SCIENCE: -- "observational science" vs "historical science" vs ... science.

At her blog Faith posted what looks like a response to this thread on 8/12/2014: The underlying foundation of Faith's attempt to dismiss science that is based upon ancient evidence is the difference between deduction and induction, but her real target is tentativity. She believes that conclusions reached through induction and the evidence used to reach those conclusions can be dismissed or disregarded simply because induction was involved.But the distinction she draws between conclusions based upon ancient versus new evidence doesn't exist. The length of a skid mark in a traffic accident to determine speed involves an induction little different from the one used to determine duration of sedimentation from the thickness of a layer and its particle size.--Percy

I'm not myself drawing any distinction between induction and inductive reasoning - have people been doing that in this thread?Faith must believe that experiments conducted today under controlled conditions with no unknowns require no inductive reasoning, but since the conclusions must inevitably rely upon theories arrived at inductively that can never be so.--Percy

It's just a reaffirmation of her position. Ignoring everything she has no answer for, she makes no attempt to address her view's self-evidently fatal problems.Faith, what you call "their usual strange misconceptions...over at EvC" are not the constructions of a small isolated group trapped in scientific misthinking. They're mainstream views shared by the community of thousands of scientists all around the world. You cannot in any accurate way characterize your views as being opposed only by those at a tiny website.Faith, it's understandable that you prefer the word "plausibility" to "probability". Plausibility is the quality of being credible or believable and is not a term with any scientific precision. Probability is a statistical measure of likelihood. To take a specific example, it is statistically extremely highly probable that tiny particulates will remain in suspension in active water, as you characterize your flood. It isn't just plausible - it's so highly probable that it nears sheer certainty.To take another specific and related example, it is statistically extremely highly probable that heavier and denser particulates will settle out of suspension before lighter and less dense particulates. Again , it isn't just plausible - it's so highly probable that it nears sheer certainty. Your view of layers being deposited without regard to weight and density is not merely improbable, it is unlikely in the extreme, or, to use your own preferred term, distinctly implausible.Or to take yet another specific and again related example, it is statistically extremely probable that footprints, burrows and nests would be destroyed in a violent flood. It is implausible in the extreme for them to be preserved and moved whole. By the way, if all sedimentary layers were formed during the flood, then how does a burrow in a non-sedimentary pre-flood layer get moved whole and then redeposited in a sedimentary layer without preserving any hint whatsoever of the pre-flood layer? Same for footprints. Those pre-flood footprints must have been made on a pre-flood surface that was not sedimentary. How does one move a footprint without preserving any of the material in which the footprint was made?--Percy

Before getting to Faith's post about the unwitnessed past I should mention that Faith has posted an "Update" to her blog post of 8/28/2014. It's at the top. It repeats that we're misinterpreting, misrepresenting and abusing her but is remarkably free of specifics. What I mean by the Unwitnessed Past was posted yesterday. There's nothing new here. She repeats the same arguments and addresses none of the criticisms. It's likely she doesn't understand the criticisms since she quotes Taq's Message 459 (which is polite and describes her position fairly accurately), then calls it too abstract to be saying anything meaningful. She still believes that past events with no witnesses cannot be subjected to meaningful analysis despite dozens of examples to the contrary, for example that a dinosaur footprint can be analyzed to determine the species of dinosaur. She thinks evidence from the past presents problems that render analysis and interpretation speculative but cannot muster any reasons why this is so, e.g.: That past events can't be repeated is a meaningless objection. No event in the past is precisely repeatable. Jack Ruby cannot murder Lee Harvey Oswald again, but the evidence that he did is indisputable. There are no dinosaurs to leave behind footprints today, but the evidence that they did in the distant past is indisputable. The ordering of layers tells us their relative ages, and radiometric dating provides their absolute ages. There are no assumptions. Faith has so far been reluctant to discuss radiometric dating or to explain how a flood sorts layers according to the concentration of daughter isotopes.Faith accurately notes that sedimentary layers are often dated using volcanic deposits contained within the same strata. She seems to believe this represents a problem but doesn't explain how.Faith also fails to provide any support for her contention that the results of radiometric dating are often inconsistent.--Percy

I don't really find much to agree with in the distinctions people have tried to draw between various fields of science. There are many tools in the scientific toolbox for gathering and analyzing the data into a tapestry of evidence upon which to base theory. Some fields lend themselves to controlled experiments to generate data (physics), others must gather much of their data from events that happened long ago (paleogeology) or over which we have no control (meteorology), but this isn't what decides our level of confidence.I don't know how one would compare our confidence in these ideas: that the charge of the proton is +1 (physics), that the universe is 13.8 billion years old (cosmology), and that the earth is 4.56 billion years old (geology). These ideas became accepted because they convinced a preponderance of scientists in the relevant fields, not because we made statistical assessments. Sometimes the scientific tool that does the convincing *is* statistics, as with the Higgs Boson, other times it is other scientific tools.--Percy

I agree with your qualitative ranking, but I was more trying to make the point that we didn't accept these ideas because we had a statistical measure of our confidence, say, that we're 99.9999% sure the charge of the proton is +1, 98.72% sure the Earth is 4.56 billion years old, and 97.36% sure the universe is 13.8 billion years old. Statistics were not a factor in the acceptance of these ideas. They only became accepted after consensuses formed in the relevant scientific communities. I don't think I can share this view. For me statistics is only a tool of science. Sometimes it's employed, sometimes not. I did provide the example of the Higgs Boson, where statistics was applied as a primary tool in identifying certain particle events as indicating its existence, but when Wegener noted the congruence of continental boundaries and the similarity of flora, fauna and geological formations on opposite sides of a broad ocean he wasn't using statistics. He didn't calculate the odds of land bridges versus his own ideas - no such statistical assessments were ever made.Or take the example of Watson and Crick's paper on the structure of DNA (A structure for deoxyribose nucleic acids), which also uses no statistics.I guess my main purpose was to quell the impression that scientific fields decide which theories or ideas to accept based upon statistical assessments of their accuracy. There's never any meeting where someone opens an envelope and announces, "Theory A has a 96.3% probability of being correct, Theory B has only a 92.8% probability of being correct, therefore our field officially accepts Theory A." There's not even a point in time that can be identified as when a consensus forms. Usually it just sort of happens and is only recognized as having happened gradually and over time. Rare are the times (such as when the universe's expansion was recognized as accelerating) that a consensus forms almost overnight.--Percy

Hi PaulK,Both you and NoNukes have replied as if you thought my focus was the use of statistics in science. It wasn't. I tried to make clear that that wasn't what I was doing, for example, noting that statistics is a tool of science, and that it's appropriate for some things in science (for instance, the Higgs Boson) and not others (for instance, Wegener's ideas and Crick and Watson's paper on the structure of DNA).I was trying to distinguish this use of statistics from what I thought HBD was saying. My reply was intended to be interpreted in the context of HBD's question about whether I agreed that some fields of science lend themselves to statistical confidence factors, that scientific fields can be divided into those that that do and those that don't. Naturally I don't agree, but obviously my use of the age of the Earth and universe confused the issue, so let me say a little more.I wasn't saying that statistics aren't involved in the determinations of age. I was saying that age determinations, even when expressed as something like 13.798 0.037 billion years (Wikipedia article on the Universe) aren't accompanied by any confidence factor that the figure is right. And the confidence factor that does sometimes accompany these figures (for instance the age figure might have been specified as 13.798 0.037 billion years at 95% confidence) is not our confidence that the figure is correct, but our confidence that the incomplete data set that was studied (invariably we have to operate with incomplete data, the induction part of science) accurately reflects the full data set. It isn't the confidence factor that the figure is right. Plus underlying assumptions (inflation, gas absorption, etc.) and fundamental measures (the cosmic distance ladder) may change. For instance, one of the references for the Wikipedia article (Planck Reveals an Almost Perfect Universe) describes a revision to the Hubble Constant, and this is unlikely to be the last revision.The concept I'm talking about, and that I thought HBD was talking about, was a confidence factor for the 13.798 0.037 billion year figure, and my position is that no such confidence factor exists. In fact, if there's anything we can be confident in it is that the figure *will* be revised as research continues.But even beyond that, we cannot say something like we are 99.99% confident that the universe is older than 13 billion years, and that we're 99.9999% confident that it is older than 10 billion years, and 99.9999999999% confident that it is older than 6300 years, because there is no way to assign such numbers. We know we're very confident the universe is very ancient, but there are no statistical measures of that confidence.--Percy

Clearly we're talking about confidence in different things. You have to ask, "Confidence in what?" It could be confidence that the figure accurately reflects a full data set. Or it could reflect the accuracy of the measurement device or method. Even when you have a figure for something like the size of the proton of .84-.87 fm, the error range is a function of the method. Any confidence factor you might derive from this range does not reflect the possibility that other methods might give other values, which is precisely what has happened with the size of the proton (see Proton size puzzle reinforced!).What error bars and confidence factors do not reflect is our confidence the figure is right. To know that we'd have to know how and why the figure will change in the future, and we can't know that a priori.Confidence in the theories and basic ideas of a field derive from a consensus that does not derive from or lend itself to measurable confidence factors. We'll never be able to say anything like, "The confidence factor for a 4.56 billion year old Earth is 99.999% while that for a 6300 year old Earth is 0.0001%, therefore we can conclude the correct value is 4.56 billion years old." It just doesn't work that way.--Percy

Me too. It's a holiday weekend, we'll probably hear from him within another day or two. What I actually said in reply to HBD was, "I don't really find much to agree with in the distinctions people have tried to draw between various fields of science." I didn't tell him he was "clearly wrong," especially since like you I'm unsure what he's saying. But you and PaulK replied and forced me to provide more details about my opinions on what I thought he might be saying before he had a chance to confirm.AbE: I included a link to the paper, here it is again: A structure for deoxyribose nucleic acids. The lack of statistics apparently was not a concern of the Noble committee.High confidence factors and tiny error bars or ranges give us confidence in the research and analysis. They do not translate into confidence factors that the research is correct.--PercyEdited by Percy, : AbE.

Here's a slide show showing that all that was needed was to make measurements of an X-ray diffraction pattern: Franklin's X-ray diffraction, explanation of X-ray pattern. You have to click on the little icon in the lower right hand corner of the slide to advance. Crick and Watson used simple math, not statistics.But this issue isn't really relevant to the main discussion. I was only providing a counterexample to PaulK's claim that "any quantitative result...will involve statistics."My simple point is only that science that involves math and/or statistics doesn't have a higher confidence factor that it is telling us things that are true about the real world than science that does not. Neither approach to science comes with any such confidence factor. What becomes accepted in science is a human process, not a mathematical one. All the math in the universe never convinced Einstein that God plays dice.--Percy

Sure, but even if it predicted Mercury's orbit to 20 digits of accuracy, is the confidence factor for GR greater or lesser than that for the idea that continents move. I'm only saying that no such confidence factors exist, and that therefore such comparisons cannot be made in any mathematical sense.--Percy

Faith responds to HBD's Message 457 in her blog post That annoying complaint about the terms "prove" and "proof": Scientists use the word "prove" all the time, but they do not use it in any mathematical sense. All they mean is that they can support their contentions with evidence. Telling scientists that "You can't prove such-and-so" makes no sense, since of course they can support their contentions with evidence. They only make claims when they have evidence. It would make no sense to make claims completely unsupported by evidence.So since it makes no sense to say "You can't prove such-and-so" for a claim that is supported by evidence, one can only assume that "prove" is meant in a mathematical sense, and then one can only respond that there is no such thing as "prove" in science. Hypotheses and theories can be supported by the evidence to varying degrees, but they can never be proven.--PercyEdited by Percy, : Focus my 1st sentence only on the word "prove".

All statistics is math, but not all math is statistics. I was merely rebutting PaulK's claim that all research with numerical results involves statistics. Okay, so don't be convinced, but the specifics of the paper I chose as an example wasn't the point, so just forget that paper. All I need to successfully rebut PaulK's claim that all research with numerical results involves statistics is a single example. Do you really think none exist? The consensus today is that Einstein was wrong, but this consensus is tentative (as are all consensuses), and whether the quantum world truly has some indeterminate and random qualities is still actively researched in some quarters. The consensus may change, but whether it does or not (and returning to my original point), a consensus doesn't form because of confidence factors that an idea is true, because such confidence factors do not exist. Accepting an idea is a subjective and human phenomenon based upon experience, background, intuition and feel, just like your rejection of the Crick/Watson paper as an example of using math but not statistics.--Percy

I didn't have an original assessment. I didn't say anything about how much statistics are used. I'm just pretty sure that the number of papers using only non-statistical math is far greater than zero.Concerning my main point about how ideas become accepted, I tried to distinguish between the statistics used in generating research results (e.g., a precise measure of the age of the Earth) from the confidence factor in an idea that I was actually talking about (e.g., the Earth is ancient) and that I don't believe exists, but evidently failed.--Percy

Yes, I guess that's about what I'm saying. Staying consistent with how you're expressing it, error bars or confidence factors aren't the whole story. For example, much of the work on MoND and dark matter (alternative theories that explain things like why rotating galaxies don't fly apart) is highly statistical, but those statistics don't help us form overall confidence factors for each theory that allow us to perform a simple mathematical comparison to decide which is correct. Consensuses don't form that way. I couldn't accept that any quantitative result will involve statistics (your Message 471), and this alternative expression is still a bit unpalatable. Does a measurement include simple counts? Anyway, if I understand what you're trying to say then I couldn't agree more that statistical approaches are very important to a great deal of research that lends itself to mathematical analysis, but I never said it wasn't, or even tried to express my opinion on the matter. Sorry if I gave a different impression by using the ages of the Earth and universe as examples.--Percy