Home » Biology, Darwinism, Informatics, Intelligent Design » Darwinism from an informatics point of view

Darwinism from an informatics point of view

progflowAs everyone knows, life in all its countless instances (organisms) involves internal instructions, as well as processors that run them. Without these instructions, no organism would be able to originate in the first place, let alone develop or survive. The discovery of these instructions – contained in DNA/RNA macromolecules and the molecular machinery that reads and writes them in biological cells – has been hailed as one of the greatest theoretical and experimental breakthroughs of the 20th century. The ID movement claims that these scientific findings have only served to highlight the weaknesses and inconsistencies of the neo-Darwinian theory of macro-evolution, according to which all species have evolved from a common ancestor, as a result of random mutation and natural selection.

The discovery of complex information processing in biology invites the question of whether there are any significant similarities between bio-informatics and the artificial informatics of computers, i.e. so-called computer science. Given that in both fields information has to be managed and processed, some similarities must of course exist. In this post, I will attempt to outline some conclusions on this topic, which lead us inexorably to the conclusion that Darwinian theory is incapable in principle of explaining the mystery of the origin of life and of species, as it claims to do.

When we consider the development of organisms and their complex internal organs and biological systems, we can easily see that these developmental sequences – and here I am talking about both ontogenetic and phylogenetic sequences – must involve complex programs, which embody decision logic about what has to be assembled, and also when and where it should be assembled. In other words, the right things need to be put in the right place at the right time, according to a precise schedule which is in some respects even more rigorous than schedules used in human engineering. For example, the development of an embryo is a process whose countless steps need to be choreographed in their most minute details by a program that is oriented towards the final result. Any error in the execution of this program may have severely deleterious consequences. The same thing can be said regarding the alleged macroevolution of new kinds of organs or even new body plans.

Given that biology and informatics both make use of programs, it will be necessary for me to say a few things about computer programming, in order to explain as clearly as possible exactly what a program is. I know that a lot of UD readers are software developers, so the points I will be making below will be very obvious to them. However, I’ll have to ask them to bear with me, as some of our readers are laypeople in these fields.

In order to process information – i.e. create software – it is necessary to create data and programs. Data is passive information: it cannot change or decide anything by itself. For example, let’s say I have a string variable (called $a) and I set it to contain the value “something” – or maybe I have a numeric variable $b which I set to contain the value 3.14. In these cases, I am neither specifying what should be done with the set values, nor when it should be done. Hence if I were to confine my work as a programmer to simply declaring the values of passive data, I would never be able to actively run a program or control any of its processes. Putting it another way: a program, in its simplest concept, is a blueprint specifying the reiteration of basic decision structures, about what to do and when to do it. A program must specify conditions and actions forming a control structure:

conditions (when to do it)
{

actions (what to do)

}

In other words, a program is active information. Since it determines conditions and actions, it has to be able to decide and organize things, and it also has to be able to create and change data. A program implies a decision hierarchy – in a word, a “logic”. It states what to do, when certain particular conditions arise. Once a program is designed, its execution by a processor can be used to control data and processes of any kind.

The simple structure described above can be repeated many times and can also be nested to create very complex structures with multiple nesting layers, such as the following example, with three nesting levels (the indentations and carriage returns have been inserted to help the reader understand the program flow, but are irrelevant per se at the level of machine code):

conditions
{

actions
conditions;
{

actions;
conditions;
{

actions;
conditions;
{

actions;

}

}

}

}

Another important concept of programming is that of the sub-function or sub-routine:

function
{

}

The main program can reference and run a sub-function as follows:

conditions
{

actions
&function

}

where “&” is the symbol for referencing.

A sub-routine is a sub-program (or “child” program) of the parent program (usually called “main”) that invokes it, which can be referenced (i.e. used indirectly, thanks to a pointer that points to it). Two important things to note about sub-functions are that they work only if they exist somewhere within the software (a very obvious point) and that they are “called” by the main program. In other words, even if we have entire libraries of sub-functions, they will be useless if they are never called: they will be “dormant software”. Thus in a sense, dormant sub-functions constitute passive information. They are passive because they still require a caller that can run them. A sub-function which is never called does absolutely nothing.

From another point of view, programming can be defined as whatever implements control of a process. Since – as Michel Behe says – the fundamental problem of biochemistry and molecular biology (and, in the final analysis, of systems biology) is the problem of control, it follows that programming is indispensable in biology, where countless complex and concurrent processes are involved. Because multiple processes are running at the same time in biological systems – a property that scientists refer to as concurrency – there must be some higher level of direction that governs them all.

It should be noted that the conclusions obtained above hold quite independently of whether an organism’s biological instructions are completely contained within its genome, or only partially. There are many (and I would count myself among them) who suspect that the genome, by itself, does not contain enough information to account for the overall biological complexity of an organism. However one thing is certain: the assembly instructions of living beings must exist somewhere, and the science of generating instructions (computer science) can help us understand their organization and fabrication.

Modern evolutionary theory proposes several unguided mechanisms in order to explain the alleged global macroevolution of species from a single common ancestor: random genetic mutations, sexual genetic recombination, horizontal gene transfer, gene duplication, genetic drift, and so on. According to evolutionary theory, the output of all of these blind processes is subsequently processed (or filtered) by natural selection, which allows only the fittest to survive and reproduce. However, as we will see below, not one of these processes is capable of generating programs. Hence they are also incapable of creating new organs, new body plans, or even new species.

The concept of the gene is fundamental to evolutionary theory in particular, and to genetics and biology in general. Despite its importance, we are still a long way from a clear definition of what a gene is. From the old definition of “recipe for a protein” to the new definition of “functional unit of the genome,” the concept of gene has evolved to the point where some researchers now openly declare that “a gene is a unit of both structure and function, whose exact meaning and boundaries are defined by the scientist in relation to the experiment he or she is doing.” In practice, this means that a gene is whatever a particular scientist has in mind when he/she is doing a particular experiment.

The argument which I am putting forward here cuts through these definitional controversies, because from my informatics-based perspective there are really only two possibilities, which can be summarized as follows: either (a) genes are data (which corresponds to the above old definition of a gene); or (b) genes are functions (which corresponds to the new definition). The key point to understand here is that the development of new organs or body-plans (macroevolution) necessarily involves new decision logic, i.e. new hierarchies of nested control structures. Specifically, the architectural complexity (at the system level) of new organs or body-plans and their embryogenesis involves assembly instructions which require advanced-level control, and hence advanced programming.

Let’s suppose that the first option is correct, and that genes are data. In this case, it can easily be demonstrated that point random mutations, sexual recombination, horizontal gene transfer and data duplication are all incapable of creating the hierarchical decision logic of the main program. In fact, data is what the main program elaborates. Data is passive, while the program is active. What is passive cannot create what is active. This is just as true for intelligently designed data as it is for the data upon which the random operations of Darwinian evolution are applied.

We can illustrate this point from another perspective, by using the analogy of the bricks in a building. If genes are data containing only “recipes for proteins,” and proteins are the “bricks” of the organism “building,” then it is obvious that genes/bricks (and the random Darwinian operations performed upon them) cannot account for the construction and assembly of the organism/building – that is, the set of rules and instructions specifying the way in which the various bricks have to assemble together, in order to yield the unity of a complete system. The building construction metaphor also helps us understand why different organisms can have almost the same genetic patrimony. Just as the same bricks can be used to construct entirely different buildings, the same genes can be used to develop entirely different organisms. In other words, in both biology and architecture, what matters are not the basic building blocks, but rather the higher-level instructions which operate upon them.

Now let’s consider the second alternative, which is that genes are equivalent to software sub-functions. This is quite a generous assumption for evolutionists to make, because it implies that genes possess their own internal decision logic, without explaining how they acquired it. In reality, the so-called “regulatory regions” of genes probably don’t warrant being described as true algorithms. But even if genes were the equivalent of software functions, then once again, random mutations, sexual recombination, horizontal gene transfer and duplication of functions would still be incapable of creating hierarchical decision logic. Why not? Because the decision logic contained in the main program is what invokes the functions (by referencing them). Just as a hammer or a drill cannot create a carpenter, the above operations on functions are incapable of creating their user.

Let us note in passing that the classic evolutionist objection that a mutation involving only a few bits (or even a single bit) is capable of triggering major changes (evolutionists typically cite homeobox genes that control some configurations of the body plan, etc.) contains another misunderstanding. For the active information for these changes still has to exist somewhere, and it must be as large as the changes require it to be. It is true that a programmer can write a very short “wrapper program” to trigger large changes, but that doesn’t mean that the changes themselves require only a little information to specify. For example, I can write a short piece of code which I choose to run on my computer – say, a word processor or a chess program. This code is a few bits long, but the word processor and the chess program are really large programs. All the function does is point to or reference them. However, the function doesn’t create the active information contained in the word processor or chess program software; rather, it simply switches control between the two. Hence there is no free-lunch creation of information whatsoever here.

Leaving aside the problems associated with defining what a gene is, it can still be shown that the random processes which evolutionary theory claims are capable of generating biological complexity, simply don’t work. They don’t work because they are, by their very nature, incapable of generating the top-down functional hierarchy of nested decision structures that is responsible for making the whole system. Since this objection to the adequacy of random processes is an in-principle objection, it is useless for evolutionists to attempt to counter it by resorting to vast amounts of time or huge probabilistic resources. The fundamental problem of Darwinism is that the greater cannot come from the less.

To sum up: Darwinism, from an informatics point of view, has absolutely zero credibility. This explains, among other things, why so many computer programmers who are interested in the ID/evolution debate are on the ID side. In their own job they have never seen a single bit of software arise gratis. Rather they have to create, bit by bit, the active information of the software applications they develop. These people are justifiably perplexed when they encounter the evolutionist claim that God did not have to write a single line of code, because biological complexity (which is far greater than any computer software) arose naturalistically. “Why no work for Him and so much work for me?” they may ask. In this post, I hope I have helped explain that God, also in this case, expects far less from us than what He Himself did and does.

  • Delicious
  • Facebook
  • Reddit
  • StumbleUpon
  • Twitter
  • RSS Feed

167 Responses to Darwinism from an informatics point of view

  1. Where in your process are the support mechanisms for dereferencing pointers, handling calls and their returns, and making comparisons in order to perform conditional logic?

    If your analogy is to be valid, you have to show that such a framework to perform processing is there.

    As far as highlighting the complexity of a chess program, the prorammer starts with the first line, and ends with the last.

    No program just appears on your hard disk, it evolves from that first line.

    If the programmer doesn’t like a function, he re-writes it, until it survives his tests.

    He also doesn’t have to write the whole program, as he can use libraries written for other programs that are suitable for use in his new chess program.

  2. What I find ironic is that the progress of science, which is supposed to back up evolutionary claims, is doing exactly the opposite: debunking evolution completely. I am talking about real science, not the storytelling, non-falsifiable, pseudoscience of the atheistic evolutionary community.

  3. Mr Niwrad,

    You write as if the entire field of genetic programming did not exist.

    Within the field of computer science, the duality of programs and data is well known. In fact you make use of it yourself when referring to unused functions. Your presentation of the ideas of program and data as logically separate entities assumes a “Harvard architecture” distinction between program and data which does not exist in the more common von Neumann architecture, and is exemplified in programming languages such as Lisp.

    It should be no surprise that Lisp was used in some of the first genetic programming systems for exactly this reason – to take advantage of the duality of programs and data. A single string in memory could be executed as a program in one phase (similar to the protein creation system in biological systems) and in another phase operated upon as data (similar to the copying of DNA during cell replication).

    You further assume that what is passive or unused can never become active or vice versa). But we know that this does happen, both in real computer systems and in biology. In computer systems we see the problem of buffer overruns leading to data being executed. This is done intentionally by computer hackers. In biology we see stop codons being mutated or lost due to copy problems, with the result that reading continues into areas of the genome that were previously passive.

    The point is that your neat logical dichotomy between programs and data does not exist in the real world, either in computer science or in biology. Therefore the conclusions you draw from it are invalid.

  4. niwrad -

    I’m in IT as a chosen profession and appreciate your analogies. However I would have to point out that whilst these analogies are useful for illustrating certain points they cannot be extrapolated indefinitely without empirical support (outside of simple analogies). I mean if we say that a sub routine must be called from the main program then we are obliged to show where the equivalents can be found within biology. A sub routine for example makes use of a structure called a heap, which is a temporary memory holding area that operates on a LIFO (last in first out) basis, thereby allowing efficient nesting of function, so once again the question will be to find the equivalent arrangement within biology. The devil is always in the detail and I would wager that the human body operates in ways that are so alien to our own models of computation and decision making that the old analogies we used to use, prove to be wrong. I think the safest analogy we can make is that at the heart of things is information, and that biological entities must process it like a black box. We are only beginning to understand the mechanics of life so we should be humble to our core (something the evolutionists are not).

  5. niwrad:

    very good post indeed. I really like the clarity with which you sum up some fundamental truths which are usually misunderstood or just ignored.

    I have tried many times to state the obvious fact that traditional protein coding genes are only the final effectors of biological information. It is obvious that, if you change a final effector, you can cause a deep change in the final result: that is the case for instance in mutations in homeobox genes, and in many single mutation diseases. But, as you state very clearly, in no way that means that the final effector contains all the information implied in the final result.

    The problem is that we know a lot about final effectors (protein genes and proteins), and very little, indeed almost nothing, about what I would call “the real code”, or “the procedures”, in exactly the same sense you use for the main program and its sub-routines. Where is this code? Where is this information? We really don’t know. And, as you very correctly state, that problem of “where is the code” has deep implications both for phylogenesis and for ontogenesis. In other words, not only we don’t know how species, phyla, body plans etc originated; we really don’t understand how a complex multicellular being originates from a single cell containing exactly the same genome as all its multiple differentiated daughter cells. Again, where is the information? Is it mainly genomic, or epigenetic? And how is it encoded?

    Today we know many things, but they are still really only a little bit of the general picture. We know the importance of transcriptomes in individual cells. But transcripptomes are really mainly the effect of transcription factors. Which are proteins. So, what controls transcription factors? What is the cause of transcription factors’ transcriptomes at any given moment?

    We know that non coding DNA and non translated RNA must have an important role. But non coding DNA remains mainly a mystery. One thing is to assume that transposons and repetitive sequences may have an important role in shaping genetic information (I comletely agree with that assumption). Another thing is to have a detailed idea of how that may happen, and of where the information is coded which controls transposons and repetitive sequences.

    The fact remains that in ontogenesis a single mass memory of data (the genome) in some way is used to realize a lot of sub-programs
    (the individual sequence of transcriptomes in each specialized cell type).

    The current explanations for how that may happen are absolutely unsatisfying, as thogh the mere sequence of genes could determine, with a little help from random outer inputs, the whole procedure. You can easily see how unlikely that is.

    There are some interesting approaches about second codes, the importance of spacial conformation of DNA segments, and so on. But the problem remains the same: the genome in each cell of a single multicellular individual (with few important exceptions) is one and only one, and in itself it is static. Somehow, living cells use it appropriately in extremely varied contexts and for extremely varied sub-routines. Which information guides them?

    I think it’s truly a pity that at present we know so little about the procedures and their information content. The more we know about that, the less credible the idea will be that such information was generated by darwinian, or neo-darwinian, or neo-neo-derwinian mechanisms.

  6. niwrad, excellent post. and from this quote of yours:,,,

    “However, as we will see below, not one of these processes is capable of generating programs. Hence they are also incapable of creating new organs, new body plans, or even new species.”

    ,,,I think You may be able to help me with something we were talking about yesterday in this post:

    A Code That Isn’t Universal
    http://www.uncommondescent.com.....-the-code/

    specifically this at comment 3 and 4:

    Canadian Team Develops Alternative Splicing Code from Mouse Tissue Data
    Excerpt: “Our method takes as an input a collection of exons and surrounding intron sequences and data profiling how those exons are spliced in different tissues,” Frey and his co-authors wrote. “ The method assembles a code that can predict how a transcript will be spliced in different tissues.

    Thus niwrad, as far as I can see, which ain’t so far, it seems that a “unique higher level programming code of active information” is generated by their methodology for each species that takes into consideration the entirety of exon and intron sequences dissimilarity found between species.,,, As you know evolutionists are very keen on highlighting the alleged +98% similarity between chimps and humans by misleadingly genomic comparisons that are very selective in “what and when” they will compare. thus I thought that someone with a programming background, as well as a basic understanding of biology, someone like you , might be able to find something to counter that +98% psuedo-fact by highlighting the difference in programming coding:

    This line of investigation, for highlighting a dramatic “programming” difference between humans and chimps, looks promising because:

    Modern origin of numerous alternatively spliced human introns from tandem arrays – 2006
    Abstract excerpt: A comparison with orthologous regions in mouse and chimpanzee suggests a young age for the human introns with the most-similar boundaries. Finally, we show that these human introns are alternatively spliced with exceptionally high frequency.
    http://www.pnas.org/content/104/3/882.full

    If you are interested and think this may worth looking into further could you let me know?

  7. correction, this should read:

    it seems that a “unique higher level programming code of active information” is generated deciphered by their methodology for each species by taking into consideration the entirety of exon and intron sequence dissimilarity found between species.

  8. Nakashima #3

    You write as if the entire field of genetic programming did not exist.

    Not for chance in the software industry genetic programming has no practical use. From the theoretical point of view genetic programming doesn’t prove unguided evolution because in genetic algorithms there is no real creation of active information.

    You further assume that what is passive or unused can never become active or vice versa. But we know that this does happen, both in real computer systems and in biology. In computer systems we see the problem of buffer overruns leading to data being executed. This is done intentionally by computer hackers. In biology we see stop codons being mutated or lost due to copy problems, with the result that reading continues into areas of the genome that were previously passive. The point is that your neat logical dichotomy between programs and data does not exist in the real world, either in computer science or in biology. Therefore the conclusions you draw from it are invalid.

    Buffer overruns in computers and DNA overruns in biology don’t cause the creation of information, rather it is likely they generate failures. Eventually you can consider the generation of crashes a form of activity, but what I had in mind is constructive not destructive activity. The distinction between data and programs is an aspect of the distinction between instructions and their processor. Both are necessary: instructions without processor do nothing; processor without instructions does nothing. To say that this distinction doesn’t exist in the real world is as to say that agents and actions don’t exist because they are the same thing. My point was that blind evolution cannot work because it can randomly rearrange data only, while the problem of biology is to create the couple data and their manager.

  9. Toronto #1 and aqeels #4

    You seem to share the objection that the informatics analogy actually has little empirical support. It is true that the job of complete reverse engineering of biological systems has still to be fully accomplished. It is the target of the biologists of the XXI century (and they will succeed only by acquiring an ID mindset). However it seems already acknowledged that the genome is sort of read/write storage medium containing symbolic data. The read/write operations are not random, rather driven by processes according to specific necessities of the cell and imply an addressing mechanism and an encoder/decoder ability. This is similar to the relation that exists between control unit and memory in a computer. The control unit must necessarily host (or anyway execute) a main blueprint or schedule governing all the sub-processes. Without a top direction that provides instructions to the bottom executers no complex system can be organized. It seems to me that, also based on these few elements, the informatics analogy is worth considering.

  10. niwrad,
    “the informatics analogy is worth considering.”

    You are in good company. So does Bill Gates:

    Bill Gates, in recognizing the superiority found in Genetic Coding, compared to the best computer coding we now have, has now funded research into this area:

    Welcome to CoSBi – (Computational and Systems Biology)
    Excerpt: Biological systems are the most parallel systems ever studied and we hope to use our better understanding of how living systems handle information to design new computational paradigms, programming languages and software development environments. The net result would be the design and implementation of better applications firmly grounded on new computational, massively parallel paradigms in many different areas.
    http://www.cosbi.eu/index.php/.....rticle/171

    of note:

    Bill Gates does not employ random number generators and selection software to devise “more evolved” computer programs.

    —————

    Every Bit Digital DNA’s Programming Really Bugs Some ID Critics – March 2010
    Excerpt: In 2003 renowned biologist Leroy Hood and biotech guru David Galas authored a review article in the world’s leading scientific journal, Nature, titled, “The digital code of DNA.” The article explained, “A remarkable feature of the structure is that DNA can accommodate almost any sequence of base pairs—any combination of the bases adenine (A), cytosine (C), guanine (G) and thymine (T)—and, hence any digital message or information.” MIT Professor of Mechanical Engineering Seth Lloyd (no friend of ID) likewise eloquently explains why DNA has a “digital” nature: “It’s been known since the structure of DNA was elucidated that DNA is very digital. There are four possible base pairs per site, two bits per site, three and a half billion sites, seven billion bits of information in the human DNA. There’s a very recognizable digital code of the kind that electrical engineers rediscovered in the 1950s that maps the codes for sequences of DNA onto expressions of proteins.”
    http://www.salvomag.com/new/ar.....uskin2.php

  11. niwrad & aqeels,

    aqeels:

    The devil is always in the detail and I would wager that the human body operates in ways that are so alien to our own models of computation and decision making that the old analogies we used to use, prove to be wrong.

    While aqeels and I appear to be on different sides on the evolution debate, we both seem to agree that your analogy doesn’t come close enough to be of any use, to the point of being misleading.

    For instance;
    - what about shared memory,
    - semaphores,
    - is there a real-time element,
    - is it a multi-processor environment?

    You have written an OP which tries to make an analogy between the process of life and data processing.

    Trying to map our model of processing on that performed by life would probably result in a very inaccurate model.

    aqeels:

    I think the safest analogy we can make is that at the heart of things is information, and that biological entities must process it like a black box.

    I have no problem with the above statement at all despite being an atheist, (but a humble one).

  12. bornagain77 #6

    As you have rightly noted, the point I stressed in my post (the difference between data and their elaboration) has something to do with phenomena as the alternative splicing, where the same or very similar genomic data can generate very different outputs depending on the mechanisms involved (for example the same genes codifying for different proteins). In particular it is obvious that the +9X% similarity between chimps and humans genomes loses much of its appeal if their elaboration is very different.

    I think that there would be a lot to say about the methods of genomic comparisons in general and between chimps and humans in particular. In other words the +9X% pseudo-fact (as you call it) can be countered per se also without taking into consideration the difference in programming coding (may be I will deal with this topic in another post).

  13. niwrad @ 11

    I agree that biologists will pursue this field. But it’s not just an ID mindset that they need but possibly a complete paradigm shift away from conventional models (like the software and hardware analogies) that may yield the greatest discoveries. I just mean that we as ID proponents need to keep an open mind and not be fooled by what we see. Otherwise I could not agree with you more!

  14. niwrad,
    I really hope you, or someone, does deal with this in a future post and would be very interested specifically in a programmers take on this: (especially given what you already outlined in this post)

    You are right, The +9x% pseudo-fact has been countered on different fronts, mainly by the inclusion of “junk DNA” segments with the finding of +90% functionality of the genome. In fact, If the entire genome is considered the “pseudo fact” of genetic similarity is cut down to 70%:

    Chimpanzee? 10-10-2008 Dr Richard Buggs
    Excerpt: Therefore the total
    similarity of the genomes (between chimps and Humans) could be below 70%.
    http://www.idnet.com.au/files/pdf/Chimpanzee.pdf

    But of more interest to you, you stated

    Without a top direction that provides instructions to the bottom executers no complex system can be organized.

    Thus exactly as you are seeing things niwrad, it is found that the bottom executers (similar genes) output different protein sequences between chimps and Humans, :

    Eighty percent of proteins are different between humans and chimpanzees; Gene; Volume 346, 14 February 2005:
    http://www.ncbi.nlm.nih.gov/pubmed/15716009

    as well the timing of execution for the genes is found to vastly different

    A Primer on the Tree of Life (Part 4)
    Excerpt: “In sharks, for example, the gut develops from cells in the roof of the embryonic cavity. In lampreys, the gut develops from cells on the floor of the cavity. And in frogs, the gut develops from cells from both the roof and the floor of the embryonic cavity. This discovery—that homologous structures can be produced by different developmental pathways—contradicts what we would expect to find if all vertebrates share a common ancestor. – Explore Evolution http://www.evolutionnews.org/2......html#more

    further note:

    The Unbearable Lightness of Chimp-Human Genome Similarity
    Excerpt: One can seriously call into question the statement that human and chimp genomes are 99% identical. For one thing, it has been noted in the literature that the exact degree of identity between the two genomes is as yet unknown (Cohen, J., 2007. Relative differences: The myth of 1% Science 316: 1836.). ,,, In short, the figure of identity that one wants to use is dependent on various methodological factors.
    http://www.evolutionnews.org/2......html#more

    Human Genes: Alternative Splicing (For Proteins) Far More Common Than Thought:
    Excerpt: two different forms of the same protein, known as isoforms, can have different, even completely opposite functions. For example, one protein may activate cell death pathways while its close relative promotes cell survival.
    http://www.sciencedaily.com/re.....134623.htm

    Human genes are multitaskers:
    Abstract: Genome-wide surveys of gene expression in 15 different tissues and cell lines have revealed that up to 94% of human genes generate more than one (protein) product. http://www.nature.com/news/200......1199.html

    off topic:

    here are a few hundred new
    Contemporary Christian Music Videos:
    http://new.music.yahoo.com/vid.....038;page=1

  15. Mr Niwrad,

    Not for chance in the software industry genetic programming has no practical use. From the theoretical point of view genetic programming doesn’t prove unguided evolution because in genetic algorithms there is no real creation of active information.

    Sorry, you can’t rescue an argument based on analogy and logic by an appeal to pragmatic considerations. Genetic programming is a well developed field in computer science, and they give out cash prizes every year for programs that have evolved capabilities better than human. Such programs certainly are examples of the creation of “real” active information – they are capable of functions the programmer did not have.

    Trivially, we know that Deep Blue played chess better than any of its programmers. We know GP systems discover novel ways of solving problems that their programmers did not know existed and can be hard put to explain. Far from being incapable of novelty, they often sample the design space more widely than a set of human generated designs do. This web page is a good discussion with references.

  16. Mr BA^77,

    Bill Gates does not employ random number generators and selection software to devise “more evolved” computer programs.

    A quick search of research.microsoft.com for “evolutionary algoirthm” suggests otherwise. You may be particularly interested in Finding a Better-than-Classical Quantum AND/OR Algorithm using Genetic Programming.

  17. Nakashima #15,16

    My friend, you would be a terrific poker gambler. I love when you bluff.

    Research.microsoft.com contains all sorts of things, included papers on
    “Pictures from the birdpark at Iguasu Falls” and “Eroticism and the night: Sensual, rhythmic and risky design”.

    So no wonder it contains articles on genetic programming too. Microsoft has produced Terabytes of software. Not a single bit of them is obtained by genetic programming. This is not a pragmatic ascertainment, rather a conceptual issue: to design a genetic algorithm that will output the final program is far more expensive than to design directly the final program. If genetic programming (or whatever machine) really create new information as you say Microsoft would not have the thousands human developers it has. Bill Gates likes to save dollars.

  18. niwrad actually Nak is a terrible poker player since we know for a fact he is going to bluff on every single hand while never holding a winning hand.

    In the following podcast, Robert Marks, a leading expert in the area of evolutionary algorithms, gives a very informative talk as to the strict limits we can expect from any evolutionary computer program (evolutionary algorithm):

    Darwin as the Pinball Wizard: Talking Probability with Robert Marks – podcast
    http://www.idthefuture.com/201.....ard_t.html

    Further note:

    A comparative approach for the investigation of biological information processing: An examination of the structure and function of computer hard drives and DNA – David J D’Onofrio1, Gary An – Jan. 2010
    Excerpt: It is also important to note that attempting to reprogram a cell’s operations by manipulating its components (mutations) is akin to attempting to reprogram a computer by manipulating the bits on the hard drive without fully understanding the context of the operating system. (T)he idea of redirecting cellular behavior by manipulating molecular switches may be fundamentally flawed; that concept is predicated on a simplistic view of cellular computing and control. Rather, (it) may be more fruitful to attempt to manipulate cells by changing their external inputs: in general, the majority of daily functions of a computer are achieved not through reprogramming, but rather the varied inputs the computer receives through its user interface and connections to other machines.
    http://www.tbiomed.com/content/7/1/3

    The Capabilities of Chaos and Complexity – David L. Abel
    Excerpt: “To stem the growing swell of Intelligent Design intrusions, it is imperative that we provide stand-alone natural process evidence of non trivial self-organization at the edge of chaos. We must demonstrate on sound scientific grounds the formal capabilities of naturally-occurring physicodynamic complexity. Evolutionary algorithms, for example, must be stripped of all artificial selection and the purposeful steering of iterations toward desired products. The latter intrusions into natural process clearly violate sound evolution theory.”
    http://www.mdpi.com/1422-0067/10/1/247/pdf

    Arriving At Intelligence Through The Corridors Of Reason (Part II) – April 2010
    Excerpt: Summarizing the status quo, Johnson notes for example how AVIDA uses “an unrealistically small genome, an unrealistically high mutation rate, unrealistic protection of replication instructions, unrealistic energy rewards and no capability for graceful function degradation. It allows for arbitrary experimenter-specified selective advantages”. Not faring any better, the ME THINKS IT IS LIKE A WEASEL algorithm is programmed to direct a sequence of letters towards a pre-specified target.
    http://www.uncommondescent.com.....n-part-ii/

    Accounting for Variations – Dr. David Berlinski: – video
    http://www.youtube.com/watch?v=aW2GkDkimkE

    LIFE’S CONSERVATION LAW – William Dembski – Robert Marks – Pg. 13
    Excerpt: Simulations such as Dawkins’s WEASEL, Adami’s AVIDA, Ray’s Tierra, and Schneider’s ev appear to support Darwinian evolution, but only for lack of clear accounting practices that track the information smuggled into them.,,, Information does not magically materialize. It can be created by intelligence or it can be shunted around by natural forces. But natural forces, and Darwinian processes in particular, do not create information. Active information enables us to see why this is the case.
    http://evoinfo.org/publication.....ation-law/

    Conservation of Information in Computer Search (COI) – William A. Dembski – Robert J. Marks II – Dec. 2009
    Excerpt: COI puts to rest the inflated claims for the information generating power of evolutionary simulations such as Avida and ev.
    http://evoinfo.org/publication.....nt-reason/

    Evolutionary Synthesis of Nand Logic: Dissecting a Digital Organism – Dembski – Marks – Dec. 2009
    Excerpt: The effectiveness of a given algorithm can be measured by the active information introduced to the search. We illustrate this by identifying sources of active information in Avida, a software program designed to search for logic functions using nand gates. Avida uses stair step active information by rewarding logic functions using a smaller number of nands to construct functions requiring more. Removing stair steps deteriorates Avida’s performance while removing deleterious instructions improves it.
    http://evoinfo.org/publication.....gic-avida/

    The Problem of Information for the Theory of Evolution – debunking Schneider’s ev computer simulation
    Excerpt: In several papers genetic binding sites were analyzed using a Shannon information theory approach. It was recently claimed that these regulatory sequences could increase information content through evolutionary processes starting from a random DNA sequence, for which a computer simulation was offered as evidence. However, incorporating neglected cellular realities and using biologically realistic parameter values invalidate this claim. The net effect over time of random mutations spread throughout genomes is an increase in randomness per gene and decreased functional optimality. http://www.trueorigin.org/schneider.asp

    “There has been increasing recognition that genes deal with information processing. They have been referred to as “subroutines within a much larger operating system”.”
    Kirk Durston

  19. As for Nak’s bluff that the deep blue chess program proves that computers are more intelligent than people:

    “GilDodgen
    One of my AI (artificial intelligence) specialties is games of perfect knowledge. See here:

    worldchampionshipcheckers.com

    In both checkers and chess humans are no longer competitive against computer programs, because tree-searching techniques have been developed to the point where a human cannot overlook even a single tactical mistake when playing against a state-of-the-art computer program in these games.

    On the other hand, in the game of Go, played on a 19×19 board with a nominal search space of 19×19 factorial (1.4e+768), the best computer programs are utterly incompetent when playing against even an amateur Go player.

    When it comes to true intelligence — even something as seemingly simple as evaluating the grammatical correctness or meaning of a trivial sentence — the best computer programs are less than worthless.

    I turn off my Microsoft Word grammar checker, because it is wrong almost all of the time, and its suggestions are almost universally laughably stupid.

    The notion that random errors filtered by natural selection created the human mind — which is capable of creating language and interpreting it — is laughably stupid raised to the 768th power.”
    http://www.uncommondescent.com.....ent-353454

    further note:

    The Capabilities of Chaos and Complexity: David L. Abel – Null Hypothesis For Information Generation – 2009
    To focus the scientific community’s attention on its own tendencies toward overzealous metaphysical imagination bordering on “wish-fulfillment,” we propose the following readily falsifiable null hypothesis, and invite rigorous experimental attempts to falsify it: “Physicodynamics cannot spontaneously traverse The Cybernetic Cut: physicodynamics alone cannot organize itself into formally functional systems requiring algorithmic optimization, computational halting, and circuit integration.” A single exception of non trivial, unaided spontaneous optimization of formal function by truly natural process would falsify this null hypothesis.
    http://www.mdpi.com/1422-0067/10/1/247/pdf
    http://mdpi.com/1422-0067/10/1/247/ag

    —————-

    Brooke Fraser- “C S Lewis Song”
    http://www.youtube.com/watch?v=GHpuTGGRCbY

  20. Mr Niwrad,

    At the risk of sounding repetitious, your original argument was an analogy, and based on a dichotomy between data and program. My reply showed that this logical distinction was false. A reply of the kind “But you can’t make much money with it!” is inappropriate.

    By directing your attention to the Microsoft issue, you bypass replying on NASA’a antenna building or the Humie awards. That is unfortunate, since the antenna example was a more direct response to your assertion that GP isn’t practical.

    But since it is fun, let’s talk about Microsoft a little more. You write:
    to design a genetic algorithm that will output the final program is far more expensive than to design directly the final program.

    If by this you mean the user interface, the use of operating system resources, etc. then I agree with you. EAs are economically attractive when they can test many alternatives quickly and cheaply, relative to some other method. For example, NASA didn’t actually build thousands of antennas, they simulated their responses using another standard piece of software. They only built a few of the high performing champions. It is difficult to design a UI this way, because you have to test a UI with real people.

    Is Microsoft wrong for using multiple coding methods? Will GP only be successful when Bill Gates uses it to brush his teeth? No.

  21. Mr BA^77,

    As for Nak’s bluff that the deep blue chess program proves that computers are more intelligent than people:

    That is not a very accurate gloss of what I said. Programs (such as Deep Blue) can acheive competencies that their programmers do not have. Sooner Emeritus notes on another thread that Blondie24 might be a more apt example, since it evolved its competency. This contradicts what Mr Niwrad was proposing.

  22. Toronto and aqeels

    Nobody claims that the information processing in cells is identical to the actual information and robotics technology. Don’t go to search for Intel CPUs or Maxtor hard-disks inside the cells. Don’t ask if the biological instructions are written in C or Java either. However, beyond the details, any kind of information processing involves some common principles.

    When you have to create organisms (able to live, self-reproduce, self-repair, survive, etc.) you necessarily have to elaborate symbolic instructions. To do that you need at least a memory, some input output devices, a universal constructor, a control unit, a language.

    Not for chance the same things are necessary to create robots. Note that actually robots don’t have the above advanced features of life (for this reason they can miss the universal constructor). However, for what they do, they just imply the above informatics principles. The implementation of these principles is designed by computer science, robotics and AI engineers.

    Hence the aim of my post was: whether robotics/informatics is designed the same biology has to be designed because both implement the same principles. Since Darwinism denies any design my conclusion was that Darwinism is necessarily untrue from the informatics perspective.

    Any difference in details doesn’t change this conclusion because it is based on principles not on details.

  23. niwrad @20,

    When you have to create organisms (able to live, self-reproduce, self-repair, survive, etc.) you necessarily have to elaborate symbolic instructions.

    Why symbolic?

    I can process an analog input voltage with a digital processor using symbolic instructions or with an analog process that doen’t need any symbolic content at all.

    If I put this in a black box, you couldn’t tell which process it was.

    If you opened the cell and showed me a digital CPU or equivalent that translated chemical inputs into symbolic information, proceesed it, and then converted symbolic values into chemical outputs,I would buy into your analogy, but you need to show me that.

    What you have done is base a conclusion on an analogy which is so far removed from reality that the conclusion you draw, is no longer valid.

  24. So Toronto are you saying that any argument with a premise that biological life is computable should be rejected.

  25. No.
    See my response @11.

  26. I just read 11. It seems like you mean ‘Yes’:

    You have written an OP which tries to make an analogy between the process of life and data processing.

    Trying to map our model of processing on that performed by life would probably result in a very inaccurate model.”.

    To slighltly reword the above, do you also ascribe to the following:

    “Trying to map any existing model of data processing on that performed by life would probably result in a very inaccurate model.”

  27. What about this statement:

    Trying to map our model of processing on that performed by the weather would probably result in a very inaccurate model.”

  28. Well, you probably do agree with that. But if weather simulations for example are not accurate its not because in principle computers are incapable of accurately simulating it. Let’s put it this way – I don’t of any scientist (at least not any of those who sign those Greenhouse gas accords) who would even for a second entertain the idea that computers in principle are incapable of simulating it adequately. Rather the existing programs are inadequate. But know one’s saying you can’t accurately model the weather in principle using instructions, memory a processor, etc.

  29. I’m not really addressing the OP. Just for the record the very first thing I noticed was the sharp distinction betwen data and instructions and then Nakashima pointed that out.

  30. Nak, your antenna example, (that you are so fond of repeating though you have been shown why it fails to generate functional information) is an excellent example of the ability of computers to find an optimal solution within a very well defined search space. Much like Deep Blue’s ability to place chess by searching through a well defined search space. What you are adverse to ever mention is that ZERO functional (logical, mathematical, algorithmic, programming) information is generated in finding that optimal solution within that well defined search space.

    John 1:1
    In the beginning was The Word (Logos),

    of note: the Greek word Logos is the root word from which we get our modern word “logic”

    Nak if you believe the antenna example generates functional algorithmic information, Please write it up in peer review so as to falsify Abel’s null hypothesis:

    The Capabilities of Chaos and Complexity: David L. Abel – Null Hypothesis For Information Generation – 2009
    To focus the scientific community’s attention on its own tendencies toward overzealous metaphysical imagination bordering on “wish-fulfillment,” we propose the following readily falsifiable null hypothesis, and invite rigorous experimental attempts to falsify it: “Physicodynamics cannot spontaneously traverse The Cybernetic Cut: physicodynamics alone cannot organize itself into formally functional systems requiring algorithmic optimization, computational halting, and circuit integration.” A single exception of non trivial, unaided spontaneous optimization of formal function by truly natural process would falsify this null hypothesis.
    http://www.mdpi.com/1422-0067/10/1/247/pdf
    http://mdpi.com/1422-0067/10/1/247/ag

  31. Nak, here is something for you to listen to as you write up that peer review to falsify Abel’ null hypothesis:

    Cricket Chirping
    http://www.youtube.com/watch?v=CQFEY9RIRJA

  32. JT @28,
    What we are talking about is the architecture of the “processor” of the information.

    In other words, does this information processor contain:

    - what we believe is equivalent to an ALU?

    - what we believe is equivalent to an instruction decoder?

    - what we believe is equivalent to a program counter ?

    - etc.

    We are not talking about the use or end result of any software we load and execute.

    niwrad is trying to say that we can see a lot of commonality between the way biology processes information and the way human-made computers process information.

    I am asking him to show me those common elements, not simply assert, but show me.

    If he can’t, he shouldn’t try to draw any conclusions from what appears to be a flawed analogy.

  33. Mr BA^77,

    What you are adverse to ever mention is that ZERO functional (logical, mathematical, algorithmic, programming) information is generated in finding that optimal solution within that well defined search space.

    What is this functional information you speak of? Your ideas are intriguing to me, and I would like to subscribe to your newsletter.

    Seriously, functioning better as an antenna doesn’t demonstrate an increase in functional information to you? I thought you liked to use “functional information” in the sense of Hazen, Szostak, et al?

  34. toronto @ 32

    The Cell – A World Of Complexity Darwin Never Dreamed Of – Donald E. Johnson – video
    http://www.metacafe.com/watch/4139390

    Bioinformatics: The Information in Life – Donald Johnson – video
    http://vimeo.com/11314902

    On a slide in the preceding video, entitled “Information Systems In Life,” Dr. Johnson points out that:

    * the genetic system is a pre-existing operating system;
    * the specific genetic program (genome) is an application;
    * the native language has codon-based encryption system;
    * the codes are read by enzyme computers with their own operating system;
    * each enzyme’s output is to another operating system in a ribosome;
    * codes are decrypted and output to tRNA computers;
    * each codon-specified amino acid is transported to a protein construction site; and
    * in each cell, there are multiple operating systems, multiple programming languages, encoding/decoding hardware and software, specialized communications systems, error detection/correction systems, specialized input/output for organelle control and feedback, and a variety of specialized “devices” to accomplish the tasks of life.

    Cells Are Like Robust Computational Systems, – June 2009
    Excerpt: Gene regulatory networks in cell nuclei are similar to cloud computing networks, such as Google or Yahoo!, researchers report today in the online journal Molecular Systems Biology. The similarity is that each system keeps working despite the failure of individual components, whether they are master genes or computer processors. ,,,,”We now have reason to think of cells as robust computational devices, employing redundancy in the same way that enables large computing systems, such as Amazon, to keep operating despite the fact that servers routinely fail.”
    http://www.sciencedaily.com/re.....103205.htm

  35. bornagain77 @34,
    With your links, you are simply asserting the same things as niwrad, and making the same analogy.

    You still haven’t shown me the equivalent architecture in a cell which corresponds to that architecture in computers made by humans, to validate your analogy.

    For instance, if I asserted that since dogs have hearts just like humans do, therefore their circulation systems must work something like ours, that would be an analogy that you might say is flawed.

    If I responded with a link that asserted the same thing, that would not be enough evidence to convince you my analogy was valid.

    If instead I supplied you with an x-ray of a dog’s heart and one of a human’s heart, there’s something you could evaluate.

    Show me that level of evidence for biological processor architecture, not a link to another assertion of the same type.

  36. From OP

    Data is passive, while the program is active. What is passive cannot create what is active.

    A program P does different things based on the data given to it as input. So in that sense the data itself is a program, telling P what to do and what not to do. So this really seems to be a fundamental oversight in the author’s conception. This is what is relevant – not the mere fact that he employs a computational analogy (and he doesn’t have to point out for example where the equivalent of an ALU in an organism. Also multiple processors – irrelevant. A Turing Machine doesn’t have multiple processors. Details of interprocess communication (e.g. semaphores, etc.) irrelevant, doesn’t need to be addressed)

    But also the following from the OP is erroneous:

    But even if genes were the equivalent of software functions, then once again, random mutations, sexual recombination, horizontal gene transfer and duplication of functions would still be incapable of creating hierarchical decision logic. Why not? Because the decision logic contained in the main program is what invokes the functions (by referencing them). Just as a hammer or a drill cannot create a carpenter, the above operations on functions are incapable of creating their user

    First of all, it wasn’t clear to me why “the main program” (in an organism) is off limits to random change. But regardless, even subprograms called by the main program can totally subvert what the main program is doing. Take a unversal program -Its behavior is totally contingent on the behavior of the sub programs it calls.

    But to use his example of a carpenter and the idea that tools can’t create a carpenter: Say you have a carpenter with internal logic like, “Look at screw head. Determine if screw head is phillips or standard. if standard go get standard drill bit and insert into drill otherwise go get phillips drill bit and insert into drill.”

    OK say you have another carpenter who is considerably less adapted, and never attempts to discern the type of screw head but just uses whatever bit is in the drill. And lets say the drill bit changes into some sort of universal drill bit that can handle either phillips head or standard. All of a sudden our unqualified carpenter is now qualified to handle any type of screw head based simply on the tool he was using changing (without any input from him.)

  37. Toronto #23, 32

    Why symbolic?

    When we have to store data and instructions (or in general information) we need to use a code. As general definition, a code is a mapping between two realities of different level, the true reality and the correlated reality. The latter is also called a “symbol” of the former.

    Examples:

    A language is a code mapping things to words. Words are symbols of things. If I write “tree” I am only pointing to the real tree. I am not showing or offering a real tree.

    The genetic code maps DNA codons to amino acids. For instance the codon “TAC” corresponds to Tyrosine. This characters string “TAC” is a symbol for Tyrosine.

    In informatics, according to the ASCII code the string “01000001″ represents the character “A”.

    Only intelligence can create codes because symbolism implies abstraction, which is unreachable by material processes for definition.

    Codes are typical of informatics and appear in biology. Besides they don’t stay in computers and in cells to do nothing. They are used by computers and cellular machinery to accomplish tasks. Just this simple fact proves that the analogy between the two fields is sensible.

  38. niwrad @37,
    The cell process might be analog.

    If this is the case, no “symbolic” data is necessary.

    I’ve been working with embedded real-time systems for 32 years and there are a lot of Evo’s here who also have a good handle on what a processor is and how it works.

    Show me that your analogy is valid by showing me the cell’s equivalent architecture that would correspond with the architecture design of human-made computers.

    If you can’t show me support for a level of indirection, you can’t claim “symbolic” codes are being used.

    Notice that I am concerned with the architecture.

    If the architecture is not the same, then neither is the manner in which the information is stored or used, and your analogy fails.

  39. About symbolic realities in cells:

    Obviously the DNA mass memory in protein coding genes is symbolic: it uses a symbolic code, the genetic code, to carry the information for protein syntesis. That is well known and acknowledged, and I am always surprised that there are still so many darwinists who try to deny it.

    But that is not the only example. I would like to mention, for instance, that the whole cell signaling process is in essence symbolic.

    Now I don’t want to get trapped in a long philosophical discussion about what symbolic means or does not mean. I just want to stress that in complex multicellular organisms, like humans, there is a long distance biochemical control of body reactions, where some central organ, let’s say an endocrine gland, coordinates an integrated response of different cell types and organs.

    How does that happen? Through biochemical “messengers”, like hormones and cytokines, which are released in a long distance communication pathway (usually the blood). When the messenger molecule reaches the target cells, it is selectively recognized by cell membrane receptors. Then a very complex and integrated network of intracellular signaling cascades, usually implying many different proteins, transmits the signal from the cell membrane to the nucleus.

    Here, the signal activates a complex a coordinated response, absolutely specific for the signal itself, which translates into a new asset of transcription factors, and therefore a new transcriptome. In other words, new genes are activated and transcribed, while others may be silenced. In other words, a local sub-routine is activated by the central signal.

    The transcripted m-RNA is still modulated at the splicing level, and finally the proteins are synthesized (again in the cytoplasm) which will really effect the behavioural responce of the cell to the original signal.

    Another important point is that the same biochemical signal (the same hormone, for instance) activates different responses in different cell types, but those responses are all perfectly integrated and coordinated to realize a general response of the organism to the stimulus which in the beginnig originated the response of the central endocrine gland (the hormone release).

    Is that a good example of a coordinate functional response to an environmental stimulus, realized through biochemical symbolic messengers and differentiated and integrated processing machinery?

  40. JT #35

    My point was that data cannot create the decision logic (main program) that overarches them. It is obvious that data, when are inputted, can change the decisions results. But to change the decisions results is an entirely different thing than to create the decision logic itself. In this sense I said that data are passive, i.e. they don’t create the decision logic from nothing. Since the question at issue was if Darwinian processes acting on data (genes) could create new programs (macroevolution implies new programs) my conclusion was that they cannot.

    The above note also answers your objection about the carpenter and his tools. The universal drill bit you have imagined could help the job of a carpenter but one cannot claim that it creates the carpenter. Since my whole point is about creation of programs, processors and carpenters my conclusion that they cannot be created by the things upon which they operate remains valid. And since blind evolution pretends exactly to do that (to create bottom-up the entire systems) the conclusion of impossibility applies to evolution too.

    By the way I never said that “the main program (in an organism) is off limits to random change” as you seem to affirm. It is Darwinism to say that mutations on genes are the driving force of evolution. Obviously all parts of any complex system may change due to entropy and other factors. However these entropic changes never increase the organization of the system, rather they destroy its organization.

  41. 41

    gpuccio,

    The standard response repeatedly given by ID opponents is that we can see mechanically how this signalling network works, therefore there is no symbollic nature to it.

    Tab A slips into Slot B and thats all there is to it.

  42. Upright BiPed:

    I would answer that we can equally see how a PC running Windows is just the sum of electric events where one state changes to another, therefore there is no symbolic nature to it.

    I believe that symbols, functions, purposes, and in general meanings are perceived only by conscious beings. We, as conscious beings, recognized the meanings and the finalities which have been encoded in mechanistic events by other conscious beings. Machines are not aware of finalities, of functions or of meanings.

    So, if an extreme materialist really does not believe in consciousness and in its specific modalities, then he will not believe in symbols, meanings or finalities. But he should use the same way of reasoning both for biological information and for human generated information (language, software, and so on).

    As for me, I do believe in consciousness. I perceive it in myself, so for me it is a fact.

  43. [38]:

    I just want to stress that in complex multicellular organisms, like humans, there is a long distance biochemical control of body reactions, where some central organ, let’s say an endocrine gland, coordinates an integrated response of different cell types and organs.

    How does that happen? Through biochemical “messengers”, like hormones and cytokines, which are released in a long distance communication pathway (usually the blood). When the messenger molecule reaches the target cells, it is selectively recognized by cell membrane receptors. Then a very complex and integrated network of intracellular signaling cascades, usually implying many different proteins, transmits the signal from the cell membrane to the nucleus.

    Here, the signal activates a complex a coordinated response, absolutely specific for the signal itself, which translates into a new asset of transcription factors, and therefore a new transcriptome. In other words, new genes are activated and transcribed, while others may be silenced. In other words, a local sub-routine is activated by the central signal.

    The transcripted m-RNA is still modulated at the splicing level, and finally the proteins are synthesized (again in the cytoplasm) which will really effect the behavioural responce of the cell to the original signal.

    Another important point is that the same biochemical signal (the same hormone, for instance) activates different responses in different cell types, but those responses are all perfectly integrated and coordinated to realize a general response of the organism to the stimulus which in the beginnig originated the response of the central endocrine gland (the hormone release).

    Is that a good example of a coordinate functional response to an environmental stimulus, realized through biochemical symbolic messengers and differentiated and integrated processing machinery?

    As far the importance of centralized control –

    I’m thinking about a city and how incredibly multifaceted it is, and it is also extremely integrated, involving all sorts of complex communication between various parts, various rules and regulations. But you have the airport, garbage collection, the school system, the traffic system, stop lights, hospitals, restaurants, the jail. But you have complex interactions between all these parts. Any new entity that want’s to integrate into this system has to do so by adhering to a number of regulations, filing of reports, orchestrated communication and the like.

    And the question would be to what extent is that entire system centrally controlled or planned in advance, or alternatively to what extent do different pieces come into it of their own accord and either integrate or fail to integrate. And to what extent are new rules being added on the fly without any sort of consideration for their long term impact. And to what extent are entities when integrating into the system blindly following rules currently existing in the system.

    But I also think about centralized control in totalitarian or Marxist regimes (e.g. Cuba, North Korea) and how invariably that leads to incredible statis and stagnation and lack of flexibility and death.

  44. Upright BiPed, gpuccio,

    If I find in a human body, a molecule consisting of an oxygen atom and two hydrogen atoms, that “is” water, it is not a “symbol” for water, despite the fact that “our” symbol for it, and I really want to stress the “our” part, is “H20″.

  45. JT:

    In the example of a city we certainly have both centralized control and individual adaptation. There should be no doubt about that.

    But, you see, we are talking of intelligent individuals who take part in a general environment. Both the central control and the individual adaptation are conscious and intelligent. They don’t happen blindly (although sometimes, when we see some of the results, we would like to think so!).

  46. Nak please read niwrad @ 37

    nak as to the fact that Abel is very specific in he is talking of “logical” algorithmic (i.e. abstract encoded) information, that is the line in the sand you must cross to falsify his null hypothesis. Care to call me when your done with that peer review?

    Cricket Chirping
    http://www.youtube.com/watch?v=CQFEY9RIRJA

    In your assertion that functional information was found in the optimum antenna search as per Hazen; Szostak, as I stated before, I accept that a optimal solution to the problem was found within a well defined search space even without measuring the string of Functional Information Bits(FITS). I don’t usually deny the obvious as Darwinists constantly do with ID. Yet nak even in this “simulated” example for the finding of functional sequence of information, within a well defined search space, when we look to biology to find a corroborating example of the generation of such a string of functional information bits we are almost always immediately referred to the Nylonase. But in Nylonase we can readily apply a measure, whereas we can’t readily apply a measure in a computer, to see if the “gain” in functional information, as per Szostak, for the Nylonase really did increase the inherent algorithmic complexity of the parent bacteria of the Nyloase by comparing the adapted Nylonase to the parent bacteria in a fitness test in the parents native environment, to see if the Nylonase are genuinely more fit (functional) than the parent strain.

    Is Antibiotic Resistance evidence for evolution? – “The Fitness Test” – video
    http://www.metacafe.com/watch/3995248

    And as you well know when the nylon is removed from the environment the parent bacteria quickly out compete the nyloase bacteria. Thus the Nylon example fails to falsify the principle of genetic entropy. And even when we look closer at Nylonase we find that it is worse than even the fitness test revealed:

    Nylon Degradation – Analysis of Genetic Entropy
    Excerpt: At the phenotypic level, the appearance of nylon degrading bacteria would seem to involve “evolution” of new enzymes and transport systems. However, further molecular analysis of the bacterial transformation reveals mutations resulting in degeneration of pre-existing systems.
    http://www.answersingenesis.or.....n-bacteria

    As well nak even though I would be willing to give you the fact that random evolution could possibly generate functional information per Szostak, but not the algorithmic (encoded) information as per Abel, It seems that in biology the search space for finding a solution of functional information is extremely larger than your cherished antenna example from computers:

    The Case Against a Darwinian Origin of Protein Folds – Douglas Axe – 2010
    Excerpt Pg. 11: “Based on analysis of the genomes of 447 bacterial species, the projected number of different domain structures per species averages 991. Comparing this to the number of pathways by which metabolic processes are carried out, which is around 263 for E. coli, provides a rough figure of three or four new domain folds being needed, on average, for every new metabolic pathway. In order to accomplish this successfully, an evolutionary search would need to be capable of locating sequences that amount to anything from one in 10^159 to one in 10^308 possibilities, something the neo-Darwinian model falls short of by a very wide margin.”
    http://bio-complexity.org/ojs/.....O-C.2010.1

    Estimating the prevalence of protein sequences adopting functional enzyme folds: Doug Axe:
    Excerpt: Starting with a weakly functional sequence carrying this signature, clusters of ten side-chains within the fold are replaced randomly, within the boundaries of the signature, and tested for function. The prevalence of low-level function in four such experiments indicates that roughly one in 10^64 signature-consistent sequences forms a working domain. Combined with the estimated prevalence of plausible hydropathic patterns (for any fold) and of relevant folds for particular functions, this implies the overall prevalence of sequences performing a specific function by any domain-sized fold may be as low as 1 in 10^77, adding to the body of evidence that functional folds require highly extraordinary sequences. http://www.ncbi.nlm.nih.gov/pubmed/15321723

    and nak this is JUST ONE friggin protein!!! We ain’t even got to the part where Darwinists have got to explain the origination of body plans>

    This inability for the DNA code to account for body plans is also clearly shown by extensive mutation studies to the DNA of different organisms which show “exceedingly rare” major morphological effects from mutations to the DNA code.

    Stephen Meyer – Functional Proteins And Information For Body Plans – video
    http://www.metacafe.com/watch/4050681

    The Origin of Biological Information and the Higher Taxonomic Categories – Stephen Meyer”Neo-Darwinism seeks to explain the origin of new information, form, and structure as a result of selection acting on randomly arising variation at a very low level within the biological hierarchy, mainly, within the genetic text. Yet the major morphological innovations depend on a specificity of arrangement at a much higher level of the organizational hierarchy, a level that DNA alone does not determine. Yet if DNA is not wholly responsible for body plan morphogenesis, then DNA sequences can mutate indefinitely, without regard to realistic probabilistic limits, and still not produce a new body plan. Thus, the mechanism of natural selection acting on random mutations in DNA cannot in principle generate novel body plans, including those that first arose in the Cambrian explosion.”
    http://eyedesignbook.com/ch6/eyech6-append-d.html

    note to toronto you better tell these people that computers ain’t got nothing to do with cells before they waste to much money on such hair-brained ideas:

    Nanoelectronic Transistor Combined With Biological Machine Could Lead To Better Electronics: – Aug. 2009
    Excerpt: While modern communication devices rely on electric fields and currents to carry the flow of information, biological systems are much more complex. They use an arsenal of membrane receptors, channels and pumps to control signal transduction that is unmatched by even the most powerful computers. http://www.sciencedaily.com/re.....091834.htm

  47. 43 gpuccio: Its a tenant of your faith that there is some magic called “intelligence” or conscious intelligence” that make everything work. When a beuracracy is in place, that is a complex series of rules and regulations that govern people’s behavior – that set of rules and regualtions is not conscious or intelligent and yet it exhibits profound control. People in their delay lives, organzations in their delay lives function by following various types of complex rules that are context sensitive. either active or inactive depending on the context. I don’t think its correct to treat humans by default as embued with some sort of special magic thing you call “intelligence” (rather its all rules and impulses and imperatives and objectives, objectives determined by primal impulses, etc.

  48. delay = daily

  49. Only intelligence can create codes because symbolism implies abstraction, which is unreachable by material processes for definition.

    Now _that_ is circular reasoning!

  50. 45 Yes I understand the we call what people are “intelligent”. I guess ID is really and truly committed to the concept of consciousness as the transcendent thing that is integral to everything (though not apparently definable.) I will be off for a while.

  51. My point about the city would be, what if a human being was like a city, evolving over time without an overall plan in advance, with complex entities coming together and over time become more integrated.

  52. niwrad:

    Only intelligence can create codes because symbolism implies abstraction, which is unreachable by material processes for definition.

    This sounds very much like arguments that are made by several other ID proponents here, to the effect that material processes can’t create abstractions. I’m no metaphysician, but this argument has never worked for me.

    Every abstraction already exists in some space of abstractions. In the context of these ID discussions, creating an abstraction obviously means imbuing it with a stronger type of existence. That seems to always entail forming something physical (or at least not demonstrably non-physical) on which the abstraction supervenes.

    The difference between the genetic code and other conceivable codes is that the genetic code is a fitting model of observed biochemical processes, and other codes are not. When we talk about the creation of the genetic code, we’re really talking about the initiation of those biochemical processes.

  53. Hi GP!

    Always good to hear from you.

    And your point on the long range signalling — note the difference from messaging [So far as I can tell there is no code made up from hormones etc] — that triggers local genetically coded activity in cells.

    Onlookers:

    I find it astonishing in a digitally literate age that when we see 3-letter, 4 state elements, specifying 64 codes; and we see them all used in genes [up to a couple of dialects], and we see that between the table elements there is not an interpolation of half one acid half the next that can go into the AA chain being coded for, people try to deny the existence of a code that is discrete state and symbolic.

    Moreover, the action of transcription [and splicing] then insertion into the ribosome and sequencing step by step to give an AA chain to produce a functional protein, is plainly algorithmic.

    When one is forced to find strained objections like we are seeing to the obvious from the just [again] linked video, that is telling.

    GEM of TKI

  54. @Rob

    -”When we talk about the creation of the genetic code, we’re really talking about the initiation of those biochemical processes”

    Can you please clarify exactly what you mean by that? Are you saying that undirected biochemical processes actually create codes?

  55. R0b #49

    The difference between the genetic code and other conceivable codes is that the genetic code is a fitting model of observed biochemical processes, and other codes are not.

    Also the informatics codes can be seen as fitting models of observed physical processes (e.g. the digital signals in the electronic circuits of a computer). I don’t see the difference.

    The key point is that biochemical and physical processes per se can only obey the biochemical and physical laws. Codes are additional high level rules respect the low level laws of physics and chemistry. How can material processes create high level abstract rules when they only obey to low level material laws? It seems here we have a double absurdity: (a) what obeys begins to rule (the opposite of obeying); (b) what works at the low level begins to work at the high level (mismatch of reality layers).

    From the natural laws can only emerge what is potentially contained in them. Biological and informatics codes are not potentially contained in those laws then there must exist a higher source that imposes these abstract codes upon the material supports.

    To deny this conclusion is as to claim that a cards deck can create the poker rules by itself.

  56. JT:

    I appreciate your reflections. Anyway, form my point of view, it’s a tenant of my experience that there is consciousness and intelligence in human beings. Faith has nothing to do with that.

    Just to repeat what should be obvious, the existence of consciousness in myself is a direct perception, absolutely undeniable, and is the basis for eny other experience of reality I can have.

    The existence of a similar experience in other human beings, instead, is an inference. But it is an inference shared practically by everybody (except maybe solipsists), and based on a very sound analogy (other human beings are very similar to me, behave more or less like me, and express in their behaviour all the objective characteristics that my consciousness expresses in my behaviour).

    So, faith has nothing to do with believing in the existence of consciousness and intelligence: it’s just a question of experience and sound inference.

  57. Hi, KF!

    It’s a pleasure to be here again with old friends.

    Regarding hormones and cytokines, I agree with you that they are signals, rather than proper messages, but my point is that they behave, in a way, like pointers in a program, being able to activate the right target in the right cell. And another point is that the signaling molecule has nothing to do with the final effect. It is just “recognized” by a receptor, which then again, through indirect pathways, elicits the final response. I don’t know if the ability of the receptor to recognize a specific biochemical pattern, and to couple that recognition to a complex response with a clear final purpose, can be considered a code or not, but for a practical man like me it smells of intelligent programming as much as an explicit code.

  58. Good GOD! Are people really questioning the existence of consciousness? It sort of reminds me of a debate involving Michael Shermer and Stephen Meyer. Shermer was confused as to what intelligence was and Meyer jumped on him for it. It was quite hilarious and shows how far away from reality people like Shermer are.

  59. If I have to infer that others are conscious next thing you know i’ll have to infer that I can put one foot in front of the other every time I take a step.

  60. Phaedros:

    That others are conscious is “technically” an inference, in the sense that it is a knowledge which is different form the direct perception of consciousness that each of us has in himself. But I agree that it is such a universal inference, that we have built all our map of reality on it.

    I have no intention to doubt in any way the existence of consciousness in others. I am absolutely sure of it (as, I hope, is anybody else here). But my point is that such a certainty is anyway based on an inference based on analogy. Therefore, inferences based on analogy cannot be such a bad thing after all, and that is certainly good news for ID :)

    Even Wikipedia seems to agree:

    The problem of other minds has traditionally been regarded as an epistemological challenge raised by the skeptic. The challenge may be expressed as follows: given that I can only observe the behaviour of others, how can I know that others have minds? The thought behind the question is that no matter how sophisticated someone’s behaviour is, behaviour on its own is not sufficient to guarantee the presence of mentality. It remains possible, for example, that other people are actually nothing more than meaty automata (or “Philosophical zombies” as the term for this example stands). Perhaps the main argument offered against this possibility in the history of philosophy is the argument from analogy; it can be found in the works of J.S. Mill, A. J. Ayer, and Bertrand Russell. However it must also be noted that the argument from analogy has faced scrutiny from the likes of Malcolm who have issues with the ‘one case’ nature of the argument.

    But the real power of the inference lies in the absolute certainty of our direct perception of consciousness, and of all the related phenomena, in ourselves. That certainty even the most extreme reductionist will never be able to really doubt, if not for a smart intellectual game.

    So, in the end, I have to disagree with Douglas Hofstadter and the likes of him: I am not a strange loop. I am certainly a conscious being: all the rest we can discuss. And strange loops are, indeed, very extreme inferences, and sometimes not very realistic…

  61. @Phaedros

    -”Good GOD! Are people really questioning the existence of consciousness?”

    We call them eliminativist materialists. There’s loads of them roaming the streets. :P

  62. gpuccio-

    I understand what you were saying, but there are some who seem to be unable to understand consciousness as a causal factor. Thanks for pointing how important analogy is, someone had to do it. If we say that all analogies are false, as has been stated on here before, then that would throw out a lot of basic human understanding wouldn’t it?

  63. above:

    -”When we talk about the creation of the genetic code, we’re really talking about the initiation of those biochemical processes”

    Can you please clarify exactly what you mean by that? Are you saying that undirected biochemical processes actually create codes?

    Yes, I said that poorly. What I’m trying to flesh out is what it means to create a code, or for a code to exist. If, billions of years ago, the physical processes of gene expression hadn’t started, would the genetic code exist?

  64. So in essence you’re saying that physics/chemistry facilitate the code but do not in it and of themselves create it? Is that correct?

  65. Nakashima #49

    Now _that_ is circular reasoning!

    No Nak. I try to explain why.

    After all this question about material and non material (abstract) has something to do with our present topic. Let’s consider the elements on the table: (a) matter (or energy); (b) the physical laws that rule matter. In a sense this distinction is similar to the distinction data/program of my post. In fact matter is analogous to data while the physical laws are analogous to instructions. Both matter and data cannot change the overarching decision logics of instructions. To think otherwise would mean that matter can change the laws of physics and so far as I know nobody believe that is reasonable. If matter cannot change the laws of physics to greater reason matter cannot create them.

    Now let’s consider an additional layer of rules, a layer even beyond the layer of the physical laws. This is exactly the case of codes overimposed on matter (to follow the examples I provided in comment #37, the words of a language on things, the genetic code on molecules, the ASCII code on electronic signals, the poker rules on a card deck and so on).

    Let’s sum up the hierarchy of levels: level 0 = matter; level 1 = physical laws; level 2 = additional codes/rules/laws. Whether matter, which is at level 0, cannot create the “simple” physical laws that stay at the hierarchical level 1, how could it create the laws that stay at level 2? It seems to me it is a contradiction that what is at level 0 (matter) doesn’t create level 1 but do create level 2, which has a hierarchical rank higher than level 1.

    This contradiction is unavoidable when one believes that matter and the physical laws can create the multiple layers of abstraction that intelligence is able to impose upon them.

  66. Mr Niwrad,

    Thank you for making a longer explanation of your ideas.

    I agree with you that I don’t know of an instance where matter changes physical law. However this shows the limit of your analogy with programs and data, because it certainly is possible for data to change the program. As I’ve pointed out a few times on this thread, We can interpret the contents of memory as either data or program as we choose and there is no hard line between them.

    As for your proposed hierarchy, you haven’t put forward any reason to think it maps to the real world at all. Does level 2 have the same relation to level 1 that level 1 has to level 0? No. A deck of cards is physically controlled by the laws of physics and chemistry, but the laws of physics and chemistry are not controlled in any sense by the laws of poker.

    Further, you have only asserted what it is that you are trying to prove, by lumping the genetic code in with ASCII and poker. The fact of the matter is that there is peer reviewed research containing experimental data confirming that some amino acids have associations with the RNA triplets of their codons that is much larger than the chance hypothesis.

    No one would argue that a deck of cards can only play poker, and we do know that the rules of poker change, with variations coming and going in popularity. Similarly, we attempt to understand how the genetic code has been formed and changed over time.

  67. Nakashima #66

    Thank you Nak for your devoted interest on this post.

    I have never said that the poker’s rules *control* the laws of physics. I have spoken about over-imposition of abstract layers. The question is about who creates what (or better said, who cannot *create* what). Nobody can deny that the laws of physics (layer 1) cannot create the poker’s rules (layer 2). This is the thing that matters. Because it leads us to understand that the poker’s rules (or in general whatever code) are imposed on the underlying layers by intelligence.

    You deny the standard paradigm inputs-program- outputs. According to this paradigm program is fixed. Inputs and outputs are variable. It is true that outputs can change depending on inputs by mean of the intermediate action of program. But this doesn’t mean that inputs=program as you claim.

    Syntactically the difference between data and program is clear: data don’t contain control structures and decision branches. My reasoning was: since new biological architectures imply new control structures then unguided evolution, which changes data only, cannot create new biological architectures.

    Semantically the difference between data and program is deep. From Wikipedia:

    The main difference is in the level of abstraction being considered. Data is the lowest level of abstraction, information is the next level, and finally, knowledge is the highest level among all three. Data on its own carries no meaning. In order for data to become information, it must be interpreted and take on a meaning.

    The level of information is where programs operate. The level of knowledge is where the final cause and its teleological goal (to guide the creation of new organisms) stay.

    Since you don’t like my reasoning (because, differently from me, you believe in unguided evolution) you have no better chance than to deny the difference between data and program. Nevertheless this difference is syntactically and semantically sound and is part of the necessary hierarchy involved in any artificial and natural design.

  68. Mr Niwrad,

    The level of information is where programs operate.

    Sorry, that is just another assertion. There are ideas about levels of abstraction and programs in computer science, they relate to the “virtual machine” or set of abstractions captured in the verbs of the program.

    What is the text of a program, except data at one stage, and an executible object at another?

  69. Nakashima #68

    What is the text of a program, except data at one stage, and an executable object at another?

    Your distinction, based on the programming stages, is somehow “orthogonal” to mine. The classic flow of compiled programming is: sources writing (what you call “text”), compilation of the executable, execution. You call “data” the sources, but they already contain the control structures, then the sources are just the program. Hence your ascertainment doesn’t refute my distinction between data and program.

    A key point of my post was the distinction between things and decisions about things, when and where. Any construction (biological developments included) necessarily involves decisions about things, when and where. To decide (or program/schedule) entails intelligence, because only intelligence can choose among the different decision branches. Blind evolution can only change (damage) things without choice. Blind evolution doesn’t make decisions and as such can construct nothing.

  70. Nakashima-

    Could you spend your time making a more frivolous argument? To try to argue that there is no difference between data and program is utterly absurd. It is quite obvious to anyone that there is a huge difference between a piece of data and a set of instructions as to what to do with that data. I for one can’t even read anything you write most of the time as it is mostly pointless semantic quibbling.

  71. Phaedros @ 70,
    Nakashima is not making a frivolous argument.

    Processors of the ‘x86′ family can run code from data space.

    This means that DNA may also not have CODE/DATA restrictions.

    Interpreters, (early BASIC, P-CODE, etc.), are programs whose executable code resides entirely in DATA space.

    The point is that we cannot map our idea of processing systems onto biology and expect that we can have a workable analogy.

    Biological processors might be more like a million bit-wide neural net with such a slow clocking rate that trying to model it as a sequential processor, like those that we have designed, would be completely invalid.

  72. niwrad @ 8

    The distinction between data and programs is an aspect of the distinction between instructions and their processor. Both are necessary: instructions without processor do nothing; processor without instructions does nothing.

    So do instructions fall under the category of data or programs? Perhaps Phaedros can tell us, since the difference is obvious to him.

  73. niwrad:

    only intelligence can choose among the different decision branches

    So when a wave function collapses such that a particle goes through slit A instead of slit B, that indicates intelligence?

  74. R0b-

    Instructions would constitute a program. Whether it’s one instruction, although even the simple “Hello World” program has a few, or many. If you want to call data something like, “bits of information” then information is made up of data. It is still, it seems to me, quite obvious that there is a distinction to be made between the two.

  75. then instructions are made up of data**

  76. As Niwrad is saying, simple data doesn’t tell itself what to do.

  77. Mr Phaedros,

    ((lambda (arg) (+ arg 1)) 5)

    The above is Lisp, one of the oldest computer languages. Which parts are program, and which parts are data?

  78. Mr Niwrad,

    Your contention in this thread has been that programs are active, data is passive, and that there is a logical barrier between programs and data. All I’ve been doing is showing that this is a limited understanding of how computers work, and therefore any analogy or further argument based on this fails.

    Now you’ve added intelligence into the soup, and would have me believe that my thermostat is amking intelligent decisions. Sorry, I think you are taking the discussion in an unhelpful direction.

  79. R0b #72

    So when a wave function collapses such that a particle goes through slit A instead of slit B, that indicates intelligence?

    By saying “only intelligence can choose among different decisions” I mean that intelligence, when decides, is not constrained by something, has freedom. The particles in the two-slit experiment obey the laws of quantum mechanics, and then are not properly free. In other words, the particles cannot be said intelligent only because go through slit A instead of slit B.

    In general is not free and cannot decide what is ruled by the physical laws. Differently, intelligence can choose (and design) because is not under such dominion. Intelligence, in its highest meaning, overarches the physical laws.

  80. Phaedros:

    Instructions would constitute a program.

    then instructions are made up of data**

    So are instructions active, like programs, or passive, like data?

  81. Nakashima-

    ((lambda (arg) (+ arg 1)) 5)

    “The above is Lisp, one of the oldest computer languages. Which parts are program, and which parts are data?”

    You tell me. I don’t anything about Lisp. However, if, say, lambda, is a variable and that phrase is some sort of function then whatever number you put in for the variable would be “data” and the function would be an “instruction” used to determine what is done with that data. The entire thing would be a program, if it is indeed some type of function, but I doubt that it would do much without more instructions. Those instructions have to be organized by a mind in order to perform a specific test. The entirety of the instructions would be the program. I mean I’m not sure what is complex about this….?

  82. Phaedros:

    As Niwrad is saying, simple data doesn’t tell itself what to do.

    Of course it doesn’t. It makes no sense for X to tell X anything, since X already knows.

    Most instruction processors look up machine instructions in a table in order to get the corresponding microcode. This is logically equivalent to:
    switch (x)
    {
    case “AAA”:
    // Do Ascii Adjust for Addition
    break;

    case “AAD”:
    // Do Ascii Adjust for Division
    break;


    }

    When the above logic is implemented in hardware, we call x an instruction. If it’s implemented in software, as above, is x an instruction, or is it “simple data”?

  83. Toronto #71

    The point is that we cannot map our idea of processing systems onto biology and expect that we can have a workable analogy.

    The idea I wanted to express was that any construction implies assembly instructions. Whether the construction must be a self-construction it necessarily involves storage memory. This is a general principle and it is useless to note (as you do) that our actual informatics can be different in details respect the biological information processing. Actual informatics implies instructions, biological developments imply instructions and this is what matters here.

    Evolutionists in this thread seem even to think that biological developments need no kind of instructions whatsoever.

  84. Phaedros @ 81,
    Nakashima and ROb are trying to show that data and code, even to the type of processing architectures we humans have come up with, does not have that sharp distinction that Niwrad is trying to claim for his analogy.

    [ROb]: When the above logic is implemented in hardware, we call x an instruction. If it’s implemented in software, as above, is x an instruction, or is it “simple data”?

    What ROb is showing at 82, is that the line gets really foggy if that process happens indirectly, e.g., an Intel x86 family processor “emulating” an Intel x86 family processor running entirely in RAM, (i.e. DATA space).

  85. niwrad:

    The particles in the two-slit experiment obey the laws of quantum mechanics, and then are not properly free. In other words, the particles cannot be said intelligent only because go through slit A instead of slit B.

    What law of QM constrained the particle to go through slit A?

    Your appeals to intelligence and freedom are rife with metaphysical assumptions that I have never understood, so I won’t pursue you down that path.

  86. Niwrad @ 83,
    You are trying to map a human designed processing model onto that used by actual biological processes.

    Please see Nakashima @ 3,

    [Nakashima:] The point is that your neat logical dichotomy between programs and data does not exist in the real world, either in computer science or in biology. Therefore the conclusions you draw from it are invalid.

  87. Nakashima #78

    Your contention in this thread has been that programs are active, data is passive, and that there is a logical barrier between programs and data.

    The logical distinction that you cannot deny is between things and instructions/decisions about things. You can also put on the table declarative languages but the above logical distinction remains. In biology it is not matter of validating theorems, rather of constructing systems. To prove that my argument fails (as you claim) you should show how to program the construction of a system without instructions/decisions about materials in the space/time (and you cannot).

  88. R0b #84

    Your appeals to intelligence and freedom are rife with metaphysical assumptions that I have never understood, so I won’t pursue you down that path.

    Have you never understood if *you* are intelligent and free?

    I won’t pursue you down the path of physics because the topic of this post is about the instructions to construct complex systems, which are a software layer beyond the physical laws.

  89. niwrad @ 87,

    The logical distinction that you cannot deny is between things and instructions/decisions about things.

    What about self-modifying code?

  90. This thread brings back memories of other discussions that I have had with Darwinists on the matter of what is logically possible or not possible. On multiple occasions, I have emphasized the point that nothing can come into existence without a cause, which is another way of saying that something cannot come from nothing. At other times, I have focused on various corollaries with respect to this same law of causality, not the least of which is the fact that a cause cannot give what it does not have to give. If the effect was not first present in the cause in some way, it cannot become manifest. I have further emphasized that this law of causality is derived from the equally compelling law of identity.

    Again, I remind everyone concerned, both ID supporters and the anti-ID critics, that Darwinists do not accept these logical principles or the practical guidance that they offer for arriving at reasonable answers to all the questions that have been put on the table. We have heard every objection imaginable, beginning with claims that the law of causality does not apply to the real world and ending with inquiries about the definition of the “something” that cannot come from nothing. Naturally, if something can come from nothing, then clearly an information code can emerge from physical laws or from matter itself. Is this reasonable? Of course not.

    —As niwrad explains, “From the natural laws can only emerge what is potentially contained in them. Biological and informatics codes are not potentially contained in those laws then there must exist a higher source that imposes these abstract codes upon the material supports.

    —To deny this conclusion is as to claim that a cards deck can create the poker rules by itself.”

    As clear as this point ought to be, Darwinists simply do not believe it because they reject the principle of causation that informs it. For them, there is no reason why biological and informatics codes that are not potentially contained in laws cannot emerge from those laws. For them, nothing is impossible: Rational universes create themselves, life comes from non-life, and mind emerges from matter. What is to prevent it? Indeed, they can find no reason to rule out the prospect that poker rules could emerge from a deck of cards. Granted, they will concede that we “have no evidence” that any such event has ever occurred, but they will not assert, in principle, that such an event could not happen. To do so, would be to acknowledge causality as a law, which would be tantamount to abandoning their materialistic paradigm, which requires selective causality. Thus, when presented with a reasoned argument that draws on the limits of that which is possible in a logical sense, they simply tear down reason’s wall and ask, “why not?” That same thing is happening right here on this thread and there really isn’t much that can be done about it except to call attention to it.

    Press them on the point and they will change their story and insist that the conditions really were present in the cause all along. That shifting strategy doesn’t work, of course, because intelligent agents, who can decide to create or not to create cannot, at the same time, be a law, which can only do what it does.

    The bottom line is that it seldom advances the cause of rational discourse to discuss science with those who dispute the law of causality without at least making the point that the discussion is not founded on a rational basis. Science is, after all, a search for causes. What is the point of probing the possible causes of anything with someone who thinks that events can occur without causes? That would be like trying to solve a murder with someone who believes that not all murders require a murderer. Just about the time you have built your argument to the point at which you can declare with confidence that A must have committed the murder because neither B or anyone else could have done it, your Darwinist friend, if he doesn’t like your conclusion, will likely plead that it may be one of those murders that occurred without a murderer.

  91. StephenB @ 90,

    I believe life had a cause.

    I believe that cause came from within nature.

  92. Mr Niwrad,

    I agree that there is a categorical differeince between the substance of the univese and the laws that govern it (modulo the possibility that a Theory Of Everything could discover that there are no free parameters). My dispute with you is whether there is a valid analogy between these two categories and the ideas of program and data as used in computers. This analogy is the basis for your argument. It is flawed. Program and data are not categorically different. If a program reads or writes the bits at a memory address, they are data. If the same memory address becomes the content of the instruction pointer, those same bits are program. It is as simple as that.

  93. —Toronto: “I believe life had a cause.”

    —”I believe that cause came from within nature.”

    Your soltution begs the question and simply denies causality at the next level. Big Bang cosmology shows that nature is an effect from another cause. It did not or could not bring itself into being. I have never met a Darwinist who does not deny causality at some level.”

    Further, I anticipated your comment, since it is quite common. That is why I wrote this:

    “Press them on the point and they will change their story and insist that the conditions really were present in the cause all along. That shifting strategy doesn’t work, of course, because intelligent agents, who can decide to create or not to create cannot, at the same time, be a law, which can only do what it does.”

  94. Nakashima #92,

    Program and data are not categorically different. If a program reads or writes the bits at a memory address, they are data. If the same memory address becomes the content of the instruction pointer, those same bits are program.

    How can the same bits be data and program at the same time for the same compiler? They cannot be both when elaborated by the same code. For example the same bits string cannot mean “$a=3.141592653589″ (data) and “if ($a==3) {$b=4}” (decision) when codified by the same code. You need two different codes to get two different meanings from the same bits string.

    But this way we are led again to the very question of this post: the creation of codes. Codes are what can interpret data. As such they are unreachable by blind evolution, which can change (damage) data only, but cannot create the interpreter of data, which overarches them.

  95. StephenB @ 93,
    We are actually in almost total agreement yet you refuse to acknowledge it.

    Why can you have an uncaused first cause, (i.e., God), and I can’t?

  96. niwrad @ 94,

    How can the same bits be data and program at the same time for the same compiler?

    Please read all of the following.

    The compiler has nothing to do with it at the system level.

    I can write the scheduler for an OS in assembler meaning there is no compiler.

    I can load and locate assorted standard linked Microsoft “.exe” modules and start them off as seperate tasks without knowing what type of compiler produced them so long as they’re in standard PE format and thus re-locatable by my loader.

    The point is that my loader can map the address spaces of two different tasks such that task A generates data that task B has mapped as code space.

    Task B doesn’t realize that the code it is executing, (Task B code), was actually generated by Task A in what A considers it’s data space, (Task A data).

    Your analogy doesn’t hold up.

  97. Toronto #96

    The analogy between informatics and biology, which you deny while many agree with, is based on the fact that only instructions/decisions can make organization. Your negation of the informatics-biology analogy implies you believe that biological organization can be created without instructions/decisions. A biological development is not a one-step photocopy process. It entails countless steps and at every step needs countless decisions about what, when and where to assembly. It is beyond me how you think to organize a biological development without instructions/decisions. Instead it is clear to me why you (and other evolutionists) deny this analogy: it is a nail in the coffin of Darwinism because informatics is designed cent per cent (and you as a programmer understand it).

    You assembler example doesn’t change one bit of this situation.

  98. Toronto #95

    Why can you have an uncaused first cause, (i.e., God), and I can’t?

    Simply because yours is not a first cause. In fact yourself defined it as “coming from within nature” (#91). As can a thing “within nature” be the cause of nature? Causa sui fallacy…

  99. Mr Niwrad,

    How can the same bits be data and program at the same time for the same compiler?

    I was going to quote for you from the lessons in IBM 360 Assembler Language that I took in college, since my professor was quite proud of how easy and useful it was to write self modifying code in BAL. However, I see that Wikipedia has a page on Self Modifying Code which would be better for you to learn from. You might note that on self modifying code has evolutionary uses, according to that page.

  100. niwrad @ 97,
    You said this:

    How can the same bits be data and program at the same time for the same compiler?

    I answered with this:

    [Toronto @ 96:] The point is that my loader can map the address spaces of two different tasks such that task A generates data that task B has mapped as code space.

    I invite you to pick apart my post @ 96.

    There is no need to appeal to anyone as either my comment at 96 is valid or it’s not.

    If there is something wrong with it, you should be able to show me where I went wrong.

    You have not convinced me, ROb or Nakashima, i.e., people that seem to be familiar with computing concepts, that your analogy is valid.

    If your analogy was actually workable, we three should be the easiest to convince, not the hardest.

    Specifically, it is your analogy, not someone elses, that is flawed.

  101. niwrad @ 98,
    —Toronto: “I believe life had a cause.”

    —Toronto: ”I believe that cause came from within nature.”

    Where do I say nature came from nature?

    I don’t know how I can in any way simplify this.

  102. —Toronto: “We are actually in almost total agreement yet you refuse to acknowledge it.

    —“Why can you have an uncaused first cause, (i.e., God), and I can’t?”

    — “I believe life had a cause.”

    —Toronto: ”I believe that cause came from within nature.”

    —Where do I say nature came from nature?

    To come from “inside nature” is to come “from nature.” We are discussing the cause of nature. Nature cannot bring itself into being, either “from itself” or from “inside itself.”

    Also, nature, since it is material, is always changing. The cause of that change must come from the outside. Imagine a pan full of water changing from cold to hot. The cause of that change must come from the outside, either from the flame on the range top or from some other source of heat. The pan of water cannot warm itself. The change most certainly cannot come from the inside of the pan. Nothing can be the cause of its own change. Further, the first cause of all other changes must, itself, be unchanging. If it could change, that change would have to come from another cause. Also, the first cause must be immaterial, because only non matter is unchangeable. That is one of many reasons why nature cannot qualify as a first cause.

  103. StephenB @ 102,

    Good, we agree life had a cause.

    The only thing left to determine is what that cause was.

    Imagine that an ID proponent puts an envelope on a table followed by one from an Evo advocate, both containing data about their cause.

    1) Are you allowed to treat the envelopes differently?

    2) If the envelopes must be treated identically, then either envelope can contain the identity of the first cause.

    If you insist they can be treated differently, then you have jumped a layer of abstraction which makes any logical debate impossible.

  104. StephenB @ 102,

    We are discussing the cause of nature.

    No we are not discussing the cause of nature, we are discussing the cause of life.

    If you want me to prove the cause of nature instead, you have to prove the cause of God.

  105. Nakashima #99, Toronto #100

    Your reference to advanced programming techniques (as self-modifying code, mixing code and data space, etc.) doesn’t help your cause of defending blind evolution. In fact my thesis was that evolution cannot create instructions. If evolution cannot create instructions to greater reason is unable to create instructions that create/modify instructions (as in self-modifying programming). What cannot create code a fortiori cannot create meta-code.

    So, in your attempt of defending evolution, you worsen your position by putting advanced programming on the table, because advanced programming needs more intelligence than simple programming. In a sense self-modifying programming implies an additional abstraction layer. For intelligence no problem insofar intelligence is, beyond many other things, what can create multiple hierarchical layers of abstraction upon matter. Unfortunately for evolution these same layers become additional impossibility layers.

    To avoid a possible misunderstanding, I add that evolution, beyond damage data, can well damage instructions (and eventually meta-instructions) of course. In this sense I agree with you that evolution “operates” on instructions too.

  106. Toronto #104

    You wrote in #95:

    Why can you have an uncaused first cause, (i.e., God), and I can’t?

    There your desire was to have your own uncaused first cause (different from God).

    Then in #104 you wrote:

    No we are not discussing the cause of nature, we are discussing the cause of life.

    Since the uncaused first cause is obviously also the cause of nature, in #104 you contradict your #95.

  107. Mr Niwrad,

    The Lisp example I gave earlier could have been written before 1960. IBM 360 assembler language is also not an advanced programming technique. But I take that you finally agree that there is no fundamental distinction of a categorical nature between program and data in computers, only a distinction in common practice and pedagogy today. As such the basis for your initial analogy is disproven.

    No one is arguing that evolution creates instructions, in the sense of creating new natural laws. The instruction set of our cosmic computer is fixed. Evolution only exploits quirks of physics and chemistry that have existed from the beginning.

    At another level, insertions, deletions, and mutations certainly do create new sequences of instructions. All forms of recombination create new sequences of instructions. Whether these operations have clear analogs in terms of computers is merely a limitation of the analogy. The field of genetic programming does make the analogy strongly.

  108. niwrad @ 106,
    Your first cause

    …”is”….
    God.

    My first cause

    …..”is”…..
    Nature.

    Since they are;
    “peers at my level of abstraction”,

    how can “I” possibly have created a contradiction?

  109. niwrad @ 105,

    Your reference to advanced programming techniques (as self-modifying code, mixing code and data space, etc.) doesn’t help your cause of defending blind evolution.

    It wasn’t meant to defend blind evolution.
    It was in response to:

    [niwrad:] How can the same bits be data and program at the same time for the same compiler.

    It was meant to show that your analogy, that “human designed computer systems are like biological processes”, is not valid.

    Any conclusions you derive from your analogy will also be invalid.

  110. —Toronto: “My first cause

    …..”is”…..
    Nature”

    I have already provided a number of reasons why nature cannot be a first cause. You simply ignored them. There are many more. At some level, Darwinists always reject causality and, in the process, reason itself.

  111. StephenB @ 110,
    You’ve left out that my Nature and your God have a peer-to-peer relationship when it comes to my definition of a first cause.

    In that case, if my first cause is in any way lacking the required capabilities, then yours is lacking in the same way.

    Please try to understand at what level of abstraction this point is being made at.

    We are talking at the level of “Schrodinger’s Cat” .

    It is as pointless to talk specifics about nature as it is to ask whether the box containg his cat was properly ventilated.

    Despite the fact that a physicist was talking about Quantum Mechanics, there was no physics or math required for the level of abstraction that the experiment occupies.

    My description about God or your description of nature is meaningless for the point that is being debated.

    The actual point is, do we call the first cause God, and say it communicates with us,

    OR…,

    do we call it Nature and say it doesn’t communicate with us.

  112. StephenB,
    Please see my posts at 103 and 104.

  113. Nakashima #107

    The Lisp example I gave earlier could have been written before 1960. IBM 360 assembler language is also not an advanced programming technique.

    Don’t equivocate Nak. Meta-programming is an advanced programming technique not in historical sense, rather in conceptual sense.

    But I take that you finally agree that there is no fundamental distinction of a categorical nature between program and data in computers, only a distinction in common practice and pedagogy today.

    You have misunderstood my friend, I do not agree at all. The distinction between data and program is a particular case of the more general distinction between agent and acted, subject and object, referencer and referenced, observer and observed, etc. Also in self-modifying software there is a code that writes and a code that is written. So, despite that both are instructions the distinction between writer and written holds and is what matters (and is far more than a mere pedagogical issue). The distinction between active and passive exists of course also in informatics because it is a general principle. This principle is what refutes at its root any illusion of self-action in general (and in particular of self-assembly).

    At another level, insertions, deletions, and mutations certainly do create new sequences of instructions. All forms of recombination create new sequences of instructions.

    One of the claims of my post was that such operations, which you cite, have analogous counterparts in informatics and biology (and justify, among other things, the informatics-biology analogy I support). My conclusion (different from yours) was that such operations on pieces of code are unable to create the hierarchical decision logic that uses and references them to construct new systems. In other words to create new junk sequences of instructions is a “horizontal” process, so to speak. To create functioning hierarchical decision logic is a “vertical” job that only intelligence can do.

  114. 114

    Bioluminescence is a phenomenon that is difficult for evolutionists to explain because it occurs across phylums. It can be found in potato plant leaves, foxfire fungi, squids, fireflies, the black dragonfish and sharks (and other places as well). Bioluminescence is a double-edged sword. It brings camouflage, a meal, or a mate in the correct context but would actually be deadly if it occurred in the wrong species (as with a mouse on a dark night). From a programming point-of-view, we can look at specific capabilities withing flora and fauna as being akin to subroutines in computer program. They bear the marks of an Intelligent Programmer because they are task-specific and are sprinkled disparately across phylums. Just as “just-in-time” factory inventory systems show the mark of intelligent planning, so the “just in time” appearance of bioluminescence among specific species shows the work of Divine Intelligence.

  115. Toronto #109

    It wasn’t meant to defend blind evolution. It was in response to: [niwrad:] “How can the same bits be data and program at the same time for the same compiler”. It was meant to show that your analogy, that “human designed computer systems are like biological processes”, is not valid.

    Interesting. Don’t you defend blind evolution?

    The informatics-biology analogy is supported by many Darwinists too. For example Richard Dawkins writes:

    Genetics has become a branch of information technology. Really the genetic code is digital, exactly in the same sense in which computer codes are. It is not a vague analogy, rather a real truth. (A Devil’s Chaplain, chap. 1)

    Unfortunately they don’t accept the unavoidable consequence of this analogy: no information technology can be created by blind evolution.

  116. niwrad @ 115,

    The informatics-biology analogy is supported by many Darwinists too.

    Here is my response to aqeels @ 11.

    [aqeels:] I think the safest analogy we can make is that at the heart of things is information, and that biological entities must process it like a black box.

    [Toronto:]I have no problem with the above statement at all despite being an atheist, (but a humble one).

    What you are promoting is not yet developed enough to be useful as an analogy.

  117. —Toronto: “You’ve left out that my Nature and your God have a peer-to-peer relationship when it comes to my definition of a first cause.”

    I have explained why it is not a “peer-to peer” relationship. You have not addressed those reasons but rather continued to claim that nature can be a first cause, not to mention that fact that you never defined nature in the first place.

    —“It is as pointless to talk specifics about nature as it is to ask whether the box containshis cat was properly ventilated.”

    Perhaps that is because you never bothered to define nature. What you are really talking about, I gather, is the physical universe, but, of course, a well-formed definition leads to clarity, and clarity is the Darwinists greatest enemy.

    —“My description about God or your description of nature is meaningless for the point that is being debated.”

    We do not have to describe God or nature, for that matter, to know what can and cannot qualify as a first cause.

    —“The actual point is, do we call the first cause God, and say it communicates with us,
    OR…,
    do we call it Nature and say it doesn’t communicate with us.”

    No, the point is that you deny the status of the causeless cause by reducing it to nature and then asking nature to assume that role. Nature, because it changes, is material, and began in time, cannot be the first cause, which must be unchanging, non-material, and eternal. Ignoring that point will not make it go away.

  118. …Toronto: “we are not discussing the cause of nature, we are discussing the cause of life.”

    We are discussing both. The two points are inextricably tied together. To say that nature caused life is to say that nature contains the life principle within it, which is to elevate it to first cause status. In order to avoid the clumsy claim that the physical universe is synonymous with the life principle and the laws which govern it, you use the imprecise term, “nature,” which can mean almost anything. Thus you seek to avoiding the precise term, “physical universe,” which would help clarify the futility of your argument.

    –”If you want me to prove the cause of nature instead, you have to prove the cause of God.”

    If God had a cause, God wouldn’t be God. If the first cause had a cause, it would not be a causeless cause. That point is well worth absorbing.

  119. StephenB @ 118,

    you use the imprecise term, “nature,” which can mean almost anything.

    No, between the two of us it means what I told you it means.

    You many times have asked for a clear definition when debating a point.

    I’ve given you mine, but instead of accepting it, you replace it with your definition.

    Since I’ve said that my first cause, we’ll call it “Nature”, is a peer of your first cause “God”, you must use it as I’ve defined it in order to contest any point I make.

    If you change that definition, you are no longer having a debate with me as much as winning a debate with yourself.

  120. 120

    StephenB,

    You have a big fan! 8)

    http://www.youtube.com/watch?v=Doqv-5KSBEE

    Notice he doesn’t provide an actual argument, he just laughs like anyone else who has nothing at all to say. Silly, silly young man.

  121. 121
    CannuckianYankee

    Clive,

    Re: 120

    Careful now, you might be the next to be lampooned on YouTube. :)

    Do you think he is ‘Silly Silly’ of Amazon fame? He tends to do the same thing in his ‘reviews’ of ID books.

  122. What the…He just read the comment… Is he trying to get UD publicity?

  123. —Toronto: “No, between the two of us it means what I told you it means.”

    You have not defined nature.

    —”You many times have asked for a clear definition when debating a point.”

    Yes, that’s true.

    –”I’ve given you mine, but instead of accepting it, you replace it with your definition.”

    No, you have not defined nature. So I can hardly replace something that wasn’t defined with another definition.

    –”Since I’ve said that my first cause, we’ll call it “Nature”, is a peer of your first cause “God”, you must use it as I’ve defined it in order to contest any point I make.”

    First you say that I replaced your definition of nature. Now you are telling me that you appropriated my definition of God as first cause and called it nature. Please decide which is the case.

    What is a peer of God? Is that your definition of nature? Are you now saying that “nature” is the first cause of the physical universe in the same way that God is said to be the cause of the physical universe? Are you saying that nature is an immaterial, unchanging, eternal reality as is God. Or, are you saying that is nature synonymous with the physical universe, which is material, changing, and begain in time?

    Please make a decision on your definition of nature since it is getting quite late in this discussion.

  124. Clive @120: Thanks for the tip and the link. That was fun.

  125. Mr Niwrad,

    To create functioning hierarchical decision logic is a “vertical” job that only intelligence can do.

    It is a nice assertion, but you haven’t said anything to back it up, either in the opening post or since then. Looking at your very conventional programming examples,

    if( condition ){ action }

    taking this as a gene, duplicate it

    if( condition ){ action }
    if( condition ){ action }

    and transpose it

    if( condition ){if( condition ){ action }
    action }

    Voila, hierarchy! We didn’t even do any mutation yet. Boy, this is easy.

    Are you going to argue that this is somehow harder in biology than in programming?

  126. StephenB,
    I have never changed my definition of nature.

    I have pointed out to you in a way that leaves no wiggle room, that my Nature is a peer of your God meaning there is nothing your God can do that my Nature can’t.

    I have tried to be as scientific as possible, that our two terms are at a peer level.

    Every time I told you that my definition makes nature a peer with God, you would come back with a comment that effectively said, “but nature is not a peer of God”.

    So I’ve given you my definition more than once, and you claim now that I’ve changed it.

  127. Nakashima #125

    Your cut&paste example is the iteration and nesting of a void skeleton (or generic template). It doesn’t work for at least three reasons. (1) The structure and nesting is random because generated by unguided processes. (2) Does who grant that the structure is syntactically correct? (3) It lacks the specificity (or specification) about all conditions and actions. This specification constitutes in ID terms one of the necessary components of the system’s CSI.

    The perfect match of #1,2,3 is what makes the structure work. This reasoning holds for computer programming and biological information processing.

    To provide an analogy, it would be as you offer a skeleton of the chapters and paragraphs structure of a book with a random plot and without words. Unfortunately a book with random plot and without words (or with random words) has no value or meaning. Analogously a random program template as yours, without conditions and actions (or with dummy conditions and actions not tuned to obtain the desired functionality), cannot work.

    Besides, your solution is based on the gratuitous assumption that genes contain decisions and are program modules syntactically nestable in hierarchical structures. As I said in my post, this is a concession very generous to Darwinism.

  128. —Toronto: “I have pointed out to you in a way that leaves no wiggle room, that my Nature is a peer of your God meaning there is nothing your God can do that my Nature can’t.”

    Let’s expore that. Does your “nature” have an intellect and a will? Does your “nature” have the capacity to decide whether to create or not create? Is your nature omnipotentent and omniscient? Is your nature an eternal, unchanging, non-material pure spirit? Does your nature integrate into one person the qualities of truth, life, goodness, beauty, and love ? Does your nature have the capacity to love its creatures? Can your “nature” take the form of a human and become a sacrificial lamb?

  129. niwrad @ 127,
    Nakashima should not have to provide a higher level of detail than you did in your initial example in the OP.

    If more detail is required for a discussion at the level you selected, then you should provide that relevant extra detail first.

  130. StephenB @ 128,
    You are presenting to me an identity and a list of qualities for your intelligent designer and I appreciate that openness.

    If my intelligent designer “is” your intelligent designer, why would it be “necessary” for that entity to display the attributes you listed, in order to possess the capabilities required to create life?

  131. —Toronto: “If my intelligent designer “is” your intelligent designer, why would it be “necessary” for that entity to display the attributes you listed, in order to possess the capabilities required to create life?”

    You are evading my questions again as I knew you would. You said that your nature is a “peer” to my designer. I explained to you more than once that such a condition is not possible because the idea of nature as first cause violates the law of causality and I also explained why.

    Yet you continue to claim otherwise. I am, therefore, asking you to back up your claim and acknowledge that your “nature” [your "peer to my God," as you put it], really is a peer–that it contains the same specific qualities and attributes of God that I listed.

    Of course, you will not do that because you do not mean what you say, that is, you already know that nature cannot possess all those qualities in spite of your claims to the contrary.

    These discussions always end the same way. I explain that Darwinists deny causality, a Darwinist takes exception, I prove the point, the Darwinist begins to get it and promptly resorts to misdirection, refusing to answer all relevant questions–even those that can be answered with a yes or no.

  132. —Toronto: …”why would it be “necessary” for that entity to display the attributes you listed, in order to possess the capabilities required to create life?”

    According to the principle of causality, a cause cannot give what it does not have to give.

  133. StephenB @ 131,
    A “peer” to X is something that operates at the same “level” as X not because it appears to be identical to X in operation.

    You can’t say that communication protocol stacks don’t have peer relationships simply because one is implemented over a serial port and the other over Ethernet.

    Any two Congressman are peers despite the fact that they vote in opposition to each other since one is a Democrat and the other a Republican.

    What you have done is to suggest that because I don’t believe the designer exhibits the human-like qualities in your list, then my designer is not a peer of your designer.

    Islam is a peer of Christianity as they are both fairly modern religions.

    If a Muslim doesn’t answer all the questions in your list the way you believe, does that mean that Islam is not to be considered a peer of Christianity?

    We are talking about the power to create life, not issues of worship.

    My designer is a peer of yours and is under no obligation to tell either of us why he created life if he was actually even conscious in the same way we are.

  134. StephenB @ 132,

    According to the principle of causality, a cause cannot give what it does not have to give.

    Your list is simply a set of human-like attributes you have mapped onto your designer.

    You couldn’t claim that Ford is not a peer of Chevrolet simply because the Ford owner won’t answer yes to your question, “But can your car take you to your daughter’s first soccer game?”

    You again demonstrate an inability to stay within the level of abstraction our debate is in at any given time.

  135. Mr Niwrad,

    As I said in my first comment on this thread, you write as if the field of genetic programming did not exist. The issues you raise have been raised by others and dealt with before. Genetic programming could not exist if they had not been dealt with. But it does exist, and hands out cash awards every year for programs that have evolved competencies, not just greater than the programmer, but greater than any human being.

    I would like to continue this discussion, but I am getting married on Friday, so I no longer have much free time. Sorry, and best wishes.

  136. StephenB @ 132,

    According to the principle of causality, a cause cannot give what it does not have to give.

    With your definition of cause, a straw could not break a camel’s back since it does not have the required mass.

    However, if it happens to be the last straw, it would be the cause.

    I think a better definition of a cause would be;

    “A cause triggers a state transition in a system”.

    This would be more useful to engineers, computer programmers, mathematicians, etc.

  137. Toronto #129

    …you should provide that relevant extra detail first.

    No. “Affirmanti incumbit probatio”. The burden of proof is on the shoulders of who affirms. I deny that unguided evolution can create software systems. You affirm it can. You should provide the proof, in concepts and details, of your affirmation.

  138. niwrad @ 136,
    1) You are the one that is affirming computer systems are like biological processes.

    2) You provided a certain level of detail as proof.

    3) Nakashima replied at the same level of detail that you set.

    4) Why should he provide a higher level of detail than you when you are the one with the burden of proof, since it was your affirmation in the first place?

  139. 139

    Toronto,

    You again demonstrate an inability to stay within the level of abstraction our debate is in at any given time.

    I think you might be so abstract that no one can follow you. I don’t think it’s StephenB’s problem of abstraction, I think you might have too much of it, and not enough grounding.

  140. Clive Hayden @ 139,

    I think you might be so abstract that no one can follow you.

    At times I do exactly that.

    When it comes to clarity, I guess I’m a work in progress.

    I’ll try to take a little more time before I hit “Submit Comment”.

  141. Toronto #138

    The discussion with my friend Nak wasn’t about the fitness of the informatics-biology analogy. The proof is the very fact that Nak followed me in this analogy by offering an example of genetic program generated by evolution.

    Our discussion was:

    a) I denied evolution can create programs.

    b) Nak replied that evolution can.

    However to avoid a boring ping-pong of accusations between us, I offer you sort of armistice until Nak is back. Ok?

  142. —Toronto: “:A “peer” to X is something that operates at the same “level” as X not because it appears to be identical to X in operation.”

    Try to sell yourself on your own idea. You claim that nature can “operate on the same level” as the first cause, but you obviously do not think so when I challenge the point.

    —“You can’t say that communication protocol stacks don’t have peer relationships simply because one is implemented over a serial port and the other over Ethernet”

    Irrelevant.

    —“What you have done is to suggest that because I don’t believe the designer exhibits the human-like qualities in your list, then my designer is not a peer of your designer.”

    The qualities I listed are necessary qualities for a “first cause,” not human qualities.

    This dialogue is very humorous indeed. Among other things, I explain why a first cause must be immaterial. You say your “nature” can be a first cause, insist that it is a peer of my first cause, and yet you refuse to answer when I ask you if it your “first cause” is immaterial, which is just one of the many attributes needed. What does that say about your intellectual honesty?

    —“If a Muslim doesn’t answer all the questions in your list the way you believe, does that mean that Islam is not to be considered a peer of Christianity?

    —“We are talking about the power to create life, not issues of worship.”

    Yes, we are talking about the power to create. So, why are you talking about religion? I didn’t introduce the subject. You are chasing your own tail.

    —“My designer is a peer of yours and is under no obligation to tell either of us why he created life if he was actually even conscious in the same way we are.”

    You keep making that claim, but when I call you on it and ask you about your designer’s attributes, you refuse to answer.

    [According to the principle of causality, a cause cannot give what it does not have to give.]

    —“Your list is simply a set of human-like attributes you have mapped onto your designer.”
    What does my list have to do with the legitimacy of the principle? Do you realize how irrelevant your comments are becoming? No, I am sure that you do not.

    —“You couldn’t claim that Ford is not a peer of Chevrolet simply because the Ford owner won’t answer yes to your question, “But can your car take you to your daughter’s first soccer game?”

    If, as a Chevrolet owner, I told the Ford owner that my automobile had four wheels, a crankshaft, and an engine, and if he refused to tell me on request if his automobile had those same qualities, I would not accept his assessment that his car could transport anyone. Further, I would know that his refusal to answer was an indication that he was hiding something for the purpose of avoiding debate, just as I know that your refusal to answer constitutes the same thing.

  143. –Toronto: “You again demonstrate an inability to stay within the level of abstraction our debate is in at any given time.”

    I think it is clear who is confused and who is not. Anyone who understands his own abstractions can answer any relevant question that an inquirer will ask. I have no problem with your questions and have answered all of them. You avoid my questions like the plague. By the way, you are very wise to do so because your ideas could not survive even one round of scrutiny.

  144. —”With your definition of cause, a straw could not break a camel’s back since it does not have the required mass.”

    So you think that all effects are the result of a singular cause do you? By the way, what do you understand my definition of a cause to be?

  145. StephenB @ 144,
    Here’s what I said @ comment 136,

    —”With your definition of cause, a straw could not break a camel’s back since it does not have the required mass.”

    which made you reply with this,

    So you think that all effects are the result of a singular cause do you? [..snip]

    which I don’t think is valid in light of my very next sentence @ comment 136 which said,

    However, if it happens to be the last straw, it would be the cause.

    clearly indicating that all effects are not the result of a single cause.

  146. StephenB @ 142,

    Yes, we are talking about the power to create. So, why are you talking about religion? I didn’t introduce the subject. You are chasing your own tail.

    Here you are @ 128 introducing religion,

    Can your “nature” take the form of a human and become a sacrificial lamb?

    unless I am mistaken and this is not a reference to Jesus.

  147. StephenB @ 142,

    You avoid my questions like the plague.

    Any questions that concern the character or describe the designer in any way, like the following,

    Let’s expore that. Does your “nature” have an intellect and a will? Does your “nature” have the capacity to decide whether to create or not create? Is your nature omnipotentent and omniscient? Is your nature an eternal, unchanging, non-material pure spirit? Does your nature integrate into one person the qualities of truth, life, goodness, beauty, and love ? Does your nature have the capacity to love its creatures? Can your “nature” take the form of a human and become a sacrificial lamb?

    will be avoided like the plague.

    This is in keeping with the spirit of ID as being a scientific endeavour and not one of theology.

  148. StephenB @ 143,

    I think it is clear who is confused and who is not.

    Agreed.

  149. StephenB @ 142,

    What does my list have to do with the legitimacy of the principle?

    Nothing, so why make the list?

  150. StephenB @ 142,

    What does that say about your intellectual honesty?

    Does a question like this narrow the gap between us or widen it?

  151. StephenB @ 142,

    —“You can’t say that communication protocol stacks don’t have peer relationships simply because one is implemented over a serial port and the other over Ethernet”

    Irrelevant.

    It’s not irrelevant since it is an example to you, of a true peer-to-peer relationship.

    The serial port driver and the Ethernet driver work entirely differently despite providing the same service to the upper levels of the protocol stack.

    The only characteristic they share is the use of electricity and yet they are peers of one another, just like our respective designers are.

  152. 152

    Interesting conversation. If niwrad can spare the time, I’d like to ask a few simple questions.

    1) What is meant by the claim that nature is blind?

    2) Is this a scientific claim, or a purely philosophical one? And if scientific, how can it be tested?

    I ask this because I see a problem with your argument, but my criticism comes from exact opposite perspective of most of those arguing with you. You say,

    Modern evolutionary theory proposes several unguided mechanisms in order to explain the alleged global macroevolution of species

    My response: How in the world can we tell if these mechanisms are guided or not? How can we tell whether or not they were implemented knowingly by a designer? (And incidentally, defenders of genetic programming will have to admit that hard proof exists that designers are capable of implementing these things. Otherwise, that field of genetic programming would not exist.)

    My own view is it seems that this question is philosophical rather than scientific – it’s like a character in Sims 3 speculating whether the laws that govern his world are guided or not (How could he tell?). But at the same time, assuming said Sim could instantiate these laws in a program, the only limited evidence he would have available would universally point to ID. Anything else would be baseless speculation, almost by definition.

    I guess my criticism would best be viewed as coming from a competing ID perspective, in the broad sense.

  153. Toronto,
    If I may make a suggestion. StephenB apparently has found you in a hole of your own making. He has handed you both a rope and a shovel in the form of several questions. The rope can be accessed by answering the questions. The shovel is accessed by squirming left and right traveling along a somewhat goofy set of assertions about abstractions. ACCESS the ROPE before you reach the earth’ core! You are perilously close.

    Others,
    I know next to zip about programming, but one thing seems to stand out in the whole data/instruction debate, and I’d say niwrad has the upper hand. Consider the most advanced chess playing programs. I’d bet that they could be written in some of that fancy code where what is data and what is program seems obscure. In fact some programs assign values to positions, and those values could be construed as something that was unprogrammed, or at least unforeseen, more than just data. After all, the program may be considering positions that the prgrammers have not explicitly listed. However the program cannot “give what is doesn’t have to give”. It will never do what humans have done for millenia: change the rules of chess.

    I am unsure if this is helpful to your conversation, perhaps if only to demonstrate what one non-techie layman is getting out of it. On the other hand, I am sure that StephenB is trouncing poor Toronto.

  154. F/N:

    A little bit of grounding, courtesy Merriam-Webster:

    _____________________

    >> Main Entry: 1cause
    Pronunciation: \?ko?z\
    Function: noun
    Etymology: Middle English, from Anglo-French, from Latin causa
    Date: 13th century

    1 a : a reason for an action or condition : motive b : something that brings about an effect or a result c : a person or thing that is the occasion of an action or state; especially : an agent that brings something about d : sufficient reason
    2 a : a ground of legal action b : case
    3 : a matter or question to be decided
    4 a : a principle or movement militantly defended or supported b : a charitable undertaking >>
    _______________________

    Of course, one may provide a diverse meaning, but one has to then explain and warrant it.

    Going a bit further:

    1: Wiki’s admission against interest:

    >> Causality is the relationship between an event (the cause) and a second event (the effect), where the second event is a consequence of the first.[1] . . . . Though cause and effect are typically related to events, candidates include objects, processes, properties, variables, facts, and states of affairs; characterizing the causal relationship can be the subject of much debate.

    According to Sowa (2000),[5] up until the twentieth century, three assumptions described by Max Born in 1949 were dominant in the definition of causality:

    1. “Causality postulates that there are laws by which the occurrence of an entity B of a certain class depends on the occurrence of an entity A of another class, where the word entity means any physical object, phenomenon, situation, or event. A is called the cause, B the effect.
    2. “Antecedence postulates that the cause must be prior to, or at least simultaneous with, the effect.
    3. “Contiguity postulates that cause and effect must be in spatial contact or connected by a chain of intermediate things in contact.” (Born, 1949, as cited in Sowa, 2000)

    However, according to Sowa (2000), “relativity and quantum mechanics have forced physicists to abandon these assumptions as exact statements of what happens at the most fundamental levels, but they remain valid at the level of human experience.”[5] >>

    –> What that proviso means is that for instance the various quantum erasure variants on Young’s double slit experiment are pointing to a domain beyond the commonly observed world in which effects we see in odd temporal orders have their causal roots.

    –> In fact, our observation is after the fact of the whole event.

    2: Catholic Enc:

    >>Cause, as the correlative of effect, is understood as being that which in any way gives existence to, or contributes towards the existence of, any thing; which produces a result; to which the origin of any thing is to be ascribed . . . Although the ideas of cause and of causality are quite obviously among the most familiar that we possess, since they are involved in every exercise of human reasoning, and are presupposed in every form of argument and by every practical action, a very great vagueness attaches to the popular concept of them and a correspondingly great ambiguity is to be found in the use of the terms expressing them . . . .

    As a matter of fact, all mankind by nature attributes to certain phenomena a causative action upon others. This natural attribution of the relationship of cause and effect to phenomena is anterior to all philosophical statement and analysis. Objects of sense are grouped roughly into two classes–those that act and those that are acted upon. No necessarily conscious reflection seems to enter into the judgment that partitions natural things into causes and effects. But when we proceed to ask ourselves precisely what we mean when we say, for example, that A is cause and B effect, that A causes B, or that B is the result of A, we raise the question of causality. Whatever answer we put forward, it will be the statement of our conception of causation. It will be the expression of our judgment as to the actual relationship between A and B involved in the conception of the one as cause and of the other as effect . . . .

    “Consider”, a Scholastic would say, “a substantial change–that is to say, a change in which one substance, made known to the understanding by its qualities, ceases to be what it was in the instant A, and becomes, in the instant B, another substance. In order that such a change should be possible, four things are necessary: namely,

    1. the thing that is changed;
    2. the term, or manner of being, or essence, that is induced in that which is changed;
    3. the active agent that produces the change, or accomplishes the existence of the new term, manner of being, or essence; and
    4. the motive, or reason why this latter acts.

    There is also, though it cannot be reckoned as a cause, the terminus a quo, or the original determinative of the thing changed, which passes out of being with the advent of the newly induced term. These four necessary things, since they produce the final result by a mutual action and interaction, in which they give being to it considered as result, are its causes. They are to be discovered, moreover, wherever and whenever any change takes place, not only in substantial, but also in accidental, changes, or mere changes of qualities.” Consider the two cases, the one of accidental, the other of substantial, change. A cube of wax is moulded by the hand into a sphere. The wax, as permanent substratum of the change of figure, is considered to be the matter, or material cause. The spherical figure supervening upon that of the cubical, is the induced formal cause. The moulder, or fashioner of the sphere, is the efficient cause. The final cause is to be sought for in the intention of the moulder. The substance of the wax remains throughout the entire process of the moulding. It is affected only accidentally by the operation. Consequently the example is one of accidental change, and gives us no more than an accidental formal cause. But in cases of substantial change, such as, for example, the electrolysis of water, the induced formal cause is a substantial one; and, moreover, since the substance of the water does not remain after the change has taken place, the material cause cannot be other than a subject, or permanent substratum, that is neither water nor oxygen and hydrogen taken together. In such a case, it is called primordial, or first matter, and is conceived as being a subject potential to information by any and all formal causes. It is a potentiality, but, as a permanent substratum, or determinable entity, is capable of receiving new substantial determinations in the place of that which actually denominates it. It cannot exist alone, but exists only as informed, or actuated by a formal cause. It is not eternal, but created, or, more properly, concreated with substantial form. >>

    –> What the current debates on cause have to reckon with as background

    GEM of TKI

  155. PS: Wiki lets its neutrality mask slip at the end of its causality article, allowing us a peek into the exact self-referentially absurd attempt to deny causality that SB has highlighted, also revealing astonishing ignorance of the issue of the existence of a continent, evidently finite past duration observed cosmos entailing the existence of a necessary being as its causal ground [debates are on what that necessary being is, not its reality]:

    >>1. If the Universe has caused things to happen, what have caused the cause? Nobody/Nothing. This is logically acceptable.

    2. When did the cause of the first cause take place? The only acceptable answer would be: Never. From this follows that if the cause of the first cause never took place then there can be no cause. Since if the cause was caused in some specific time, then we can no longer think of the universe as eternal. Eternity implies no start and no end.

    The solution to this problem does not lie in God having caused the cause, since then we would need to ask the same question concerning God. Therefore we must understand that the only logical explanation is that cause and effect are “appearances of the senses” which humans as limited beings in time and space experience, but from an eternal point of view nothing really happens, thus the Universe is void of cause and thus of effect. The Universe just IS.”[5 -- i.e Sowa 2000] >>

    1 –> The logic of cause-effect bonds entails that in a world where contingent beings exist and the observed world as a whole is contingent, two classes of being are possible: contingent ones that have a beginning and may go out of existence, and necessary ones that do not have a beginning nor can they have an end.

    2 –> Unsurprisingly, Wiki misrepresents the Cosmological argument. In a simple form, that is more properly:

    1. Some contingent beings exist. (E.g.: us, a tree or a fruit, an artifact, the planets and stars, etc. — anything that might not have existed, i.e. is caused.)

    2. Contingent beings do not exist by themselves – that is in part what “contingent” means – so they require a necessary being as their ultimate cause.

    3. If any contingent being exists, then a necessary being exists.

    4. Thus, there exists a necessary being, the ultimate cause of the existence of the many contingent beings in the cosmos.

    3 –> The logic here is that there are commonly encountered things whose existence points beyond themselves to their causes. In a world of such caused entities, we face the issue of something that underlies and grounds the existence of such a world, i.e by necessity something that is not part of a chain of caused beings.

    4 –> Formerly, this was believed by materialists to be an eternal universe of matter in some form. But the demise of the Steady State universe theory has led tot he loss of scientific grounds, even as the rise of the Big Bang cosmology points to a universe we observe with a strictly finite duration from the singularity, usually estimated at 13.7 BYA.

    5 –> Little bangs have little bangers and big bangs, logically, should have much the same. Unless, we are prepared to assert that things can come into being out of nothing — no time, no space, no energy, no matter, no agency etc — for no reason under no forces acting. (On what evidence?)

    6 –> The patent counter-intuitiveness [or is it outright absurdity?] of and utter want of evidence for this has led to the rise of much speculation about multiverses and oscillating universes [which last have been killed off for now yet again by the discovery of flatness in the observed universe].

    7 –> For instance we hear much about quantum fluctuations or occasionally about oscillating closely aligned parallel branes — which imply an assumed underlying stratum that is in effect eternal.

    8 –> In short, trace the cause-effect chain long enough and ask what must underlie and support it, and we see that there is something that is implicit under the chain. Ultimately, something that is implicitly or explicitly eternal. (And notice, we are not here discussing the first in a claimed infinite successive chain [which courtesy Hilbert's hotel runs into the most patent absurdities itself], but the ground on which the chain rests, as did Aquinas et al, ever so long ago. And, on our observation and reasonable inferences therefrom, the chain of successive events in our observed universe has a terminus 10 – 20 BYA, even on the conventional timelines.)

    9 –> God, if he is, is just such a necessary and eternal being, so the question “who caused God” posed above is absurd on its face.

    10 –> The sufficient explanation for a being with a beginning is its cause; for a necessary being, it is its essential nature AND its necessitation on the logic of cause in not only the successive sense but the sense of grounding a chain of cause-effect bonds.

    11 –> The onward blatant attempt to deny the fact of cause by Sowa, as cited, brings to mind the saying that one may debate explanations but one does not dispute facts:

    1: Get two boxes of ordinary safety matches.

    2: open one and strike a match against the strip, observing the flame: heat + Fuel + air –> Fire, a combustion process.

    3: Sweep a second match in the air next to the strike strip. No flame as no heat.

    4: Try striking the match under water (or in a tank of non-oxidising gas like Nitrogen or Carbon Dioxide): friction is still tere but no flame as air is removed.

    5: Get the second match box out and strike it, getting it to burn a bit. Tilt up so the match tries to burn the already burned part, The flame fgoes out.

    6: So, we see that the fuel, heat and air are each necessary for and are jointly sufficient to initiate and then support a fire, per mechanical necessity. Cause-effect by factual example.

    7: In the wider context we see that someone decides to strike the match under set circumstances, and does so for a purpose, so we see intelligent purposeful cause in action in a simple scientific context. But in principle an accident could make the strike happen.

    8: we also observe that effects are initiated and sustained by causes, that some are necessary, that others are sufficient, that accidents may act and that agents may act. (The existence of a crime of arson shows that we can credibly distinguish the two.)

    ___________________

    So, again, if we are in a meeting hall and a BANG happens, every head will turn to see what caused the little bang [especially here in M'rat]. It is credible that we have had a big bang, at least on scientific evidence and reasonings. If little bangs have causes, why then is it argued — apart form metaphysical biases and preferences — that big bangs do not have causes? –G

  156. Nullasalus #152

    Good questions.

    What is meant by the claim that nature is blind?

    Nature (the universal existence) has not in itself its own principle. It depends totally on the Being. Being has intelligence and knowledge. Therefore to be “blind” applied to nature is a metaphor for not to be intelligent and not to know per se (eyesight is a traditional symbol of knowledge and the ability to know).

    Is this a scientific claim, or a purely philosophical one? And if scientific, how can it be tested?

    It is a metaphysical claim. As such it can be known directly by mean of universal intellect. Reason, which is a by-product of intellect and an individual indirect faculty of knowledge, when is honest, can only agree at its level. Science, which is based on reason (and experiments), when is unbiased, can agree too. In this case we can say that such metaphysical claim becomes also a scientific claim. If for scientific test we mean all processes where reason (and eventually experiments) are involved to investigate nature, then in a sense the claim is scientifically testable (admitted that reason doesn’t turn against intellect).

    How in the world can we tell if these [evolutionary] mechanisms are guided or not? How can we tell whether or not they were implemented knowingly by a designer?

    If we were sure that biological macroevolution from a common ancestor happened (and we are not), we would be sure that its evolutionary mechanisms were guided and implemented knowingly by a designer. In fact if life implies information technology (IT) and advanced engineering (AE) to greater reason macroevolution is an IT&AE problem raised to the n-th power. ID theory proves that just IT&AE implies CSI, then intelligence. If macroevolution happened it was certainly driven by intelligence because an IT&AE^n problem is far beyond chance and necessity.

  157. Tim @ 153,

    I know next to zip about programming, but one thing seems to stand out in the whole data/instruction debate, and I’d say niwrad has the upper hand. Consider the most advanced chess playing programs. I’d bet that they could be written in some of that fancy code where what is data and what is program seems obscure.

    Yes, that line between DATA and CODE is not as fixed as niwrad wants it to be.

    While you say you know zip about programming, your intuition in this case, which is not too far off the mark, is pulling you to my side of the debate.

  158. kairosfocus @ 155,

    If little bangs have causes, why then is it argued — apart form metaphysical biases and preferences — that big bangs do not have causes? –G

    I agree with you on this statement.

    The only difference between StephenB’s and my position, is that he believes the first cause must have all the qualities and motives he ascribes to his vision of God, while I believe that in this debate, the only thing of concern is the designer’s ability to create life.

    Whether the designer is still around and has been joined by other beings, or has simply disappeared, or is content to just observe, is besides the point.

    Whatever caused the universe, clearly “did” cause the universe, but whatever form it presents to us now, if any, is irrelevant to that first act.

  159. 159

    niwrad,

    I have nothing to add, but I thank you very much for your reply, which I’ll be mulling over.

  160. I should take time out to explain to the poor readers who are trying to follow Toronto’s non-reasoning overtures what exactly he has been doing.

    Recall that I presented several attributes that a first cause must possess. I didn’t go through the proof for each one because I knew that Toronto would simply ignore them. So, I explained only one attribute and why it is necessary. The physical universe, that is, all of empirical reality, is changing. Anything that changes must be changed by something else. Nothing can cause its own change. Thus, a pan of water cannot change itself from cold to hot. The change must come from the outside, meaning that the heat from the range top causes the water to change from cold to hot. Anything that changes is changed by something else. That chain cannot go on indefinitely. Hence, we come to the causeless cause, which is unchanging and unchangeable. If it could change, that would mean that something else was causing the change and it would not be a first cause.

    Among some of the other attributes required by first cause, we could list the following:

    The first cause must have 1) An intellect and a will, 2) It must have the capacity to create or not create, 3) It must be eternal, 4) and it must be immaterial.

    Toronto did not, at first, contest the point that these attributes are necessary for a first cause. Quite the contrary, @106 he claims that because his “nature” is a “peer” of “God,” nature can do anything that God can do.

    —As he puts it, “Why can you have an uncaused first cause, (i.e., God), and I can’t?”

    Notice, by the way, that he introduces religion into the discussion for the first time. I said nothing about God. More important, he introduces the concept of a “peer to peer” relationship.

    —“In that case, if my first cause is in any way lacking the required capabilities, then yours is lacking in the same way.

    Got that? If his first cause lacks ANYTHING, then my first cause must also lack those qualities.

    And again, he asserts:

    —“Since I’ve said that my first cause, we’ll call it “Nature”, is a peer of your first cause “God”, you must use it as I’ve defined it in order to contest any point I make.

    It is on that basis that I asked Toronto to confirm that his “nature,” as he defined it, contained all the qualities, capacities, and attributes of a first cause. It his nature does not have the same attributes as my “first cause,” it doesn’t qualify as a peer. That should be obvious.

    Remember, that Toronto called the first cause God and said “nature,” his first cause, can do “anything that God can do.” So, I asked the obvious questions: “Does your “nature” have all the attributes that would enable it to do anything that God can do. Does it have an intellect and a will? Does it have the capacity to decide whether to create or not create? Is your nature omnipotentent and omniscient? Is your nature an eternal, unchanging, non-material pure spirit? Does your nature integrate into one person the qualities of truth, life, goodness, beauty, and love ? Does your nature have the capacity to love its creatures? Can your “nature” take the form of a human and become a sacrificial lamb?

    No answer.

    Not only does he completely ignore the questions, indicating that he does not mean what he says, he begins his campaign to change the definition of “peer” from equal to not equal.

    —“A “peer” to X is something that operates at the same “level” as X not because it appears to be identical to X in operation.”

    Notice how the term “peer” is getting downgraded to something that “operates at the same level,” whatever that means.

    —“My designer is a peer of yours and is under no obligation to tell either of us why he created life if he was actually even conscious in the same way we are.”

    So now his “designer,” can do all the things my first cause can do even though it has none of the attributes necessary to do them. Incredible. This is classic Darwinism. [The denial of causality]

    He confuses the issue further by saying,

    —- “Your list is simply a set of human-like attributes you have mapped onto your designer.”

    Notice now that Toronto disputes the fact that attributes are necessary at all. At the same time, he will not address my substantive arguments indicating that they are, indeed, necessary.

    —“What you have done is to suggest that because I don’t believe the designer exhibits the human-like qualities in your list, then my designer is not a peer of your designer.”

    The morphing argument continues: Now, suddenly, the attributes for a first cause, which he first argued can be found in “nature” by calling nature a “peer” of the first cause, and then later claimed are not present in nature at all, have finally been downgraded to the level of “human-like” qualities.

    Then, we get this:

    —“The serial port driver and the Ethernet driver work entirely differently despite providing the same service to the upper levels of the protocol stack.”

    How nice. What does that have to do with the argument about FIRST causes?

    —“The only characteristic they share is the use of electricity and yet they are peers of one another, just like our respective designers are.”

    A “peer” first cause cannot have the luxury of not having first cause characteristics.

    Then, we get this:

    —“If a Muslim doesn’t answer all the questions in your list the way you believe, does that mean that Islam is not to be considered a peer of Christianity?

    —“We are talking about the power to create life, not issues of worship.”

    So, he injects religion into the discussion yet a second time, while reminding me that we should not be talking about “worship.” Remarkable.

    —”With your definition of cause, a straw could not break a camel’s back since it does not have the required mass.”

    I asked him what he understands my definition of a cause to be. No answer.

    Then we get this:

    —“Here you are @ 128 introducing religion,

    —“Can your “nature” take the form of a human and become a sacrificial lamb?

    —“unless I am mistaken and this is not a reference to Jesus.”

    Obviously, this is an untrue statement. He introduced the subject of religion on two occasions [as I pointed out earlier] and insisted that his nature can do anything the GOD can do. I had said nothing about God. So, naturally I am following up on yet another of his claims for which he will not hold himself accountable, namely that nature can do anything that GOD can do [his words, not mine].

    Then, as a special bonus, after having introduced religion into the discussion, he justifies his unwillingness to answer all my question, saying that he will not answer religious questions. Unbelievable.

    Finally, he sums up his long record of non-responses with this incredible statement:

    — “This is in keeping with the spirit of ID as being a scientific endeavour and not one of theology.”

    Would it be worthwhile to explain to Toronto the difference between metaphysics and physics? Do you suppose he could grasp it? Would it help if I pointed out again that all the religious references came from him and all my answers that contained a religious element were responses to his religious questions? Probably not.

    Notice also that anyone with even a modicum of intellectual curiosity would have avoided all these obfuscations and tried to confront my one specified argument. Since everything in nature is changing, and since nothing can change itself, the first cause must be unchangeable. Then, I could have gone on to explain WHY the first cause must also be impersonal, eternal, etc. But, of course, such a discussion would be possible only with someone who can present and follow reasoned arguments.

  161. StephenB @ 160,

    No answer.

    That is an unfair thing to say to those of us on the moderation queue, who can’t respond in a timely fashion.

    I have to carefully choose what I invest time in replying to since there is no guarantee that anyone will see it.

    I also see that you have already made your closing statements to the jury and would thus like to thank you for the discussion.

  162. So according to Toronto nature is God’s peer…

    What exactly is the difference between paganism and atheism again? It’s getting increasingly difficult to differentiate between the two lately…

  163. @Stephen

    Forth line from the end… Did you mean to say ‘personal’ rather than ‘impersonal’?

  164. above @160, yes, of course. Thank you for catching that.

  165. Amazing.

  166. I think what may be interesting, would be an open blog where I would allow both sides to present OP’s and to battle it out head to head.

    If enough of you are interested in contributing, let me know and I’ll set one up.

  167. StephenB @ 160,

    But, of course, such a discussion would be possible only with someone who can present and follow reasoned arguments.

    It is June 1 @ 12:48 PM and I would love to present an argument.

Leave a Reply