Hold on, the naturalistic origin of life is pretty plausible based on current understanding (Miller-Urey showed amino acid development would be very plausible, and from there there are a number of AIUI chemically sound models for those building blocks naturally forming self-replicating organisms or pseudo-organisms).
Are you genuinely arguing that its probability is so low that it would be less productive to investigate naturalistic abiogenesis mechanisms than it would be to look for new hypotheses? Or, alternately, do you have a more specific idea of what minimum level of probability it takes for a hypothesis be “plausible” rather than “being taken on faith”?
There is some truth to the claim that even atheists currently take some things “on faith”. The naturalistic origin of life, for example.
Hold on, the naturalistic origin of life is pretty plausible based on current understanding.
And that claim by you is based on … what exactly? Experiments you have performed? Books you have read explaining the theory to your satisfaction with no obvious hand waving? Books like the ones we all have read describing Darwin’s theory of evolution through natural selection? Or maybe you have encountered a section in a library filled with technical material beyond your comprehension, but which you are pretty sure you could comprehend with enough effort? For me, something in this category would be stereo amplifiers—I’ve seen the books so I know there is nothing supernatural involved, though I can’t explain it myself.
From what you write below, I’m guessing your background puts you at roughly this level regarding abiogenesis. Except, the difference is that there is no library section filled with technical material explaining how life originated from non-life. So, I think you are going on faith.
(Miller-Urey showed amino acid development would be very plausible, and from there there are a number of AIUI chemically sound models for those building blocks naturally forming self-replicating organisms or pseudo-organisms).
No there aren’t. There is not a single plausible theory in existence right now claiming that life originates from amino acids arising from a Miller-Urey type of process. There are no chemically sound models for creating life from Miller-Urey building blocks.
There are some models which have life starting with RNA, and some which have life starting with lipids, or iron-sulfide minerals or even (pace Tim) starting with clay. But you didn’t mention those more recent and plausible theories. Instead you went on faith.
Are you genuinely arguing that its probability is so low that it would be less productive to investigate naturalistic abiogenesis mechanisms than it would be to look for new hypotheses?
No. Not at all. I don’t have a clue as to what it would even mean to look for, let alone investigate a non-naturalistic hypothesis.
What I am saying is this: Suppose I have before me a theist who claims that a Deity must have been the cause of the Big Bang. “Something from nothing” and all that. Suppose further that my own version of atheism is so completely non-evangelical and my knowledge of cosmology so weak that I say to him, “Could be! I don’t believe that a Deity was involved, but I don’t have any evidence to rule it out.”
So that is the supposition. Next, suppose he says “Furthermore, I think the Deity must have been involved in the origin of life, back 3+ billion years ago. From what I know of chemistry and biochemistry, that could not have been spontaneous.” I would tell him that I too know quite a bit about chemistry and biochemistry, and that there are many, many clues indicating that life based on RNA probably existed before modern life based on DNA and amino acids and proteins. There is strong evidence that life in its current form came from something more primitive.
Now, suppose he says “Yes, I understand all that evidence. But you still have no theory to explain how life based on RNA might have started. All you have is handwaving. My belief, since I already believe in a Creator-Deity for the Big Bang, is that a Creator-Deity was also involved in the origin of the RNA World.” If he said that, then I would answer, “Ok, you can believe that, but since I don’t already believe in a Creator Deity, I would prefer to believe that the first living organism on earth arose by some unknown natural process. In fact, I have some ideas as to how it might have happened.”
Yes, I said “prefer to believe” this time. I don’t have a good explanation for life’s origin, though I have spent a good deal of my spare time over the past 30 years looking for one.
I think what’s happening here is that the theist’s preexisting belief about a creator god is causing them to privilege the hypothesis of divine RNA-creation.
The trouble is, you seem to be privileging it too. The way you’ve set up the scenario makes it seem like there are two hypotheses: (1) goddidit, (2) some unknown natural process.
But (2) is actually a set of zillions of potential processes, many of which have a far better prior than Yahweh and the Thousand Claims of Scripture, even if we can’t actually choose one for sure right now. Taken together, all their probability mass dwarfs that of the goddidit hypothesis.
You don’t have to know all the answers to say “you’re (almost certainly) wrong.”
I don’t see any reasons why (2) - unknown natural process—gets to benefit from being a “set of zillions of potential processes, many of which have a far better prior” and (1) - goddidit—does not.
If you want to sum the probability of a hypothesis by performing some weighted sum over the set of zillions of it’s neighbors in hypothesis space, that’s fine. But if that is your criteria, you need to apply it equally to the other set of hypothesises you are considering—instead of considering only one specific example.
Bostrom’s simulation argument gives us one potential generator of ‘goddidits’, and a likely high prior for superintelligent aliens gives us another potential generator of ‘goddidits’. Either of those generators could spawn zillions of potential processes which have far better priors than Yaweh, but could look similar.
None of this leads to any specific conclusion—I’m just pointing out an unfairness in your methodology.
I don’t see any reasons why (2) - unknown natural process—gets to benefit from being a “set of zillions of potential processes, many of which have a far better prior” and (1) - goddidit—does not.
You’re quite right. When I said this, I was thinking of “goddidit” as a set of very specific claims from a single religious tradition, which I should’ve stated.
Mmm actually you did state it as a fairly specific claim. I’m just saying one can’t fairly compare highly specific complex hypothesizes vs wide general sweeps through hypothesis-space. This is itself a good argument against the specific “yawheh did it”, but not against the more general “goddidit” which you originally were referring to:
I think what’s happening here is that the theist’s preexisting belief about a creator god is causing them to privilege the hypothesis of divine RNA-creation.
You’re right—but there is another side to this coin. An atheist has a top-level belief (or it’s negation) which sends down cascading priors and privileges naturalistic hypothesizes. So far this has worked splendidly well across the landscape.
But there is no guarantee this will work everywhere forever, and it’s at least possible that eventually we may flip or find an exception for the top-level prior—for example we may eventually find that pretty much everything has a naturalistic explanation except the origin of life—which turns out to have been seeded by alien super-intelligence (ala Francis Crick) - for example.
This is itself a good argument against the specific “yawheh did it”, but not against the more general “goddidit”
[...]
An atheist has a top-level belief (or it’s negation) which sends down cascading priors and privileges naturalistic hypothesizes. So far this has worked splendidly well across the landscape.
But there is no guarantee this will work everywhere forever, and it’s at least possible that eventually we may flip...
I entirely agree. While I don’t know of any good reasons to think the origin of life was not a happy accident, it is not inconceivable a priori (simulations, seeding etc.).
When I describe myself as an atheist (which I try not to do), I really mean that
(1) all the anthropomorphic creation myths are really laughable,
(2) there’s not much positive evidence for less laughable creators, and
(3) even if you showed me evidence for a creator, I would be inclined toward what I will call meta-naturalism—i.e., still wanting to know how the hell the creator came to be.
Basically, I doubt the existence of gods that are totally ontologically distinct from creatures.
Bostrom’s simulation argument does NOT give us a generator of “goddidits” regarding the origin of life and the universe, because implicit in the question “How did life originate?” is a desire to know the ultimate root (if there is one), and us being in a simulation just gives us some more living beings (the simulators “above”) to ask our questions about. Where did life in the universe “one level above us” come from? Where did our simulator/parent universe originate?
There is nothing unfair in dismissing “A MIRACLE!” in comparison to the set of plausible naturalistic processes that could explain a given phenomenon. And to second SarahC, it’s somewhat incoherent to talk about non-naturalistic processes in the first place. You need to be very clear as to what you’re suggesting when you suggest “god did it”. But, no one here is suggesting that, so I’ll stop tangenting into arguing against theists that don’t seem to be present.
You don’t have to know all the answers to say “you’re (almost certainly) wrong.”
Well, I certainly don’t have to know all the answers in order to think that. But my brand of atheism tells me that I ought to have at least some of the answers before saying that.
There are no chemically sound models for creating life from Miller-Urey building blocks. [...] There are some models which have life starting with RNA, and some which have life starting with lipids, or iron-sulfide minerals or even (pace Tim) starting with clay. But you didn’t mention those more recent and plausible theories.
Thanks for catching me in this error. I was very vaguely familiar with those theories, but not enough to realize that they require source materials not available from Miller-Urey building blocks.
The problem I see is not so much with the source materials or “building blocks”. It is putting them together into something that reproduces itself. When Miller performed his experiment, we had no idea how life worked at the mechanical level. Even amino acids seemed somehow magic. So when Miller showed they are not magic, it seemed like a big deal.
Now we know how life works mechanically. It is pretty complicated. It is difficult to imagine something much simpler that would still work. Putting the “building blocks” together in a way that works currently seems “uphill” thermodynamically and very much uphill in terms of information. IMHO, we are today farther from a solution than we thought we were back in 1953.
But, isn’t the issue not only the amount of information required but also the amount of time and space that was available to work with?
To pick one scientific paper which I think summarizes what you’re talking about, this paper discusses “[...][t]he implausibility of the suggestion that complicated cycles could self-organize, and the importance of learning more about the potential of surfaces to help organize simpler cycles[...]”.
The chemistry discussed in that paper is well above my head, but I can still read it well enough to conclude that it seems to fallaciously arrive at probabilistic-sounding conclusions (i.e. “To postulate one fortuitously catalyzed reaction, perhaps catalyzed by a metal ion, might be reasonable, but to postulate a suite of them is to appeal to magic.”) without actually doing any probability calculations. It’s not enough to point out that the processes required to bootstrap a citric acid cycle are unlikely; how unlikely are they compared to the number of opportunities?
Am I missing something important? The above is my current understanding of the situation which I recognize to be low-level, and I present it primarily as an invitation for correction and edification, only secondarily as a counterargument to your claims.
It is important to realize that Orgel is a leader of one faction (I will resist the temptation to write “sect”) and he is critiquing the ideas of a different faction. Since I happen to subscribe to the ideas of the second faction, I may not be perfectly fair to Orgel here.
Orgel does not calculate probabilities in part because the ideas he is critiquing are not specific enough to permit such a calculation. Furthermore, and this is something you would need some background to appreciate, the issue here isn’t a question of a fluke coming together somewhere here on earth of the right ingredients. It is more a matter of a fluke coming together of laws of chemistry. Orgel is saying that he doubts that the cycle idea would work anywhere in this universe—it would take a suspiciously fine-tuned universe to let all those reactions work together like that. It is a reasonable argument—particularly coming from someone whose chemical intuition is as good as Orgel’s.
I think Orgel is pretty much right. The reductive citric acid cycle is a cute idea as the core of a metabolism-first theory, but it is probably too big and complicated a cycle to be realistic as the first cycle. Personally, I think that something simpler, using CO or HCN as the carbon source has a better chance of success. But until we come up with something specific and testable, the “metabolism first” faction maybe deserves Orgel’s scorn. The annoying thing is that our best ideas are untestable because they require enormous pressures and unsafe ingredients to test them. Damned frustrating when you want to criticize the other side for producing untestable theories.
Orgel was fair to the extent that he also provided a pretty good critiqueto his own faction’s ideas at about the same time. But it is possible that Sutherland’s new ideas on RNA synthesis may revive the RNA-first viewpoint.
If you really dig watching abiogenesis research, as I do, it is an exciting time to be alive. Lots of ideas, something wrong with every one of them, but sooner or later we are bound to figure it all out.
“Ok, you can believe that, but since I don’t already believe in a Creator Deity, I would prefer to believe that the first living organism on earth arose by some unknown natural process. In fact, I have some ideas as to how it might have happened.”
That’s about what I would say in the same situation, though I might go on to say that I “prefer to believe” in that hypothesis because its probability of truth seems high enough, although it is not as probable as more established theories such as common descent.
Let’s taboo “faith” from here out because otherwise I think we’re likely to fall into a definitional argument. My next question is: what actions do you feel are justified by the probability of the naturalistic abiogensis hypothesis, and why isn’t the flat statement “Life on Earth came about naturally” part of that set?
Let’s taboo “faith” from here out because otherwise I think we’re likely to fall into a definitional argument.
Ok by me.
My next question is: what actions do you feel are justified by the probability of the naturalistic abiogensis hypothesis,
Actions? I don’t need no steenkin’ hypothesis to justify actions… [Sorry, just watched the movie]
To be honest, I don’t see that it makes much difference to me whether life on earth arose spontaneously, or by directed panspermia, or as a once-in-a-multiverse fluke, or by the direction of some Omega running a sim. My actions are the same in any case. It is a fascinating question, though, even if it matters so little.
and why isn’t the flat statement “Life on Earth came about naturally” part of that set?
You are asking why I am not justified in coming out and saying it? But I am justified. I am justified in saying flatly that life on earth came about naturally. The statement is justified by my
By “taboo” I meant this LW meme which requires that you not just replace taboo’ed words by alternate symbols, but with working definitions. So, I’m still curious about your last paragraph: how is that statement justified? Why do you, why should you, feel comfortable saying and believing it? Should that comfort level be greater than you’d have saying “Earth-life was created by directed panspermia”?
Well, recall that the tabooed word is one which I sought to apply both to the theist “goddidit” and to the atheist “unknown-natural-processes-didit”. So what definition fits that word?
So how about this: “I make that statement because no other possibility fits into my current worldview, and this one fits reasonably well”. Or, if the taboo be removed, “I can’t prove it to your satisfaction. Hell, I can’t even prove it to my satisfaction. Yet I believe it, and I consider it a reasonable thing to believe. I guess I am simply taking it on faith.”
Why not just have an amount of belief proportional to the amount of evidence? That is, wouldn’t it be more rational to say “I think naturalistic self-organized abiogenesis is the most plausible solution known, and here’s why, but I’m not so confident in it that I think other possible solutions (including some we haven’t yet thought up) are implausible” and skip all this business about worldviews and proof? Proof isn’t really all that applicable to inductive reasoning, and I’m very skeptical of the idea that “X fits with my worldview” is a good reason for any significant amount of confidence that X is true.
Why not just have an amount of belief proportional to the amount of evidence?
Because, as a Bayesian, I realize that priors matter. Belief is produced by a combination of priors and evidence.
Proof isn’t really all that applicable to inductive reasoning,
Sure it is. Proof is applicable in both deductive and inductive reasoning. What you probably meant to say is that proof is not the only thing applicable to inductive reasoning.
I think that you will find that most of the reasoning that takes place in a field like abiogenesis has more of a deductive flavor than an inductive one. There just is not that much evidence available to work with.
and I’m very skeptical of the idea that “X fits with my worldview” is a good reason for any significant amount of confidence that X is true.
Well, then how do you feel about the idea that “X does not fit with my worldview” is a good reason for a significant amount of skepticism that X is true?
Seems to me that just a little bit ago you were finding a nice fit between X = “Miller-Urey-didit” and your worldview. A fit so nice that you were confident enough to set out to tell a total stranger about it.
Well, then how do you feel about the idea that “X does not fit with my worldview” is a good reason for a significant amount of skepticism that X is true?
I was thinking of “worldview” as a system of axioms against which claims are tested. For example, a religious worldview might axiomatically state that God exists and created the universe, and so any claim which violated that axiom can be discarded out of hand.
I’m realizing now that that’s not a useful definition; I was using it as shorthand for “beliefs that other people hold that aren’t updatable, unlike of course my beliefs which are totally rational because mumble mumble and the third step is profit”.
Beliefs which cannot be updated aren’t useful, but not all beliefs which might reasonably form a “worldview” are un-Bayesian. Maybe a better way to talk about worldviews is to think about beliefs which are highly depended upon; beliefs which, if they were updated, would also cause huge re-updates of lots of beliefs farther down the dependency graph. That would include both religious beliefs and the general belief in rationality, and include both un-updateable axiomatic beliefs as well as beliefs that are rationally resistant to update because a large collection of evidence already supports them.
So, I withdraw what I said earlier. Meshing with a worldview can in fact be rational support for a hypothesis, provided the worldview itself consists of rationally supported beliefs.
Okay, with that in mind:
Seems to me that just a little bit ago you were finding a nice fit between X = “Miller-Urey-didit” and your worldview. A fit so nice that you were confident enough to set out to tell a total stranger about it.
My claim that Miller-Urey is support for the hypothesis of life naturally occurring on Earth was based on the following beliefs:
The scientific research of others is good evidence even if I don’t understand the research itself, particularly when it is highly cited
The Miller-Urey experiment demonstrated that amino acids could plausibly form in early Earth conditions
Given sufficient opportunities, these amino acids could form a self-replicating pseudo-organism, from which evolution could be bootstrapped
Based on what you’ve explained I have significantly reduced my confidence in #3. My initial confidence for #3 was too high; it was based on hearing lots of talk about Miller-Urey amino acids being the building blocks of life, when I had not actually heard of specific paths for such formation that are confidently accepted by experts in the field as plausible.
Okay, so my conclusion has been adjusted (thanks!), but to bring it back to the earlier point: what about worldviews? Of the above, I think only #1 could be said to have to do with worldviews, and I still think it’s reasonable. As with your stereo amplifier example, even though I may not know enough about a subject to understand the literature myself, I can still estimate fairly well whether people who do claim to know enough about it are being scientific or pseudo-scientific, based on testability and lack of obviously fallacious reasoning.
Mis-application of that principle led me to my mistake with #3, but I think the principle itself stands.
Beliefs which cannot be updated aren’t useful, but not all beliefs which might reasonably form a “worldview” are un-Bayesian. Maybe a better way to talk about worldviews is to think about beliefs which are highly depended upon; beliefs which, if they were updated, would also cause huge re-updates of lots of beliefs farther down the dependency graph.
Yes.
Beliefs have hierarchy, and some are more top-level than others. One of the most top-level beliefs being:
a vast superintelligence exists
it has created/effected/influenced our history
If you give high weight to 1, then 2 follows and is strengthened, and this naturally guides your search for explanations for mysteries. A top-level belief sends down a massive cascade of priors that can effect how you interpret everything else.
If you hold the negation of 1 and or 2 as top-level beliefs then you look for natural explanations for everything. Arguably the negation of ‘goddidit’ as a top-level belief was a major boon to science because it tends to align with ockham’s razor.
But at the end of the day it’s not inherently irrational to hold these top-level beliefs. Francis Crick for instance looked at the origin of life problem and decided an unnatural explanation involving a superintelligence (alien) was actually a better fit.
A worldview comes into play when one jumps to #3 with Miller-Urey because it fits with one’s top-level priors. Our brain is built around hierarchical induction, so we always have top-level biases. This isn’t really an inherent weakness as there probably is no better (more efficient) way to do it. But it is still something to be aware of.
But at the end of the day it’s not inherently irrational to hold these top-level beliefs. Francis Crick for instance...
But, I don’t think Crick was talking about a “vast superintelligence”. In his paper, he talks about extraterrestrials sending out unmanned long-range spacecraft, not anything requiring what I think he or you would call superintelligence. In fact, he predicted that we would have that technology within “a few decades”, though rocket science isn’t among his many fields of expertise so I take that with a grain of salt.
A worldview comes into play when one jumps to #3 with Miller-Urey because it fits with one’s top-level priors.
I don’t think that’s quite what happened to me, though; the issue was that it didn’t fit my top-level priors. The solution wasn’t to adjust my worldview belief but to apply it more rationally; I ran into an akrasia problem and concluded #3 because I hadn’t examined my evidence well enough according to even my own standards.
The scientific research of others is good evidence even if I don’t understand the research itself, particularly when it is highly cited
Yeah, it sure sounds like a reasonable principle, doesn’t it? What could possibly be wrong with trusting something which gets mentioned so often? Well, as a skeptic who, by definition rejects arguments which get cited a lot, what do you think could be wrong with that maxim? Is it possibly something about the motivation of the people doing the citing?
What could possibly be wrong with trusting something which gets mentioned so often?
The quality of the cites is important, not just the quantity.
It’s possible for experts to be utterly wrong, even in their own field of expertise, even when they are very confident in their claims and seem to have good reason to be. However, it seems to me that the probability of that decreases with how testable their results are, the amount and quality of expertise they have, and the degree to which other experts legitimately agree with them (i.e. not just nodding along, but substantiating the claim with their own knowledge).
Since I’m not an expert in the given field, my ability to evaluate these things is limited and not entirely trustworthy. However, since I’m familiar with the most basic ideas of science and rationality, I ought to be able to differentiate pseudo-science from science pretty well, particularly if the pseudo-science is very irrational, or if the science is very sound.
That I had a mistaken impression about the implications of Miller-Urey, wherein I confused pop-science with real science, decreases my confidence that I’ve been generally doing it right. However, I still think the principles I listed above make sense, and that my primary error was in failing to notice the assumption I was making re: smoke → fire.
Excellent summary, I think. I have just a few things to add.
… it seems to me that the probability of that decreases with how testable their results are …
A claim that the (very real) process that Miller discovered was actually involved in the (also very real, but unknown) process by which life originated is pretty much the ultimate in untestable claims in science.
… the amount and quality of expertise they have …
In my own reading in this area, I quickly noticed that when the Miller experiment was cited in an origin-of-life chapter in a book that is really about something else, it was mentioned as if it were important science. But when it is mentioned in a book about the origin of life, then it is mentioned as intellectual history, almost in the way that chemistry books mention alchemy and phlogiston.
In other words, you can trust people like Orgel with expertise in this area to give you a better picture of the real state-of-knowledge, than someone like Paul Davies, say, who may be an expert on the Big Bang, but also includes chapters on origin-of-life and origin-of-man because it helps to sell more books.
The point of tabooing a word isn’t to replace it with a mark. The point is that it forces one to expand on what one means by the word and removes connotations that might not be shared by all people in a discussion.
Hold on, the naturalistic origin of life is pretty plausible based on current understanding (Miller-Urey showed amino acid development would be very plausible, and from there there are a number of AIUI chemically sound models for those building blocks naturally forming self-replicating organisms or pseudo-organisms).
Are you genuinely arguing that its probability is so low that it would be less productive to investigate naturalistic abiogenesis mechanisms than it would be to look for new hypotheses? Or, alternately, do you have a more specific idea of what minimum level of probability it takes for a hypothesis be “plausible” rather than “being taken on faith”?
And that claim by you is based on … what exactly? Experiments you have performed? Books you have read explaining the theory to your satisfaction with no obvious hand waving? Books like the ones we all have read describing Darwin’s theory of evolution through natural selection? Or maybe you have encountered a section in a library filled with technical material beyond your comprehension, but which you are pretty sure you could comprehend with enough effort? For me, something in this category would be stereo amplifiers—I’ve seen the books so I know there is nothing supernatural involved, though I can’t explain it myself.
From what you write below, I’m guessing your background puts you at roughly this level regarding abiogenesis. Except, the difference is that there is no library section filled with technical material explaining how life originated from non-life. So, I think you are going on faith.
No there aren’t. There is not a single plausible theory in existence right now claiming that life originates from amino acids arising from a Miller-Urey type of process. There are no chemically sound models for creating life from Miller-Urey building blocks.
There are some models which have life starting with RNA, and some which have life starting with lipids, or iron-sulfide minerals or even (pace Tim) starting with clay. But you didn’t mention those more recent and plausible theories. Instead you went on faith.
No. Not at all. I don’t have a clue as to what it would even mean to look for, let alone investigate a non-naturalistic hypothesis.
What I am saying is this: Suppose I have before me a theist who claims that a Deity must have been the cause of the Big Bang. “Something from nothing” and all that. Suppose further that my own version of atheism is so completely non-evangelical and my knowledge of cosmology so weak that I say to him, “Could be! I don’t believe that a Deity was involved, but I don’t have any evidence to rule it out.”
So that is the supposition. Next, suppose he says “Furthermore, I think the Deity must have been involved in the origin of life, back 3+ billion years ago. From what I know of chemistry and biochemistry, that could not have been spontaneous.” I would tell him that I too know quite a bit about chemistry and biochemistry, and that there are many, many clues indicating that life based on RNA probably existed before modern life based on DNA and amino acids and proteins. There is strong evidence that life in its current form came from something more primitive.
Now, suppose he says “Yes, I understand all that evidence. But you still have no theory to explain how life based on RNA might have started. All you have is handwaving. My belief, since I already believe in a Creator-Deity for the Big Bang, is that a Creator-Deity was also involved in the origin of the RNA World.” If he said that, then I would answer, “Ok, you can believe that, but since I don’t already believe in a Creator Deity, I would prefer to believe that the first living organism on earth arose by some unknown natural process. In fact, I have some ideas as to how it might have happened.”
Yes, I said “prefer to believe” this time. I don’t have a good explanation for life’s origin, though I have spent a good deal of my spare time over the past 30 years looking for one.
I think what’s happening here is that the theist’s preexisting belief about a creator god is causing them to privilege the hypothesis of divine RNA-creation.
The trouble is, you seem to be privileging it too. The way you’ve set up the scenario makes it seem like there are two hypotheses: (1) goddidit, (2) some unknown natural process.
But (2) is actually a set of zillions of potential processes, many of which have a far better prior than Yahweh and the Thousand Claims of Scripture, even if we can’t actually choose one for sure right now. Taken together, all their probability mass dwarfs that of the goddidit hypothesis.
You don’t have to know all the answers to say “you’re (almost certainly) wrong.”
I don’t see any reasons why (2) - unknown natural process—gets to benefit from being a “set of zillions of potential processes, many of which have a far better prior” and (1) - goddidit—does not.
If you want to sum the probability of a hypothesis by performing some weighted sum over the set of zillions of it’s neighbors in hypothesis space, that’s fine. But if that is your criteria, you need to apply it equally to the other set of hypothesises you are considering—instead of considering only one specific example.
Bostrom’s simulation argument gives us one potential generator of ‘goddidits’, and a likely high prior for superintelligent aliens gives us another potential generator of ‘goddidits’. Either of those generators could spawn zillions of potential processes which have far better priors than Yaweh, but could look similar.
None of this leads to any specific conclusion—I’m just pointing out an unfairness in your methodology.
You’re quite right. When I said this, I was thinking of “goddidit” as a set of very specific claims from a single religious tradition, which I should’ve stated.
Mmm actually you did state it as a fairly specific claim. I’m just saying one can’t fairly compare highly specific complex hypothesizes vs wide general sweeps through hypothesis-space. This is itself a good argument against the specific “yawheh did it”, but not against the more general “goddidit” which you originally were referring to:
You’re right—but there is another side to this coin. An atheist has a top-level belief (or it’s negation) which sends down cascading priors and privileges naturalistic hypothesizes. So far this has worked splendidly well across the landscape.
But there is no guarantee this will work everywhere forever, and it’s at least possible that eventually we may flip or find an exception for the top-level prior—for example we may eventually find that pretty much everything has a naturalistic explanation except the origin of life—which turns out to have been seeded by alien super-intelligence (ala Francis Crick) - for example.
I entirely agree. While I don’t know of any good reasons to think the origin of life was not a happy accident, it is not inconceivable a priori (simulations, seeding etc.).
When I describe myself as an atheist (which I try not to do), I really mean that
(1) all the anthropomorphic creation myths are really laughable,
(2) there’s not much positive evidence for less laughable creators, and
(3) even if you showed me evidence for a creator, I would be inclined toward what I will call meta-naturalism—i.e., still wanting to know how the hell the creator came to be.
Basically, I doubt the existence of gods that are totally ontologically distinct from creatures.
Bostrom’s simulation argument does NOT give us a generator of “goddidits” regarding the origin of life and the universe, because implicit in the question “How did life originate?” is a desire to know the ultimate root (if there is one), and us being in a simulation just gives us some more living beings (the simulators “above”) to ask our questions about. Where did life in the universe “one level above us” come from? Where did our simulator/parent universe originate?
There is nothing unfair in dismissing “A MIRACLE!” in comparison to the set of plausible naturalistic processes that could explain a given phenomenon. And to second SarahC, it’s somewhat incoherent to talk about non-naturalistic processes in the first place. You need to be very clear as to what you’re suggesting when you suggest “god did it”. But, no one here is suggesting that, so I’ll stop tangenting into arguing against theists that don’t seem to be present.
Well, I certainly don’t have to know all the answers in order to think that. But my brand of atheism tells me that I ought to have at least some of the answers before saying that.
Different strokes for different folks.
Thanks for catching me in this error. I was very vaguely familiar with those theories, but not enough to realize that they require source materials not available from Miller-Urey building blocks.
The problem I see is not so much with the source materials or “building blocks”. It is putting them together into something that reproduces itself. When Miller performed his experiment, we had no idea how life worked at the mechanical level. Even amino acids seemed somehow magic. So when Miller showed they are not magic, it seemed like a big deal.
Now we know how life works mechanically. It is pretty complicated. It is difficult to imagine something much simpler that would still work. Putting the “building blocks” together in a way that works currently seems “uphill” thermodynamically and very much uphill in terms of information. IMHO, we are today farther from a solution than we thought we were back in 1953.
But, isn’t the issue not only the amount of information required but also the amount of time and space that was available to work with?
To pick one scientific paper which I think summarizes what you’re talking about, this paper discusses “[...][t]he implausibility of the suggestion that complicated cycles could self-organize, and the importance of learning more about the potential of surfaces to help organize simpler cycles[...]”.
The chemistry discussed in that paper is well above my head, but I can still read it well enough to conclude that it seems to fallaciously arrive at probabilistic-sounding conclusions (i.e. “To postulate one fortuitously catalyzed reaction, perhaps catalyzed by a metal ion, might be reasonable, but to postulate a suite of them is to appeal to magic.”) without actually doing any probability calculations. It’s not enough to point out that the processes required to bootstrap a citric acid cycle are unlikely; how unlikely are they compared to the number of opportunities?
Am I missing something important? The above is my current understanding of the situation which I recognize to be low-level, and I present it primarily as an invitation for correction and edification, only secondarily as a counterargument to your claims.
It is important to realize that Orgel is a leader of one faction (I will resist the temptation to write “sect”) and he is critiquing the ideas of a different faction. Since I happen to subscribe to the ideas of the second faction, I may not be perfectly fair to Orgel here.
Orgel does not calculate probabilities in part because the ideas he is critiquing are not specific enough to permit such a calculation. Furthermore, and this is something you would need some background to appreciate, the issue here isn’t a question of a fluke coming together somewhere here on earth of the right ingredients. It is more a matter of a fluke coming together of laws of chemistry. Orgel is saying that he doubts that the cycle idea would work anywhere in this universe—it would take a suspiciously fine-tuned universe to let all those reactions work together like that. It is a reasonable argument—particularly coming from someone whose chemical intuition is as good as Orgel’s.
I think Orgel is pretty much right. The reductive citric acid cycle is a cute idea as the core of a metabolism-first theory, but it is probably too big and complicated a cycle to be realistic as the first cycle. Personally, I think that something simpler, using CO or HCN as the carbon source has a better chance of success. But until we come up with something specific and testable, the “metabolism first” faction maybe deserves Orgel’s scorn. The annoying thing is that our best ideas are untestable because they require enormous pressures and unsafe ingredients to test them. Damned frustrating when you want to criticize the other side for producing untestable theories.
Orgel was fair to the extent that he also provided a pretty good critiqueto his own faction’s ideas at about the same time. But it is possible that Sutherland’s new ideas on RNA synthesis may revive the RNA-first viewpoint.
If you really dig watching abiogenesis research, as I do, it is an exciting time to be alive. Lots of ideas, something wrong with every one of them, but sooner or later we are bound to figure it all out.
That’s about what I would say in the same situation, though I might go on to say that I “prefer to believe” in that hypothesis because its probability of truth seems high enough, although it is not as probable as more established theories such as common descent.
Let’s taboo “faith” from here out because otherwise I think we’re likely to fall into a definitional argument. My next question is: what actions do you feel are justified by the probability of the naturalistic abiogensis hypothesis, and why isn’t the flat statement “Life on Earth came about naturally” part of that set?
Ok by me.
Actions? I don’t need no steenkin’ hypothesis to justify actions… [Sorry, just watched the movie]
To be honest, I don’t see that it makes much difference to me whether life on earth arose spontaneously, or by directed panspermia, or as a once-in-a-multiverse fluke, or by the direction of some Omega running a sim. My actions are the same in any case. It is a fascinating question, though, even if it matters so little.
You are asking why I am not justified in coming out and saying it? But I am justified. I am justified in saying flatly that life on earth came about naturally. The statement is justified by my
By “taboo” I meant this LW meme which requires that you not just replace taboo’ed words by alternate symbols, but with working definitions. So, I’m still curious about your last paragraph: how is that statement justified? Why do you, why should you, feel comfortable saying and believing it? Should that comfort level be greater than you’d have saying “Earth-life was created by directed panspermia”?
Well, recall that the tabooed word is one which I sought to apply both to the theist “goddidit” and to the atheist “unknown-natural-processes-didit”. So what definition fits that word?
So how about this: “I make that statement because no other possibility fits into my current worldview, and this one fits reasonably well”. Or, if the taboo be removed, “I can’t prove it to your satisfaction. Hell, I can’t even prove it to my satisfaction. Yet I believe it, and I consider it a reasonable thing to believe. I guess I am simply taking it on faith.”
Why not just have an amount of belief proportional to the amount of evidence? That is, wouldn’t it be more rational to say “I think naturalistic self-organized abiogenesis is the most plausible solution known, and here’s why, but I’m not so confident in it that I think other possible solutions (including some we haven’t yet thought up) are implausible” and skip all this business about worldviews and proof? Proof isn’t really all that applicable to inductive reasoning, and I’m very skeptical of the idea that “X fits with my worldview” is a good reason for any significant amount of confidence that X is true.
Because, as a Bayesian, I realize that priors matter. Belief is produced by a combination of priors and evidence.
Sure it is. Proof is applicable in both deductive and inductive reasoning. What you probably meant to say is that proof is not the only thing applicable to inductive reasoning.
I think that you will find that most of the reasoning that takes place in a field like abiogenesis has more of a deductive flavor than an inductive one. There just is not that much evidence available to work with.
Well, then how do you feel about the idea that “X does not fit with my worldview” is a good reason for a significant amount of skepticism that X is true?
Seems to me that just a little bit ago you were finding a nice fit between X = “Miller-Urey-didit” and your worldview. A fit so nice that you were confident enough to set out to tell a total stranger about it.
I was thinking of “worldview” as a system of axioms against which claims are tested. For example, a religious worldview might axiomatically state that God exists and created the universe, and so any claim which violated that axiom can be discarded out of hand.
I’m realizing now that that’s not a useful definition; I was using it as shorthand for “beliefs that other people hold that aren’t updatable, unlike of course my beliefs which are totally rational because mumble mumble and the third step is profit”.
Beliefs which cannot be updated aren’t useful, but not all beliefs which might reasonably form a “worldview” are un-Bayesian. Maybe a better way to talk about worldviews is to think about beliefs which are highly depended upon; beliefs which, if they were updated, would also cause huge re-updates of lots of beliefs farther down the dependency graph. That would include both religious beliefs and the general belief in rationality, and include both un-updateable axiomatic beliefs as well as beliefs that are rationally resistant to update because a large collection of evidence already supports them.
So, I withdraw what I said earlier. Meshing with a worldview can in fact be rational support for a hypothesis, provided the worldview itself consists of rationally supported beliefs.
Okay, with that in mind:
My claim that Miller-Urey is support for the hypothesis of life naturally occurring on Earth was based on the following beliefs:
The scientific research of others is good evidence even if I don’t understand the research itself, particularly when it is highly cited
The Miller-Urey experiment demonstrated that amino acids could plausibly form in early Earth conditions
Given sufficient opportunities, these amino acids could form a self-replicating pseudo-organism, from which evolution could be bootstrapped
Based on what you’ve explained I have significantly reduced my confidence in #3. My initial confidence for #3 was too high; it was based on hearing lots of talk about Miller-Urey amino acids being the building blocks of life, when I had not actually heard of specific paths for such formation that are confidently accepted by experts in the field as plausible.
Okay, so my conclusion has been adjusted (thanks!), but to bring it back to the earlier point: what about worldviews? Of the above, I think only #1 could be said to have to do with worldviews, and I still think it’s reasonable. As with your stereo amplifier example, even though I may not know enough about a subject to understand the literature myself, I can still estimate fairly well whether people who do claim to know enough about it are being scientific or pseudo-scientific, based on testability and lack of obviously fallacious reasoning.
Mis-application of that principle led me to my mistake with #3, but I think the principle itself stands.
Yes.
Beliefs have hierarchy, and some are more top-level than others. One of the most top-level beliefs being:
a vast superintelligence exists
it has created/effected/influenced our history
If you give high weight to 1, then 2 follows and is strengthened, and this naturally guides your search for explanations for mysteries. A top-level belief sends down a massive cascade of priors that can effect how you interpret everything else.
If you hold the negation of 1 and or 2 as top-level beliefs then you look for natural explanations for everything. Arguably the negation of ‘goddidit’ as a top-level belief was a major boon to science because it tends to align with ockham’s razor.
But at the end of the day it’s not inherently irrational to hold these top-level beliefs. Francis Crick for instance looked at the origin of life problem and decided an unnatural explanation involving a superintelligence (alien) was actually a better fit.
A worldview comes into play when one jumps to #3 with Miller-Urey because it fits with one’s top-level priors. Our brain is built around hierarchical induction, so we always have top-level biases. This isn’t really an inherent weakness as there probably is no better (more efficient) way to do it. But it is still something to be aware of.
But, I don’t think Crick was talking about a “vast superintelligence”. In his paper, he talks about extraterrestrials sending out unmanned long-range spacecraft, not anything requiring what I think he or you would call superintelligence. In fact, he predicted that we would have that technology within “a few decades”, though rocket science isn’t among his many fields of expertise so I take that with a grain of salt.
I don’t think that’s quite what happened to me, though; the issue was that it didn’t fit my top-level priors. The solution wasn’t to adjust my worldview belief but to apply it more rationally; I ran into an akrasia problem and concluded #3 because I hadn’t examined my evidence well enough according to even my own standards.
Yeah, it sure sounds like a reasonable principle, doesn’t it? What could possibly be wrong with trusting something which gets mentioned so often? Well, as a skeptic who, by definition rejects arguments which get cited a lot, what do you think could be wrong with that maxim? Is it possibly something about the motivation of the people doing the citing?
The quality of the cites is important, not just the quantity.
It’s possible for experts to be utterly wrong, even in their own field of expertise, even when they are very confident in their claims and seem to have good reason to be. However, it seems to me that the probability of that decreases with how testable their results are, the amount and quality of expertise they have, and the degree to which other experts legitimately agree with them (i.e. not just nodding along, but substantiating the claim with their own knowledge).
Since I’m not an expert in the given field, my ability to evaluate these things is limited and not entirely trustworthy. However, since I’m familiar with the most basic ideas of science and rationality, I ought to be able to differentiate pseudo-science from science pretty well, particularly if the pseudo-science is very irrational, or if the science is very sound.
That I had a mistaken impression about the implications of Miller-Urey, wherein I confused pop-science with real science, decreases my confidence that I’ve been generally doing it right. However, I still think the principles I listed above make sense, and that my primary error was in failing to notice the assumption I was making re: smoke → fire.
Excellent summary, I think. I have just a few things to add.
A claim that the (very real) process that Miller discovered was actually involved in the (also very real, but unknown) process by which life originated is pretty much the ultimate in untestable claims in science.
In my own reading in this area, I quickly noticed that when the Miller experiment was cited in an origin-of-life chapter in a book that is really about something else, it was mentioned as if it were important science. But when it is mentioned in a book about the origin of life, then it is mentioned as intellectual history, almost in the way that chemistry books mention alchemy and phlogiston.
In other words, you can trust people like Orgel with expertise in this area to give you a better picture of the real state-of-knowledge, than someone like Paul Davies, say, who may be an expert on the Big Bang, but also includes chapters on origin-of-life and origin-of-man because it helps to sell more books.
The point of tabooing a word isn’t to replace it with a mark. The point is that it forces one to expand on what one means by the word and removes connotations that might not be shared by all people in a discussion.