|
Author
|
Topic: Mechanically Specific Relativity
|
KBC1963
Member
Member # 1868
|
posted 22. January 2006 08:25
I would like to share with everyone here a paper that I wrote that I believe gives us another arguement to defeat Darwins concept of slight changes over time. This paper was posted in a forum that was mostly comprised of evolutionist and I am happy to say that they were unable to counter my arguement with any intelligent response. They were however very forthcoming with personal attacks on my lack of actual scientific background. If by chance anyone see's an error in my post please enlighten me as to the error so I can improve my arguement.
I have been designing mechanical systems for many years now and one of the most understood concepts of people in my profession is what I call "MECHANICALLY SPECIFIC RELATIVITY". All complex mechanical systems share this rule or law of form whether they are living or not. Let me show you what this concept is by considering what would be required to change a mouse circulatory system into something larger since it is assumed according to evolutionary theory that man evolved from something the size of a mouse by slight modification.
The circulatory system consists mainly of these mechanical components:
1) heart
2) ateries
3) veins
4) bones
5) brain
Each of these mechanical components is a sub system of an overall complex system and each of them is highly specific relative to every other component in the system. To better clarify this consider that the heart is fairly specific in its size compared to the size of the organism, it must pump a fairly specific amount of blood in order for the organism to function correctly and its size is also relative to the size of the arteries that convey that volume of blood to the veins and the veins are relative in size and quantity to the arteries in order to convey the volume of blood originally pumped by the heart. The bones in turn must regulate a specific amount of blood cell production in order to keep the total volume just right for the system volume, and lastly but most importantly the brain which regulates each of these sub systems must keep them all operating in a highly harmonious rythem, It must also be considered that each of these components must also grow from infancy to adulthood at a highly specific rate which is ultimately a function of the DNA's preprogrammed control.
Now what would it take to increase the size of the organism?
How much of the genome controls all the aspects of this system?
Could a slight change to any of the sub systems occur without a relative change to each of the other components?
It should be readily apparent that any slight change in any sub system would unbalance the relativity of this system and the only logical conclusion would be that the entire system would have to change in a highly specific amount to accomodate a change in organism size. Each of the components has specific length, height, & width all related to one another and specifically defined by the DNA and any change in the system must specifically account for each of these features in a highly specific manner and to assume that random minor changes would account for an organisms size change from mouse to man without upsetting its balance along the way is ludicris in the extreme.
Darwins theory violates the law of mechanical relativity and should be easily seen for the flawed concept that it is and I invite anyone to give me a reasonable explanation of how evolution can overcome the mechanically specific relativity that every intelligent mechanical designer considers within every complex system.
IP: Logged
|
|
Atom
Member
Member # 1840
|
posted 22. January 2006 17:11
Hey,
I find your idea very intuitive and interesting. I'm formulating a response. It may take a minute, as I'm attempting to be thorough. But I agree with your premise. As a software systems engineer, I have dealt extensively with interdependent component systems.
Atom
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 22. January 2006 19:23
This seems related to a discussion we had here about three years ago. LINK
IP: Logged
|
|
KBC1963
Member
Member # 1868
|
posted 23. January 2006 06:31
Yes Irving that link does indeed have much of the same points as my post and is even more involved.
I am hoping that just as irreducible complexity is a main tenet of intelligent design we could also add mechanically specific relativity as a tenet based on our knowlege as intelligent designers of what is required to make changes to complex systems. I say that complex mechanical systems are impossible to change one component at a time without changing everything involved so maybe this is a form of irreducible complexity but in any event I know this law of form exists and if those people who have the ability to publish and are respected were to add this concept to their existing arguement then I think it would be even more convincing overall.
IP: Logged
|
|
David L. Hagen
Member
Member # 323
|
posted 23. January 2006 22:16
KBC Excellent perspective.
Encourage you to study
Adrian Bejan's
Constructal Theory
Especially:
Shape and Structure, From Engineering to Nature
Bejan demonstrates how the structure of lungs, arteries and veins etc. are configured to minimize pumping power. He also shows how heart rate etc are correlated.
Bejan perfunctorily attributes this to evolution without addressing any of the foundational mechanisms. However, energy minimization is an obvious design argument. I am exploring how these configurations are achieved regulated systems, and the corresponding design parameters involved. See my response to your other post. A Request for More Complex Systems Analysis
See also Dr. G's discussion of how critically important control of oxygen, thirst and hunger are:
Glicksman: Wouldn't it be wonderful if we never had to experience hunger or thirst
The brain controls the heart pumping and breathing, and without oxygen, brain cells die within a few minutes.
(PS Is your paper listed in full above or do you have a link to it elsewhere?
[ 23. January 2006, 22:30: Message edited by: David L. Hagen ]
IP: Logged
|
|
KBC1963
Member
Member # 1868
|
posted 23. January 2006 23:35
Thankyou David for your response.
Wow, Adrian appears to be a virtual god in the mechanical world, from degree's to books he seems to have done it all. I read some of the link you gave me and as of this moment I am incapable of responding to his arguement but I will read it completely and then respond.
You say that he attributes everything to evolution and I can only say that this is depressing because those with the best understanding should be the most able to debunk error and inaccuracy and with his obviously broad background it should be a snap to layout foundation to overview, In my view an explanation that avoids foundational explanation and logic allows error to continue. A good designer details his house from the ground up leaving no single part questionable and with this in mind I seek to fill in my foundations based on physics and laws of form that are undeniably found within any functioning design.
You ask "PS Is your paper listed in full above or do you have a link to it elsewhere?" and yes that is my horribly short and terribly written paper in its entirety that was composed in one sitting to respond to a thread in another forum.
I must apologize for my shortcoming as a writer but my strong suit is engineering and my presence here is not to become a published intellectual but maybe in some small way I may give those that do have the genetic gift of writ another item to add to the truth that they are trying to solidify into ID.
IP: Logged
|
|
Atom
Member
Member # 1840
|
posted 26. January 2006 11:18
Hey KBC and others,
I really like your idea about MSR. I read through the link provided, Irving, and I see you've brought up scalability in discussion before, so let's see if we can move the discussion forward a little.
After reading KBC's original post, here's what I took as the main lines of his argument:
You have a system where...
1) It consists of parts
2) Each part is a subsystem of a larger system.
3) Each part is highly specific relative to *every* other part of the system. (This constraint may be relaxed, and become more generalized, I think)
4) The entire system may also be a highly specified subsystem relative to another larger system.
With these ideas in hand, I propose a first draft formalization of the idea. This rough approximation is not meant to be a 100% accurate model which captures everything about MSR; it is only here as a starting point for building a formal definition of it. Call it a 50% model. Keep what works, discard and supplant the rest.
Formalization:
Begin with a multi-part system that is subject to one or more external constraints, which we may define as the total system function constraint. The system consists of N parts. Among these N parts there can exist intrasystem constraints. Let m denote the total number of distinct intrapart constraints.
Each constraint can be defined as as a funtional relationship between parts, defined by a series of functions: f1 - fm. These can be simple ratios or more complex relationships.
Define ei as the perturbation space around each function fi, where i is between 0 and m. This is the range of percent the input values of fi can vary, while keeping the Total System Function constaints satisfied. (This takes into account both neutral changes in fi and changes that increase optimization). For example, if a given subsystem can be 0.05 percent smaller or 0.10 percent larger while not affecting the external system contraint(s), it is said to have an ei of 0.15.
Now MSR may be defined as the ratio:
m/(e0 * e1 * ... * em)
for all i between 0 and m.
Now, the above is a first approximation. In formulating the above ratio, I tried to have the MSR function maintain the following properties:
1) The larger m, the larger the positive measure of MSR.
2) The smaller each ei, the larger the MSR measure. This is our intuition: the smaller the region each internal constraint can vary by, the higher the degree of specific relativity.
Now, there may be a better formula to capture the relative properties, so I invite you to refine my formula if need be.
The important aspects of my formalization are that the entire system is subject to an external constraint, and that each subsystem ni can affect this external constraint. In evolutionary terms, each subsystem can maintain or increase fitness of the total system. But since fitness is a notoriously vague word, I will use different terminology, which is hopefully more specific.
Also, each subsystem is subject to a number of intrasystem constraints, which they have to satisfy for the total system function constraint to be satisfied. Satisfaction in my terms refers to maintaining current output, or optimizing the output to a higher degree.
Any reduction in optimization or failure to satisfy the constraint means that Darwinian mechanisms will not be able to select for the proposed change, and it will be subject to either being weeded out by more optimized competitors or subject to genetic drift.
I think this captures what KBC was attempting to show in why high MSR systems are a challenge to Darwinian mechanisms.
One last point, a possible counterargument against MSR being a stumbling block for RM+NS is that perhaps a system exists where one fi has a large ei region, and changing this value causes the e regions of the other parts to change as well. Then those parts change, which causes further changes in e regions, thus allowing for net change over time. To such a suggestion, I would say such a system may be logically possible but would need to be demonstrated in detail before such an objection can take force. Simply saying "there may be a system" is vague and is similar to the objections held against IC: too vague to refute, but also too vague to hold force.
Atom
[ 26. January 2006, 12:51: Message edited by: Atom ]
IP: Logged
|
|
Atom
Member
Member # 1840
|
posted 26. January 2006 12:56
As a post script thought,
If MSR raises a large barrier for RM+NS change, this may be another mechanism which helps to explain stasis of organisms over time. The MSR would act as a conservative constraint, much as IC has been proposed for helping to maintain stasis. If FDOCC's idea about there being a conservative force among basic types of organisms, where they can alter only in variation not speciation, MSR may be a candidate for such a conservative force.
Just a thought.
Atom
IP: Logged
|
|
David L. Hagen
Member
Member # 323
|
posted 26. January 2006 14:39
Good scoping start Atom and KBC.
This MSR begins to scope out the possible Design Function space.
I am not sure what m gives that is not already in ei. What about just the product of all m of 1/ei ?
Bits MSR Information
As the next step suggest converting this to bits of MSR information by taking the negative log base 2. i.e.,
MSR = the sum over m functions fi of
the negative of log base 2 of ei
where ei = variation in input parameters to fi.
Is m required to normalize this? Maybe some of the math wizzes could help out here.
This MSR would then be a quantified subset of the larger Design Information of the system.
quote:
ei - is the range of percent the input values of fi can vary, while keeping the Total System Function constaints satisfied.
Rather than "percent", I recommend defining the perturbation as the fraction of the change in the input parameters of function fi over the maximum possible change of that parameter within the design space which you specified as that required to maintain the external constraints.
Just to make sure I understand you. Where we have only one perturbed parameter ei per function fi -
Then, on declining perturbations ei is the change code:
(Xmean-X)/(Xmean-0)
, i.e. where del X < Xmean.
For increasing perturbations, we define ei as code:
(X-Xmean) /(Xmax-Xmean)
This generalizes to
ei = Del X / Max Change
code:
= (X-Xmean)/(Xxtrm-Xmean)
Where the extreme parameter
Xxtrm = Xmin if X< Xmean, and
Xxtrm = Xmax if X>Xmean.
Coming up with Xmin and Xmax are at least particular to the application.
To generalize we could possibly take some maximum function value as Xmax, and a minimum function value as Xmin=0.
With dimensions, these functions could take on the values of the maximum dimension and 0. However, these may pose the difficulty of possibly making the relative perturbation appear smaller and thus have more specified information than possible actual design contstraints.
1. Varying extremes
The larger challenge is that they may be a function of all the other ei if we require that the overall external constraints be maintained? Because then code:
Xxtrm(i)
are functions of the other Xxtrmj.
2. Multiple parameters per function.
What about function with multiple input parameters? Do we need to take the next step of having a double sum over nj parameters ei of m functions fj? Can these be simplified to one input parameter per function? Or do we need to assume that as an approximation to start with?
3. Discrete perturbation space.
The next challege is to look at how to transform this to functions of parameters that have only discrete values. e.g. atoms with perscribed interatomic dimensions and angles.
Look forward to your thoughts Atom on how to develop these issues, expecially where they address atoms! ![[Smile]](smile.gif)
IP: Logged
|
|
KBC1963
Member
Member # 1868
|
posted 26. January 2006 22:57
Hey Atom,
I notice your comment about "I see you've brought up scalability in discussion before, so let's see if we can move the discussion forward a little"
This my kind sir directly up my alley, as you may have noticed in that thread there came up a question of an acurate meaning for the term scalability. This I can define for everyone here and leave no mystery as to its usage.
I use scaling as a part of 3D drafting using Autocad 2005 in the course of design.
The word scale does have a number of uses in different contexts but I will focus only on its usage as applied to mechanical systems either living or not. When drawing geometric shapes and models I can design a bolt the size of a needle and if lets say I find that I need a similar bolt but of a size thats much greater or smaller I can hit a command to scale this object and I can change the size of my bolt relatively up or down by percentage numbers such as 2x {= twice as big on all three axis x,y,z Length, width, height} and when it performs this function it changes the size of the object relatively on each axis, so a rectangular box having a shape that is 5" long by 2" wide by 1" deep when I scale it up 2x it becomes 10" long by 4" wide by 2" deep.
Sooo... To scale something either up or down is to proportionately change its size in all three planes at once.
This is the engineering meaning of the word. now its use in virtual thought and mechanical applications is pretty straightforward and fairly easy to envision, however when attempting to apply it to living systems it cannot possibly apply as no living system can increase all its component sizes all at once as this would require an entire rewrite of the genome in order for this to happen. A better word in this case would be "miracle" because the necessary changes would be so mathematically stringent as to rival infinity for chance of random possibilty.
So when dealing with living systems we should never attempt to appy scalable as this implies a global organism change of size relative to every component within it. When dealing with evolution it can at best make minor incremental changes so the more operative word would be expandability because truley any size change would involve adding mass to individual components one at a time at best and that is a stretch as well because every bone must have millions of bits of information lockedup within the genome that allows it to be built accurately in length, width and height, lets call this its 3 dimensional matrix which is stored in a compressed or cryptic way within the DNA. this is part of what I am talking about when I say something is highly complex. try to imagine for a moment how many lines of code it would require to define the positions for all the cells that make up just a fingernail with millions of cellular structures.
In autocad it requires 3 numbers to define a position for every individual component in an object so if an individual cells position needs 3 numbers to define it location in model space we may now understand what junk DNA is all about. Well whether its numbers or some other convention beyond my tiny mind the empirical facts remain that each and every component in a living system defines specific shapes with billions of cells and they hold that shape fairly accurately.
[ 26. January 2006, 23:05: Message edited by: KBC1963 ]
IP: Logged
|
|
Atom
Member
Member # 1840
|
posted 27. January 2006 00:39
Hey David and KBC,
quote:
I am not sure what m gives that is not already in ei. What about just the product of all m of 1/ei ?
Yes, as I experimented with different formulas I settled on m/ei, but didn't realize that m would be inherent in 1/ei as well. Oversight on my part. I think 1/ei would be better as well, since a component which could vary 100% would have an ei of 1.00 (per my original definition), thus including it in 1/ei would not affect our measure of MSR, which coincides with our intuition. (A component that is not specified at all, and can vary fully, should have no MSR, and should not add to our measure of total MSR.)
quote:
Rather than "percent", I recommend defining the perturbation as the fraction of the change in the input parameters of function fi over the maximum possible change of that parameter within the design space which you specified as that required to maintain the external constraints.
I think I see where you are going with that. My initial motivation to use percentages was based on two considerations:
1) We are assuming a stable system, which currently satisfies the external constraint(s). Thus, we will have different values xi for each function fi, and these xi will be real numbers. For example, the total volume of blood produced by the skeletal system could be the number of liters. We can empircally calculate these intial xi values, and so they will be known.
2) Given that we know current values of xi, we can then either theoretically or empircally adjust the value of a given xi, both positively and negatively, and mark the resulting improvement or degradation of the total system. (See if it optimizes or at least maintains current levels of total function). At some point, in a tightly integrated system, the change in xi will result in degradation of function. The smallest value xi can take on before degradation, and the largest value, will mark the endpoints of our range. This range for each fi (centered around xi) should also be able to be determined.
Given these considerations, I thought a percentage would be simple to derive from the known values of xi (initial value), xmin (lower bound where performance begins to degrade), and xmax (also where performance degrades). Percentages are also unit independent due to their existence as ratios, another consideration, since choice of units should not affect our measure of MSR.
I see that you want to take this fraction of change and divide it by total possible range. I think I didn't make clear what I was saying by perturbation range (ei). This range ei *is* the total range, with xmax and xmin denoting the points where the external constraint is no longer satisfied. From what you wrote, I'm assuming you thought I was speaking of the individual changes we make to xi in determining ei. Let me know if I'm missing something in your line of thought.
As for multiple parameter inputs for each fi (x0-xk), I considered that, but thought it would distract from the first approximation and add too much detail too soon. Eventually, we could generalize the fi to be as high a degree function as we like, but sticking with simplicity as our guide, we could consider only:
A) What the subsystem produces (i.e., its output), which will be an input to the larger total system, and in our terminology is xi. For example, the total volume of blood produced by the skeletal system would be x0, and would equal some number of liters. This would be one input (xi) to the total system (the circulatory system). By varying the volume of blood produced, we vary xi, and at some point either too little or too much blood produced will cause the total system function to degrade. The percent change we can decrease blood volume added to the percent change we can increase it by will mark our ei for that component.
B) The function fi, which in this example case would be the ratio of blood produced to the internal volume capacity of the arteries and veins. I'm not a physiologist, but I will assume our fi in this case would be a simple 1 to 1 ratio: f(x) = x.
With this, I think we can get a pretty good idea of what MSR means in a formal sense. Adding detail to it is desired, but we should always keep our intuition intact as to how these numbers should behave, what properties they must maintain, and try to arrive at the simplest formulation that captures all the relevant properties of our desired metric.
As for discrete values, my initial thought is that we can enumerate the total number of possible configurations, and then test to see how many configurations maintain or increase total function. The ratio of functional configurations to total configurations should provide a meaningful number, though adding it to our above measure seems to me a little like apples and oranges. I will leave it to the more mathematically inclined among us to see if the intuitive desired properties would still remain intact if such values were allowed within our MSR equation.
KBC I agree that scalability can refer to volume increase in size, but with your more generalized approach we may be taking into consideration constraints which are not just limited to physical size, such as pressure, electrical charge, etc. I think the MSR principle, if formalized genrally enough, may become applicable to all kinds of constrained optimization problems.
Sorry if this was long. David, I look forward to your clarification and refinement of my ideas. I feel that one of us may have missed something in the other's formulation, so if possible, I'd like a clarification of what you intended with your change of ei. I don't think I fully grasp where you are going with your formulation, and I'd like to.
Atom
[ 28. January 2006, 11:06: Message edited by: Atom ]
IP: Logged
|
|
Atom
Member
Member # 1840
|
posted 27. January 2006 10:56
Another postscript set of observations...
When comparing IC and MSR, I think MSR may have some benefits over IC. Firstly, MSR only needs a degradation of system function, not a total function loss. So MSR needs to show reduced efficiency, where as IC needs to demonstrate total loss of function.
Secondly, MSR is inherently recursive. Each "part" of an MSR system can by definition also be an MSR system, where the External Constraint (Total System Function) of the subsystem will equal its output into the larger system (xi). Therefore, finding a subsystem that is also MSR would not force us to redefine our core or even blush; it would actually add another layer of complexity that would then need to be explained.
Some differences, which may be drawbacks for MSR, are that MSR needs at least two systems to pose an *evolutionary* challenge. It is a valuable metric on its own, helping to define natural limits of gradual change, but to say that it challenges RM+NS we would need to have at least two systems: one starting MSR system, and another with different internal and external constraints which it supposedly evolved into. KBC's example of a small mouse size animal gradually transforming into a large human sized one via RM+NS is a good example, when dealing with their respective circulatory systems.
Atom
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 27. January 2006 22:07
quote:
This I can define for everyone here and leave no mystery as to its usage.
Perfectly fine, but you'll also noticed I attempted to address scalability from an architectural standpoint. I proposed the concept that the interface of a repeatable, modular component is an example of specified complexity.
[ 27. January 2006, 22:08: Message edited by: Irving ]
IP: Logged
|
|
David L. Hagen
Member
Member # 323
|
posted 27. January 2006 22:56
KBC
quote:
when attempting to apply it to living systems it cannot possibly apply as no living system can increase all its component sizes all at once as this would require an entire rewrite of the genome in order for this to happen.
Appreciate you raising this very important issue of relative scaling. I think it has major potential. Just a caution on different types of scaling vs genome information. Cf
1 Mammalian scaling: Variation from shrew to great blue whale.
2 Growth scaling: Variation from conception to maturity.
3 Racial scaling: Variations due to migration / genetic segregation.
e.g. 150 cm pygmies to 200 cm khambas.
4 Gender scaling: Variations due to chromosome/hormonal expression.
Categories 2, 3 and 4 suggest that scaling does not necessarily require a massive change “throughout the genome.” Growth suggests scaling due to variations in hormonal expression.
Category 1 shows allometric not isometric scaling. See:
White & Seymour Allometric scaling of mammallian metabolism. J Expt. Biology 208, 1611-1619.
Bejan shows much scaling fits minimizing energy use. Bejan & Marden Unifying constructal theory for scale effects in running, swimming and flying J. Expt. Biology 209, 238-248.
This principle naturally fits a design argument. (Bejan gives no mechanism to explain how it is addressed by evolution.)
A key design objective is to minimize the design information required. From this, we obtain design information embedded into growth regulation parameters, not massive genomic change.
From this one or a few parameters such as hormone expression may be used to control major variations in growth.
Water Control: Body water content is one parameter that could be used to scale blood volume in the body. Dr. G addresses how water is controlled using a sensor in the hypothalamic cell together with a regulatory network. See:
Dr Howard Glicksman September 17, 2003 He Who Cannot Control His Water Will Not Survive
Blood Regulation: Dr. G also discusses blood regulation. See
Dr Howard Glicksman March 15, 2004 Why Blood is Red and other Bedtime Stories
Kidney cells that secrete erythropoietin can detect oxygen concentration in the blood. The erythropropoietin triggers production of red blood cells with hemoglobin in the bone marrow.
(Atom - I'll get back on the math later.)
[ 27. January 2006, 23:21: Message edited by: David L. Hagen ]
IP: Logged
|
|
KBC1963
Member
Member # 1868
|
posted 27. January 2006 23:10
Irving,
Yes I did see your attempt:
"Scalability of Design is very sophisticated and is not something I suspect would result from random mutation and selection. The thought being that random mutation wouldn't continually develop the same structure over and over--coupled with the concept that without pre-design, replicating components detract from function rather than enhance it."
The only thing missing was the example to show why empirically it wasn't going to work.
Let me now supply the evidence. When scientists were mutating fruit fly's they had all sorts of mutations occur such as the Hox gene mutations that literally copied entire structures such as legs and applied them in different locations, do you remember these? well no matter how well they were replicated they lost their function which is primarily because when a structure is copied the mutation would also have to copy the control system that operates it in order to make it function and the obvious results they got showed that the mutation was incapable of accurately copying the other Mechanically Specific Relative part of that system. We have learned that these systems have more than just a physical shape, they also have a controller located outside the Hox gene.
You also said:
"As a "brainstorm" I'd like to toss these ideas around as engineering predictions of design. Nuture the idea and see how it develops.
I personally don't know how either of these concepts may apply to living organisms...though I suspect they might. I'm resisting the attempt to identify real-life examples right now as I'd like to interest people into actually brainstorming something first (then argue if life proves it out later)."
And so we shall continue.
Atom,
I am highly impressed with your thoughts on the MSR idea and at this point you are pulling math that is hard for me to follow but I was hoping that the germ of the idea could be applied in some way and darned if you didn't hit the ground running with it. We are kinda coopting right? seems kinda evolutionary don't you think? hehehe
IP: Logged
|
|
|
Printer-friendly view of this topic