Return to Website

Number Watch Web Forum

This forum is about wrong numbers in science, politics and the media. It respects good science and good English.

Number Watch Web Forum
Start a New Topic 
View Entire Thread
Re: Is AGW a "scientific" theory?

As Dawkins has pointed out, evolution is very easy to falsify. If anyone ever finds a single bone from a single 4 billion year-old rabbit fossil, evolution is finished as a theory.

I'm wondering whether AGW has anything equivalent to the 4 billion year-old rabbit fossil.

Re: Re: Is AGW a "scientific" theory?

You're talking about the famous "Precambrian rabbit fossil" there. I've heard that was invented to fob off a 'Popper zealot'. Popper has quite a substantial 'fan club'. A few years ago the Sp!ked website ran a survey called something like "If you could teach the world one thing", where people were asked to impart some piece of knowledge to others they thought was crucial. The Popper falsification idea, to my surprise, turned out to be one of the most popular pieces of knowledge. So I can understand that Dawkins may feel obliged to come up with something to satisfy the Popper followers.

The follwing link gives a fairly good discussion about whether it really would falsify evolution:


"Would anachronistic fossils disprove evolution?

Evolutionary biologist Richard Dawkins said that the discovery of fossil mammals in Precambrian rocks would "completely blow evolution out of the water."[14] However philospher Peter Godfrey-Smith doubted that a single set of anachronistic fossils, even rabbits in Precambrian, would disprove the theory of evolution. The first question would be whether the alleged Precambrian rabbits were really fossilized rabbits. Alternative interpretations include incorrect identification of the "fossils", incorrect dating of the rocks and a hoax such as the Piltdown Man was shown to be. Even if the Precambrian rabbits turned out to be genuine, they would not instantly refute the theory of evolution, because that theory is a large package of ideas, including: that life on Earth has evolved over billions of years; that this evolution is driven by certain mechanisms; and that these mechanisms have produced a specific "family tree" that defines the relationships between species and the order in which they appeared. Hence Precambrian rabbits would prove that there were one or more serious errors somewhere in this package, and the next task would be to identify the error(s). Benton pointed out that in the short term scientists often have to accept the existence of competing hypotheses, each of which explains large parts but not all of the observed data. "

In terms of falsifying AGW, one possible scenario that I think the AGWers might accept would be that background CO2 levels suddenly started to go down significantly.

Re: Re: Re: Is AGW a "scientific" theory?

As a follow-up to my previous post on the Precambrian rabbit,it might be useful to note (though it's going even more off-topic) that the Precambrian rabbit situation has actually occured in physics.

There have been two equivalents of the Precambrian rabbit in the theory of light (and electromagnetic radiation in general). One is that the carrier medium for light, "the aether", can't be detected and is no longer assumed to exist. The wave theory of light is however still retained because it appears to be successful in what it does, but it can't be correct if the waves travel in nothing. Something has to be waving. The second Precambrian rabbit is the realisation that light shows particle-like behaviour in certain situations, and this has been absorbed into something called 'particle-wave duality'. But the particle or corpuscular theory of light was actually falsified in an experiment in the 19th century by Foucault when he showed that light travels at a slower speed in water than in air. The whole idea of particle-wave duality sticks two fingers up at Popper.

Now there is a mechanism for propagation of light which doesn't need an aether and which could explain particle-wave duality, but it's too far-fetched to contemplate. This mechanism assumes light doesn't actually travel, it gets from atom A to atom B as an oscillating force by action-at-a distance. Newton's gravity uses the idea of action-at-a distance, but for light it would be a lot more complicated because light has (or appears to have) a definite speed associated with it and unlike gravity not all matter is transparent to light. This mechanism is called 'retarded action at a distance' and atom A would somehow have to 'know' how far atom B was away from it to adjust the time delay in sending light to B. On top of that, if an object that isn't transparent to light was inserted between A and B, atom A would have to recognise this situation and not send anything to B.

Given that precedent from the world of physics, I think evolution theory could handle a Precambrian rabbit situation.

Re: Is AGW a "scientific" theory?

I think the point of the falsification argument is that it is at least possible to falsify evolution in principle. Of course, if you accept that it is never possible to know something with 100% certainty it is also not possible to falsify anything with 100% certainty - so the question is one of degree. A rabbit fossil as old as life itself would falsify evolution, notwithstanding the caveats that you mentioned and that one would be obliged to attach to such falsification. I am looking for how AGW can be falsified in principle. If there is no way to falsify AGW in principle, that seriously undermines the value of the hypothesis.

I see AGWers investing a lot of effort in trying to persuade me that they are right. I don't see them investing much effort in trying to persuade themselves that they are wrong - and back when I were a scientist, my alleged poverty of ejumucation notwithstanding, I did the exact opposite.

Re: Re: Is AGW a "scientific" theory?

"I see AGWers investing a lot of effort in trying to persuade me that they are right. I don't see them investing much effort in trying to persuade themselves that they are wrong - and back when I were a scientist, my alleged poverty of ejumucation notwithstanding, I did the exact opposite."

One of the common explanations for that particular observation is one that drifts somewhat into conspiracy theory territory. In the past few decades there has been a rise in what might be called "ethical science" or "Green leaning science". This webpage about a UK organisation called "Scientists for Global Responsibility" (SGR) gives some details about ethical science.


One of the career paths that ethical scientists like to go into is climate science. The founder of SGR, Stuart Parkinson, is himself a former climate scientist. The Climate Research Unit, Hadley Centre and Tyndall Centre are all listed as organisations that are considered to be 'ethical' on SGR's ethical employment contacts webpage.


Re: Is AGW a "scientific" theory?

A little knowledge is a dangerous thing, or so it is said, and the problem one might see is that pronouncements are often made with certainty when in fact we have no easy or reliable way to measure out total 'knowledge', thought one might be talking from a point of 100% familiarity with the present state of knowledge.

Evolution is often presented as something happening over unimaginably long time scales - but is that true? Is the ability to measure accurately back in time so good that it can remove any doubt about estimates of measurement?

In what we think of as geological time scales there can be quite obvious evolutionary development in a species, or not much at all. If one thinks of human development as something that has occurred over, say, a 1 million year period such developments could have occurred more than once in the life of the planet and be lost in the noise. Even in what we might think as recent times.

Given the same kind of environment the potential for extremely similar evolutionary development without a continuous link for the species between the periods could not be ruled out.

On the other hand Creation proof would require the identification of a actor that had the ability to create directly in some way OR by running a process from raw materials and having some controlling influence (perhaps imperfect) over the subsequent evolution.

The identification of such an actor would give a strong boost to creation as a concept but still not exclude evolution as a truth. It would simply be a truth after creation.

I have no reason to suppose the 'humans' have not been around several times and then disappeared again in the geological past. Proof of that, or even the potential for it to be likely, would require luck since any evidence we are likely to understand would mostly have been destroyed by ice ages and other planetary events. We cannot recreate the experiment and expect to survive it to observe the result so have to work on the balance of probability. That's a judgement call. So it is with the AGW hypothesis, IMO.


Re: Is AGW a "scientific" theory?

I had my concerns that this might end up focusing on something other than AGW, but here goes.

We observe life and seek an explanation as to why we observe life. The "hypothesis space" is essentially infinite - limited only by reality, and within reality limited only by human imagination and ingenuity. But we need only concern ourselves with those hypotheses which have so far been stated, and we can conclude (A) one of the hypotheses is sufficiently close to the actual truth to be useful, (B) none of the hypotheses are of any use and someone needs to do some more work. Assuming (A), we can then state reasons as to why we find certain hypotheses to have a close to 0% probability of being correct and why we find one of them to have a >>0% probability of being correct, or at least not wrong.

No one would claim that the current understanding of evolution is absolutely correct and will remain unchanged for all time, but the weight of expectation is that it is far more likely that the theory will move incrementally towards a perfect description of reality in ever smaller steps without ever getting to its destination, than that God will come down from his heaven on judgment day and tell us we have got it all wrong.

AGW seems to me to pull a number of stunts which bypass this process altogether. Firstly, what observation are we trying to explain? It seems to me the observation that AGWers are trying to explain is largely something that they predict we will observe over the next 100 years rather than something we have already observed (and that many of the expected observations such as hurricanes inundating cities built below sea level in hurricane-prone areas, are pretty much dead certs over a 100-year period.) Secondly, if there is a observation that needs explaining (e.g. warming between 1980 and 1998 which almost no one disputes actually happened), what are the other ideas in town? Proponents of evolution at least acknowledge the existence of crackpot religious ideas and the more patient evolutionists expend considerable effort on explaining why the crackpot religious ideas have a close to 0% probability of being correct. AGWers simply dismiss alternative hypotheses without a second thought (in much the same way that religious crackpots dismiss evolution without a second thought). Even if the suggestion that that warming was not anthropogenic in origin is totally crackpot, someone should be able to provide a thoughtful explanation as to why.

Your point about humans potentially having evolved many times is taken. It's a possibility, but a rather remote one. Evolution works by applying selection to (effectively) random genetic changes and it is vanishingly unlikely that near enough the same thing will happen twice, especially at a distance of time with very different circumstances.

Re: Re: Is AGW a "scientific" theory?


I agree. With everything you wrote including the vanishingly small potential that there is likely to have been almost identical instances of human evolution separated by periods of huge change that resulted in different basic circumstances existing at possible starting points.

Very unlikely indeed. But still difficult to disprove.

I could extend it and say there may have been 'similarities'.

Perhaps humanity has not so much evolved from a previous distinctly different biological state more than once but may have evolved once and become highly developed as a social and knowledgeable civilisation (in a modern way we could understand) and then failed for some traumatic reasons - too much carbon sequestration perhaps or maybe galloping temperature tipping points, any reasons, does not matter what - and resulted in ALMOST total destruction but some survived and returned to regenerate another version of 'civilisation' developed over time from a new blank mind. (Or not ...)

Alternatively humanity has not existed previously but creatures with similar abilities to invent and adapt to a greater extent than other creatures and perhaps use tools and some equivalent of science for their own betterment .... and so on.

As with AGW there is always the potential that some event might prove to be a decisive proof, one way or the other, that such events did in fact occur, but the chances of that are very very small based on out current knowledge, skills and potential for radical discoveries. AGW is the same, in my opinion, other than the effective battle ground is the future rather than the past. Whilst that might suggest a greater potential for improving and refining the knowledge available to come closer to refining the hypothesis in much shorter geological time scales it may equally prove to pose more scientific questions (based on new knowledge) than are answered.

Perhaps the problem lies not with science and scientific principles but with the concepts of working with them reliably detached from any other influences so that over time the target question at least remains the same (as it likely does most of the time with fundamental historically oriented questions.)

But that won't happen so it will come down to a moving target of what people are trying to prove. Or rather many targets I would suspect. None of them are likely to be absolutely provable without large error bars and nor are the moving targets, billed as 'future', likely to be disprovable in a reasonable (if any) period of elapsed time. The entire theory, from a social context, seems to rely on that being the case 0 in which event it can hardly be considered scientific.