|
Author
|
Topic: Structure of Scientiifc Anlalysis
|
warren_bergerson
Member
Member # 262
|
posted 03. March 2003 13:43
Frances,
Just a reminder, the subject of this thread is ‘standards for evaluating theories’. The discussion of ‘universal genetic code’ was intended to provide a broad overview of the results that might result from the standards I proposed.
There is no disputing the fact that individuals working in evolutionary biology have speculated on possible explanations for DNA and the evolution of DNA. There is also no disagreement on the fact that some authors have labeled proposed explanations as scientific hypothesis and/or scientific theories.
The questions of concern here, and questions which you have not addressed, are:
1. Do the scientists offering potential explanations of phenomena such as universal genetic codes, actually assert that the explanations offered constitute valid, predictive, mathematical scientific theories logically compatible with some overall predictive model or theory of evolutionary change processes? and
2. If the proposed explanations are in the form of predictive theories, would the proposed theories pass the standards proposed?
3. If the proposed explanations will not pass the standards I proposed, are there some other explicitly formulated objective standards which can be satisfied?
It might in the discussion here be useful to distinguish between 1)direct- formally presented models, theories and hypothesis presented and explicitly supported by an individual scientists, and 2)indirect or secondary interpretations of models, theories and hypothesis. As Cornelius has pointed out, there would appear to be significant differences between the direct models, theories, and speculation available in the literature, and the secondary interpretations of these theories.
Again a reminder, the subject here is standards for evaluating theories and the potential impacts of specific types of standards.
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 04. March 2003 01:28
Warren: Your questions are good ones but they are somewhat puzzling since without knowing their answers you seem to already have reached the conclusion "One goal of the proposed standards is to eliminate or reject theories which do not meet rigorous standards. This, it appears would include most of the evolutionary and genetic theories currently accepted as valid."
I guess you may have been jumping to conclusions rather than follow the actual trail of evidence. I have found this somewhat frustrating with many of your postings in which you make many assertions about what RM&NS or Darwinism can or cannot do without much supporting evidence.
Cornelius: You ask for examples of predictions and falsifications of various hypotheses for the genetic code. I once again refer you to the references I posted in this regard.
Your redefinition of evolution to be "species arising through natural means" also seems to be somewhat ad hoc and in contradiction which much of the scientific literature. This makes a discussion somewhat complicated. It would be helpful if we can at least work from common definitions and try to avoid such non sequiturs as confusing non-evolutionary mechanisms with non-natural. Such rethorical arguments make a discussion quite difficult and only serve to undermine one's argument when exposed.
I am somewhat confused by your arguments about the universal genetic code as well as its evolution. Common descent surely would predict that we can trace back the code to a universal ancestral form and evolution explains why some variants are found in the code.
Yes evolution as a mechanism involves most commonly variation and selection, not to be confused with what is refered to as the fact of evolution which is the observed common ancestry. I believe that the confusion about what evolution predicts versus what the mechanisms of evolution predict may be at the foundation of our disagreements. The fact of evolution, that is common descent would predict a universal genetic code which can be traced back through time. Evolutionary mechanisms explain variations of these codes. Understanding the restrictions of variations in these codes, not surprisingly, explains why the genetic code has remained quite well preserved over time.
As far as the RNA world and Knight is concerned, see my final quote from a lecture Knight gave.
quote:
These calculations let me figure out how surprising it should be to find particular motifs. Here's a plot of the length of the motif against the number of random sequences you'd need to search to have a 99% chance of finding it, for the solid lines, and a 1% chance of finding it, for the dashed lines. The isoleucine motif has two modules and 12 bases, and you'd need to search about 100,000 sequences to find it; the hammerhead has three modules and 11 bases, and you'd need to search about 1000. Note that this is just for the sequence requirements: there's no guarantee that they'd fold into the right structure. However, the amazing thing is that this is about a billion-fold fewer sequences than the 10^15 typically used in SELEX!
He has also some very interesting papers on this topic. As I said, check out his website.
[ 04. March 2003, 01:40: Message edited by: Frances ]
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 04. March 2003 02:22
Frances:
You write: "Common descent surely would predict that we can trace back the code to a universal ancestral form"
I would not say this is surely the case. The hypothesis of common descent does not, in itself, predict such traceability. Since the code can evolve (see below) then there could be different codes. Similarities between the different codes could have been lost if the evolutionary rates were sufficiently high relative to the time since divergence. And of course, evolution need not limit itself to a single common ancestor. So there is no prediction that we can trace back the code to a universal ancestral form.
You wrote a few posts back: "The theory of evolution once again does not require there to be a unique code."
Agreed. The bottom line is evolution does not hypothesize that the code poofed into existence. It is supposed to have evolved from simpler codes (somehow). Since evolution holds that the code can evolve, it therefore could easily accommodate a world where there is not a universal code, but instead are multiple codes. This fact, coupled with the fact that evolution cannot explain the evolution of the code beyond speculation, suggests to me that the universality of the code would not be cited as strong, compelling evidence for evolution. Yet you and evolutionists say it is strong, compelling evidence. To my way of thinking there is a disconnect here. All I'm asking is for you to fill me in on the missing steps of the proof.
As for my definition of evolution, I'm sorry you interpreted it as some sort of rhetorical ploy. I'm simply trying to reflect the theory as best I can. It is clear from the literature that {RV+NS} is a subhypothesis that evolutionists would do without if a better one came along. The more important tenet of evolution is that the species must arise via natural means, and right now {RV+NS} is considered to be the predominant mechanism of change. For example, evolution critics have been dinged for being ignorant of the true theory of evolution for referring to {RV+NS} too much. And of course, Darwin himself made as much clear:
"Whether the naturalist believes in the views given by Lamarck, by Geoffroy St. Hilaire, by the author of the ‘Vestiges,’ by Mr. Wallace or by myself, signifies extremely little in comparison with the admission that species have descended from other species, and have not been created immutable: for he who admits this as a great truth has a wide field open to him for further inquiry." [C. Darwin, Quoted in John C. Greene, Science, Ideology, and World View, p. 52, University of California Press, Berkeley, 1981.]
So please don't think badly of me for using "species arising through natural means" as a definition for evolution. When you say Knight has non-evolutionary hypotheses for the origin of the DNA code, I understand you are probably using a more narrow definition of "evolution." That is fine, but you made the point in order to counter my claim that Knight presupposes evolution. Surely, you're not saying that because Knight considers code origination mechanisms not involving {RV+NS} that he is therefore not presupposing evolution?
This is germane. For example, in this thread you have made several statements which appear to make little sense unless they are interpreted as being from within the paradigm (ie, you are presupposing evolution). For example, you said evolution is a fact, and you said the RNA world hypothesis is practically a certainty. And you said Knight has supplied powerful explanations for the origin of the code and you have said the universality of the code is strong and compelling evidence for evolution. Do you see what I'm getting at? There is some sort of incommensurability going on, as these statements, on the face of it, make little sense. This is no slight on you, as we've discussed here in this thread, these statements are common in the evolution literature. Perhaps the explanation is that you and evolutionists are presupposing the truth of the theory. If not, then what?
--Cornelius
[ 04. March 2003, 03:11: Message edited by: Cornelius G. Hunter ]
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 04. March 2003 11:57
I think I see where the confusion lies. When I refer to the universal genetic code, I am refering to the code used by the last common ancestor. Common ancestry predicts that when looking back we should be able to trace back to a common ancestor.
Looking forward from the common ancestor we see that the genetic code did evolve, although slightly because it may not be a surprise that the genetic code is well conserved just like many of the hox genes. Thus the fact of evolution, that is common descent would predict that we can trace back the code to such an ancestor, the theory of evolution can explain variations and indeed it seems that the variations on the genetic code share a common ancestry.
You then asserts that whatever evolutionary mechanisms played a role since the LCA must have played a role in the 'evolution' of the genetic code. Here is where the risk of equivocation on the term evolution runs the risk of clouding the topic. Lets define some terms so that we all speak the same language.
Fact of evolution: Common descent
Theory of evolution: Neo-Darwinism
So how did the genetic code come about? Necessity?, accidental? or a combination? Various hypotheses stress various aspects.
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 04. March 2003 12:11
In introducing this thread, I proposed defining structured scientific analysis standards in terms of requirements for defining structure, variables, and causal relationships. The proposal was based on the logic that if 1)you defined rigorous objective standards for defining variables, etc. then 2)you could determine if existing models and theories in fields like genetics or evolutionary biology meet the defined standards and 3)if existing models and theories do not satisfy standards then develop new procedures and models etc..
It will be noted that this ‘establish objective standards’ approach is often quite effective in applied science and business applications. If you can establish explicit objective standards, and if there is an appropriate incentive for meeting standards, then you frequently see very rapid progress toward meeting and exceeding the defined standards. If you look at the discussion here, it seems very doubtful ‘objective standards for defining variables, etc.’ can ever result in agreement on whether existing theories and models meet standards. Without the ability to reach agreement on such fundamental issues, there is no hope that the standards defined could produce progress toward better models and theories.
It might be interesting to explore why Cornelius, Francis and myself can not seem to reach agreement on such an elementary issue as ‘the existence of a valid mathematical predictive model and/or theory of genetics and genetic change’. To address this question we would need to look at human decision making behavior. How can three people look at the same set of publicly available information and come to 3 different/contradictory decisions or interpretations? Any interesting question but probably a topic for another thread.
SUMMARY
Even a limited number of responses suggests that ‘standards for defining and testing variables, structure and causal relationships’ are unlikely to be productive in resolving disagreements relating to theories and models of evolution and/or genetic change. IMO, the basic technique of ‘defining and applying objective standards’ can under appropriate conditions be effective. The ‘problem’, again IMO, with attempting to establish objective standards for defining scientific models and theories, is the issue of how these standards are to be interpreted, and even more, an issue of who gets to be the authority on interpreting standards.
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 04. March 2003 12:21
Frances:
I see no problem with defining the universal genetic code as the code used by the hypothetical last common ancestor, and I don't think your using this definition is adding any confusion. What you are explaining is the evolutionary view of how to interpret the evidences (ie, variations in the code are due to small-scale evolution since the last common ancestor). Yes, I'm familiar with the evolutionary explanation. What I don't understand, and you seem to have difficulty explaining, is why the universality of the code is such powerful, compelling evidence, given that evolution does not require such universality and that it can't explain how the code arose, beyond speculation.
Frances wrote: "You then assert that whatever evolutionary mechanisms played a role since the LCA must have played a role in the 'evolution' of the genetic code."
I made no such assertion. What I asserted is that evolutionary theory could accommodate multiple codes just as easily as a single universal code.
Frances wrote: "Let's define some terms so that we all speak the same language. Fact of evolution: Common descent; Theory of evolution: Neo-Darwinism"
Do you understand that evolution is a fact only to those who believe in it? And can you explain why the universality of the code is such powerful, compelling evidence, given that evolution does not require such universality and that it can't explain how the code arose, beyond speculation?
--Cornelius
IP: Logged
|
|
Noel Rude
Member
Member # 516
|
posted 04. March 2003 15:42
Whereas Frances may be more convinced of common descent than the facts warrant, I think he is right to point out the difference between a theory of common descent (evolution) and the materialist's explanation (Darwinism).
But as for Darwinism predicting a common DNA code, I have to agree with Mr. Hunter: one is hard pressed to imagine that Darwinism even predicts that there should be a code in the first place. Even common descent is not a prediction except in so far as we are forced, by the sheer unlikelihood of our materialist theory to produce anything, to postulate no more than a single origin. On the other hand, common descent would be predicted by any theory of Intelligent Design that was based on what we know about design (meaning from human artifice).
And given that there is a genetic code and that materialists are forced into postulating a common origin (re the rarety of serendipitous chance), Darwinism does not predict any stability in the code over time. Nor does it predict instability.
As an analogy ... some linguists suspect a common origin for all human languages, but if that origin was as far back in time as we suspect then there is no reason that any evidence (cognate forms) of that "Proto-World" should survive. All languages are quite alike (clauses, semantic roles, discourse coherence, etc.) but this derives from biological innateness, functional imperatives and/or the Platonic realm -- it does not exemplify features conserved from a common origin down through eons of historic change.
Hope this isn't straying too far afield from this thread ... but for those of us who admit to the philosophical basis of our theories of origins ... might it not be helpful to distinguish what those variant philosophies predict?
The Western theological tradition, with its extreme transcendance and utter otherness, though it demanded a beginning it nevertheless saw no need for development -- there was no prediction of evolution -- except that as theology evolved into Deism no place was found for creation either and thus the question of origins could be ceded to the materialists.
Materialism, on the other hand, most naturally opts for steady state theories -- except that geology-paleontology insisted on a history of development and later cosmology even asked us to believe in a beginning. In this case materialism -- though it doesn't predict evolution -- it must come up with a theory to account for it. There is also the matter that it seems less odd to let chance work its magic over time than in one fell swoop. Historic development (the geologic column, star factories, etc.) is more in line with the reality we see.
But ... as I said ... any theory of ID based not on the pronouncements of theology but on what we observe in ourselves also predicts evolution. The question then: Which kind of evolution -- one more in line with technological evolution (which we can study) or serendipity fueling success (about which we can tell stories) -- best fits the facts?
[ 04. March 2003, 17:54: Message edited by: Noel Rude ]
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 05. March 2003 02:26
Noel,
May I point out some errors in your arguments. First of all Darwinism is not a materialistic explanation but a scientific explanation. I understand that the concept of materialism/naturalism is often confused with methodological naturalism and I do not fault you for this but I want you to realize that there are some differences.
As far as your suggestion that Darwinism does not predict that there be a code, I am somewhat lost for words. While Darwinism itself does not predict any particular form of code, it does work from the observation that there is some form of information storage which is passed on between generations and which can vary.
I am not sure what you mean by your claim that "Even common descent is not a prediction except in so far as we are forced, by the sheer unlikelihood of our materialist theory to produce anything, to postulate no more than a single origin. On the other hand, common descent would be predicted by any theory of Intelligent Design that was based on what we know about design (meaning from human artifice)."
Both assertions seem to be somewhat lacking in supporting detail but I am sure that this is an oversight that can be corrected. I personally would love to hear more about how you reached these conclusions.
Your use of terms such as materialist and other non sequiturs tends to distract from the discussion but I am willing to entertain the possibility that they have relevance thus I encourage you to elaborate in some more detail on how you reach these conclusions.
Given the fact of common descent, a common universal code, looking backward is inevitable. That looking forward from the LCA, the code did slightly change is somewhat surprising at first since one would expect the code to be quite well conserved but can be explained once realizing the mechanisms involved. Nevertheless the fact that these codes can be traced back to a common ancestors seems quite relevant.
I understand that your main expertise is linguistics, a topic which is also dear to my heart. Even linguistics seems to have been able to detect similar evolutionary patterns, not surprisingly I might add, in languages.
Certainly I find some comfort in your suggestion that ID also predicts evolution, after all the facts are hard to argue. Could you explain this ID prediction in more detail?
Also I wonder what you mean by technological evolution and serendipity? How does ID and Darwinism fit into these categories? I would love to hear a linguist perspective on these intricate biological concepts. As Mike Gene suggests, we may learn something from such cross polination of ideas.
I am also looking forward to share with you my best understanding of evolution and biology and although, like you, I am not an expert, I feel that we can both contribute to this discussion.
Perhaps you may start to explain your use of the term materialism and "by the sheer unlikelihood of our materialist theory to produce anything". I am sure that my limited knowledge of biology may be helpful to you.
In Christ
F
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 05. March 2003 10:24
The subject matter of this thread is ‘possible’ standards for formal structured scientific analysis. I suggested as a starting point, standards would include criteria for precisely defining and testing variables, causal relationships and structure. Frances has presented us with an excellent example of assertions that might pass as a theory in evolutionary biology, but that satisfy none of the proposed criteria for rigorous structured scientific analysis. Frances provides no precise definitions of the variables involved, no testable predictive models of causal relationships and no precise definition of structure. Frances simply provides ‘descriptive speculation’ and labels it as a scientific theory. Frances sites a number of sources in the literature to support his descriptive speculation, but he does not demonstrate that any of these sources have actually proposed formal scientific theories or models of the type Frances suggests.
Reinterpretations and misinterpretations of the literature raise serious difficulty in trying to evaluate scientific models and theories. To address this issue I suggest that the following behavioral standards be added to the requirements for rigorous structured scientific analysis.
EXPLICIT FORMULATION- The most recent or current form of a scientific model and/or theory with its associated definitions must be explicitly formulated and available for review, comment and analysis. It would be expected that for complex models such as those needed to model complex biological systems, the most recent form of the model would change rapidly as new information is generated. Maintaining a publicly available current version would likely require the use of emerging technologies.
RESOLVABILITY- (A variation on the testability requirement)- The purpose of science is to develop models and theories on which informed observers can reach agreement or consensus. Consistent with this goal, any model or theory proposed must be in a form on which informed observers can reach agreement or a form on which agreement can be reached based on identifiable future information.
It would appear that the existing life sciences would not satisfy these behavioral requirements and the requirements for defining variables, causal relationships, and structure. By the standards proposed here, genetics and evolutionary biology would not qualify as true, rigorous forms of structured scientific analysis. The standards proposed here are satisfied by sciences such as electrical engineering and mechanical engineering.
The standards I am proposing here for ‘true, rigorous’ scientific analysis are essentially engineering standards. I am suggesting that engineering or applied science, not theoretical science, is the basis for ‘real’ science. Theoretical science, from this perspective, simply serves a support role for the ‘real’ or true applied sciences. Although ‘applied science drives science’ perspective is clearly not the conventional view, there are long term historical justifications for the view (science developed or evolved as practical problem solving). This perspective would suggest that a science does not qualify as a real or true science until it has reached the stage where it has a rigorous form of engineering.
It may be useful to reconsider for a moment the impact of the proposed ‘engineering standards’ on the development of scientific model or theory of a ‘universal genetic code’.
A universal genetic code model starts with the observations that 1)all known life forms contain genetic material, 2)horses develop from cells with horse genes, humans from cells with human genes etc. . From these observations one can generate a descriptive theories of the general form:
1. All the phenotypes of an organism are determined by an organisms genetic material, and
2. All evolutionary changes in life forms are determined by changes in genetic material produced by random variation and natural selection operating on genetic material.
If you attempt to translate these ‘speculative descriptive’ explanations into formal predictive theories, you run into all sorts of technical and logical problems most of which have been discussed at great length. The problems include:
1. INCOMPATIBILITY WITH DARWIN- The model ‘assumes’ that ‘evolutionary change processes’ developed after the development of DNA and complex cells.
2. INCOMPATIBILITY WITH KNOWN FACTS- If you substitute human DNA into shark eggs or shark DNA into human eggs, you will not produce humans and sharks as predicted by the theory. Evolution clearly involves more changes in DNA.
3. INABILITY TO DEFINE PHENOTYPE
4. INABILITY TO DEMONSTRATE EXISTENCE OF GENOTYPE TO PHENOTYPE MAP
5. INABILITY TO DEFINE RANDOM VARIATION AND NATURAL SELECTION
6. INABILITY TO FORMULATE PREDICTIVE CHANGE IN GENETIC MATERIAL MODELS
As I have pointed out on a number of occasions, variations on the universal genetic code model or theory are not the only ‘speculative descriptive’ theory which fits the ‘horses from horse genes observation’. The life forms as ‘information generating machines’ model or theory fits the facts as well or better than the universal genetic code model.
If we consider an analogy to music, the universal genetic code model suggests that DNA stores the tunes or music that define life forms. The life forms as information processing model suggests that DNA stores the notes or ‘keys on the piano’ (or actually some of the notes and keys) and the information processing capacity generates the actual songs or music of life.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 05. March 2003 16:01
- INCOMPATIBILITY WITH DARWIN You don't need "DNA and complex cells" in order to get differential success in self-duplication. And if you're talking about UGC, the code didn't have to arise via evolution necessarily; evolution just preserved its universality.
- INCOMPATIBILITY WITH KNOWN FACTS The phenotype of the parent is important--specifically, the parent must be able to load the egg with appropriate transcription factors and such, and must provide appropriate metabolic support. Look up "maternal effect". There is nothing about this that is incompatible with evolution, just with radical jumps where a frog gives birth to, say, a lizard. That would be unlikely to work.
- INABILITY TO DEFINE PHENOTYPE phenotype The form taken by some character (or group of characters) in a specific individual. Source: Griffiths et al., Genetic Analysis, 5th ed..
- INABILITY TO DEMONSTRATE EXISTENCE OF GENOTYPE TO PHENOTYPE MAP See Mendel for a demonstration of heritability. See anything with transgenic Green Fluorescent Protein to note that there is genotype to phenotype mapping (in a broad usage of the term "mapping"; it's not 1-1).
- INABILITY TO DEFINE RANDOM VARIATION AND NATURAL SELECTION I've already demonstrated a detailed mathematical model for an aspect of random variation (matrix of nucleotide transitions). But we can have random variation: change in the genetic material of an organism via a stochastic process, and natural selection: differential numbers of viable offspring depending on phenotype and environment.
- INABILITY TO FORMULATE PREDICTIVE CHANGE IN GENETIC MATERIAL MODELS This is true of any stochastic process, unless you use statistical predictions. For example, if you find oobles and wumpies, which have a recent common ancestor, and zybogs which are much more distantly related, you expect the genes of oobles and wumpies to be similar to each other, even if oobles and zybogs live in the same habitat. (E.g. dolphin vs. bat vs. tuna).
Despite all these problems with your claims, I still agree that the UGC fits many possible hypotheses about the origin and nature of life. However, if you know that there is a genetic code, given that organisms are complicated and the code cannot be vast, evolution predicts that it would be fixed over related organisms. Thus it would predict universality over organisms that we could tell were related (e.g. all vertebrates). Beyond that, it could go either way depending on how many separate abiogenic events there were on the planet, and how many of those survived until now. Given that evolution selects for the fittest organisms, you would expect that evolved organisms would outcompete any abiogenic chemical reaction, and thus there should be very few independent origins. So the UGC is consistent with evolution, but the genetic code could have been such that it was inconsistent with evolution, and evolution wouldn't predict that the code be strictly universal. That is at least better than other models (e.g. spontaneous generation of life forms) that don't predict anything about the universality of code. (Under spontaneous generation ("poof!"), every species could be given a different code. That would be a fantastic way to keep species reproductively isolated.)
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 06. March 2003 11:18
Rex,
The issues here are 1)proposals for sets of standards appropriate for performing and presenting structured scientific analysis and the question and 2)the impact of proposed standards on the development of models and theories. As I pointed out, developing a UGC theory which would satisfy the standards being proposed would require resolving at least six complex technical issues.
It is not clear if you are suggesting that ‘any or all of the six technical issues could be resolved based on the proposed standards’. If this is the claim you are making then it would interesting to see if you can in fact support such a claim for any one of the six issues. If this is your claim, pick one of the issues and see if you can in fact satisfy the standards proposed. The starting point for demonstrating that the proposed standards are being satisfied would be to show that the proposed formal resolution of issue has been explicitly expressed in the literature.
Again, the issue here is conformity to explicitly stated standards. I am proposing for discussion, structured analysis standards which are essentially the engineering standards. Based on the examples considered so far, it does not appear that the models and theories developed and presented in evolutionary biology can meet the explicitly formulated ‘engineering standards’.
The inability of biological models and theories to meet engineering standards is not surprising. The ability to satisfying engineering standards would imply the ability to engineering and/or reverse engineer biological systems. While we clearly have the ability to reverse engineer some of the functionality of biological systems under limited conditions, these reverse engineering capabilities are relatively limited.
IMO, the inability of the life sciences in general and evolutionary biology specifically to satisfy engineering standards is significant to the development of a science of biological design. Engineering standards, it can be argued, define the minimum benchmark for a true, rigorous science. If a new science can be developed which develops models and theories of biological systems which meet engineering standards, then the new science in effect displaces the existing life sciences.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 06. March 2003 19:20
Oh, that's easy then. We don't understand biology well enough to manipulate it the way we manipulate electrical circuits, for instance. Come back in a few hundred years and look again.
We're still learning--haven't you noticed? It takes a lot of work to get to the point you described.
There are cases where it can be done, and many where it can't. Can you meet your own criterion when it comes to analyzing the impact of microfabrication defects in high-temperature superconductivity?
UGC is a hard problem, because we have very little evidence left to go on. But, sure, I can tackle one of the six. I defined phenotype. Maybe this isn't precise enough for you. What do you want to see?
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 07. March 2003 07:51
Rex,
Quote: INABILITY TO DEFINE PHENOTYPE phenotype The form taken by some character (or group of characters) in a specific individual. Source: Griffiths et al., Genetic Analysis, 5th ed..
Quote: . I defined phenotype. Maybe this isn't precise enough for you. What do you want to see?
In a UGC type model or theory, phenotype would refer to "the state of an organism at a point in time on which ‘selection operates’". For the sake of simplicity, we can start by defining phenotype as ‘A state of an organism at a point in time which is essential to survival and reproduction’. This eliminates consideration of ‘states’ that may be somewhat beneficial and somewhat harmful and concentrates of states that are essential to survival.
As has been discussed elsewhere, this set of survival essential states or phenotypes can be defined at the level of the individual cell and in terms of sequences of ‘changes in status’ of regulatory genes. In order to model just ‘survival essential phenotypes’ it would be necessary to either define and model ‘phenotype’ at that level of detail, or to demonstrate that some alternative approach to phenotype was viable.
Phenotype is a set of properties on which selection operates. At a minimum, defining phenotype requires defining the ‘sets of properties at points in time’ which are essential to survival. You have not provided a definition with sufficient detail or precision to make it possible to identify and model point in time survival essential phenotypes.
It does appear possible to define and model the ‘set of phenotypes defining an organism’ in terms of a sequence of cell level changes in status of regulatory genes. However, as also discussed elsewhere, such a definition would not be compatible with a UGC type model or theory since the phenotypes will involve vastly more information than the genetic material.
Quote: Oh, that's easy then. We don't understand biology well enough to manipulate it the way we manipulate electrical circuits, for instance. Come back in a few hundred years and look again.
I have proposed five criteria or minimum standards for evaluating scientific models and theories. Based on the discussion it is reasonable to suggest that:
1. The proposed standards are routinely used and applied in the engineering sciences.
2. No one has been able to demonstrate that any of the models or theories generated in evolutionary biology can satisfy the minimum standards proposed.
3. It is unclear what portion of individuals involved in the biological sciences have the knowledge and experience to apply engineer standards in evaluating models and theories.
At a time in the not too distant past, scientists and philosophers were willing to recognize that fact that ‘science’ had yet to succeed in developing valid, predictive, models and theories of life forms. The life sciences, for reasons no one could fully explain, had failed to develop the types of predictive theories which were the defining characteristic of the physical sciences.
Today, it appears, we 1)question the predictive precision of theories in the physical sciences, and 2) are willing to accept rather dubious claims for scientific theories and models in the life sciences. The engineering type standards discussed here, provide, I suggest, solutions to issues in both the physical sciences and the life sciences. In the physical sciences, engineering standards provide a pragmatic alternative to the unworkable ‘scientific theory as absolute truth’ ideal. In the life sciences engineering standards provide an objective basis for distinguishing ‘descriptive speculation’ from ‘scientific theory and model’.
The engineering standards proposed here provide an objective and verifiable basis for determining what qualifies as a scientific model or theory. The standards, it appears, are likely to disqualify essentially all existing evolutionary models and theories. The standards, it also appears, may bring in to question the qualifications of many individuals to evaluate scientific models and theories. It will be noted that standards which disqualify all existing theories and models are only of academic interest unless and until models and theories can be developed which can pass the proposed standards. I believe that such models and theories are possible, but that is a subject for another thread.
The purpose of this thread has been to point out that objective verifiable minimum standards for scientific theories and models can be developed.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 09. March 2003 02:03
Warren,
Where did you propose these five criteria? I found four, or seven, or three; six complaints, etc.. If you could quote them again, it would be helpful.
Please define "the engineering sciences". This is an odd term since engineering is not really a science; although there are many scientific considerations, there are plenty of non-scientific ones too.
Also note that engineering is, in most cases, extremely simple and incredibly well understood compared to biology. Although it is extremely important to use quantitative methods to advance the understanding of biological processes, it is nonsensical to advocate that advances can be made only approaching it as an engineering problem (in your apparent definition of an engineering problem). For example, you can notice that Prozac makes people feel less depressed without having a precise mathematical definition of depression and a complete architectural model of the brain.
Now, I fully agree that modeling is an excellent way to verify that one's hypotheses capture the essential elements of a process under examination. But this is not necessary to make progress; many wrong ideas can be easily ruled out by qualitative observations. Tailoring the precision of your model to the availability of data is key.
So for instance, when you say,
quote:
It does appear possible to define and model the 'set of phenotypes defining an organism' in terms of a sequence of cell level changes in status of regulatory genes.
I have to ask: is this data available? The answer is: no. We cannot measure the precise levels of thousands of gene products simultaneously yet. We're becoming able to measure the level of thousands of different mRNAs, but that is clearly nowhere near good enough since there are many types of regulation that occur after transcription.
I then have to ask: is it enough to simply make measurements? Well, no. In order to have a model, you have to understand something about the functional interactions between the components of the model. (You have mentioned something similar.) Do we understand the interactions between all the essential proteins in any organism? Absolutely not.
So it seems as though you are asking for a model to be constructed prematurely. This is why I say to wait a few hundred years. Or you could use more abstract or less accurate models for now.
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 09. March 2003 12:25
Rex,
The following are the five standards I previously proposed plus the new ‘define environmental conditions’ standard-
1. DEFINE VARIABLES- A scientific explanation must provide precise definitions that make it practical to identify and quantify cause and effect variables.
2. MODEL AND TEST CAUSAL RELATIONSHIPS- A scientific explanation must provide quantifiable and testable models of cause and effect relationships
3. DEFINE STRUCTURE- A scientific explanation must provide a verifiable and testable structure for the interrelationship between different causal relationships.
4. DEFINE ENVIRONMENTAL CONDITIONS
5. EXPLICIT FORMULATION- The most recent or current form of a scientific model and/or theory with its associated definitions must be explicitly formulated and available for review, comment and analysis. It would be expected that for complex models such as those needed to model complex biological systems, the most recent form of the model would change rapidly as new information is generated. Maintaining a publicly available current version would likely require the use of emerging technologies.
6. RESOLVABILITY- (A variation on the testability requirement)- The purpose of science is to develop models and theories on which informed observers can reach agreement or consensus. Consistent with this goal, any model or theory proposed must be in a form on which informed observers can reach agreement or a form on which agreement can be reached based on identifiable future information.
Quote: Please define "the engineering sciences". This is an odd term since engineering is not really a science; although there are many scientific considerations, there are plenty of non-scientific ones too.
The conventional academic view is that abstract or theoretical science is ‘real science’ and ‘applied science’ is an almost incidental side benefit of real science. This traditional perspective has its origins in the nobility-commoner class structure[ This class structure is readily apparent in the British education system where, I have been told, for a long time ‘engineering’ wasn’t even taught in the top universities.] There is an alternative, if less popular, viewpoint that suggests that engineering, applied science, or practical/beneficial application is the driving force behind ‘real science’ and abstract science serves only a secondary support function. The subject here is not, however, the history of science or the right way to view science. The issue here is objective verifiable standards for performing rigorous formal scientific analysis.
I have proposed what I label engineering standards. Biological systems are complex systems. Engineering provides highly effective techniques for modeling and analyzing complex systems. There is, therefore good reason to consider engineering standards as appropriate for the analysis of complex biological systems. It should also be pointed out that no one has yet to identified an alternative set of objective verifiable standards which can be applied to areas such as evolutionary biology.
Quote: it is nonsensical to advocate that advances can be made only approaching it as an engineering problem
Really? I recognize that there has been a ‘pre-scientific’ period of a few hundred years when we have accumulated informal descriptive information. To move science beyond informal observations and speculation, you need the ability to formulate rigorous verifiable models and theories. Engineering would appear to offer one possible approach to moving toward formal structured analysis. No one is offering any alternative approach to formal, objective, structured analysis.
Quote: For example, you can notice that Prozac makes people feel less depressed without having a precise mathematical definition of depression and a complete architectural model of the brain.
There is a very fundamental difference between the ‘trial and error- see if it kills them or cures them’ analysis used in medicine and the ‘pencil and paper engineering analysis’ used to evaluate the designs of bridges.
Quote: I have to ask: is this data available? The answer is: no. We cannot measure the precise levels of thousands of gene products simultaneously yet. We're becoming able to measure the level of thousands of different mRNAs, but that is clearly nowhere near good enough since there are many types of regulation that occur after transcription.
The issue here is not ‘perfect knowledge of phenotypes’, but models and theories logically consistent with what is currently known. Phenotype is ‘a set of traits impacting survival and the likelihood of survival.’ Based on currently knowledge, we know that complex organisms will not survive if cells (with a limited margin for error) do not have the appropriate set of traits at points in time. We also know that transitions from one ‘survival essential form or another’ involve changes in regulatory genes. A definition of phenotype is inadequate or misleading if it does not properly reflect current knowledge.
Quote: I then have to ask: is it enough to simply make measurements? Well, no. In order to have a model, you have to understand something about the functional interactions between the components of the model. (You have mentioned something similar.) Do we understand the interactions between all the essential proteins in any organism? Absolutely not. So it seems as though you are asking for a model to be constructed prematurely. This is why I say to wait a few hundred years. Or you could use more abstract or less accurate models for now.
Two points. First, I am proposing for discussion ‘standards’ for differentiating ‘valid rigorous, models and theories’ from ‘invalid, non-rigorous models and theories’. As I have stated, I believe that current models and theories would not meet the proposed minimum standards. While I believe I can introduce models which meet these standards I have not yet done so.
Second, satisfying engineering standards does not require complete knowledge. Engineer standard models and theories are progressive. Models satisfying engineering standards can be developed based on limited ‘current level of knowledge’. The models and associated theories are modified and refined as new knowledge is acquired and the starting point models are tested and validated.
IP: Logged
|
|
|
Topic Closed
Printer-friendly view of this topic
