The God of Science
- Thread starter 786
- Start date
Evolutionary theory doesn't depend on perfect randomness, that is only the source of arbitrary mutations upon which natural selection can work. If mutations are somewhat less than perfectly random, they are still the result of natural processes and do not conflict with a naturalistic explanation of the emergence of species.
I was doing some googling and ran accross this website.
From randommutation.com,
According to the experts, Darwinian evolution is driven by a combination of Random Mutation and Natural Selection:
"Cumulative selection is the key to all our modern explanations of life. It strings a series of acceptably lucky events (random mutations) together in a nonrandom sequence so that, at the end of the sequence, the finished product carries the illusion of being very very lucky indeed." -Richard Dawkins, The Blind Watchmaker
Do Your Own Darwinian Evolution Experiments with the Random Mutation Generator
The website makes a good "Case for Intelligent Evolution" for all you intelligent designer believers.
IMO, we humans are the ones making the cumulative selections Dawkins is speaking about and are the only Gods around.
Last edited:
Anyway, randomness as a source of variation does not control the outcome. Natural selection controls the outcome, and that is not random. It almost doesn't matter how the variations came to be.
Currently I'm leaning towards the expression and environmental reinforcement of normal genetic variation instead of random mutation. So far everything we know about random mutation shows it to be unhealthy for the recipient. Cells have repair mechanisms to suppress it and the actively seek out and destroy cells which mutate.
However species do have natural variations from individual to individual. In a stable period with large populations, these variations tend to average out. But in times of hardship certain variations may prove more successful in the new environment and the less successful ones die off creating genetic drift towards the more successful variation.
Thus it is the cycle of die back during climatic change, or the expansion into new territories which drives evolution.
However species do have natural variations from individual to individual. In a stable period with large populations, these variations tend to average out. But in times of hardship certain variations may prove more successful in the new environment and the less successful ones die off creating genetic drift towards the more successful variation.
Thus it is the cycle of die back during climatic change, or the expansion into new territories which drives evolution.
Yo folks lets not lead the discussion astray to "drift vs natural selection"- So I agree with GeoffP- leave it until I at least respond to them.
But definitely it is an interesting topic, you are welcome to start a new thread regarding drift vs natural selection. I think there has been research done that shows genetic drift to have more effect on evolution than Natural selection- I'm not too familiar with the work but I know it has been done.
Peace be unto you
But definitely it is an interesting topic, you are welcome to start a new thread regarding drift vs natural selection. I think there has been research done that shows genetic drift to have more effect on evolution than Natural selection- I'm not too familiar with the work but I know it has been done.
Peace be unto you
Hi 786:
Yes, it can, given further experimentation. "Pseudorandomness", presumably, would be that variance belonging to real, additional factors such as other genetic loci, interacting loci, environmental factors or gradients. Variance not explainable by any other effect would be truly random. For the purposes of any test, however, all unexplainable factors are usually represented as random variance. I would expect, however, that deviations from normality in the residuals of the original test would suggest either heteroscedasticity, OR - under other discrete distributions - the presence of underlying, unassayed variables.
Yes, it can, given further experimentation. "Pseudorandomness", presumably, would be that variance belonging to real, additional factors such as other genetic loci, interacting loci, environmental factors or gradients. Variance not explainable by any other effect would be truly random. For the purposes of any test, however, all unexplainable factors are usually represented as random variance. I would expect, however, that deviations from normality in the residuals of the original test would suggest either heteroscedasticity, OR - under other discrete distributions - the presence of underlying, unassayed variables.
Ok... but how will you perform "further experimentation" if you don't know the other factors for which to test?
Agreed.
True, that is if you knew all the "other effects. If we don't know all the "other effects", which we practically can't (or at least know that we know) and I don't think a scientist will come out and say "we know everything about this...." for this specific reason, and thus asserting "truly random" would be pre-mature.
The fact that they are represented as "random variance" doesn't mean they are. Secondly you are affirming that "all unexplainable factors" would be deemed random even though we don't know all the factors that effect it, which could very likely be due to some type of "control" and not randomness- you can test further to see if it is truly random but we can't because we don't know all the factors to test for in the first place. So in other words by performing an experiment you aren't really testing for randomness but for "control"- anything other than that would AUTOMATICALLY be deemed random- in other words everything unexplainable is assumed to be random until proven otherwise- I find this assumption unscientific-
Anyways, why can't we assume everything to be controlled until proven otherwise? Because as far as I can tell- when science discovers new things they show them in the context of control, science always shows control when we know more about a subject- so if the scientific trend is as such then assuming everything to be controlled until proven otherwise is more valid in my eyes- and in this case you would actually be testing for randomness. But this is difficult- because you always have the problem of pseudorandomness- in which case that would mean you can't really test for "true randomness".
Also as James R (I will respond to him later) said- the simplest explanation is more accepted- I would have to argue that "control" is more simple than "randomness" because control can be explained in the context of physical laws while randomness is basically an unknown force which is not explained by physical laws. If we take Chaos Theory, it never really shows "true randomness" but pseudorandomness, because it is still "controlled" but the event is so "sensitive to the initial conditions" that we have this pseudorandomness-like appearance.
Which you could also expect under pseudorandomness I believe. And on a second thought, do we have the data from the history to actually perform the statistical analysis on the mutations of the past? I don't believe we do. Only statistical tests can be done are what we can observe now- Do you know of a research that has done the statistical tests on natural mutations which they observed? I would like to see that paper...
Peace be unto you
PS-Note: I would also like to point out that there has been increasing evidence that all mutations are not entirely 'random.' There has been increasing research showing adaptive mutations which the organism controls itself to a degree due to environmental stress factors-. You can probably find recent work if you look at online databases like PubMed and others.
Here is an example (very interesting stuff): Adaptive Amplification and Point Mutation Are Independent Mechanisms: Evidence for Various Stress-Inducible Mutation Mechanisms - The introduction lists some other related work as well
Last edited:
In nature random simply means that events are unpredictable not that their cause is unknown. A mutation for example is an interaction between natural events that coincide. Our inability to predict is reflected only by our inability to track the large number of variables involved. If we had sufficiently advanced metrics and measuring capabilities then prediction would be possible. The best we can do until then is measure the rate of occurrence and state a probability estimate.
But this is fully testable, however, the duration of the tests may be significant, but that does not negate the science involved.
In the case of Gods there are no known metrics involved that we can track or measure. In this regard the proposition for the existence or not existence of gods is truly untestable, i.e. there is nothing to measure.
But this is fully testable, however, the duration of the tests may be significant, but that does not negate the science involved.
In the case of Gods there are no known metrics involved that we can track or measure. In this regard the proposition for the existence or not existence of gods is truly untestable, i.e. there is nothing to measure.
I agree but that is definitely not how people understand random. They seem to know the cause- the cause is chance- out of the blue type stuff. As seen by replies to this thread it can be concluded that the concept of "random" is misunderstood because of the underlying implication that build up in the minds of the people when they hear/look at the word "random".
That is why to eliminate this misunderstanding the word random should be changed to something more appropriate like "unpredictable" "uncertain" or some other form.
Peace be unto you
Well, that's why you need to control other factors as much as possible. In the case of genetics, you could use clonal organisms with site-directed mutations or deletions. If you're trying to map unknown genes, generally people use genetic markers with narrow inter-marker intervals to detect single genes while controlling for the background effects of other genes. If it's an environmental factor you're testing, you'd use clonals or highly inbred strains. Granted, some variance from background genes might occur in the latter two scenarios, but even in site-directed mutagenesis you still get residual variance. It's a simple fact that random variation does occur.
We can certainly say something along that line in cases like the ones described above. Although, in point of fact, any effect detected in any analysis is only detected at a probability of (1-a). At the usual 5% significance threshold, then, the likelihood of any individual gene, say, existing is only 95%. The same is true in any linear analysis. So we can't even say that the factors themselves are real with 100% certainty, although we might approach it (say, to 99%, which I not uncommonly find). How can we say there is no such thing as random variance when we can't even say for sure there is an effect?
I disagree: in many cases, this is undoubtedly so. See above.
Possibly – but, other factors should in most cases segregate randomly with sufficiently large sample size, which is one of the reasons we test (or SHOULD test) for statistical power prior to any analysis, and perform repeated analysis where the null is rejected. Naturally, as n decreases, we expect more spurious results, both due to small sample size of the coded factors, and of non-random background factors.
Well, it would be worse to assign a control to it that we can’t prove. But anyway, this isn’t always the case: clear patterns in data, such as periodicity, centrality or lack thereof, modality and general distribution can infer a great deal about control.
Rather: it is preferentially the cases which show control get published. Negative results are often hard to get into the literature. And, in each case, a percentage of variance is assigned to random error, or at least to the residual. In gene mapping, even after we control for one gene, we often find no additional evidence for other genes in complete genome-wide surveys. The control of phenotype is never so accurate.
Simple: Brownian motion.
Chaos theory, IMHO, is poop.
Natural mutations are very hard to follow. There is lab work which has tracked the course of mutations, though, and in each case there’s still within-population variation. Or even within-subject variation.
Now, if you can find within-subject variance – in an individual with 100% relationship to itself, so to speak – what does that tell you?
Yes, I use it often, though I don't actually really get into nonrandom point mutation too often. My work is busy enough surveying what's actually there. Now: we can't say that all mutations are entirely random, but we also can't say that all mutations are not random, or even disadvantageous in certain circumstances or even overall.
We can only try to control all variables (if we knew them all in the first place). The random variance "does occur"- which is true but how much of that random variance is due to unknown variables and how much is due to pseudorandomness can not be known- at least I don't know how you would know it- so to say it "does occur"- although observable doesn't really show if it truly is random.
I'm not trying to say or prove that there is no "random variance", I'm trying to show exactly what you just said. We can't be sure "there is an effect"- true- but likewise we can't be sure it is due to randomness that we observe the variance- but we also accept it as random, why not as an effect? We are preferentially choosing to go with random, even though the evidence would be equal because of the unknown effect of pseudorandomness.
I don't think there is anything without doubt especially randomness.
Still you can't differentiate with high certainty that it isn't pseudorandomness because all your tests would appear random.
I agree, I was simply trying to say that we take the other assumption (of randomness) but not of control. But it just means its presumed.
Brownian motion although 'random' can not be explained truly by physic laws, if it could then it would only be shown to be control because the forces that "push" it to jump one way are still controlling it- the difficulty is actually measuring each force- which theoretically could be done but not practically- and so the 'random' motion again is 'random' only because of the unknown magnitudes of the forces that act on it- which leads to a very unpredictable motion.
Well I do find it odd in some respects but I find it very correct on the fact that everything is controlled by the physical laws and that the unpredictability is there due to the sensitive nature of the conditions. But I guess its a difference of opinion between you and me.
They would be considered random- but like I said before you can't differentiate the effects to be truly random.
Correct, but why chose either side? Just change the word. Again I'm not trying to prove that randomness exists or does not exist but simply that the concept itself is confused by many, secondly there is no definitive way of differentiating pseudo-randomness, and since all mutations are not entirely random then to use it in general terms then when people read the theory they get the wrong idea. I'm only asking for a change of terms which can hopefully better represent the knowledge we actually have.
Peace be unto you
Well, let's try this road: clonal organisms still show residual variance under common environment. Can this be unexplained control factors?
Not precisely sure what you mean here, but this isn't so: with sufficient sample size and phenotypic differentiation in a mapping population (n = 1000, a = 2sigma or so, for example), you should be able to detect genes with as little as 2-3% effect, at a p-value of 99% or higher. That's pretty high certainty.
With respect to test being considered, it is random where there's no clear pattern to the residuals.
Well, this is a very good point: but higher order-math still is very descriptive of stochastic processes - http://en.wikipedia.org/wiki/Wiener_process.
Correct, but why chose either side? Just change the word. Again I'm not trying to prove that randomness exists or does not exist but simply that the concept itself is confused by many, secondly there is no definitive way of differentiating pseudo-randomness, and since all mutations are not entirely random then to use it in general terms then when people read the theory they get the wrong idea. I'm only asking for a change of terms which can hopefully better represent the knowledge we actually have.
Peace be unto you
And to you. The ultimate question is whether you think anything at all can be random either in process or outcome. Can a coin really produce a random distribution or no? Is the process non-random? And the ultimate question: all right, so what if it weren't?
I know this is off topic but do you know what east Korea is?
GeoffP knows what East Korea is better than anyone.