Understanding through Discussion


Welcome! You are not logged in. [ Login ]
EvC Forum active members: 78 (8896 total)
Current session began: 
Page Loaded: 03-22-2019 7:52 AM
43 online now:
PaulK, RAZD, Tangle (3 members, 40 visitors)
Chatting now:  Chat room empty
Newest Member: WookieeB
Post Volume:
Total: 848,541 Year: 3,578/19,786 Month: 573/1,087 Week: 163/212 Day: 5/25 Hour: 0/0


Thread  Details

Email This Thread
Newer Topic | Older Topic
  
Prev1
2
Author Topic:   Probability-based arguments
PaulK
Member
Posts: 14749
Joined: 01-10-2003
Member Rating: 1.7


Message 16 of 27 (485187)
10-06-2008 7:23 AM
Reply to: Message 1 by boysherpa
10-04-2008 9:13 PM


quote:

Example 1: This takes place over tens of millions of years, and is therefore possible (implying a number of trials over which random mutations may operate)
Example 2: The odds of a jawless fish becoming "jawed", with fully operating nerves, etc., are found to be 1:10^some big number, and is therefore not possible

As they stand both arguments are wrong, and I wouldn't call either "probability-based". The first is not completely without merit - it is valid to point out that the timescales are relevant to any attempt to calculate the probability, giving more time for the needed variations to appear and for natural selection to act on them.

The second is worse - it is dishonest, since it is only an argument from personal incredulity dressed up with a worthless guess at the probability. (It might be possible to make a very rough estimate of the correct figure, but it would be a lot of work and require some very detailed knowledge - I personally doubt that it would be worth attempting).


This message is a reply to:
 Message 1 by boysherpa, posted 10-04-2008 9:13 PM boysherpa has not yet responded

    
Wounded King
Member (Idle past 2171 days)
Posts: 4149
From: Edinburgh, Scotland
Joined: 04-09-2003


Message 17 of 27 (485211)
10-06-2008 11:26 AM
Reply to: Message 15 by boysherpa
10-05-2008 11:55 PM


Re: Not equivalent
I give credit to Behe's book "Darwin's Black Box" for asking good questions, although I am not sure his analyses was thorough or that he conferred before release.

I find it hard to see in what way 'Darwin's black box' asks good questions except perhaps to the extent that it poses the question 'How can complexity evolve?' In what way is it anything other than a reworking of the original argument from design?

Your own rating of the response by "the scientific community" seems to be as pulled from thin air as the probability estimates in your OP. To what extent have you looked for response in the scientific literature? What shortcomings did you feel they had?

Can you really point out an example of a response from "the scientific community" equivalent to the "hundreds of millions of years" argument. I can imagine seeing it on a message board but most of the people arguing on EvC message boards are not representative of "the scientific community".

If you look at his more recent work, such as 'Edge of Evolution', you will see that Behe makes exactly the same sort of probability arguments that you are lamenting, see http://www.arn.org/docs/behe/mb_smu1992.htm . Indeed his peer reviewed research paper (Behe and Snoke, 2004) on protein evolution is based around developing a simulation allowing him to produce yet more specious probability calculations.

He asked reasonable questions. TSC did not give reasonable responses

Again examples definitely needed to support this assertion, from both ends. What reasonable questions were asked? What responses do you consider to have been unreasonable?

It is also disingenuous to ignore that the reason many people treated Behe as a crackpot was not the questions he asked but the answers he came up with for them, essentially 'God did it', and his pronouncement that coming up with such reasoning "must be ranked as one of the greatest achievements in the history of science", rivalling "those of Newton and Einstein, Lavoisier and Schrodinger, Pasteur, and Darwin".

Making Behe out to be a martyr rather than a crackpot makes me seriously question your judgement.

TTFN,

WK


This message is a reply to:
 Message 15 by boysherpa, posted 10-05-2008 11:55 PM boysherpa has responded

Replies to this message:
 Message 18 by boysherpa, posted 10-11-2008 7:41 PM Wounded King has responded

    
boysherpa
Junior Member (Idle past 3448 days)
Posts: 19
From: Lomita, CA
Joined: 10-04-2008


Message 18 of 27 (485801)
10-11-2008 7:41 PM
Reply to: Message 17 by Wounded King
10-06-2008 11:26 AM


Re: Not equivalent
slow down, WK
Making Behe out to be a martyr rather than a crackpot makes me seriously question your judgement.

The propensity to make personal judgments is often a sign of weakness of intellectual position, but I digress. No, I don't give a hoot about Behe's thesis or intelligent design leanings. The complexity question is the only reason I bring it up. As far as examples of how the "millions of years" response is used, actually, I tend to see that in popular science quite regularly. Remember, my interest here is in probability. However, when I have looked at scientific literature the results are mixed. For some macro-scale elements, good results are found, such as the evolution of the eye. The jawed fish is relatively weak. The upright human debate is more speculation than anything else. There simply is not enough data available to explain, since inference is the chief tool. I admit to being less well-versed to the world of biochemical evolution, but the responses to Behe's book seemed more diatribe than substance, from a scientific viewpoint.

However, an objective observer is not satisfied with merely a sparse fossil record to explain complex transitions. (Relatively recent publications re mammalian inner ear tracing do show some exciting lineage, and bear promise to real evolutionary proof rather than inference).

In a probabilistic sense, the evolutionary scenarios I would pull from Behe's questions are the following:
1. How to account for the fact that an existing system can adapt to another purpose, in other words, each feature does not have to evolve from scratch in parallel and independently.
2. How to model the probability that parallel required system elements will be in place at the required time (the jawed fish example works at the macro scale, but several biochemical system work at the micro scale).
3. General Note: The power of the biochemical resources available for evolution is extraordinary. Simple probability which does not recognize the cumulative capabilities of this "system" and its self-organization capabilities can never capture the concept of evolutionary change.


This message is a reply to:
 Message 17 by Wounded King, posted 10-06-2008 11:26 AM Wounded King has responded

Replies to this message:
 Message 19 by PaulK, posted 10-12-2008 12:55 PM boysherpa has responded
 Message 20 by Wounded King, posted 10-14-2008 10:17 AM boysherpa has responded

  
PaulK
Member
Posts: 14749
Joined: 01-10-2003
Member Rating: 1.7


Message 19 of 27 (485853)
10-12-2008 12:55 PM
Reply to: Message 18 by boysherpa
10-11-2008 7:41 PM


Re: Not equivalent
quote:

As far as examples of how the "millions of years" response is used, actually, I tend to see that in popular science quite regularly.


I very much doubt that. I've never seen it.

And Behe's argument is very weak. He only deals with the indirect routes to evolving irreducible complexity by writing them off as "improbable". Yet decades earlier, Mueller had predicted that evolution would produce irreducible complexity.

In the Dover case Behe wrote off the work on the evolution of the immune system because it wasn't detailed to a point that is completely impractical.

The question is, why bother with probability-based arguments at all, when we have no adequate basis for the calculations ?


This message is a reply to:
 Message 18 by boysherpa, posted 10-11-2008 7:41 PM boysherpa has responded

Replies to this message:
 Message 22 by boysherpa, posted 10-18-2008 11:30 PM PaulK has responded

    
Wounded King
Member (Idle past 2171 days)
Posts: 4149
From: Edinburgh, Scotland
Joined: 04-09-2003


Message 20 of 27 (485985)
10-14-2008 10:17 AM
Reply to: Message 18 by boysherpa
10-11-2008 7:41 PM


Re: Not equivalent
As far as examples of how the "millions of years" response is used, actually, I tend to see that in popular science quite regularly.

As PaulK suggests, this is the very opposite of supporting your contention, you are simply reiterating the bare assertion. Plus there is the question of whether 'Popular Science' is representative of the 'scientific community' in any way whatsoever.

I admit to being less well-versed to the world of biochemical evolution, but the responses to Behe's book seemed more diatribe than substance, from a scientific viewpoint.

And again. Please stop just making the same bare assertions. Can you give an example where the 'scientific community' particularly failed to address Behe's argument?

Remember, my interest here is in probability. However, when I have looked at scientific literature the results are mixed. For some macro-scale elements, good results are found, such as the evolution of the eye. The jawed fish is relatively weak. The upright human debate is more speculation than anything else.

What do any of those things have to do with probability? If you think you can make a probability argument based on any of those things you are making exactly the sorts of error you started out talking about. In the absence of a time machine inference is the only tool we can use on these questions. Any probability estimate will surely be the result of extensive inference itself?

However, an objective observer is not satisfied with merely a sparse fossil record to explain complex transitions. (Relatively recent publications re mammalian inner ear tracing do show some exciting lineage, and bear promise to real evolutionary proof rather than inference).

I think you need to explain your thinking here a bit more clearly. What exactly would 'real evolutionary proof' look like? Exactly how would it nor involve inference? When does drawing a connection between fossils stop being a matter of inference? Does extending what we see during the development of extant species to cover extinct species not rely heavily on inference of the genetic basis of the development of extinct species whose genetic makeup is almost universally unknown to us?

I don't believe 'real evolutionary proof' of the sort you ask for can be drawn from the fossil record. Indeed I doubt any probabilistic model based approach could ever provide it. Our 'proofs' such as they are must come from the observations we can make on extant organisms and we have to rely on inference of principles from such research to apply them to what we see in the fossil record.

As to your questions ...

1. How to account for the fact that an existing system can adapt to another purpose, in other words, each feature does not have to evolve from scratch in parallel and independently.

This depends to some extent on whether you wish to retain the original 'purpose' or not. In any case each answer to this question is going to be highly dependent on the actual case. There are a multitude of answers ranging from the loss of need for the original function, the addition of a novel function while retaining the previous function, the duplication of genetic elements allowing the parallel evolution of a separate duplicate system while retaining the function of the original. I would suggest that in most cases it is far easier to explain the development of many traits by co-option and modification of redundant elements of existing systems than by the de-novo development of such traits. Are you thinking of something specific here?

2. How to model the probability that parallel required system elements will be in place at the required time (the jawed fish example works at the macro scale, but several biochemical system work at the micro scale).

What possible reason would you have to do this? This seems a pointless exercise in the sort of numbers out of a hat navel gazing I thought you were originally objecting to. What 'required time' is there? Are you trying to work out the probability that one particular course of evolutionary history, the one we ourselves are a product of, would occur? Why? What possible value does such a calculation have? The chances are obviously going to be astronomical, but that doesn't make them meaningful, it is exactly like the premise of the tornado in a junkyard argument. or Yockey's essentially meaningless probability estimates for a specific amino acid sequence to appear de-novo.

3. General Note: The power of the biochemical resources available for evolution is extraordinary. Simple probability which does not recognize the cumulative capabilities of this "system" and its self-organization capabilities can never capture the concept of evolutionary change

And actual evolutionary biologists and geneticists are using it every day in their research, as they were at the time Behe first published 'Darwin's Black Box'. Behe's apparent unfamiliarity with the actual current extant research was one of the major criticisms leveled at the book, i.e. on the evolution of the blood clotting cascade and other supposedly IC systems.

Molecular evolution is perhaps an area where we can think of our inferences as being strongest since they are based on a large body of experimentally derived observations. But even then the estimate of an ancestral gene or protein sequence is still surely just an inference. The probabilities involved surely come from some level of inference inherently like those represented in Bayesian and Maximum-likelihood calculations of phylogeny.

TTFN,

WK


This message is a reply to:
 Message 18 by boysherpa, posted 10-11-2008 7:41 PM boysherpa has responded

Replies to this message:
 Message 21 by boysherpa, posted 10-18-2008 11:24 PM Wounded King has responded

    
boysherpa
Junior Member (Idle past 3448 days)
Posts: 19
From: Lomita, CA
Joined: 10-04-2008


Message 21 of 27 (486358)
10-18-2008 11:24 PM
Reply to: Message 20 by Wounded King
10-14-2008 10:17 AM


Re: Not equivalent
Wounded King, you seem to understand the point of my question - but you get distracted very easily. Truly, I am uninterested in your debate over Behe's positions. It is like the tide fighting the sand - lots of action, but just wears after a while.

Here are some items where I do not seem to be clear:
1. The probabilistic model I am investigating has nothing to do with proving/disproving a particular instance of evolution or evolution in general. It is a mathematical modeling investigation, and it is challenging to collect the models and data.
2. It is not clear that folks understand what a probabilistic model is versus a descriptive model. A descriptive model cannot predict whether something will happen again. A descriptive model may or may not be correct as to the occurrence it covers. A probabilistic model must work with existing examples, predict new instances, and give insights into operating mechanisms.
3. Think carefully about your statement "What possible reason would you have to do this?" If you understood developments in Physics and Chemistry over the past 150 years, you would have your answer. I admit, this is a daunting task. Complex is not equivalent to inexplicable (Faraday's Laws - deterministic) or unpredictable (quantum mechanics - probabilistic).

As to the issues of "evolutionary proof", you have yourself stated what I have observed - inference is the only proof available for evolutionary change. At the macro level, scarcely more than sparse fossil records fuel that inference. As to whether this is quantitative proof, hardly. It is merely a reasonable inference. Any day a better explanation for the data could be offered, and this does happen. However, the record, the science, and the inferences all constitute data.

My question is deals with all of that data. Can we use this data to build probabilistic models able to reasonably predict.

Consider the example of the eye evolution (Details are not important for this question). Probabilistic models need multiple runs of experiments to collect statistical data (like rolling a six-sided die many times). The fossil record normally represents one such experiment. This is why the probability exercise seems futile - not for the reasons you suggest. But, should we be able to outline the fundamental elements required for the steps of eye evolution, this then would be like imagining the rolling of a that die given its structure and correlation assumptions. There are numerous examples of "lower" eye structures in living organisms which can serve as independent "experiments" in our development of the model. This is how the probabilistic model could be developed. Your "tornado in a junkyard" comment seem to express more the complexity of the task than an understanding of how such modeling could be done.

The biological sciences has relatively recently been able to rise above qualitative modeling. Using such models allows for the ability to predict. This is the big difference between biological complex system sciences and those in the so-called "hard sciences".


This message is a reply to:
 Message 20 by Wounded King, posted 10-14-2008 10:17 AM Wounded King has responded

Replies to this message:
 Message 24 by Wounded King, posted 10-20-2008 1:07 PM boysherpa has responded

  
boysherpa
Junior Member (Idle past 3448 days)
Posts: 19
From: Lomita, CA
Joined: 10-04-2008


Message 22 of 27 (486359)
10-18-2008 11:30 PM
Reply to: Message 19 by PaulK
10-12-2008 12:55 PM


Re: Not equivalent
OK, i will accept the challenge. Give me accessible references for evolution and I will see if they use the "millions of years" terminology.

By "accessible" is meant

Something able to be accessed without cost or undue trouble - I am not spending much effort on this. Also, this should deal with evolution in a more popular sense - if it is a purely scientific journal then the vast majority of the population will not read it. When I published I knew that no more than 3 people read my papers, and I forced one person to read them.

Keep in mind that I am a supporter - I am doing this to kill this ridiculous argument.


This message is a reply to:
 Message 19 by PaulK, posted 10-12-2008 12:55 PM PaulK has responded

Replies to this message:
 Message 23 by PaulK, posted 10-19-2008 4:19 AM boysherpa has responded

  
PaulK
Member
Posts: 14749
Joined: 01-10-2003
Member Rating: 1.7


Message 23 of 27 (486361)
10-19-2008 4:19 AM
Reply to: Message 22 by boysherpa
10-18-2008 11:30 PM


Re: Not equivalent
quote:

OK, i will accept the challenge. Give me accessible references for evolution and I will see if they use the "millions of years" terminology.

I don't have to give you references. YOU have to provide them since YOU claim that they exist.

Also it is not a matter of terminology. You need to back up your claim that this argument is frequently used:


Example 1: This takes place over tens of millions of years, and is therefore possible (implying a number of trials over which random mutations may operate)

quote:

Keep in mind that I am a supporter - I am doing this to kill this ridiculous argument.

The evidence suggests that you are falsely claiming that popular science books often use a bad argument. That's not something a supporter would do.


This message is a reply to:
 Message 22 by boysherpa, posted 10-18-2008 11:30 PM boysherpa has responded

Replies to this message:
 Message 26 by boysherpa, posted 10-31-2008 12:31 PM PaulK has responded

    
Wounded King
Member (Idle past 2171 days)
Posts: 4149
From: Edinburgh, Scotland
Joined: 04-09-2003


Message 24 of 27 (486427)
10-20-2008 1:07 PM
Reply to: Message 21 by boysherpa
10-18-2008 11:24 PM


Not getting anywhere.
Truly, I am uninterested in your debate over Behe's positions.

Then why bring him up? Don't you think people are going to question your actual grasp of an issue when you say your questions are based on those of the principle proponent of ID, someone with a poor track record of following current evolutionary theory himself, and saying how you feel the scientific community failed to address them? You

3. Think carefully about your statement "What possible reason would you have to do this?" If you understood developments in Physics and Chemistry over the past 150 years, you would have your answer. I admit, this is a daunting task.

Why do you insist on these vacuous non-answers? Are you incapable of responding clearly? Do you not understand why the concept of a 'required time' is a nonsense unless you are operating under some very bizarre assumptions regarding the way evolution works?

What purpose, other than fatuous arguments from big numbers, does calculating the probability for the specific elements of any particular system coming together at a particular time serve? In what way does it differ from Hoyle's tornado in a junkyard or Hubert Yockey's pointless calculation of the probability of the exact modern sequence of cytochrome C enzyme forming essentially de novo?

There are numerous examples of "lower" eye structures in living organisms which can serve as independent "experiments" in our development of the model. This is how the probabilistic model could be developed.

This is as clear as mud. Unless you believe all of the eye structures come from completely seperate evolutionary lineages how can you call them independent? The fact that many of the same genes are involved in eye development in vertebrates and invertebrates argues against the independence of these 'experiments'.

The only are where I can see any chance of such an approach working would be a highly controlled cell culture type of experiment. A high selective pressure, such as an antibiotic, on a population with a thoroughly sequenced genome. With a detailed enough knowledge of the biochemistry of the bacterium and the rates of mutation it is susceptible to we might be able to predict the likeliest routes to antibiotic resistance or strategies of avoidance that it might evolve.

Outside of such a limited scope scenario I just don't see how we can hope to have a full enough picture of all the variables involved to make any sort of even tenuously reliable prediction.

TTFN,

WK


This message is a reply to:
 Message 21 by boysherpa, posted 10-18-2008 11:24 PM boysherpa has responded

Replies to this message:
 Message 25 by boysherpa, posted 10-31-2008 12:27 PM Wounded King has not yet responded

    
boysherpa
Junior Member (Idle past 3448 days)
Posts: 19
From: Lomita, CA
Joined: 10-04-2008


Message 25 of 27 (487426)
10-31-2008 12:27 PM
Reply to: Message 24 by Wounded King
10-20-2008 1:07 PM


Re: Not getting anywhere.
WK, there is something in the style and content of your reply that is troublesome. I quote directly:


3. Think carefully about your statement "What possible reason would you have to do this?" If you understood developments in Physics and Chemistry over the past 150 years, you would have your answer. I admit, this is a daunting task.

Why do you insist on these vacuous non-answers? Are you incapable of responding clearly? Do you not understand why the concept of a 'required time' is a nonsense unless you are operating under some very bizarre assumptions regarding the way evolution works?

You left off the last line of the paragraph which does give a clear example of how quantitative and predictive modeling have been successful. Then, you call my response vacuous and unclear. Motive is questionable here. So, you want some specifics. How about radio wave emission from AC current in a wire (as modeled quantitatively by Faraday's laws). How about probability amplitude models (Schroedinger's equation) being able to predict the structure of simple atoms (predictive probabilistic model). This is what was referred to in the sentence you dropped.

But, to deal with the topic, the point is that quantitative modeling is important, even if you don't understand why. The hard sciences have had enormous success with quantitative modeling, while the evolutionary sciences have been limited to mostly descriptive modeling (although significant progress has been made since computation power has allowed, hmm, complex problems such as genetics to be addressed).

Predictive and quantitative models allow you to move forward. Descriptive models do not.


This message is a reply to:
 Message 24 by Wounded King, posted 10-20-2008 1:07 PM Wounded King has not yet responded

  
boysherpa
Junior Member (Idle past 3448 days)
Posts: 19
From: Lomita, CA
Joined: 10-04-2008


Message 26 of 27 (487428)
10-31-2008 12:31 PM
Reply to: Message 23 by PaulK
10-19-2008 4:19 AM


Re: Not equivalent
PaulK, you misread my reply.

I was suggesting that you provide acceptable publications and I will find the references. It was an attempt to avoid the situation of providing references which you would claim are not credible because of source quality.

Sheesh, you people are touchy. Must be the forum. Argument prone.


This message is a reply to:
 Message 23 by PaulK, posted 10-19-2008 4:19 AM PaulK has responded

Replies to this message:
 Message 27 by PaulK, posted 10-31-2008 3:56 PM boysherpa has not yet responded

  
PaulK
Member
Posts: 14749
Joined: 01-10-2003
Member Rating: 1.7


Message 27 of 27 (487454)
10-31-2008 3:56 PM
Reply to: Message 26 by boysherpa
10-31-2008 12:31 PM


Re: Not equivalent
quote:

I was suggesting that you provide acceptable publications and I will find the references. It was an attempt to avoid the situation of providing references which you would claim are not credible because of source quality.

I note that you are ignoring the other important point. That it is not terminology that is of concern, but of the specific argument you refer to.

And it is still up to you to provide the sources that you claim use this argument. It seems a little odd for you suddenly be concerned that you may have been relying on "low quality" sources. (The more so. given the number of popular books on evolution written by scientists who have actually worked on evolution). Why are you so reluctant to actually give real examples ?


This message is a reply to:
 Message 26 by boysherpa, posted 10-31-2008 12:31 PM boysherpa has not yet responded

    
Prev1
2
Newer Topic | Older Topic
Jump to:


Copyright 2001-2018 by EvC Forum, All Rights Reserved

™ Version 4.0 Beta
Innovative software from Qwixotic © 2019