Quote:
Originally Posted by CraigS
Ok .. Science in general can be distinguished into:
(i) theory and;
(ii)observation/experiment (observation, for short).
For observations (ie: the data), tests to assess whether the data is independently verifiable could be split into two parts:
i) Test #1 (Pass/Fail): say, a hard-copy printout, or release of raw electronic data giving unmanipulated access to others;
ii) Test #2 (Pass/Fail): a thorough description of (a) how to set up the instruments, tune them, and; (b) how to recreate the events leading to the generation of the data.
For theory (or hypothesis), tests to assess whether the theory is independently verifiable might simply be:
iii) Test #3 (Pass/Fail): Publication of the idea .. ie: everyone gets to the ability to check the details of the theory, until they can replicate those details with precision.
Self-consistency of a theory is also vital:
Determination of a theory's self-consistency has to be based on a "Pass" result to test #3. It's entirely possible to have a theory which is objective and independently verifiable, and for that theory to be internally inconsistent. But, unless it is both objective and independently verifiable, it might be impossible to work out that it is internally inconsistent.
Mathematics thus performs the function of assessing the self-consistency of propositional outcomes, outlined as results.
=================================== ================
In the case of Speigel's paper:
i) Test#1 result is a 'Pass', as the empirical data underpinning the assumption that abiogenetic life arose rapidly on Earth, is readily available for review by anybody (they have tabulated it);
ii) Test#2 result is a 'Pass' as the conditions for recreation of the measurements of the evidence that abiogentic life arose rapidly on Earth is readily available to everyone;
ii) Test#3 result is a 'Pass' because I found their paper ... I could read it, and I could educate myself sufficiently to understand what Bayesian Inference is, (I thus understand their methodology). More importantly, the paper has been submitted to Proceedings of the National Academy of Sciences, USA and will undergo rigorous peer-scrutiny.
Self-consistency
Each step in the Bayesian process has been clearly outlined in the paper. The mathematical steps in their analysis is open for all to scrutinise. They have tabulated their model parameters and provided sensitivity analyses applicable to pessimistic and optimistic values. They have published the algorithms used and provided results so they can be checked. The results have been bounded into upper and lower categories.
The conclusion is that their analysis is self-consistent.
None of the authors has made any assertions as to whether the analysis results in TRUTH. The assumption that this analysis attempts to portray TRUTH, lies entirely with those attempting to decry this work as 'subjective'.
These 'decriers' have provided no scientific basis whatsoever for their 'opinions' and are thus not practising Science.
Cheers
|
You can write all you like, Craig, it makes no difference to the outcome of the argument, either way.
Let me tell you why your rant here has failed....
Your dataset, for a start, is based totally on assumptions made in the past that are only what scientists
believe to be the case based on what
scant evidence they actually have. I should know a bit about this, Craig, considering it's what I studied at Uni. I'm not coming at this from some completely different field or without any prior knowledge in this, you know. Unlike yourself. Getting back to the matter at hand, those tabulated values are nothing more than guesses/assumptions of what might have been the case. The models are based on this because they have no hard evidence to go by. Most of those early trace fossils have been argued over for years as to whether they're actually real fossils of algae/bacteria that are present in those early rocks. In any case, those rocks have been so heavily metamorphosed, it's almost impossible to tell what's what from possible biogenic or purely abiogenic sources. The only thing that can be said for certain is that it happened, and probably rather early on.
Their data is subjective right from the start.
My own position on this is that life did arise early on this planet after its formation. Most likely after the first 100-200 million years or so. Despite their results and whatever you might think, there is a good chance that given the right conditions and all things being equal, life will arise elsewhere in a similar manner. Though, there is no evidence either way to prove/disprove this. It is a subjective answer based on subjective data.
As for your second test, they fail this miserably. They have not gone and collected their own data. All they have done was used previously published data (where it could be found) and entered that into their own analysis. The data was based on rather fluid assumptions to begin with, so their own conclusions will be affected by this. In other words, their conclusions will be nothing more than guesstimated opinions, not rigorous results that you would expect otherwise from a scientific paper. It was nothing more than a gedankenexperiment...a thought exercise, with only marginal applicability. More a topic for conversation and further discussion and
maybe a pointer to some further research.
Independent verification does come via publication, because that is where others get to see their paper and then to mull over it and try to replicate the results. So in that regard they pass the third test. However, any scientist worth their salt would see this paper for what it is and not even try to verify the results. Because, that is not what this paper is about. Scientists know the ins and outs of the various methods of statistical analysis and know that some are better than others for the jobs they perform. I've already stated in earlier posts the pitfalls of Bayesian analysis and why it's not seen by many as being an ideal method for some areas of scientific data analysis. To reiterate the main reasons as to why this form of statistics, in this particular case, is not ideal....
1. Quantifying prior beliefs into probability distributions is not simple. First, we haven’t all thought much about our prior beliefs about most things, and, even if we have some beliefs, those aren’t usually condensed into a probability distribution on a parameter.
2. We might not agree with colleagues on the prior distribution.
3. Even if we can find a formula for the distribution describing our prior beliefs about the parameter, actually doing the probability calculations to find the posterior distribution using Bayes Theorem maybe more complex than we can do in closed form.
....basically, in order to find the posterior calculations (the answers), a prior knowledge or belief about the distribution of the data is required. If that prior distribution is already based on tenuous assumptions, then the answer you get will be the same. You can even get completely different answers from the same set of data, using the exact same equations. All depending on what prior parameters we care to use at the time. They don't even have to vary that widely to get different results. Many scientists don't like Bayesian analysis for these reasons as it's not an objective method of analysis.
Just because anyone can pickup a paper and read it doesn't mean it automatically passes independent verification, and it hardly means it will just because someone thinks they understand what it says because they've read up a bit on it. You have to actually have done the science before, in order to verify it, or at least have studied in the field and then understood the implications of that science in the paper being read. You wouldn't get a neuroscientists writing an in depth critique of a journal paper written by a nuclear physicist, simply because he hasn't the required learning, or experience, under his belt to be able to do so...and vice versa. Even if they do understand some of what's being written.
If a theory or hypothesis is plagued by subjectivity in its methods of analysis because its underlying maths is this way itself, then self consistency is moot.
As far as truth is concerned, at no time at all did any of us here arguing the counterpoint to your position, Craig, ever say that we had a handle on the truth. It's you, in fact, that have brought this point up. I don't know what the truth is and most likely will never get the chance to find out. All I have argued is that given the universality of the physical laws governing the universe and the fact that life evolved on this planet, that all things being equal, life should evolve elsewhere in the universe and especially on planets where the conditions are similar to what they were on the early Earth.
Not that it did or has, or that it exists at present. But that given all the circumstances etc, it most
likely will be.
The ridiculous thing about all of this is that the paper itself even said this. However, they took the negative stance on the matter where it concerned their results. And they're entitled to do so. However, I and many others, would beg to differ despite conceding that their conclusions might be correct.
.
