On the Origin of Life and Falsifiability

It is a splendid argument, because if something is impossible, then the probability of it happening is 0 -- or, at least, extraordinarily close to 0. So, for example, you can argue that your door has x*y dimensions, and the elephant would be too large to fit through this doorway. You can then point to the structural integrity of the walls surrounding the doorway, citing its Young's modulus and so on. Then you can say that the elephant's musculature and biomechanics would not have allowed for this Proboscidean creature to generate sufficient force to squeeze through the doorway. All of these would be very rigorous arguments that would rule out the possibility that the elephant was in your house while you shopped.

If you can propose or cite equally rigorous experiment or series of experiments that could potentially falsify the RNA world or metabolism first scenarios, then I will gladly concur that these models are falsifiable.

Because in the case of lithopanspermia our reasoning is very well-grounded in the radiobiology of microbes. A dose of 5,000 Gy will kill a whole population of E. coli -- and this can be experimentally verified -- while it will not kill a population of D. radiodurans. So now we can determine what kind of radiation resistance would be required for a population of microbes to make it through space and land on Earth. We can ascertain this down to the level of what specific protein parts would be needed.

On the other hand, what experiments would establish that the RNA world model couldn't happen because chemistry?

First, extraordinary radiation resistance a la D. radiodurans and B. subtilis is largely conferred by extra gene copies of particular genes (which are found among many bacteria), or increased expression of such genes, or redundant metabolic pathways.

Second, my argument isn't that core genes like ABC transporters is indicative of survival in space That isn't the point; I'm not presently attempting to muster evidence for panspermia. Instead, I am proposing a route to potential falsification. A progenote population lacking proteases, ABC transporters, nucleases, and a variety of catabolic enzymes would be rather quickly exterminated by galactic cosmic rays. So for panspermia to be correct the initial population of microbes must necessarily have the above protein parts.

Okay. So then, in your view, the RNA world is -- by virtue of its general nature -- not falsifiable except through the accumulation of extensive problems. Yet lithopanspermia is also a rather general model, but as I have argued (contentiously, given objections raised by Dr Adequate), it is falsifiable in the Popperian sense. Thus it would seem to me your primary objection here is that Popperian falsificationism is not an ideal criterion for demarcation of science and non-science. Is this basically correct?

That's a pretty terrible way to falsify a hypothesis, since there's no experiment that can potentially demonstrate that RNA replicators are impossible.

Because I am genuinely curious if you have discovered assumptions built into lithopanspermia which I am not aware of. However, you appear to prefer a more combative than explorative approach to discussion.

So you think there might be particular assumptions in lithopanspermia but you can't establish that. Mmk.

There is evidence that there was a considerable number of life-friendly planets at the time of Earth's origin, which means that abiogenesis would have had more time to occur. This is a rather straightforward argument that panspermia increases the probability of abiogenesis.

How does this address my question that the evidence for panspermia is historical in nature, in contrast to the evidence for the RNA world and metabolism first models? You're response seems to boil down to: "Well, panspermia doesn't address abiogenesis." That appears to be a bit irrelevant to my question. Please elucidate further so I understand what your argument here is.

No, but that's not what I'm saying, is it? Do you believe that pursuing panspermia research is a dead-end for pragmatic purposes?

Edited by Genomicus, : No reason given.

What do you mean by "generally falsifiable" in contrast to "falsifiable scenarios"?

Karl Popper anticipated the Duhem-Quine thesis in his formulation of the falsifiability criterion for scientific hypotheses. The criterion of falsifiability does not exist in a vacuum; it's part of a broader methodological ecology in which the following conventions are adopted:

"Scientific hypotheses are falsified when we accept existential statements that contradict them. We should, however, demand, when possible, that results are reproducible and inter-subjectively testable. If additional experiments are impracticable, such as in the case of natural experiments, then we should at least want to specify conditions of reproducibility. Stray or fleeting results which fail to satisfy these criteria, while perhaps hinting at a problem or inspiring new research, aren’t to be regarded as falsifications.

"We should refrain from what Popper calls ‘conventionalist stratagems’. That is, ad hoc maneuvers to evade falsification, such as changing definitions, accusations of deceit, dismissing results as observational errors or equipment malfunction. Bare appeals to doubt, or the mere logical possibility of false auxiliary hypotheses aren’t acceptable objections to an apparent falsification.

"If we wish to defend scientific hypotheses from a putative falsification, then we must structure our objections as independently testable hypotheses, or modifications to hypotheses should increase their degree of falsifiability. In other words, if an objection merely papers over the problem, serving no purpose other than to quarantine or excuse the apparent failure and thereby reduce explanatory content, then it has no scientific merit." From: here.

Auxiliary models can only patch up a hypothesis in crisis for so long, and these auxiliary models must increase the degree of falsifiability of the hypothesis in question; otherwise, the auxiliary models increasingly move the general hypothesis away from the domain of science.

By your argument, however, it seems that only the weight of difficulties can bring down the edifice of a hypothesis. But this seems like a pretty poor criterion of demarcation between science and pseudoscience. A "weight of difficulties" can be found in creationism (for which there does not appear to be a compelling falsification scenario), so is creationism therefore science in some way?

Yes, provided that somewhat rigorous quantification takes place, instead of merely asserting that "Phenomenon X is implausible."

But the burden of proof isn't really on me. A hypothesis should be falsifiable, at least if we adopt Popper's view of what constitutes a scientific hypothesis (a view, incidentally, that has been adopted by quite a large number of philosophy of science experts). No one should have to prove the negative view that a hypothesis is not, in principle, falsifiable. Instead, the specific details of how a hypothesis could potentially be falsified must be outlined if a hypothesis is to be considered scientific in the strictest sense.

That's quite significant, actually. If, on the one hand, we have a hypothesis (lithopanspermia) that's directly falsifiable -- this is certainly preferable to a model full of knowledge gaps that render it unfalsifiable. This is not to say that these gaps will not be filled or that exploratory science of this model should not take place; there is a good deal of value associated with any serious attempt to resolve the enigma of how life arose. But a model full of knowledge gaps that make it unfalsifiable does not necessarily qualify as a scientific hypothesis compared to a model which is plainly falsifiable.

The "hypothesis" that Jews caused epidemic diseases would never have qualified as a scientific notion as there was no supporting evidence for this view. I am not aware that there ever existed a scientific consensus of the above idea because it was quite patently non-scientific and not backed by any evidence (popular myth notwithstanding). So you may want to come up with another example with which to frame your argument.

Wait. So in your effort to demonstrate that lithopanspermia is not falsifiable, you're divining the possible existence of analogs of proteases, ABC transporters, and nucleases, which would have then been lost under mysterious selective pressure that somehow allowed such a loss to not affect the reproductive fitness of the population in a negative way, and wherein no homologs were preserved; and then cells evolved proteases, ABC transporters, and nucleases. I must be misunderstanding your argument, because otherwise you're straining credulity at its seams.

Could you please find another way to phrase your argument so it's clearer to me? Thanks.

The RNA world hypothesis:

Free-floating nucleotides stochastically assembled into short sequences with catalytic properties --> Primitive catalysis propelled the generation of ever-longer nucleotide sequences --> Base-pairing coupled with thermal and catalytic activity allowed for polynucleotide replication --> Replication lead to molecular evolution and increased efficiency of RNA catalysts --> Catalytic RNAs allowed for extensive peptide bond formation among amino acids --> Primitive protein molecules thereby emerged, which eventually evolved into the sophisticated protein machinery of modern cellular life. The RNA-to-protein system was also upgraded to a DNA --> RNA --> protein system.

The Lithopanspermia hypothesis:

Biological life emerged on another planet, rich with organic chemical resources around the time of the origin of the Earth --> Much of this microbial life evolved sophisticated machinery for resistance against ionizing radiation --> Ejecta from this planet following an impact contained members of these microbes --> This ejecta traveled through space and eventually struck Earth at approximately 3.6 Ga.

How again is the RNA world model more general than lithopanspermia?

So then if not by exhaustive trial, how?

And the reason panspermia has lower theoretical content in relation to historical claims is what?

Well, I certainly agree with you that research needs to be proposed before it's funded. But I'm also saying that such research should be proposed for panspermia largely because the alternative models for the origin of life are considerably unfalsifiable, and therefore more exploratory than pragmatic in a scientific sense.

Sure, and falsifying the RNA world model does not in itself falsify abiogenesis. If you have a falsification scenario for the RNA world model, by all means present it. I will then immediately concede that the RNA world model is falsifiable.

Really? Let's take the age of the Earth, for example. No matter how much evidence is presented that the Earth is approximately 4.5 Ga, the creationist can always resort to a divine mechanism (and they have) which accounts for observations of isotope-based dating. You will find that creationists can always account for a particular observation by inserting a divine mechanism, so the weight of the evidence seems to be of no weight at all! The blade that slices creationism away from the domain of science must therefore be the criterion of falsifiability. It is falsifiability, not the weight of evidence, that most starkly divides science from pseudoscience.

I'll just reiterate what I said to PaulK in Msg 57:

Edited by Genomicus, : No reason given.

Hey caffeine,

You are correct with regards to the above studies. I cited these in the OP so as to steer the discussion away from attempts to argue that panspermia could not have happened. However, while there is a considerably large volume of research indicating panspermia is feasible, there are also numerous studies that suggest it actually happened in biological history. For starters:

• Molecular clocks of prokaryotes based on 32 well-conserved proteins present in both bacteria and Archaea suggest that bacteria diverged ~4.1 Gyr ago and Archaea diverged at ~4 Gyr (Battistuzzi et al. 2004). Yet the most ancient, definitively microbial fossils are found in Archaean rocks about 3.5 Gyr in age.

  This observation is in complete accord with the hypothesis of lithopanspermia. First, evidence suggests that there was a maximum number of habitable planets in the Milky Way when the Earth began its formation about 4.6 Gyr ago (van Bloh et al., 2003). So the emergence of biological life in our galaxy would have been most probable around this time. Under the lithopanspermia hypothesis, cellular life would have then evolved for several hundred million years before having the capacity to survive transport through space in planetary ejecta. Thus, life should be older than the first microbial fossils on Earth, and younger than about 4.6 Gyr -- precisely what this molecular clock analysis indicates.

  Opponents of lithopanspermia may respond with a technical criticism: the molecular clock observations can be waved away by positing rapid molecular evolution during the divergence of bacteria and Archaea. Yet this is not as convincing an argument as it may appear; first, use of different calibration points and minimum bounds than what was used in an earlier study (Sheridan et al., 2003) still closely match the molecular clock results of that earlier study, which would not be expected if the molecular clock results are due to a significantly elevated rate of molecular evolution early in prokaryotic evolution; and second, Battistuzzi and colleagues employed a Bayesian local clock, which would be more accurate than a global clock as it does not assume an identical molecular clock rate across all tree branches.

  This discrepancy between molecular clock analyses of life's most basal lineages and the paleontological record has puzzled some researchers; however, the lithopanspermia model elegantly explains this observation in a single stroke.

• The gene set of the LUCA was sufficiently large enough to allow for transport through space of microbes. In other words, the bulk of phylogenetic and genomic evidence invariably point to the conclusion that the LUCA was not a progenote (which would have been unable to survive the ionizing radiation of galactic cosmic rays), but instead had a sophisticated repertoire of protein parts, including several of the core proteins of bacterial flagella (here I am bringing up bacterial flagella only as a way of highlighting the not-so-primitive state of the LUCA).

I did, in fact, read Wallis' 2003 paper prior to writing up the OP. It is indeed speculative, but it's not "wildly" speculative IMHO -- unless you would also grant that label to the RNA world model. The distribution of Aib-polypeptides in the paleontological record are tentatively suggestive of a biogenic origin; further work would have to be done in order to shed more light on this.

References

Battistuzzi, et al., 2004. A genomic timescale of prokaryote evolution: insights into the origin of methanogenesis, phototrophy, and the colonization of land.

van Bloh, et al., 2003. Maximum number of habitable planets at the time of Earth's origin: new hints for panspermia?

Sheridan, et al., 2003. Estimated Minimal Divergence Times of the Major Bacterial and Archaeal Phyla.

What specifically did I say that was irrelevant?

And I would say that "Base-pairing coupled with thermal and catalytic activity allowed for polynucleotide replication" is quite specific. You're going to have to make a better case that the RNA world is more general than lithopanspermia.

And how are you going to show that there is some restriction that cannot be overcome without exhaustive trial?

That doesn't actually answer my query, though, and looks rather circular. Let me try again: why is it that lithopanspermia has lower theoretical content than, say, the RNA world model? What is the intrinsic property of lithopanspermia that gives it more historical content (see, e.g., my discussion of molecular clocks in my response to caffeine's post) than the RNA world model? Is it that there are quite a few knowledge gaps when it comes to the RNA world, in contrast to lithopanspermia?

This sort of comment might get a layperson thinking that the dozens and dozens of scientific papers published on molecular clock analyses of prokaryotic phylogeny is little more than evolutionary biologists dreaming stuff up. I find this argument suspect as it seems to imply that the aforementioned problem has not been tackled by those in the field, when in fact the paper I cited -- and many others -- understood the nature of the problem and properly addressed it with the appropriate methodology.

Consider the following approaches to resolving the problem meaningful calibration points for prokaryotic phylogeny:

(1) The use of calibration points based on lineage-specific biomarkers, such as okenane (associated with gammaproteobacteria) and chlorobactane (found in the green sulfur bacteria at about 1.6 Ga).

(2) The use of node age constraints as calibration points, thereby providing information as to the minimum absolute ages of the nodes in the tree (which can be pretty easily done with Bayesian software like MrBayes).

(3) The use of eukaryote divergence times as inferred by paleontology and extrapolating to prokaryotic lineages while controlling for rate differences among prokaryotes and eukaryotes.

(4) The use of statistical tests (e.g., likelihood ratio test and relative rate test) of primary structure data to determine the extent to which a global molecular clock holds for a particular set of sequences. And if a global molecular clock does indeed hold, then only a few calibration points are needed.

(5) The use of local molecular clocks based on Bayesian and maximum-likelihood approaches to increase the reliability of molecular clock estimates even when there are few calibration points and when a global molecular clock is rejected.

That depends on the bacterial fossils under consideration, though.

Understood, but:

(1) There are quite a few other biomarkers other than oxygenic photosynthesis tied to prokaryotic clades that can yield fairly realistic calibration points, as mentioned above.

(2) We can test for useful calibration points based on whether removing these calibration points significantly alter the node ages. If they do not, then they are pretty accurate calibration points.

(3) You keep mentioning that certain parts of this are "complex" or "complicating." That may be an idiosyncrasy of your diction, in which case I wholly understand, but I am sure you do acknowledge that the complexity of a field of knowledge does not mean it cannot be understood. There is a very robust literature on a wide range of biomarkers, for example, as much energy has been put into elucidating the details of such markers.

That argument looks pretty ad hoc when you combine the observation of certain biomarkers with molecular and paleontological arguments. E.g., these biomarkers are often found in geological sites alongside other biomarkers; taken together, these all point to having been produced by organisms not at all unlike the modern bacterial clades they are thought to be ancestors of.

Since Cavalier-Smith's (arguably weak) transition analyses can't be used to establish absolute dates of origin, his figure in the paper (from which I believe you got the 850 million years) cannot be taken to be any sort of evidence for absolute ages.

More later.

I can get through most paywalls, actually, so if there's ever a paper you'd like me to dig up in full, just pass the word on to me.

Brasier et al., 2006, is indeed an interesting article and an insightful read. However, I suspect you are misunderstanding my argument (no doubt due to my late night attempts at communication!). My argument here is not in any way the idea that life on Earth would have had too little time to emerge (this is a line of reasoning that some researchers have used to justify panspermia, but is not my approach here). Rather, my argument can be summarized thusly:

(1) The oldest, definitively prokaryotic fossils go back roughly 3.5 Ga. Even this may be generous, however, and life on Earth may have emerged at an even later date.

(2) Molecular clock analyses based on (a) multiple protein sequences from multiple prokaryotic taxa, and (b) different molecular clock approaches with a variety of calibration points in each study all converge on a more-or-less unanimous conclusion: that prokaryotes are more than 4 billion years old, and less than 4.5 billion years old.

(3) This is complete accordance with lithopanspermia, wherein biological life arose some 4.5 Ga ago on other planetary surfaces -- when there was a maximum number of habitable planets in the Milky Way -- and subsequently evolved sophisticated molecular machinery over the course of hundreds of millions of years before being transported to Earth. In other words, if lithopanspermia is correct, then life should be older than the dates for the first paleontological indications of fully-formed microbial life. And this is what multiple molecular clock analyses seem to suggest.

I'd be interested in seeing if there's actually much of an argument behind your florid prose describing this paper.

And not one Wallis is making per se. He proposes the following two scenarios as possible explanations for the geologic distribution of Aib peptides: (1) that fungi from space landed on Earth, and that the genomes of these fungi encoded genes for Aib polypeptides, or (2) an unknown kind of microbial organism landed on Earth, equipped with genes for Aib polypeptides; these genes were subsequently transferred to fungal genomes.

Your attempt to refute (1) by pointing to the long branches of fungi in eukaryote phylogenies fails to consider that it's hypothetically possible for fungi-like organisms to land on Earth with that lineage going extinct -- but not before some of their genes were laterally transferred to other fungi clades and certain prokaryotes.

So what really needs to happen is this: first, analyze genes encoding Aib polypeptides and see if there are indicators of having been acquired through horizontal gene transfer (e.g., through GC-content analysis); second, conduct relevant molecular clock analyses to determine when these genes first arose.

Anyways, what's your explanation for the particular distribution of Aib in the K/T transition?

Edited by Genomicus, : No reason given.

More precisely, what we methodologically rule out when it comes to falsifying a hypothesis are auxiliary hypotheses that do not increase the degree of falsifiability of that hypothesis, as stated previously:

"If we wish to defend scientific hypotheses from a putative falsification, then we must structure our objections as independently testable hypotheses, or modifications to hypotheses should increase their degree of falsifiability. In other words, if an objection merely papers over the problem, serving no purpose other than to quarantine or excuse the apparent failure and thereby reduce explanatory content, then it has no scientific merit."

So we methodologically rule out gods because that auxiliary hypothesis merely papers over the problem and does not increase the degree of falsifiability of the hypothesis under consideration.

Once again, then, it is not a matter of the "weight of evidence." This is not a very good line of demarcation between science and pseudoscience, as no matter how heavy the weight of evidence might be against a particular hypothesis, auxiliary hypotheses can always be conjured. Thus, what matters here is falsifiability: do auxiliary hypotheses increase the degree of falsifiability or do they simply function as a kind of quarantine?

Edited by Genomicus, : No reason given.