Tuesday, February 19, 2013

Six impossible things before breakfast

Occam's razor is involved in the following assertions being protected from the burden of proof:

1) Random mutations and natural selection is necessary and sufficient to explain the origin of species from other species.

2) Abiogenesis occurred on Earth

3) Abiogenesis requires the conditions of a Planet (We don't know how abiogenesis happens, we haven't demonstrated it happening, so how can we presume what it needs?)

4) Pre-life is more fragile than life

5) pre life has been extincted by life because life is superior ( we don't yet know what came before life and see 2 through 4 we have no gnosis of where that might be, so we cannot possibly know the results of the two coexisting)

6) pre life has an x need for mass flux/energy flux/reproduction.....  We do not know of the process, so we certainly cannot know quantitatively of the need for any particular feature of the process.





20 comments:

Chris Fellows said...

Okay, I'll bite.

1. Yes, and rightly so. Sticking with the simpler model until and unless it has been demonstrated to fail is the way to progress. This is a good and proper use of Occam's Razor for the benefit of science. I see no evidence whatsoever in the data to question this assertion.

2. I don't think anyone seriously considers this assertion to be protected from the burden of proof. It is irrelevant, though, since all models of the chemical origin of life are applicable to any vaguely Earthlike world.

3. Again, I don't think anyone seriously considers this assertion to be protected from the burden of proof by Occam's razor. You need something rather than nothing for things to happen. Planets are the best environment we know of where there is something rather than nothing.

4 & 5. I argue that pre-life and life are part of the same evolutionary process that continues today. The process works to optimise individuals, species, and ecosystems for survival vis-a-vis others. The fate of pre-life is just that of the Pleistocene megafauna writ large. This seems like a no-brainer to me, but it is an arguable point. Occam's razor has nothing to do with my argument.

6. Again, this is an argument by extension from what we see around us that has nothing to do with parsimony. Complex processes occur rapidly in environments with lots of energy and matter moving around; look at the biological complexity of tropical rainforests cf. tundra. You can argue about it as much as you like; it is not "protected from the burden of proof". If you think a complex and interesting process such as life is more likely to arise in an environment where mass and energy fluxes are low, knock yourself out: in a purely pragmatic sense, you will be pushing cometary waste out of a steep gravitational well with a sharp stick, because this intuition of mine is nigh-universal. Not because of parsimony; and not because it is necessarily true.

Marco Parigi said...

1. 2 words - false dichotomy. There are plenty of people like Lennox that demonstrate quite conclusively why it fails. Unfortunately, they are mainly of a creationist bent. I assert that speciation could not happen without at least a strong form of neo-Lamarckism AND horizontal gene transfer being involved. You are too busy, with the rest of mainstream science, keeping the dogs of creation at bay.

2. This one and the next one are related and feed off each other. Scientists are free to speculate on geogenesis without the perennial question. "Do you have ANY evidence of this?", while with absolutely every other theory it is constant. I did discover it to be informally protected through my research which mentioned parsimony in the Wikipedia article about it. There is NO evidence for Geogenesis.

3. Here, scientists confuse what is good for life is good to also incubate life. Besides, in the early solar system planets were not what they are today. Again, scientists are wishing in evidence that just isn't there. Not knowing the process means not knowing what environment is good for the process.

4 & 5. Without agreeing on 1, I can hardly agree with you on the grounds that I don't think we understand the process of evolution the same way. I do agree that it is the same process as what I think is happening with regular evolution. In which case, there will be evidence of pre life, just as there is for even the most complete class extinctions we know about.

6. This one is one intuition versus another, not really parsimony. It still requires an idea of the process, timing and what the selection driving force is, given that we agree it is an evolutionary process. You cannot just address metabolism and forget about reproduction. Sure, one has to address where the energy is coming from and going. Mass transfers are secondary. Still, we can't lock ourselves into one line because it is a shared intuition.

Chris Fellows said...

1. No, no, no, no, no. Lennox et al. demonstrate nothing. Why do you think they do? Yes, horizontal gene transfer surely plays a larger role than is recognised; but in most cases, I am certain that random mutations are sufficient and necessary to explain speciation.

2. Okay, if people invoke Occam's razor to protect this assertion from the burden of proof,
then they are using it wrong. But see what I said before about it being irrelevant to the scientific question.

3. Early solar system??? How is this relevant? There were Earth-like planets when our solar system was a diffuse cloud of gas.

"Not knowing the process means not knowing what environment is good for the process"; see my first response to 6.

4&5. Your neo-Lamarckist claptrap still presumably leads to "fitter" organisms, so is subject to the same logical trend as conventional sensible evolution.

You say: "In which case, there will be evidence of pre life, just as there is for even the most complete class extinctions we know about."

Wikipedia says: "For macroorganisms, the Cambrian biota completely replaced the organisms that populated the Ediacaran fossil record." (http://en.wikipedia.org/wiki/Ediacara_biota)

6. Like I said in a comment on your previous post, reproduction is secondary and irrelevant to the origin of life. You can, and should, ignore it if you want to make any progress. Read "The Black Cloud" again for a good account of life with no need for reproduction. We are culturally just conditioned to think of reproduction as the defining feature of life since we discovered DNA and the 'Central Dogma' and it was so awesome.

Marco Parigi said...

"Natural selection on random mutations leading to speciation" I have no need for that hypothesis. Every single piece of fossil evidence and genetic data can be scientifically accounted for without it. It has not been experimentally verified, and its engineering equivalent is only useful in optimisation, with strictly controlled parameters at that. It's the God hypothesis for Darwinists.

Marco Parigi said...

If a theory is solid enough, it can cope with the burden of proof thus not requiring Ockhams razor. If it is not solid enough, then we shouldn't be trying to say that it is, by shifting the burden of proof.

Chris Fellows said...

Every single piece of fossil evidence and genetic data can be scientifically accounted for without it.

As we have argued over and over and over and over again, your neo-Lamarckism has no mechanism. That is why it should be rejected, and is not a credible scientific account for the evidence.

Marco Parigi said...

Every single piece of fossil evidence and genetic data can be scientifically accounted for without it.

As we have argued over and over and over and over again, your neo-Lamarckism has no mechanism. That is why it should be rejected, and is not a credible scientific account for the evidence.


To get back to my original point - I am not saying we should "Rule in" Larmackism, but that we should stop protecting "random mutations resulting in speciation" from the burden of proof.

The can means "is possible to be" ie. I don't have the mechanism, but I know information about which mutations may be suitable for which environmental stresses is being discovered and stored *somehow*, and being invoked when those environmental stresses occur. It is not unreasonable to suppose a chain of causality without a specific mechanism

Chris Fellows said...

I will do a quick gymnastic maneuver here to be true to my conception of how science works and get out of the trap of using language in the same way you are using it.

There is no "burden of proof" in science.

I can assume anything. Any bizarre thing at all. I don't need to "prove" it. All I need to do is compute the consequences of my assumption - in this case, 'natural selection on random mutations'. If I find it explains observations adequately, I am done. That's it. That's all I need to do.

If you find dobservatiomns that my hypothesis doesn't explain, then there is a problem. Only then.

Until then, the bar for your alternative model is that it be an *equally good* explanation for the data. Which it isn't.

You need to postulate a credible mechanism, consistent with what we know about molecular biology (summarised in that book I lent you), for "information about which mutations may be suitable for which environmental stresses is being discovered and stored *somehow*, and being invoked when those environmental stresses occur". Until then your theory will be (rightly) ignored, like Wegener was (rightly) ignored on continental drift. Or Raving-Looney was (rightly) ignored on space beavers with laser eyes.

Marco Parigi said...

All I need to do is compute the consequences of my assumption - in this case, 'natural selection on random mutations'. If I find it explains observations adequately, I am done.
I can play ball - this is exactly the way Lennox plays. The consequences of models and experiments which simulate this shows that it can explain "subtractive" mutations (losing a limb or function of an eye), and *perhaps* optimisation of one tiny function isolated from every other function, but not enough to show the observed progression of species jumps.

You find it explains things, and subjectively it does, but it does not stand up to the rigour of models that strictly adhere to the assertion.

The bar is set high for no apparent reason other than sticking it to creationists.

Marco Parigi said...

You need to postulate a credible mechanism, consistent with what we know about randomness and statistics(summarised in the God's Undertaker), for "truly random" mutations observed at a molecular level to translate to "beneficial mutations" that we can observe, with selection as the only informational feedback mechanism.
Until then, you are just repeating what you have been taught in first year biology, and are (rightly) suffering from believing people in authority, and disbelieving people because of their religious beliefs.

Chris Fellows said...

Lennox *cruelly abuses* randomness and statistics in a stupid way that is irrelevant to our question. FYI, I never did any first year biology. But the mechanism is all there in the 3rd year molecular biology book I referred to.

Our DNA is being continually damaged by radicals produced by metabolic processes, leading to various sorts of errors which you can read about.

Most of this damage is neutral, so if it gets through it just leads to composition drift. Some of this damage is negative.

Now, there are two things you need to remember about how DNA works to generate us and flies etc.:

*It is a hierarchical system; genes do not encode 1:1 for infiritesimal properties'; there are regulatory genes, and developmental genes, and higher order genes for their regulation and develpopment.

*It is a complex system; this means that the only way we can know what a change in the program will do is to *run the program*, and you cannot say that any arbitrarily small change will not have an arbitrarily large effect. We can see this very clearly in terms of some genetic diseases, where a change in one residue in one protein can make a gigantic difference to the entire phenotype.

Step 1: A tiny change induced by an entirely random mutation causes a large change in the phenotype of an organism. This is the butterfly effect, which all complex systems are prey to.

Step 2: This change in the phenotype causes other organisms to be more likely to reproduce with the first organism.

Step 3: There is no step 3. The first step is random mutation, the second is natural selection. Because butterfly effect, the steps can be arbitrarily large. The limit on the size of the steps is simply compatibility between the organism and its potential mates: remember we live in a world with ligers. There is no line between "microevolution" and "macroevolution"; they are just weather and climate.

We know that most species are stable, and that most species are dead ends, so it is probable that these "felix culpa" butterfly events are very rare.

Lennox, like 'climate scientists', is ignoring the poor old butterfly and all we know about complex systems and trying to shoehorn the universe into a plodding simplicity it just does not have.

Marco Parigi said...

The whole of your argument would be "equally" convincing and have just a high bar if you left the words "entirely random" out of step 1. Why is it unreasonable to require a demonstration of a causal link between any entirely random mutation and any observed speciation, whether in a real life, or a simulated model.

In the genetics book there was a lot of different mechanisms which modified DNA involving transposons etc.We know viruses go in and change DNA in ways we don't understand. we understand barely a millionth of what happens there. Why is a catch all "random" description placed on the ones that make a difference? A complicated coding change will "look" no different to a random one. We don't know how this beneficial mutation got there, we know that a lot of mutations are just random. Must be random *Unless Otherwise Proven*ie. getting back to my original point, that deny it or not the burden of proof has been shifted. We can call it random without having to prove it. If we want to say that random will not do it, we have to prove a *specific* alternative to randomness, when we know a tiny fraction of what in some cases is self modifying code, does.

Chris Fellows said...

First, I reassert that there is no such thing as burden of proof in science. A model with random mutations can fit the data. (Of course, all mutations are to some extent non-random, because certain stretches of DNA, and certain environments, will be more susceptible to mutation than others.) So there is no need to bring in any weird and wonderful mechanism to introduce genetic changes.

The whole of your argument would be "equally" convincing and have just a high bar if you left the words "entirely random" out of step 1

Not necessarily. For example, the argument would be entirely implausible and ridiculous if I were to modify the words 'entirely random' so that they would cover what you appear to envision, i.e., a non-random mutation in the sense of a mutation whose outcome is 'known' to be beneficial. This is because you cannot know the outcome in advance of this magnitude of change in a complex system. You need to run the programme to find out what it does; you need to generate the phenotype to know what the change in the genotype entails. "Randomness" is not protected from the "burden of proof" by some special dispensation by St. William: instead, all common-or-garden definitions of "non-random" are ruled out because they reverse causality.

Marco Parigi said...

I am not sure what you mean when you say that the model fits the data. There is no data in your explanation, and models that try to simulate it, don't back up what you are saying in any quantitative sense.

I didn't want to replace the words, but leave them out. I am not sure what you mean that non-random implies a reversal of causality. You know that I mean non-causality-reversing type of non-randomness. A non-randomness from knowledge stored within that is not perfect but is better than blind guesses. It means a filter of mutations rather than going in with a gene splicer and hacking randomly.

I don't imply a reversal of causality, I imply that other things like the Weismann barrier, or the central dogma, or the proof against Lamarkism are also not strictly correct. None of these are proven, and there are even counterfactuals. There are even experiments that have demonstrated mutations that are not random in the sense that I mean.

Marco Parigi said...

I have been looking at the Wikipedia article on directed mutagenesis and I detect a bit of denial in the biological sciences hierarchy. The proposal of the mechanism of random mutations and selection predates our understanding of molecular biology. Is it falsifiable? Why is it scientific to say:
Random mutations and selection explains all speciation.
Without proof, while ANY other competing theory is responded with : Explain the mechanism in detail, otherwise, keep quiet. It's not even enough to say "All I want to show is that random mutations do NOT explain this phenomena"

It appears that it is protected from the falsifiability criterion until an alternative mechanism is PROVEN. It is not enough that an alternative mechanism is required by the evidence.

Chris Fellows said...

There is no data in your explanation, and models that try to simulate it, don't back up what you are saying in any quantitative sense.

The data is *the sum of our knowledge* about life; the phenotype of everything we know and the genotype of everything we know. That data can be explained by natural selection on random mutations (plus, as you know I know, horizontal gene transfer and symbiosis; but predominately in bacteria, plants, and ancient prehistory).

The models are irrelevant, because we don't know enough about these systems to model them properly. They are like models of bicycles without the wheel.

I am happy with stress increasing mutation rates; I am happy with increased mutation rates in some tissues/genes than others; I am happy with horizontal gene transfer, and I have even (as long ago as 2005) conceded the possibility that different entire phenotypes could be stored, and 'switched on' under certain environmental cues, so that polar bears could be 'extinct' and then spontaneously re-emerge from grizzly bears, or a salamander might spontaneously give birth to Tony Blair.

What else do you want, if not a reversal of causality? Hmmm?

BTW, did you have a look at that cosmology video I linked in a comment on your last post? I'm keen to argue about something else now... ;)

Marco Parigi said...

You know I don't even acknowledge cosmology as a proper science.

What I want is an acknowledgement that Ockhams razor implies a shift in the burden of proof (research it yourself, by all means), or in your terminology raises the bar for ousting or even slightly modifying an incumbent vaguely worded hypothesis impossibly high.

Everything that I don't like about in the current state of science is encapsulated in this idea that Ockhams razor is never appropriate if it shifts the bar higher than other ideas that fit observations. The bar should be set high for things with repeatable observations backing them up, and lower where there are much less repeatable observations, or if the system is too complicated to test the model appropriately.

Chris Fellows said...

Natural selection on random variations is a fantastically fruitful hypotheses. There is *no need* to replace it with a more complex hypothesis.

That is all parsimony is: a pragmatic maxim to avoid complications that are not necessary.

You asked me to postulate a credible mechanism; I did.

Ockham's Razor sets the bar at an entirely appropriate place, so anybody who want to suggest a more complicated mechanism has to provide a clear and present rationale for doing so. This is to prevent crackpots from wasting everybody's time.

When you come back with a credible mechanism consistent with our knowledge of molecular biology for "non-random" mutations, and phenomena that it fits *better* than natural selection on random mutations, then I will acknowledge that your model is worthy of further discussion. (If I can find time between shooing all those flying pigs off my land and swigging whiskey from the 100-proof brook at the bottom of the hill.)

Marco Parigi said...

I am happy with stress increasing mutation rates; I am happy with increased mutation rates in some tissues/genes than others; I am happy with horizontal gene transfer, and I have even (as long ago as 2005) conceded the possibility that different entire phenotypes could be stored, and 'switched on' under certain environmental cues, so that polar bears could be 'extinct' and then spontaneously re-emerge from grizzly bears, or a salamander might spontaneously give birth to Tony Blair.
And
Ockham's Razor sets the bar at an entirely appropriate place, so anybody who want to suggest a more complicated mechanism has to provide a clear and present rationale for doing so.

The rationale for me is that I am severely embarassed defending the use of the word "random" for any of those non-random aspects of mutations. It is an irrational thing to not know the origin of a mutation of interest because it is beneficial and saying "must be random" It is severe "randomness of the gaps" in my view. It is equally damaging as "God of the Gaps" is for the same gaps.

It may be a fruitful hypothesis, but it is all poisoned fruit in my view. Much better to say that it is hard to trace the source of beneficial mutations, but technology should be able to eventually piece together how generalised the non-randomness (as described in the first quote above) some scientists have accepted exists is, across species generally, and whether there are even more sophisticated and intelligent sources of mutations/transposons that are generated by the workings of DNA etc.

Scientists aren't going to bother looking for anything new and interesting along these lines if the response from their peers is "that must be specific to the simpleton species you are able to do a controlled experiment on", rather than something that might be happening accross the board in some way.

Marco Parigi said...

If you look up Occam's Razor, see how it mentions a shift in the burden of proof, despite your protestations that there is none in science. Follow that link and see how it explains that "If this responsibility or burden of proof is shifted to a critic, the fallacy of appealing to ignorance is committed" Thus you can see how I have come to believe that Occam's Razor has come to validate arguments from ignorance wherever it is relied on in science in arguments such as the one we are having.