If there are free parameters in the laws of physics, then these parameters are going to affect both probabilities, i.e. the probability of life originating, and the probability of it evolving (especially evolving to a point where intelligent life appears).

Empirically we can observe that the survival and evolution of life is something that occurs with reasonably high probability, at least once life has started. Our planet has a history of thousands of millions of years in which millions of species have evolved, some of them into ever-more complex forms.

Read that last paragraph again. "Empirically, we can observe that the survival of life is something that occurs with high probability, at least once life has started". Yup, I think that we can all agree with that: the probability of survival of life is high given that a living thing survived to reproduce. Talk about empty statements! The probability of this article being posted on my blog tomorrow has a high probability, at least once I'm done posting it today.

However, the origin of life appears to be something that occurs with a very low probability. It has only ever been observed to occur once, since all existing life forms appear to be related to each other. And there is no evidence of the existence of alien life out in space. If the origin of life was at all common, we would expect to see physical evidence of at least some alien civilisations, even if the expected self-destruction rate of such civilisations was as high as 99.99%.

Another very poor argument. The universe is a big place. A really big place. To quote Douglas Adams, "Space is big. You just won't believe how vastly, mind-bogglingly big it is. I mean, you may think it's a long way down the road to the chemists, but that's just peanuts to space". It's big. And it's got a speed limit, which is amazingly low when you realize just how big it is.

Among the 100 billion stars in our galaxy, how many have planets could contain life? We really don't know. If they did contain life, how many of them are close enough that it's remotely possible that we could encounter unmistakeable signs of life from them? So far, we haven't been able to discover unmistakeable signs of earth-like planets around any star anywhere. We've barely ever managed to prove the existence of non-gas-giant planets around other stars. The distances are just so huge that it's hard to even find something the size of a planet! How sure can we reasonably be that there's no intelligent life anywhere in our galaxy? Not very sure at all. Anyone honest would admit that we have absolutely no clue. And Mr. Dorrell also conflates two different things: the probability of life, and the probability of technological intelligent life. Perhaps life is amazingly common - but the development of intelligent, technological civilization is incredibly rare. There's just no meaningful way of measuring those probabilities: we do not have the information. I can easily whip out some Bayesian calculations about the probability of intelligent life - some that argue that it's virtually impossible for there to be intelligent life anywhere where we could observe it, and some that argue that it's so common that it's inevitable that we'll eventually encounter it. And I can easily show you what's wrong with both kinds of arguments: they're all based on made-up numbers, because we just don't know enough to make an informed estimate.

Given that the quantum multiverse probably exists, our own existence does not tell us anything at all about the probability of the origin of life. A lower bound on the probability is determined by probability of atoms spontaneously forming into the simplest possible life-form (while in some environment where reproduction can take place once the initial life form exists). As low as this probability might be, the number of universes within the quantum multiverse is much greater. So if it is at all possible for life to exist, life will exist.

And once again, he's making stuff up. How may "universes" are there in the "multiverse"? How do we know that the number of multiverses compatible with the existence of life is large enough to guarantee one where life does exist? We don't know if there are other universes. If there are, we don't know how many, or how much they vary from one another. And even if they're nearly identical with our own, we don't know how likely it is for life to begin. So how can we make any meaningful assumption like this? We can't.

However, it is possible to apply a probabilistic argument to distinguish between a parametrized multi-multiverse and a non-parametrized multiverse in which a specific set of values for the free parameters has been chosen.

These two choices can be described in terms of a "God" who has created the multiverse, and what we might be able to deduce about why this God has created it.

It's only possible to apply a probabilistic argument to distinguish between a parameterized multiverse and a non-parameterized one by talking out your ass. The existence of a multiverse is not observable from within our universe. That's pretty much by definition: if it were, then the "universe" would include those other sub-universes. A multiverse isn't observable from within. If there are other universes with different values for the basic physical constants, we can't see them. So how can we probabilistically distinguish between a multiverse where there are varying constants, and one where there isn't? The real answer is "we can't". But that never stops a crackpot.

If there is only one multiverse with one specific set of free parameter values, then this implies that God has chosen that particular set of values for some purpose. And if the values are chosen from some very small set of values which allow the development of intelligent life, then it is a reasonable conclusion that allowing the development of intelligent life is the very reason why those values where chosen.

On the other hand, if God has chosen to create the full set multiverses for all possible sets of free parameter values, then we can no longer deduce from our own existence anything about why God might have create such a multi-multiverse. In particular, it might be something happening in some other part of the multi-multiverse which explains the true purpose of God creating it.

Note that now he's basically assuming the existence of God.

It's possible that the universe really is the universe - the only one, with its set of physical constants as they are. And if that's the case, then the fact that there is only one universe doesn't prove that it was deliberately created. That's what Mr. Dorrell is trying to prove. You can't just assume from the existence of one universe that God must exist.

In fact, by that argument, there's no need to go any further. Because if there is a multiverse containing multiple universes that have varying parameters, that means that the multiverse itself is a universe with its own laws, and with its own constants describing the set of sub-universes contained by it. If it's the only multiverse, then the fact that the multiverse is the only one, with one specific set of free parameter values would imply that God had created those for some purpose. On the other hand, if there are multiple multiverses, then there's a multi-multi-verse, with its own laws and constants, ....

Now, we're finally approaching the part that he calls a proof. He again couches things in mathematical-sounding terms. He's not "proving that god exists", he's "rejecting the null hypothesis of a parameterized multi-multiverse". Before getting to it, I'll just briefly point out that I've been meaning to write an article on what's wrong with null hypothesis testing; as the Bayesian folks pointed out in the last probability flame-war, null hypothesis testing is actually a piss-poor way of testing a hypothesis. Null hypothesis testing is, arguably, complete rubbish.

But ignore that. He's not even building an argument to reject a null hypothesis of a parameterized universe. He's trying to reject a null hypothesis of a Godless parameterized universe.

Even if you like null hypothesis testing, one of the important things in doing it right (or as close to right as NHT gets) is choosing an appropriate null hypothesis. He's spent the bulk of his document so far creating artificial constraints on his null hypothesis. It's not a null hypothesis at all.

To apply a probabilistic argument to the parametrized multi-multiverse, it is necessary to determine an a priori probability distribution over the N-dimensional space of free parameter values. If we consider that the sub-space where life is possible is actually a very small region (as implied by the fine-tuning argument), then it is plausible that the probability distribution, whatever it might be, is approximately uniform over this sub-space.

Once again, pulling "facts" out of his ass. We don't know that the sub-space where life is possible is small. We don't know that there is a multiverse. If there is a multiverse, we don't know that it's constrained to only vary by a finite number of specific constants. Even if there is a multiverse with those constraints, we don't know anything about the distribution of constant values over the member universes. This is completely pulled out of the air, with no reason, no evidence, no justification, and absolutely nothing resembling proof.

A secondary assumption is that the probability of the on-going survival and evolution of life is relatively insensitive to the free parameter values within that region in which the origin of life is most probable. This is plausible on the assumption that the two probabilities are largely determined by different factors. (A crude analogy is betweeen the probability of a fire starting, and the probability of a fire continuing to burn, which in many circumstances are determined by completely separate variables. For example, in the wild, the probability of a fire continuing to burn depends on the amount of dead plant material, and the general level of dryness, whereas the probability of a fire starting depends mainly on how often lightning strikes the area in question. Factors that strongly influence the occurrence of lighting storms might have little or no effect on the general accumulation of combustible material.)

And again. Same trick.

Assuming an approximately uniform distribution, and assuming a parametrized multi-multiverse, we have to conclude that the most probable set of free parameter values must be whichever set of parameter values has the maximum probability of the origin of life.

Whereas, assuming only a multiverse with a fixed set of parameter values, we can only deduce that the parameter values must allow life to originate.

I have a hard time figuring out what he's saying in that first sentence. I think that what he means is that if there's a parameterized multi-verse with uniform distribution, then the most probable set of parameter values for our universe is the one which makes the origin of life most likely, and that we can therefore assume that the parameters in our universe are the ones that make life most likely.

That's not a valid assumption at all. Assuming a uniform distribution, all universes where life is possible are equally likely. We can't assume that we're sitting at the sweet spot that makes it most likely. If you've got a parameterized multiverse with a large number of sub-universes with uniform distribution, and there's a range of constant values that produce a universe where life is possible, you'd expect a large number of sub-universes where life was possible - and of those, a large number with life. With uniform distribution, most of the ones with life won't be at the real center.

I think that he wants to argue for a normal distribution of life. But he doesn't say that, and even if he did, it wouldn't allow him to draw the conclusion that he wants. He wants the argument to say that we're at the perfect sweet-spot for life.

Now if God happens to choose a fixed set of values, but those happen to be the values which maximize the probability of life originating, then from within our own universe, we cannot distinguish this from the parametrized multi-multiverse.

But if God chooses a fixed set of values which allow life to originate, but for which the probability of life originating is much, much lower than the maximum possible value, then we can effectively reject the null hypothesis that we live in a parametrized multi-multiverse, and we can deduce that someone or something (i.e. "God"), has chosen the particular set of values that we observe in our universe.

So... If the universe in which we live has constant values which are non-optimal for life, then we can reject the idea that life came into being without a God. Ummm.... Sorry. no.

If there's an infinite number of universes (and his explanation of multiverse theories clearly seems to suggest that there are), and we happen to be living in one with physical constants that make life unlikely but not impossible... Then we're living in a world that you would predict should exist.

In summary... If we assume (for no particular reason) that we live in a bizarrely constrained sort of multiverse, and you assume (for no particular reason) that the multiverses are uniformly distributed, and you assume (for no particular reason) that we should be living in the universe that makes life most likely unless there's a God who deliberately put us in this unlikely universe, then you can conclude that the universe shows that God must have deliberately put us here.

Can you say "assuming the conclusion"?

No responses yet

  • Anonymous says:

    "The answer is obvious: because if you didn't constrain the multiverse in this way, his proof wouldn't work. So he just handwaves his way past it."
    Agree that this is absolutely bogus, but I really don't see how allowing universes without the electromagnetic force (or with extra forces) would ruin his "proof".
    The only part of the proof he really needs is where he says that:
    "Assuming an approximately uniform distribution, and assuming a parametrized multi-multiverse, we have to conclude that the most probable set of free parameter values must be whichever set of parameter values has the maximum probability of the origin of life."
    This leads to the conclusion that if we are in a low-life-probability universe then G_d made us regardless of what set of universes the argument is acting on.

  • Ambitwistor says:

    For a better probabilistic argument about God, life, and fine-tuning of physical constants, see "The Anthropic Principle Does Not Support Supernaturalism" by Ikeda and Jefferys, which is summarized in more lay terms in Jefferys' review of "The Privileged Planet". This happens to come to the opposite conclusion: "finely tuned" physical constants can undermine but never support supernatural creation.
    The argument basically boils down to: life can only exist in a naturalistic universe if the physical constants are conducive to life (which may mean "fine tuned"). Life in a supernaturalistic universe can exist for any constants, no matter how bizarre, just by divine fiat. Since naturalistic life (unlike supernaturally created life) can only exist with suitable constants, observing such suitable constants can only support natural origins. It doesn't imply supernatural origins. That would be implied if we observed that the constants aren't suitable for life yet life exists anyway.

  • AlefSin says:

    Hi Mark,
    A very interesting article as usual. Again, as usual you've systematically taken that argument apart, analyzed each segment and pointed to their weaknesses.
    However, as a regular reader of your blog, I cannot stop wondering what is your take, your point of view about God and his/her/its(their?) existence. If I'm not mistaken, you have mentioned before that you do in fact believe in God. I know that this is a personal thing and you might not want or be ready to discuss about it but I cannot stop wondering what the reasons for your belief are when you are so good at systematic, logical thinking and so good at finding underlying problems for the kind of arguments you discussed here. I mean, your reasons, your system of belief, it should be a profoundly good one to satisfy a critical mind like yours. I'm really interested to know about your view.

  • SLC says:

    1. It should be pointed out that the very notion of the multiverse was proposed for the purpose of negating the fine tuning argument. Thus, the counterargument to fine tuning goes that there may be a large number of universes, each with its own set of fundamental constants and we just happen to live in one of the universes that has a set of constants conducive to the development of life (presumably most of them don't). The problem with this argument is that it doesn't seem to provide any testable hypotheses and therefore appears more philosophical then scientific.
    2. Interestingly enough, back in the 1990s there was a debate between Carl Sagan and Ernst Mayr on the subject of the possible prevalence of intelligent life. Prof. Sagan argued that such life was possibly quite prevalent while Prof. Mayr argued that such life was possibly quite rare. As I recall, not having read the text for some years, Prof. Sagans' argument was based on the premise that any life that formed and evolved for a sufficient length of time would have a high probably of eventually producing intelligence. Prof. Mayrs' argument was based on the premise that the evolution of intelligent life had a low probability because of the number of contingencies that would be required (for instance, extinction of the dinosaurs was necessary for the development of intelligent mammals such as humans).

  • Joe N says:

    I think one could poke fun at all the multiverse theories, in as much as they masquerade as scientific theories. But hey, one man's grant money is another's late night routine.

  • Michael says:

    Hi Mark
    Great post, just one thing to add that wasn't touched on: the branching of the universes in the Many Worlds Interpretation in quantum physics is NOT the same thing as the multiverse.
    I believe the multiverse is about things on a higher level than the Big Bang, for instance the question of whether there were universes before the Big Bang, or whether our universe (all the matter/energy/events caused by the Big Bang) are part of the larger whole.
    The MWI on the other hand deals with THIS universe: its truth implies that THIS universe is in fact a tree that splits into branches that don't interfere.
    So yet another thing wrong with his argument: certainly quantum physics (at least as it stands now) deals largely with this universe (since it was built from data in this universe).

  • Joe N says:

    For what it's worth a blog devoted to "bad astronomy" -couldn't resist the similarity to the name of this blog!
    Also, as far as many-worlds quantum splitting (as opposed to multiverse quantum foam) check out Huw Price's totally unorthodox quantum interpretations as expounded in "Time's Arrow and Archimedes' Point". Its both very logical and very mind-blowing.

  • Max says:

    I would be charitable and try a different interpretation along similar lines:
    Suppose before observing the physical constants and deciding on probability of life evolving given those constants you have a prior belief about God's existence which is essentially probability p for (some form of) God exists and probability (1-p) God does not exist. Now suppose your alternatives are indeed either 1)assume multi-world quantum mechanics (you can either assume or reject multiverse, I will reject multiverse to make things simpler) or 2)assume a God-picked single universe.
    Now you measure the constants and find a probability of life evolving under those constants is q. You also observe the fact that you are alive. What should your new beliefs (you are being a rational Bayesian here) be?
    Here it gets kinda tricky. It's precisely what Scott Aronson was talking about talking about. If you do Self-Sampling Assumption then
    P(model|data)=P(data|model)P(model)/P(data), so P(God exists)=1p/p+q(1-p), P(God does not exist)=q(1-p)/(p+q(1-p)). So if we had q very small the P(God) will be pretty high (q on the order of p will give roughly P=1/2, assuming p,q both small).
    Of course it requires strong assumptions like p>0 in the first place and even worse p not much smaller then q, and existence of multiworld QM and some way to compute q and a rigid model of how the two different models work (you pointed out problems with those, especially with trying to calculate q), but with all that it does increase your belief that God exists.
    The main blow, in my mind, is the SSA, though. If you do SIA (condition on the fact that you are alive), then q is essentially, and you get no new information whatsoever.

  • Ambitwistor says:

    "It should be pointed out that the very notion of the multiverse was proposed for the purpose of negating the fine tuning argument."
    Is that true? Where was it first proposed? The term has been around some time in science fiction (Heinlein? DC Comics?). But I think the first time I remember it being used by an actual physicist was by David Deutsch, in reference to the many-worlds interpretation (contrary to Michael's claim). I think since then it has been broadened to refer to almost any kind of "parallel universe" in physics.

  • Ricahrd Eis says:

    I fail to see how, even if only this universe exists and it has unique physics, that this implies intelligent design. It merely implies that something interesting (to us) happened. Either the universe was created or it was not and no amount of probability wringing is going to change that.
    An interesting review but as the saying goes about creationism, "you do not need to be a tailor to see that the emperor has no clothes."

  • SLC says:

    Re Ambitwistor
    I can't say when the concept of the multiverse first entered discussions on cosmology but, according to Steven Weinberg, it definitely is based on the notion of fine tuning, also known as the strong anthropic principal. See articles by Frank Tipler on this subject.
    As I understand it, Tipler and his followers argue that a certain set of constants of physics are fine tuned such that they support the development of life. They argue that if the values of any of these constants were slightly different, life would not begin and evolve. They further argue that this fine tuning is evidence for the existence of god.
    This position has been challenged by Victor Stenger who has argued that the proposed set of constants is, in fact, a multivariate set and that significant changes on one of the constants could be compensated by changes in one or more of the others.
    However, others, including Prof. Weinberg, have proposed the theory of the multiverse in which a vast number of universes, each with its own set of fundamental constants, have been formed by big bangs. If this were to be true, then the Tipler argument of the improbability of a fine tuned set of fundamental constants collapses because of the large number of proposed universes. To put it another way, as the number of universes increases, the probability of one of them happening to have the right set of constants also increases.

  • Uncle Al says:

    the probability of technological intelligent life.
    The universe is rich with pond scum. So? Had the New World remained isolated from Europe and Asia would technological society ever have emerged? Folks across two continents suffused with gods never quite wrapped their minds around the wheel (and add Australia).
    In the whole of human history across the entire planet not one deity has volunteered Novocain. It is a telling omission.
    (physical reality) - (empirical reality) = faith

  • "He starts by imposing an artificial limit on the kinds of multiverses."
    He chose the wrong topology for the multiverse.
    I refer to the issue of the Tip of Calabi-Yau spaces by Hodge indices, and to other such recent theories.
    My wife and son and I have made a far more sophisticated choice in our jointly-written science fiction: "Oh, and Another Thing About the Universe."
    In measuring pairwise distance between universes, is the Topology:
    (a) a Metric Space
    (b) a Semimetric Space
    (c) None of the above?

  • Tony Jeremiah says:

    Here's a question: Would it be theoretically possible to determine whether the universe and life on earth arose through stochastic or nonstochastic processes by evaluating the estimated timeline of the following important events (BYA = Billion Years Ago):
    Big Bang: 13.73 BYA +/- .12
    Sun formation: 4.57 BYA
    Earth formation: 4.54 BYA
    Life on earth: 3.7 BYA
    Modern Humans: .000025 BYA
    In short, is it possible to ascertain whether the indicated time line for these events would be statistically different if one assumed the universe began with stochastic processes (e.g., not starting with the known universal constants), vs. non-stochastic processes (e.g., starting with the assumption of the known universal constants).

  • Reginald Selkirk says:

    However, the origin of life appears to be something that occurs with a very low probability. It has only ever been observed to occur once, since all existing life forms appear to be related to each other.

    This does not follow, even here on planet Earth. It is true that all observed life seems to share common ancestry (as judged by near-universal genetic code, etc.). However, this need not reflect a low probability of origin, or a single origin of life. There have been many "winner take all" steps in the history of life (with interesting analogies in the "evolution" of computer technology), and several large-scale catastrophes that could have wiped out some contenders.

  • Stephen Wells says:

    We probably can't even know how many different versions of a replicator were around on the early earth, because all the losing attempts were eaten by our distant ancestors billions of years ago.

  • Ken says:

    This guy may be an idiot, but if he click on his name, his web page has some really cool math games for kids. One is a shoot the number with prime factors. But if the prime factor is greater than 19 you must use the p.
    Also on his page he teaches you how to find the cube root of numbers.
    Just saying since you were talking about how to teach multiplication to kids earlier.

  • I'm not sure if this is the right thread, but heard something so Bad Math-y on the radio that I remarked on it to my wife and to another Physicist friends. And this morning I found it it my snailmailed New Scientist:
    United weights of America
    "If current trends were to continue unchecked, all American adults would be overweight or obese by 2048, say researchers at the Johns Hopkins School of Public Health in Baltimore, Maryland. By 2030, half could be clinically obese (Obesity, DOI: 10.1038/oby.2008.351)"
    [New Scientist, 9 Aug 2008, p.7]
    The radio piece said "100%" by 2048.
    All? 100% How about people who die of anorexia. Will they be obese also?
    I don't have space here to critique misuse of Math in obesity studies; I did that in last quarter's grad course on Public Health. Policymakers prefer to use non-peer-reviewed crud that supports their agenda, it seems to me.

  • Torbjörn Larsson, OM says:

    Late to the party, but FWIW:
    The post captures most of the many problems with this crackpottery. The major howler is of course the critical substitution of the creationist large number argument for abiogenesis ("probability of atoms spontaneously forming into the simplest possible life-form") in place of abiogenesis theory such as spontaneously emergent reproductive systems. The minor howler is that he assumes that a universe with life (intelligent to boot) is the only possible universe.

    However, this need not reflect a low probability of origin, or a single origin of life.

    In fact, the short time between Earth formation and first observations of life ( 4.25 Ga) argues for a high abiogenesis probability.

  • Torbjörn Larsson, OM says:

    There are some problems with science in the post though. That the MWI is not a multiverse has been noted.

    The existence of a multiverse is not observable from within our universe. That's pretty much by definition: if it were, then the "universe" would include those other sub-universes.

    Yes, that is what one would think, if physics definitions and theories are set in stone. Currently there isn't a consistent definition of multiverses. (But they are usually seen as the many possible states of non-unique cosmologies.)
    On such grounds it isn't paradoxical that physicists are currently looking for telltales of multiverses, for example when bubble universes collides.

    null hypothesis testing is actually a piss-poor way of testing a hypothesis. Null hypothesis testing is, arguably, complete rubbish. ... Even if you like null hypothesis testing,

    There may be better procedures, but it is SOP in science and engineering. There isn't anything to like about useful methods.
    [And mostly bayesians arguments amounts to crap anyway, as they use bayesian probabilities to argue against perfectly sane frequentist probabilities for philosophical reasons. Case in point is NHT, where they confuse accumulated differences with actual correlated differences.]

  • Anonymous says:

    Uups, I forgot to add to my first comment that: there are problem with the crackpot argument that no one pointed out; despite his own figure of a sharply peaked probability for life he assumes a uniform distribution. (Due, again, to his cherry picking of 'facts'.)
    And to the second comment that: "we just don't know enough to make an informed estimate" doesn't preclude biologists that note that evolution is contingent in its path dependence to suggest that intelligence is an a priori unlikely outcome.
    @ Max, #8:
    A model that runs afoul of a Doomsday argument is suspect IMHO. It probably can't count the possible worldlines correctly. (A known problem.)
    The model that Ambitwistor mentions in #2 is straightforward and produces an easy to understand debunking of erroneous religious anthropic 'finetuning' arguments.
    @ SLC, #4 & #11:
    AFAIU Weinberg made the first weak anthropic calculation for the purpose of explaining the cosmological constant value. It was specifically not aimed at "negating the fine tuning argument" of the religious type. At most, you can claim that it solved a physics finetuning problem, which is a very different affair.
    @ Joe N, #5:
    I'm sorry to burst your monoverse bubble, but multiverse theories are part and parcel of current physics. They are natural states of string and inflation theories. Search the physics arxiv, you will see many papers! Besides, your own link claims as much: "fairly mainstream".

Leave a Reply