The Probability of Life

2020 Mar 27

It so happens that we exist alive here on earth. We also find ourselves asking questions about the origins of life and our environment. How did this come to be? Could our life have occurred on earth due to natural random processes? If that were so, then probability would be involved; that is: probabilities would have made life possible. note Is it reasonable, however, to think our existence happened by chance?

Note that this is not an argument regarding Darwinian evolution. It is primarily a probabilistic argument regarding chemical abiogenesis which must occur before evolution can begin. The thought experiment will even attempt to help such a process by thinking of it as a search engine that is actively looking for life and for the information that life requires.

This will be an argument about genetic information which is then used to generate biological products. DNA which stores genes has no chemically preferred sequences; it is a pure information storage system. Therefore DNA itself does not modify the odds on the information that it carries. This has been demonstrated by encoding text, pictures and video into DNA and then accurately decoding them back out again. DNA is simply a molecular information storage system. Purposeful information just like what was in the that encoding demonstration is also what is encoded into the DNA of living things.

Probabilistic reasoning about pure information is specifically appropriate here. Information theory depends on reasoning just like this. The modern computer tech that we use all the time depends on probabilistic methods for successful information management.

Importantly, natural processes uniformly degrade information. Information systems that do not account and correct for this will fail. Self-replicating systems require built-in error detection and correction mechanisms. The argument here, however, will mostly ignore this issue and focus on the problem of finding and gaining useful information. Still, without error handling, the rate of information loss by random processes greatly exceeds all information gain. So, this information retention problem strengthens the arguments here by a substantial amount.

By conclusion, if abiogenesis is unable to happen, then evolution never would have happened afterwards. Since we are alive, if we can infer that life itself could not have happened by natural processes then we must inevitably accept that it must have come to existence by a non-materialistic means.

Living things are complex information management systems. Information of this type according to all uniform human experience only comes into existence from an intellect. Therefore, by logical reasoning, the most likely cause for this information would be from an agency external to the natural material universe. If this is so, then it is also a reasonable idea that speciation may have been caused similarly by the same external agency. And this then is not Darwinian.

A useful way to think of this is to imagine the universe itself had always been searching for life. The chances of life occurring in this universe however are finite. What then are the probabilistic resources available to our universe that could account for whether it could discover life? note Was life inevitable, or was it implausible?

The multiverse has been suggested to make our existence be more likely. See The Multiverse God. I suggest however that my arguments will make even a multiverse be probabilistically insufficient and useless.


  • It is estimated that the universe is 13.8 billion years old, which is less than 10^18 seconds. (That is a 1 with 18 zeros after it, or also written as 1e18.) note
  • The smallest unit of time that exists is the Planck time interval, of which there are less than 10^43 each second (1 with 43 zeros after, or 1e43).
  • There are an estimated 1e82 atoms in the known observable universe. note

How Many Atoms Are There in the Universe? Estimated from 1e78 to 1e82 atoms in the universe.

When we write 1e18, the 'e' marks the start of the exponent, which basically is the number of digits (the zeros) after the first whole digit in the number. So, 1e18 is written out as 18 zeros after the 1, which is 1,000,000,000,000,000,000.

The phrase “orders of magnitude” is another way of talking about this number of zeros.

Remember also: when multiplying numbers with exponents, the rule is that the exponent values are added.

There are about 1e43 increments of time in each second, and conservatively there have been about 1e61 increments of time since the beginning of the universe. note Every atom has had no more than 1e61 individual moments through all time. And cumulatively, all atoms in the known universe would have together had 1e143 moments. That is, there have been no more than this total of momentary events. If then every atom in the known universe had been involved continuously as fast as possible from the beginning in an active search for some combination of stuff that was alive, the whole universe would have had a total of about 1e143 search chances. note If the complexity of life is less than this number, then there is some probability that life would have accidentally occurred, and that life could have been found by the universe. note

Note that some people imagine that life on earth was seeded from elsewhere in the universe. (This is called panspermia.) This number of chances for life appearing in the universe completely includes all the chances of life from panspermia, so panspermia is useless for the problem of life.

4.35e17 seconds x 1.9e43 intervals x 1e82 atoms = 8.26* 10^(18+42+82) = 8.26e142 is less than 1e143 chances. This number includes all chances for life to appear over all time anywhere in the universe.

This also is generously giving each moment the ability to recognize a successful search candidate even though recognition like that is not in these momentary events. An external recognition would actually be required because the evolutionary process of natural selection does not exist before life and/or self-replication.

I will be generous and conservatively give more chances to the universe than it actually has.

So actually there would be 4.35e17 seconds since the beginning of time. Wikipedia: Planck units And actually there would be 1.9e43 Planck time intervals in a second. Together these would be a conservative total of 1e61 time increments since the very beginning.

What does it take to make a living biological organism? There are many necessary things, only one of which is a collection of essential genes. Genes, roughly speaking, are sections of information in our genetic code that can convert to the making of a protein. There is no standard count of genes because the counts are different between types of organisms. However, by some estimates it might be that as few as 256 genes would be enough for a simple organism note. It would need every last one of these unique genes to maintain life. This would mean that at minimum the organism would need 256 uniquely different functional proteins together at the same time to maintain life.

Wikipedia: Minimal genome

What are proteins? note They are biomolecules that perform functions in the cell. Proteins are the physical or chemical engines of the cell and participate in virtually every process in the cell. They start as a long chain of amino acids assembled in a specific sequence from the instructions in a gene. After a specific chain is assembled, the protein is folded into a specific three-dimensional shape that will enable it to perform functions.

Wikipedia: Protein

Proteins come in different sizes, but are commonly several hundred amino acids long. The correctness of their assembly is critical to their functionality. We know this because mutations that change the sequence in a protein consistently result in broken functions. Are there lots of sequence combinations that turn out to be useful, or however are useful protein sequences rare? We will look at this question in two ways.

One study looking at this question constructed proteins of random sequence and tested them for a minimal function note. The simple test of functionality that was used was if the random protein would attach itself to another important molecule called ATP. Of the initial random proteins, only 0.1% would bind to ATP. This means that this very primitive functionality existed in only 1 of 1000. (And all the 999 rest had no function.) If our organism requiring a minimum of 256 genes had proteins that were each only this rare, then the probability of randomly finding a complete set of genes that would make the necessary functional proteins would be at least (1/1000)^256 which is 1 out of 1e768. This unfortunately is 625 orders of magnitude too rare for the universe’s total probabilistic search resources. note.

Again, orders of magnitude is the size of the exponent in scientific notation. So 100 = 1e2 is two orders of magnitude and 13.8 billion = 13,800,000,000 = 1.38e10 has ten orders of magnitude.

The universe has 1e143 chances to find life. However, the likelihood of finding life by the measure here is 1 out of 1e768. (That is 1 followed by 768 zeros!) This gap is 768 - 143 = 625 orders of magnitude. It is 1e625 times too improbable to have happened.

Another way of studying this is to see how many mutations a protein can withstand before it looses its intended functionality. A couple of studies using this method note came to the similar conclusion that the rarity of properly working functional proteins is on the order of 1 out of 1e77. note Our theoretical all-searching universe at best then has the probabilistic resources to only ever find one isolated functional protein. Two coordinated functional proteins, however, are 11 orders of magnitude too improbable to ever be found. note And we know that most of the elementary functions of life for all organisms happen in complex coordinated systems, requiring the coordination of more than just two proteins. note

Living systems are highly integrated, and important life processes depend on coordinated function. Said another way: life fails if the protein functions are not coordinated.

There are some functions are done by single protein engines. However, there are very many cases where multiple proteins coordinate to complete a function. Hemoglobin, which carries oxygen in our blood, is made of four protein units. A more dramatic example is the bacterial flagella which is a propulsion organelle composed of 20 different proteins; and it requires every one of them for correct function.

A requirement for two coordinated proteins multiplies the odds against it. 1e77 x 1e77 which is 1e77^2 = 1e154   Since the chances in the universe = 1e143, then 154-143 = 11 orders of magnitude implausible in our whole universe.

Other papers suggest proteins are not quite so rare at 1 in 1e40. This lessor rarity could be true for selected proteins because they come in a large range of lengths. However, the argument will still hold because even this smaller value is still strikingly rare. The all-searching universe could at best find three coordinated proteins of this rarity 1e40 x 3 = 1e120 which is less than 1e143.

However, the rarity of the full genome would be 1e40^256 = 1e10240, which is 10240 - 143 = 10097 orders of magnitude implausible.

See also New Book: New Proteins Evolve Very Easily

A Dentist in the Sahara: Doug Axe on the Rarity of Proteins Is Decisively Confirmed. “… only 1 in 1e77 sequences corresponds to a functional fold/domain within the protein…”

Estimating the prevalence of protein sequences adopting functional enzyme folds “… implies the overall prevalence of sequences performing a specific function by any domain-sized fold may be as low as 1 in 1e77…”

By this more realistic metric of protein rarity (1 out of 1e77), the complete set of 256 genes to make the required usefully functional proteins had 1 chance out of 1e19712 of occurring by accident. note Compare this to the total of 1e143 available search chances in our universe (19712 - 143). This is 19569 orders of magnitude too rare for the entire lifetime of the universe. It is beyond plausible to say that this minimal genome could have occurred through time by natural processes.

1e77^256 = 1e19712

Theorists on abiogenesis suggest instead that only a single (modest) self-replicating molecule needed initially to have happened by chance. From there, evolutionary processes would be able to take over to produce the type of life organisms that we see today. Some RNA strings are capable of both storing information in the molecule and catalyzing chemical processes similar to the work of proteins. Although RNA does neither thing well, some RNA is functional in this way by some small amount. Therefore, in the opinion of some theorists this is sufficient to demonstrate a possible path from chemicals to life. note

Chemical abiogenesis is hopeful self-deception because the chemistry is adverse to success - in an RNA world or in any other chemistry. There is poisoning of the reaction with by-products, lack of purification of the reaction products and lack of protection from destructive environments. Even if some researcher intelligently determines a candidate for a self-replicating molecule, this does not make abiogenesis be more likely because the adverse chemistry still dominates in all natural contexts. So this chemical approach ignores that chemistry always works to make such a process fail.

This is even acknowledged in the famous Miller-Urey experiment. The molecular products that they were able to generate had to be rescued from the source materials to avoid ruin. No such laboratory rescue would be available to any naturalistic abiogenesis.

In addition to the raw chemistry, unless this RNA would have virtually perfect reproduction, its copies very quickly degenerate and become non-viable. This is what lab experiments always show. This is a practical re-statement of the original problem.

Note that philosophically, even if an imagined chemistry of self-replicators could be achieved, they do not change the context because they have no instinct to survive like living things do. They will just as “happily” die (that is: fail) as they will live. They won’t try to live, so they won’t survive. The “will” of living things to continue living is a non-trivial component of life.

Machines that self-replicate have never been demonstrated in spite of concerted research and the encouragement of science fiction. note A machine that could self-replicate must be able to obtain all its raw materials by itself, and then duplicate all the parts that make it as well as all the information that it contains. This requires mastery of multiple disciplines, the ability to fabricate multiple materials, and the ability to do self-assembly. This is incredibly difficult. It is no wonder that the only known self-replicating entities are living organisms. Much of the complexity of life is required to service replication. Self replication is the pinnacle feature of living things. Or, said another way: self replication demands the complexity of life.

As I heard Elon Musk say in an interview, a repeatable process to make a space rocket is an order of magnitude more difficult than making one rocket, and a machine (a factory) to make make rockets is again an order of magnitude more difficult. He was not even suggesting that the rocket was a self-replicator.

Living things can’t be wasting anything in the process of replication. Natural selection already assures this to be true (because wasted resources reduce fitness). Living things must already be at an evolutionary point of greatest simplicity; any further simplification would not be not functionally viable. Therefore this central tenet of evolution argues that no modest molecule could ever be a self-replicator.

Note also that the simplicity of a self-replicating machine will be inversely proportional to its ability to survive any modification. Said another way, the more simple the machine is, the less it is capable of surviving modifications. This can be seen in the competitions to write short self-replicating computer programs. The shorter programs will completely fail if even a single smallest bit anywhere is mutated. The same would be true of a simple replicating molecule. Changing any atom in it will stop it from replicating. It would be incapable of withstanding any modifications, therefore it would be incapable of evolving.

Evolutionary theory assumes that the entities involved can be incrementally changed from one form to another, but uniform repeated experience shows that does not work. You cannot change a good car into a good submarine by increments that consistently add functional advantage note Fitness for one category of function is never bridged to a different category of function by a single step, and therefore it will not be by a path of gradually improving fitness. The path from one high point of fitness to another goes across deep valleys, and natural selection actually prevents any crossing. Fitness only exists at rugged isolated peaks of competence that are separated by incompetence, failure and death. The path from chemical self-replicator to life is guaranteed to pass through these wastelands. Evolution is incompetent to get through.

This was not meant to be an argument about the adequacy of evolutionary processes in living things. However since evolutionary processes are suggested to function in pre-living chemistry, the problems on the theory must be addressed. The model for transition of a car to submarine is analogous to the assumed evolutionary transition from land animal to whale. The non-living thought experiment (the car to submarine) shows the extreme unlikeliness of a successful path between the two such types in living entities and the very certain lack of fitness in the intermediaries.

Changes like this had to have occurred in any path from a first self-replicating molecule to LUCA (last universal common ancestor). Therefore, the necessary path must go through states that are non-viable.

It is theorized that with a self-replicator the evolutionary processes will modify the context so that it is no longer limited by probability; it will then perform better than just by chance. However, this is a false hope because evolutionary processes are slower than chemical processes. Once evolution is involved there is both information processing time as well as chemical reaction time involved. In a proposed evolutionary path toward life, certainly it is critical to retain the information gained between generations. However, evolution must wait for mutations at the rate of generation cycles, and this is much slower than chemical reactions and it diminishes the possibility of success. Evolution becoming involved will actually make the chances become substantially smaller for the universe to successfully produce life. note

Even if a simple self-replicating molecule existed since the beginning of the universe, and even if it was trying combinations of improvements as fast as possible (at Planck intervals), it would (by definition of abiogenesis) still have to be able progress up to the complexity of life as we observe it. The same rarity argument from before still holds. There would still only be 10e143 opportunities to come up with anything in the known universe. However as previously pointed out, the rarity of life from an over-simplified view of proteins is 625 orders of magnitude too rare to be produced in our material world, and from a more realistic view of protein rarity, existing life is 19569 orders of magnitude too rare to exist.

In the thought experiment of a life-searching universe, the mechanism for recognizing successful searches was given for free without a mechanism. Here, evolution “taking over” is finally just taking on that necessary task. This means that the context has not changed for the better from the original thought experiment. In fact it is much worse because the free recognition mechanism was up to full Planck speed, but an evolutionary mechanism will have a recognition interval on the order of seconds to years and generally not do a very good job of recognizing fitness.

A self-replicating molecule to solve the problem of abiogenesis is a profoundly naive self-deception.

Next, evolutionary processes must have been capable to achieve the information that we know is contained in the common biological toolkit (CBT) which all living organisms share. note This particular information is so rare that there is only one singular toolkit for all known forms of life. This means the CBT code has never been able to evolve into something different since evolution’s last universal common ancestor (LUCA) at the very beginning of life. (Meaning all mutations away from the CBT have ended in death.) note This re-emphasizes the ruggedness of the fitness landscape. Natural selection is incapable of incrementally finding this toolkit information because it is evident that there is zero fitness without the complete CBT. This also means no fitness test could indicate incremental progress toward a functioning CBT. All fitness results would be a complete zero until suddenly (with the full CBT) everything is fit all at once.

Note that if the CBT was only something ordinary, then more than one variation or kind of toolkit should have evolved. That fact that there is only one CBT is evidence that the CBT is spectacularly special.

Essentially, that which I am calling a Common Biological Toolkit (CBT) is the functional basis for all living things. On evolutionary theory it must have existed since the last universal common ancestor (LUCA). It includes DNA expressed through RNA into proteins, metabolization of fats and sugars, ATP as an energy intermediate, a cell enclosed by a lipid membrane, ion transporters across the cell wall, cellular division, and etc.

Wikipedia: LUCA

Since it is complex, the CBT fills a substantial part of our theorized minimum genome of 256 genes. This re-argues that the minimum 256 genes must be correct all at once for life, and any supposed precursor that might be missing any part of the CBT is actually non-viable for life. However we have seen already the insane impossibility of this genome set.

The CBT itself also re-argues that any type of self-replicator needs complexity equivalent to the CBT. This argument definitely includes the chemical self-replicators. So it shows that no simple molecule could ever do this function. Insanely impossible complexity is again required right at the beginning. note

The thoroughgoing materialist will say that life as we know it could have been different if time was run again from the beginning. They don’t have any attachment to life as it exists now. Instead they speculate materialistically that with a new run though time, life could have taken a different path and a different kind of life would have emerged. Therefore they let arguments based on the practical problems of abiogenesis or of evolution just roll off their back. This is because they see observable life as just an example of the life-developing process. They don’t take the view that that the process is defined by life as it does exist in our history.

Even if every mechanism they propose for explaining life development within their process ends up being invalid because of an obstacle, that also is not perceived as a problem. These were just potential mechanisms that could be taken up or discarded, and discarding a mechanism is not perceived as evidence against abiogenesis or evolution. Evolution as Fact

However, there is zero evidence for any kind of life other than what we know. A concept for a type of life other than what is observed is only speculation, and there is no scientific evidence for it. Such speculation is non-scientific and it is metaphysical.

In fact due to the massive evidence for fine-tuning of the universe, in my thinking it is more likely that there simply are no other viable chemistries for life other than what we already know. Life depends on remarkable characteristics of carbon, oxygen, and water and others. These were established already at (or before) the beginning of our universe. Therefore, there can be no naturalistic explanation for how these physical characteristics came to be so uniquely ideal for life.

If there is no other life other than what is known, and there can be no other life other than what is known, and life is unexplainable based on what is known of naturalistic processes, then most likely life was specially made here on earth. As the character Sherlock Holmes said, “How often have I said to you that when you have eliminated the impossible, whatever remains, however improbable, must be the truth?”

So, there is no naturalistic path that could proceed to life. Natural laws do include the laws of probability, and probability indicates that abiogenesis could not have happened. The requirements are beyond all best-case naturalistic options. If natural processes then are not capable of getting to evolution’s putative LUCA, then instead obviously an abiogenesis miracle is being claimed for “science”. note

A better approach is to allow the option of non-naturalistic explanations. Life clearly looks designed. The universe clearly looks optimized for life as we know it. Both of these would suggest a super-intellect that optimized and designed the universe and life.

  • Note that a person allowing non-naturalistic explanations still has access to all the naturalistic explanations. They have a bigger and a better explanatory toolbox than those that claim to only use naturalistic explanations.
  • Note also that it is impossible to avoid metaphysical things in science. See Darwinian Metaphysics. Materialistic science depends on a bedrock of immaterial metaphysical assumptions. Misunderstanding this is profound ignorance.

Of course, “unlikely” things happen every day. A first example of unlikeliness is the color sequences of passing cars on a highway. We expect each new sequence collection to be unique - randomly unique. note Each is unlikely because it was not expected or predicted. However, none of these many random sequences is meaningful because any one sequence is as satisfactory as any other. No sequence like this is meaningful or useful unless I only need noise, and then they all serve equally well.

This is sometimes called Shannon information.

A second kind of example of unlikeliness is the sequence of all my experiences up to the present point. This sequence also was not predicted, and it also has some portion of randomness in it. The unique collection leads to the present, but it didn’t construct the present. That is, if anyone attempts to repeat the steps, they won’t recreate this moment. The sequence can be seen in retrospect, but it is not a plan sequence to be used again to create a future.

However, there is a third kind of unlikeliness. It is a sequence that has the power to predictably and reliably change the future, and it is very different from these two previous kinds. Consider the order in the process for making a house, or the sequence of events for launching a rocket. These are not sequences that could be expected to occur randomly. If these sequences are done wrong, then there is a problem and sometimes the effort is a spectacular failure. If they are done correctly, we can be reasonably confident of success. note This kind of unique sequence is not random data; it is purposeful specifications for making useful things or doing functional things. This is the kind of unlikely specification info that fills DNA and that is required for life. note

This third type of unlikely information is functionally specified information. This type of information is consistently a product of a mind.

The scientifically respected SETI project (Search for Extra-Terrestrial Intelligence) is based on this being true. It assumes that specified complex information is an indicator of intelligence and therefore the existence of this type of information is evidence of it coming from a mind.

Note that there are many more ways to make a process fail than there are to make it work. Therefore the fact that getting some types of unique sequences is easy is never an argument that it is therefore also easy to randomly get functionally specified sequences.

If we observe the existence of some object which is reliably made by a process, or we observe a complex task repeatedly being performed by a process, we intuitively interpret this to mean that an intelligence was involved to make the process. An ordered process to make or do the thing just wouldn't have happened by accident. The process came because of purposeful design. A house standing on some property and a rocket launching successfully into space are both artifacts of design. And in uniform repeated human experience, this kind of design only comes from an intellect.

All living things are filled with intricate co-dependent sequenced processes that construct and run the tiny machines of life. note This clearly looks designed and not random. Therefore we should apply this correct inference of design to the origin of life.

Note that this whole argument has ignored many other improbable things that are necessary for life. A world must exist for the life, with finely tuned parameters to make the physics, chemistry, temperature, radiation, energy, and everything else needed for the environment.

An organism living in this world needs more than genes. Even if a complete set of genes were to be found together by accident, they are insufficient for life because other highly complex systems and integrated components are needed at the same time and at the same place. (Examples: bi-layer lipid membrane, sugar code on membrane, cytoskeleton, regulatory control networks, heritable epigenetic system)

Note also that if I pull apart a living cell, I will have all the necessary parts all together at that time at one place, but there is no life. A living cell requires all the parts to also be in an integrated functional arrangement. However, outside their intact context, the components of life do not self-assemble and integrate.

Genes are necessary for life to exist, but genes alone are insufficient, and all the other necessary things are no less complicated than the genes.

In summary, the probability of life is insanely out of reach of the probabilistic resources of the universe. It is not reasonable to think that our existence happened by chance. Naturalistic abiogenesis is just not possible from materialistic or random processes. Simple reason shows this. However, we do exist and we are alive. The only reasonable approach is to allow that life came from a non-materialistic origin.

20200327 updated 20210102   © 2021 Larry Grove