Open Questions: The Origin of Life

In a previous post, I covered the claim that science proves there is no god.  This is one of the weaker arguments I hear when religion and science are discussed, because it's easy to disprove.  You simply have to challenge the claim, "How does science prove there is no God?  What experiments demonstrate this?"  However, in my experience most people who make this weak claim are actually thinking of a stronger atheistic claim, namely: "Science takes the place of religion.  It has now explained all the old gaps in our knowledge, and now we're just filling in the small gaps that remain."  In my last post, I talked about how the domains of science and religion don't generally overlap.  The popular conception is that science explains the how and religion explains the why of the world.  However, there is one major area of overlap for science and religion and that is creation, both of life and of the universe.  If science can explain the creation of the universe and of life, in other words if there's nothing for a creator to do, why bring God in at all?

For non-experts, this feels like a much stronger place to argue from.  It doesn't require us to disprove God directly, it just requires that we demonstrate God is unnecessary.  Sure, you can believe in a God, but it seems dumb if doing so requires you to ignore certain scientific facts in order to wedge him into your worldview so he can help out with creation even though he's not necessary.  If you went about adding extra steps just to God could be there, any rational person might look at you as preferring a false version of the world in place of a true one.  All so you can feel more comfortable with some made-up 'meaning' you've imposed, not on the real world, but on the made-up world you created when you chose to ignore proven scientific principles in favor of a story that can never be proven.  Here's a sample argument I've heard far too often:

Creationist: You can't prove there is no God!
Atheist: I don't need to.  I don't see any reason to believe in one to begin with.
Creationist: Okay, how do you explain how life formed?
Atheist: That was solved a long time ago.  Macromolecules formed in primordial oceans, RNA-based enzymes developed into self-replicating molecules, and the rest is history.
Creationist: Okay, but that doesn't explain how multi-cellular organisms developed!
Atheist: Look, I'm not going over all the theory point by point.  If your argument is that I have to believe in a creator until I've answered every possible scientific unknown it's pretty weak.  Isn't this how that faith you preach works?  I don't need to know everything, I just need to know the general outline and certain foundational ideas.

I'd like to break down this argument from both sides.  First to the creationist, in an attempt to point out and avoid this faulty line of reasoning.  Then to the Atheist, again in order to avoid the counterpart faulty line of reasoning.

Science doesn't prove God

The essential problem of the creationist's logic above is that it hinges on the idea that science can prove or disprove the existence of a creator.  I'm not sure that's true, at least in the negative sense.  This fundamental idea - that you can't disprove God - is frustrating to many atheists.  They just don't see any reason to believe in a creator, and claiming the burden of proof is on them to exhaustively disprove God is unfair.  I'm saying this as a person who does believe in God.  It's okay to say this, because it's honest.  The reason I believe in God is not because I did an exhaustive literature search and wasn't able to explain the prokaryotic to eukaryotic transition sufficiently.  I suspect the same goes for most believers.  They believe in God for one reason or another, but very few people I've met convert based off of what they see in the peer-reviewed literature.

The reason I'm stressing this is to point out what happened in the conversation above.  The atheist's side made a claim about a scientific explanation, and the creationist side accepted the claim but then demanded additional explanations, thereby 'moving the goal posts' in a way the atheist rightly suspected was not genuine.  If they explained the next phenomenon, the goal posts would move again, and again, and again ad infinitum.  The reason is because the believer isn't convinced by the argument that God has been 'explained away', even if they can't articulate why.

Self-Replicating RNA Doesn't Answer the Question

Now on to the atheists, and the error made in the above explanation.  It's wrong.  That theory does not explain how life first started, even if it is the one most commonly referenced by scientists.  It doesn't explain how life came to be any more than saying, "he works in the Oval Office" explains how Trump got elected.  It would therefore be wise to refrain from using it as an answer to creationists for why you don't believe.  (Besides, do you really need an answer?  It's fine to believe or not believe.  I'm not sure why science has to get involved to justify religion or atheism.  Can we just get science out of both sides already?)  This will make more sense if I break down the problems point-by-point:

  1. I want to preface this by stating that I understand natural selection and common descent are powerful explanations within evolutionary theory.  Biologists use these two concepts every day in their work, and no decent biologist has a problem with tracing organisms back to common ancestors, or rejects the power of selecting events to drive genetic changes within a population.
  2. Most explanation of early life revolve around non-biochemical arguments, but biochemistry is the place where the origin of life actually happened.  Everything outside a biochemical explanation is moot.
  3. The standard story is that randomly generated macro-molecules combined together over a long period of time to create life.  We can identify specific places near certain kinds of deep-sea vents where these conditions were present.  Since life arose, it follows that the random events near these vents serendipitously combined to create life.  This isn't hand-waving here, it's more like saying that if you roll two 6-sided die enough times eventually you'll roll a double-six.
  4. There are so many problems with this theory's ability to 'explain' how life developed, and they all arise from biochemistry:
    • RNA is assumed to be the first self-replicating molecule, but the complexity such a molecule would require is mathematically impossible in any random system.  Yes, impossible.  There's a point where the statistics make something so unlikely as to be unbelievable.  This isn't rolling a few dice a hundred times.  This would be like rolling enough 50-sided dice to fill the moon a couple hundred times and expecting them to all come up 23.  Proponents of this theory usually like to say that "it's okay if it's not a perfect match, it can be an imperfect copy and still work."  This is hand-waving all the important features of an RNAase (RNA enzyme) capable of self-replication:
      • It can't just make any other RNAase - itself a major feat.  It has to be able to create another RNAase that looks enough like itself that the new enzyme creates another new enzyme.
      • Even non-self-replicating RNAases are uncommon in nature for a reason: they don't work that well.  That's why cells default to amino-acids to do nearly all their work.  Amino acids are orders of magnitude more efficient.  So using a less efficient RNA-based system to create enzymes makes your probabilities orders of magnitude more complex.  You're trading, "but I only need to have RNA, not 20 amino acids and a host of other mechanisms in place" for "okay now it's ridiculously complicated and long".
      • The chemistry of what an actual RNAase would have to do here is mind-boggling.  It can't encode its own complex structure, and it can't run off a template like DNA or RNA polymerases do.  Why?  Because the moment that first hypothetical RNAase randomly forms there is no other self-assembled molecule hundreds (more likely thousands) of bases long that functions remotely similar to it. (That would be like taking dozens of randomly-shuffled sets of playing cards, setting the stack next to a second randomly-shuffled set, and expecting them to both be about the same number of cards and nearly identical.  Sure, they don't have to be exactly the same, but the degrees of freedom cause wrong answers to dwarf the number of 'right answers' by so many orders of magnitude it's ridiculous to contemplate.)  So if the RNAase can't work off a template, it has to self-replicate a complex series of bases - selecting chemically very similar molecules with some precision - in a way that's somehow baked into the structure of the molecule itself.  The complexity problem here is akin to the fuel/propulsion problem in rocketry, where adding fuel requires more fuel to lift the fuel itself into the air.  Here, adding complexity requires more complexity.  And the longer the chain is, the more likely a single-stranded RNA molecule is to spontaneously break and become useless.
      • There's no forcing mechanism here.  Yes, if we dropped our hypothetical self-replicating RNAase into solution with a bunch of RNA bases we can imagine it would create lots of copies of itself to the eventual exclusion of all others.  But natural selection is hypothesized to act in a step-wise manner.  Here, the first step is, "build an impossibly complex molecule".  Except that's not the first step, that's all the steps!  This is akin to saying that the way to build a car is, "step one, drive the car off the lot."  This is the real explanatory failure of this hypothesis.  It assumes the whole ridiculous result with no way of explaining how that result came about except 'random chance'.  This is not an explanation any more than 'an angel blessed it and it was so'.  Q: "How did life begin, did it appear out of nowhere?" A: "Actually, self-replicating RNAases randomly appeared out of nowhere. They created life."  No.
    • The next step in this hypothetical process again fails the same problems addressed above, except that the complexity mounts further and we still lack the kind of forcing mechanisms that will drive selection at any kind of speed.
      • Once RNA can replicate itself, it's going to use up all the natural RNA building blocks, leaving none left for further development; exponential growth happens fast, so we're stuck with whatever slow natural processes are driving production of new RNA bases to create new RNAases.  Except these RNA enzymes are unstable, so they won't stick around for millions of years of slow growth, and none of the natural processes proposed is fast enough to allow our hypothetical self-replicating RNA enzymes to create additional RNA enzymes capable of producing bases.
      • The synthesis of something as complex as RNA bases is itself a very complex process.  Eventually you're going to run into a complexity carrying-capacity barrier here.  (That barrier is likely long before you have self-replicating RNAases.)  So let's say it's not possible with one single molecule to also carry on these other essential functions.  What's going to carry them out?  It can't be another self-replicating RNAase, but if it's some complex RNA base replicator that doesn't also replicate itself then it's a flash in the pan molecule that doesn't make a difference.
    • Okay, what you need, then, is an RNAase that can copy itself and any other RNA sequence!  (Remember that if all it can do is copy other RNA, but not itself it does no good, since its stability is too low to keep at it long enough, and the probability of hitting another RNA-replicating RNAase is ridiculously low.)
      • *sigh* We now have another overly-complex molecule, capable of feats vastly more complicated than even the most sophisticated multi-subunit proteins can pull off.  There's no forcing mechanism, so again we're just wishing fully-formed processes into existence. 
      • An RNAase that can replicate any random RNA sequence is going to get bogged down replicating a lot of random garbage.  This would be harder than trying to play a symphony during a hurricane at the local dump.  The problem isn't just in finding the musicians and getting them to play together.  Even if they miraculously all got together and played perfectly the garbage flying around would completely drown out anything coherent they were trying to play.  Except imagine the dump is the size of the sun, and the musicians are scattered throughout the entire volume of the garbage-storm star.
    • Fine, we need some kind of barrier to enclose the RNAase.  That way, some randomly-assembled collections of RNAases could function without the 'storm' outside.
      • This step again creates more problems than it solves.  The problem with enclosing things inside a cell is that this both keeps things out and keeps things in.  Remember that RNA isn't stable, so unless you can get more supplies into the cell - which requires complex transport mechanisms - all you get in ten minutes is a mess of destabilized former miracle enzymes.  Meanwhile, if you can't get out of the cell, or divide, you can't create copies of yourself - or the copies have nowhere to go - so again you don't end up with something capable of persisting.  But the whole point of enclosing was to keep out the garbage, so if there's any permeability it has to be magically selective, or highly complex.  And division isn't simple either.
      • These enclosures are hypothesized to be random - again no forcing mechanisms here - where some naturally-occurring amphipathic substance encloses random strings of RNA and whatever else is floating around.  The probability that a self-replicating RNAase would get packaged with any other useful RNAase - let alone the dozens required for replication, membrane transport, etc. - is again too low to contemplate.  Remember that all these mechanisms in current-day cells are accomplished by proteins in multi-step processes requiring more than one protein for each step.  If you think chemistry is complicated, try watching cells do chemistry.  I can't even imagine a cell trying to limp along with RNA-based enzymes.
    • Remember that all this has to happen through chemosynthesis, since we can't get energy from the sun until we develop more complicated processes.  Only, for all the above to happen randomly would require more energy than all the energy in the mantle.  It would take longer than the heat death of the universe, if the entire life of the universe through to the heat death of the universe were played out once a second every second through the life of the universe we'd still not have enough time for this to happen through random chance.  We can't say these probabilities are astronomical, because astronomical numbers are far too small!

In other words, this theory is akin to saying, "Here's how life first started: 1. there was an RNA enzyme capable of self-replication 2. it could also copy other RNA 3. there were also RNA enzymes capable of creating the building blocks necessary for life 4. they were enclosed inside some kind of cell-membrane 5. and the mechanisms for transporting things across that membrane were present 6. and it all just happened to be in the same cell there together at the same time."  That's not an explanation of how life happened.  That's a short list of some of the things that would be required under one pretend scenario in order for early life to exist, with no explanation for how it all got there in the first place.  Saying that the elements were "some RNA bases could have randomly been generated in this environment, among all the other myriad chemicals present" is like saying "there's silicon and a lot of minerals on Earth, a fully-functioning iPhone could easily appear randomly somewhere."  It hand waves through the only important parts worth talking about.  It's mundane statements attached to outlandish claims about what some mundane elements could accomplish all on their own, with no forcing mechanisms.

The Correct Answer Is 'We Don't Know'

I'm going to stop here.  I was going to go on, but I don't think it would help to be exhaustive here for the same reasons that were outlined in the representative discussion above.  If you're not convinced that current 'theories' of the origin of life have no explanatory power about how life began I don't know that going into more detail will help.  (They're wild unsupported hypotheses, not theories, but people call them that so whatever.)  I'm not the first person to point out that biochemistry poses these huge theoretical problems.  A few evolutionary biologists have tried to throw together some half-baked explanations trying to apply the mechanisms that work at higher levels of abstraction down to a molecular level.  Those efforts have been disappointing, and entirely unconvincing.

(The biochemical case from the origin of life to humans still struggles when you add natural selection as a forcing mechanism.  Evolution just struggles to explain biochemistry from end to end, and we don't have a good way to resolve this yet.  It can explain some small changes here and there, but the bulk of molecular biology remains unexplained.)

The real answer to the question of how life first came to be, and how humans arose out of all that, is simple: we don't know.  All theories are woefully inadequate at a biochemical level of analysis.  It's okay to admit this, because it's honest.  Admitting you don't know something this fundamental doesn't create some prerequisite that you go and join the Catholic church, or go on a quest to find God.  It's simply an honest scientific admission.  I'd like believers to agree that God  isn't 'required' in science just because we don't know one or another fine-level detail.  I'd also like atheists to agree that God hasn't been squeezed out of the equation with flimsy explanations for difficult-to-explain phenomena.

Why?  Because when both sides can admit the evidence isn't on their side we can get back to the challenge of actually figuring out these hard problems.  Creationists don't need to be afraid that science is about to 'disprove' God, because it's really not even close.  Atheists shouldn't be afraid to question whether something like natural selection has the power to explain molecular biological complexity just because 'Darwin is sacred' and they need something to answer the creationists with.

Origin of Life vs. AI

Having laid all this out, I want to make a few brief comparisons to the current debate about artificial intelligence.

Seeing all these problems with the most popular explanation of how life first developed might cause you to throw up your hands and declare that it 'couldn't possibly have happened by chance.'  There are some scientists who make this claim at this level of abstraction (including one/two of the ones from the YouTube video earlier about biochemical challenges to evolution). I'm not one of them.

This same thing has happened with AI, which I would posit is farther along than theoretical work on the origins of human life and biochemical evolutionary mechanisms.

Early AI researchers demonstrated an AI capable of playing a flawless game of tic-tac-toe, then checkers, etc.  Critics claimed it couldn't play better than the best human chess players; this was solved, and the familiar cycle of moving the goalposts ensued.

Where's the line between general AI and a specialized AI capable of complex tasks? These are still categorically different things, and some people - myself included - are skeptical direct extrapolation of current principles of AI can possibly lead to general intelligence; there are some major things we're missing that prevent us from even approaching generalized AI.  Since we don't know what those things are, by definition we can only argue about how close we may or may not be to creating some generalized AI.

The answer, then, is not to throw up your hands and declare this to be impossible.  Whether creation of AI is possible is a known unknown, and we should be willing to openly state this.  It helps us to identify holes in our theories.  "We don't know how to create generalize AI capable of making decisions."  This is the same as for evolutionary theory and the origin of life, "We don't know how to explain biochemistry through natural stochastic mechanisms."  That doesn't mean give up, that means we admit what we don't know.

Holding Questions Open

The whole point of this series is to improve thinking overall by holding unsolved questions open in your mind.  I often see religious people seek refuge in current knowledge gaps to justify a requirement for a creator.  This is the same kind of foolishness as claiming AI is a myth because it can't play chess or identify a bird in a picture.  Eventually - hopefully - we find a way around those limitations.  The best the person who employs this goalpost-movement argument can hope for is to become irrelevant in the effort to push our knowledge of the field forward.

If you're a religious person, then, how can you argue the case for god when science is involved?  My recommendation is simple: don't be afraid of science, but don't try to use it to prove God either.  Science is constantly pushing out the boundaries of human understanding.  Regular readers of this blog know that this exploration by its very nature only serves to expand the number of new and interesting unanswered questions we have. There's no reason to be afraid scientific theory is about to invalidate your faith with some new discovery that will close off or 'complete' some domain of knowledge.  The opposite of that is happening with every discovery. Maybe some day, millennia from now, we'll start closing off domains of science as - finally - 'complete'.  As of now, we're not even close. If you want to believe in a creator, go ahead and do it.  You don't need science's permission to do so, and you don't have to ignore scientific discoveries out of fear.  Just don't make the mistake of chaining God down to some specific gap in our current understanding and you'll be able to participate fully in the ongoing scientific revolution without having to budge an inch in your belief in a creator.

The reason I brought in AI at the end here was because it's so similar to our lack of understanding about how life began.  Whether we can really create AI is still an open question.  We call certain algorithms AI because they're complex and capable of producing surprisingly useful results, but that's a far cry from creating something capable of actually thinking for itself about general problems.  I'm not saying we can't do it, or that trying isn't worthwhile.  So far, AI research has produced some great things.  All I'm saying is that we don't currently know whether it's possible, so making an unqualified statement like, "we will soon create AI" is folly.  Most of us understand that this kind of statement requires a qualifier or two to convey our uncertainty.

This same principle goes for statements like, "We know God can't exist because we've already proved we don't need him.  We understand how life formed, so what's the point of believing in a God who stood by and watched while all the action happened?"  We're closer to a future of AI overlords than we are to even a cursory understanding of how life formed.  It's foolish to make grand pronouncements about how science 'proves we don't need God'.  It's entirely understandable to make a declaration of faith that once these problems are worked out we will understand how life formed - without the intervention of a creator.  Saying so should come with serious qualifiers (just like the AI speculation) about how much we don't actually know in this area.

In my experience, tying atheism to science is as detrimental to the scientific process as inserting God into science.  Both introduce biases into the way a scientist thinks, keeping them from seeing clearly.  Whether it's a religious person expecting to see god at whatever depth they're working, or the atheist wanting to tie a ribbon around a theory about an important question, both can easily become distracted by their biases.  Both can think a problem is solved that remains unanswered.  Thus, a careful scientist does not to bring religion or atheism into their analysis of the experiment at hand, or of the field in general.  It's okay for this question to remain open.


  1. Great post. I've seen other posts and articles like this from time to time, mostly only on the fringe. It still seems that pointing out that the current best "theories" for the creation of life handwave a lot of astronomically unlikely (or worse as you point out) things is clearly well outside the Overton Window.

    In fact this is an interesting coincidence because I read one of these articles not that long ago and I was going to see if I could get you to read it and give me your opinion.

    It's a book review and in the review he focuses on a lot of problems, including proteins, and I'd be interested if you think any of the points actually have an answer.

    Completely off topic: I'm 3/4 of the way through the Lies of Locke Lymora, we need to talk about that book... I have some issues.

    1. Interesting book review. Ever since Behe published Darwin's Black Box, I've noticed a trend where these books turn into full-throated endorsements of Intelligent Design. I categorize this in the open questions section for two reasons:

      First, because as you suggest there's really no good answer to how proteins develop. There are some speculative ideas that lower the bar a little, but nothing that could even bring us close to the kind of biodiversity in molecular programming we observe in nature. For example, we see lots of protein families, where a previously-generated functioning protein plan is used as a template to create new functions. Sometimes the new function is related to the old function, and sometimes it's used for a new pathway. The problem is that this still requires a pre-cursor template to work off of. Evolution can't always be happening elsewhere. This is why I chose the origin of life question for this post. It happens prior to any replication mechanism being in place. The math people look at for crafting proteins through de novo mutation allows a protein to sit dormant in a cell, as stable DNA. It's still impossible to get there through random mutation without some forcing mechanism, but it's not nearly as outlandish as getting there through single-stranded RNA.

      Second, it's a bad habit to say, "We don't know how [concept] developed; therefore it must be ID/God." I know these guys craft reasoning why 'we should have seen this all along', but it's more philosophy than it is following the evidence where it leads. It's exactly the kind of argument that was leveled against Darwin himself prior to the neo-Darwinian synthesis. "Darwin can't explain Concept X; therefore this is where we can see God's intervention." This piecemeal back-and-forth about at what level we can observe the intervention of a creator (or wait, no, we can explain that without invoking god now - wait, now there's a new objection) is called God-of-the-gaps, and it's one of the weaker arguments for divinity.

      Incidentally, your comment inspired a whole new post where I defend Darwin - because I really feel like he needs a strong defense, even considering all of the above.

  2. Did you see this?

    1. I hadn't seen that. I'd seen the previous research about bases being made separately. Honestly this discovery, while significant, represents an advance about a millimeter along the path to explaining this (to my mind unlikely) hypothesis. A path that's light-years long at this pace.

      Yes it's a problem that these incredibly complex molecules that don't normally occur anywhere in nature organized in solution happen to be present. But the RNA world hypothesis has so many larger problems than this one that it has always been something of a footnote to my mind.

      I don't remember getting into the problem on random creation of nucleic acid synthesis above, other than to point out the transition problem. And the transition problem is more interesting to my mind. This is because I have confidence that someone working in a lab somewhere is eventually going to figure out the right admixture and combination of steps to create full nucleotides in "primordial conditions". This is less empiricism and more inference through association and creative storytelling (something that happens far too often in evolutionary biology; just because you can tell a creative story about something doesn't mean you have evidence that the story actually happened - that's NOT science).

      So I've basically assumed away a dozen problems for them already and it's still impossibly out of reach. This is because the fundamental problem of life is that anything self-sustaining is also a huge coordination problem. Current theory doesn't address this problem at all. There are no forcing mechanisms along the way. It simply can't all be stochastic from nothing but a bunch of nucleotides that happen to be there to a fully-functioning coordinating system of complex macromolecules before any selection takes place.

      And some of the elements you'd need to make random assembly possible (warmer environments to allow the inexplicable auto-assembly of nucleotides together) make the functioning RNA enzyme's persistence more of a problem. I'll be surprised if some creative chemist somewhere creates lab conditions where random auto-assembly of RNA bases 'naturally' occurs. But these demonstration exercises (still not empiricism) were never at the core of the critical problems with the RNA world hypothesis.

      Those problems are just too numerous. It's more than premature to be taking this skeleton hypothesis seriously.


Post a Comment

Popular posts from this blog

A better addiction

Covid-19: Epidemiology is useful