On machine-information metaphors

by Neil Rickert

A recent post by David Tyler at Uncommon Descent has drawn my attention to a recent paper by Pigliucci and Boudry, “Why Machine-Information Metaphors are Bad for Science and Science Education“.  I am posting this as a form of participation in the discussion of such use of metaphors.

We often use metaphors as part of our communication.  And they often work very well as communication devices.  I don’t have a problem with that.  But there is a problem when people take them as truth, rather than as mere metaphors.  Pigliucci and Boudry are particularly concerned with the use of metaphors in biology, such as the idea of DNA as a blueprint for the organism and the idea of the cell as a factory.

The machine metaphor has always seemed misleading to me.  My own experience in the world tells me that we humans are not machines.  Indeed, I see biological creatures as very different from designed things.  Perhaps the worst of the machine metaphors is the one that sees the brain as a computer.  That has always seemed doubtful to me, so I find it no surprise that AI (artificial intelligence) has made so little progress in the 60 years since Turing’s famous paper.  That we make heavy use of metaphors should already be evidence against such mechanistic theories of brain function.

In Linguistics, the mechanistic thinking from Chomsky has dominated the field for some time.  Yet it is surely mistaken.  The Wittgenstein view of language as a form of life seems to present a rather more realistic view of natural language.  In philosophy we see what seems to be an excessive reliance of logic with its mechanistic rules of inference.  Typically, within epistemology, knowledge is defined as justified true belief (or something similar), and that seems rather too mechanistic.  And then the all too frequent arguments that deny free will are based on an overly mechanistic view of the world.

David Tyler does not understand Hume’s way of thinking:

It has always surprised me that David Hume’s arguments are considered weighty. The preceding generations of scholars did have a rationale for thinking that there is a relationship between the Creator’s design and human design.

And yet surely Hume is simply pointing out what should be obvious to the observer of nature.  It should be obvious that the world is not the product of design.  For sure, I can look at Kepler’s laws of planetary motion, and see intelligent design in those laws.  But the intelligent designer was none other than Johannes Kepler himself.  If Kepler’s laws were part of the design of a mechanistic world, we would expect them to be followed exactly.  Yet they are only an approximation.  It took great skill and knowledge for Kepler to come up with a mathematically simple approximation that worked so well.  But it is not evidence of an intelligently designed world.

So, yes, I do see the overuse of such metaphors as bad for science, bad for philosophy, bad for theology.

About these ads

88 Responses to “On machine-information metaphors”

  1. Agreed.

    For instance, claiming that DNA is a like a computer code for an organism makes it possible to claim that H2O is the code for water, which is not the case.

    H2O ..is.. water.

    Like

    • While I am commenting on my own blog, I might as well respond to this.

      You need to be careful when discussing whether “H2O is the code for water”. For the character string “H2O” is code for water molecules, while the string of atoms is not. There’s a bit of ambiguity in how you expressed that.

      Apart from that ambiguity, you sum up my view of DNA rather well. I don’t have a problem with mathematically modeling DNA as information, if such modeling turns out to be useful. But I do see it as a problem when people seem to assert that DNA is part of a metaphysical category of information.

      Like

  2. There was a detailed response to this post in a comment by gpuccio over at Uncommon Descent.

    I thank gpuccio for the comment, and I may respond later.

    Like

    • This is intended as a response to the comments that gpuccio made at UD.

      For instance, I would never say that “DNA is the blueprint for the organism”, but I would definitely say that it contains specific information which the organism certainly needs.

      I see DNA as more like the key that unlocks my door. It contains a physical configuration that is used by the organism (or, in the case of the key, used by the lock mechanism.

      Again, if you mean that living things are more than designed objects, I absolutely agree. But if you mean that, in living beings, there are not parts which are obviously designed and which work as designed machines, then I have to disagree.

      When I read the ID literature, or the posts at UD, what I often see is a description of something the writer finds awe inspiring, followed by an assertion of design. If I see something that is awe inspiring, then it is appropriate to say that it is awe inspiring. To then claim design is reach a conclusion based on additional presuppositions. Science has learned, from past experience, that calling on additional presuppositions leads to serious mistakes.

      Well, the brain “is” in part a computer. It certainly has great computing power.

      I am very doubtful of that. And I do spend time studying the question.

      It’s apparent that my comments on Kepler’s laws have confused people. Those laws describe planets as following elliptic paths. If we had been using Ptolemaic assumptions (geocentrism), then it would have been completely obvious to everyone that Kepler was wrong, and the the planets were actually moving in cycles and epicycles. What this shows, is that the regularities we see are not a raw property of the universe. They depend very much on the standards we choose to follow in how we describe the universe. Copernicus showed that we can have more regularities in our descriptions if we base them on heliocentric assumptions. Part of what scientists do, is invent ways of describing things that had previiously never been described. And they have a tendency to choose ways of describing that lead to lots of regularities in the descriptions. So at least part of the regularities are the result of the intelligent work of the scientists in coming up with those ways of describing things.

      The correspondence between outer world and inner mathematical laws is certainly a cognitive problem, which cannot be easily dismissed. It has nothing to do with metaphors. It is just an unexpected fact, the basis for true science and good philosophy.

      Many people see it that way. I disagree. I setup this blog site as a place to discuss my alternative views. You might want to look at some of the other posts here, particularly those in the epistemology category. And the name “heretical philosopher” reflects the fact that I am disagreeing with the conventional wisdom.

      Like

  3. @Toronto

    Its nothing like it. H2O the molecule is just water. DNA is like a hard drive. Just as a hard drive is capable of storing data and instructions so can a hard drive. Water can’t do this at all. The language stored in DNA is used by all life forms in the same way i.e. the language – data and instructions – are interpreted in the same way by all life forms. Just as a computer program is a complicated layered set of instructions and data, so is the program (life form) contained in the DNA

    Like

    • No, DNA is not at all like a hard drive.

      Many of these online debates masquerade as disagreements about facts. But they are really disagreements about meanings.

      The word “information” has a technical meaning in Shannon’s mathematical theory of information, but that meaning really only applies as part of a mathematical model.

      “Information” also has a meaning in everyday common language, from which Shannon borrowed the term. But, in everyday common language, it does not have just one meaning. It has many different meanings. We run into trouble when we conflate these different meanings.

      With the technical meaning from the mathematical theory, it is best (in my opinion) to treat information as an abstract entity. Looked at that way, the hard drive has a physical representation of our information, but it doesn’t have information in the sense of an abstract entity.

      I tend to think of DNA as more like the cogs on a gear wheel, or like the shape of the key I use to unlock the door. They do what they do by means of their physical causal properties, and adding the term “information” does not tell us any more.

      Like

  4. nwrickert:
    You’re right. I meant the atomic structure of the molecule.

    gpuccio: “b) But what is a machine? It’s simple. A machine is an object which has been structured by a conscious intelligent designer to perform a function, to achieve a purpose.

    //…SNIP…//

    DNA is a machine. Enzymes are machines. Ribosomes are machines. They serve purposes, well defined purposes. And they are designed.”

    gpuccio has gone and done that actual thing that mis-using analogies leads to.

    By terming these structures machines, as humans use the word when talking about things we have designed, you have now made the leap to claiming these structures must have therefore been designed. It was your misuse of the “machine analogy” that led to this.

    Like

  5. I meant “gpuccio” when I used the term “you” and “your” in my closing paragraph.

    This time I spotted my ambiguity first!

    Like

  6. Andrew:”DNA is like a hard drive.”

    DNA is not like a hard drive at all. A hard drive can store any pattern of bits I desire anywhere I want to put them, in other words, a hard drive is a random storage device, while DNA isn’t.

    If the drive is used as a non-bootable device, such as a “D:” dive is typically used in a Windows based system, you could probably change any or every bit and the system booting off “C:” will still boot up and run.

    You can’t do that with DNA.

    Again, a misuse of an analogy.

    Like

  7. gpuccio is one of the less willfully ignorant denizens of the UD cesspool (not a high bar to clear) so I hope he has the intellectual courage to respond here rather than at UD where discussion is heavily censored by Clive Hayden and the other cowardly moderators.

    Like

  8. Neil sez

    It should be obvious that the world is not the product of design.

    Nice bald declaration- too bad your position isn’t science as it cannot be tested.

    IOW Neil there it is obvious that the world isn’t a result of an accumulation of accidents.

    DNA as a hard-drive? It is definitely an information storage device.

    But anyway you guys could get rid of the use of metaphors if you actually step up and start supporting the claims for your position!

    Imagine that instead of constantly bitching you actually find some positive evidemnce to support yout nonsense.

    Then maya chimes in- well maya you are just another intellectual coward of an evolutionist who hangs out in the septic tank that is atbc.

    Like

    • Maya, myself and many others are simply not allowed to reply to ID’ers on the UD site without being censored or outright banned.

      How can we be the cowards if Clive is afraid to let us have an open dialogue on your site?

      Surely your position is strong enough to be supported on UD without silencing the opposition.

      I feel “Expelled”.

      Like

      • Toronto:

        Maya, myself and many others are simply not allowed to reply to ID’ers on the UD site without being censored or outright banned.

        Most likely because you don’t have an argument and can just spew nonsense.

        Toronto:

        How can we be the cowards if Clive is afraid to let us have an open dialogue on your site?

        You don’t want an open dialog- You can’t afford an open dialog because that would mean you actually have to produce positive evidence for your position as opposed to just spewing nonsense about ID.

        Toronto:

        Surely your position is strong enough to be supported on UD without silencing the opposition.

        Opposition would mean that you actually have some positive evidence for your position.

        You don’t. All you can do is rail against ID with your misconceptions of science and reality.

        Like

    • IOW Neil there it is obvious that the world isn’t a result of an accumulation of accidents.

      “Design” and “an accumulation of accidents” are not the only possibilities.

      Like

  9. Toronto:

    A hard drive can store any pattern of bits I desire anywhere I want to put them, in other words, a hard drive is a random storage device, while DNA isn’t.

    My prediction is that we find out that DNA is an information storage device similar to a hard drive. And the genes are carriers of information just as data packets are carriers of information in data communication.

    The DNA does both- act as a storage device, it can also access that info and then act as a liason- ie a data transmitter.

    I also predict that the alleged metaphors turn out not be metaphors at all. Rather they are real descriptors of actual machines and codes.

    But I understand why you guys don’t like them- your position doesn’t have any.

    And I am sure that bothers the lot of you.

    Like

    • If DNA is an information storage device, try storing some information on it and see what happens.

      What bothers me is that you insist following generations should fear the real world as much as you do.

      Like

      • Toronto:

        If DNA is an information storage device, try storing some information on it and see what happens.

        There is information stored on it.

        Also if I gave you a blank computer disc- sans computer and interface- could you store information on it?

        Toronto:

        What bothers me is that you insist following generations should fear the real world as much as you do.

        LoL!

        Unfortunately for you I don’t fear the real world.

        The real wprld screams of design…

        Like

  10. Neil:

    “Design” and “an accumulation of accidents” are not the only possibilities.

    Well Neil what does your position say?

    Like

  11. But anyway you guys could get rid of the use of metaphors if you actually step up and start supporting the claims for your position!

    Imagine that instead of constantly bitching you actually find some positive evidemnce to support yout nonsense.

    I take it that ain’t going to happen any time soon…

    Like

    • That’s because Clive is not willing to let it happen any time soon.

      I’m glad to see you have the courage to engage that he seems to lack.

      Like

      • Toronto:

        That’s because Clive is not willing to let it happen any time soon.

        So Clive is preventing you from producing positive evidence for your position?

        Spoken like a true intellectual coward.

        Toronto:

        I’m glad to see you have the courage to engage that he seems to lack.

        Stop talking of courage as it appears you are spineless.

        Now how about some positive evidence for your position?

        Like

  12. Joe G:

    Toronto:

    If DNA is an information storage device, try storing some information on it and see what happens.

    There is information stored on it.

    Also if I gave you a blank computer disc- sans computer and interface- could you store information on it?”

    Which is the point of this post, that the misuse of metaphors and analogies leads to bad conclusions.

    1) Hard discs are blank when they are ready to use, DNA is not blank.

    2) Hard discs have an interface to store and retrieve data, DNA doesn’t.

    3) Hard discs are re-writable, DNA isn’t.

    4) Hard discs allow any sequence of data to appear anywhere on the device while DNA won’t allow that.

    Hard discs are not like DNA at all.

    Like

    • Toronto:

      1) Hard discs are blank when they are ready to use, DNA is not blank.

      It is before it is loaded with the information required to run the cell/ organism.

      Toronto:

      2) Hard discs have an interface to store and retrieve data, DNA doesn’t.

      Sure it does- we just can’t figure out how we can tap into it.

      Toronto:

      3) Hard discs are re-writable, DNA isn’t.

      DNA gets re-written every time it gets copied.

      Toronto:

      4) Hard discs allow any sequence of data to appear anywhere on the device while DNA won’t allow that.

      How do you know?

      Wow it seems that bald assertions are all you can muster.

      Like

  13. Joe G:

    Sure it does- we just can’t figure out how we can tap into it.

    That is a perfect example of a bald assertion.

    Like

  14. Maya:

    I just read your comment here.

    Believe it or not, I would be happy to spend time writing here too, but I really have not those resources. I have just spent hours answering Neil’s last comments at UD. I do prefer writing at UD. That is my place, and I want to share the discussions with the people there. I don’t think cowardice has anything to do with that. I have written on Mark Frank’s blog once, because that was a nioce place. This is probably a nice place too, but I really have not the time. I would never wite on PT, or similar places, anyway. My choice. Call it cowardice, if you like.

    I am sorry if some of you cannot write on UD. I have nothing to do with moderation, although I believe that some moderationis anyway needed. I appreciate those who come to discuss. A little less those who come just to disturb.

    Anyway, I have had many long and profitable discussions at UD, even with those who utterly disagree with me. I like those confrontations. So, I hope that you, and the others, may come there, if you want, and present your arguments.

    Like

    • “So, I hope that you, and the others, may come there, if you want, and present your arguments.”

      But we did want to come and talk to you, it was Clive who decided we couldn’t.

      You never saw the questions that were censored.

      Do you think you could have answered those questions?

      What would your reply have been to those unseen comments?

      Did you have a good response that we will never see?

      Like

  15. Joe G,

    The evidence for modern evolutionary theory is overwhelming, by any measure. There are three possibilities for your claims otherwise:

    1) You are unaware of the evidence. Ignorance is curable.
    2) You are incapable of understanding the science. Lack of intelligence is regrettable, but not an argument against that which you don’t understand.
    3) You reject the evidence because of your faith. Your irrational delusions are also not an argument against the science.

    Finally, you keep using the term “intellectual coward.” I do not think that means what you think it means. I’m happy to debate you or any other ID supporter in any open forum. UD is not such a venue. You should ask yourself why Clive is so scared of allowing free and open discussion (aside from the fact that he’s a girlyman bitch).

    Like

  16. Toronto:

    Just a brief note about your comment reporting some of my statements. In those phrases I was only giving some conclusions, not detailing the arguments which lead to those conclusions. I obviously don’t think that design is proven by my definitions. The design inference has its procedure, which I have tried to detail in my many posts. In no way it derives from my definition of “machine”, or from any other “metaphor”. It is an inference by analogy, explicitly, as I have written a lot of times.

    Like

    • gpuccio:

      “b) But what is a machine? It’s simple. A machine is an object which has been structured by a conscious intelligent designer to perform a function, to achieve a purpose.”

      “It is an inference by analogy, explicitly, as I have written a lot of times.”

      That is the point of this post, that is it possible that metaphors and analogies, can be used in a manner that hinders instead of help scientific research.

      What I see you and the ID community saying is something like this:

      SINCE A ..appears to be like something in set X,
      THEN A ..is a member of set X

      and that I believe is a misuse of a metaphor/analogy.

      It’s one thing to use a metaphor to allow you a different viewpoint of a problem, quite another to use it to justify a conclusion.

      A tool need not be designed, as a piece of shale could be used as a knife, a rock could be used get food out of it’s shell, and half a coconut shell could be used to carry water.

      A fallen log across a second fallen log is a machine, a lever.

      Please note that in all cases, these devices are natural, but are selected by the environment, which in this case is the tool-user, us.

      Like

      • That is the point of this post, that is it possible that metaphors and analogies, can be used in a manner that hinders instead of help scientific research.

        This is an important point. Scientists know, from past experience, that relying on analogies can lead to false science.

        There is a role for analogies. They are useful in description and in introducing new concepts. Einstein’s famous thought experiments were of that form. But one can never use analogy as the basis for inference.

        Like

  17. Toronto:

    You say:

    “A tool need not be designed, as a piece of shale could be used as a knife, a rock could be used get food out of it’s shell, and half a coconut shell could be used to carry water.

    A fallen log across a second fallen log is a machine, a lever.

    Please note that in all cases, these devices are natural, but are selected by the environment, which in this case is the tool-user, us.”

    That is perfectly correct. That’s why ID theory requires functional complexity to detect design.

    I will try to clarify. For the sake of discussion, I ask you to accept my operative definition of “machine” as something which has been designed by a conscious intelligent being for a purpose, to effect a function. We will then call “pseudo-machines” those systems which can effect a function, but which have not been designed. Your “natural devices”. I am just giving names, not implying anything else.

    Now, let’s say that our point is to be able to distinguish machines from pseudo-machines. IOWs, to affirm a design origin for something which appears to be a machine, because it can perform a function. In ID language, we have objects which are functionally specified, but we don’t know if they are designed.

    If we have direct or indirect evidence of the involvement of a conscious designer, obviously, we need no inference. We just know. In the same way, if we have evidence that the object was formed in a system where no designer was involved, or simply needed, in the same way we need no special inference.

    But in many cases we just don’t know. We have the object, and we try to infer, form the object, if it was designed.

    Now, it is a groos misrepresentation of ID to say (as it could seem even in your post) that just because an object is functionally specified it must have been designed. Nobody in ID has ever made such a silly statement.

    Your “pseudo-machines” are certainly functionally specified (we can describe an explicit function for them), but as you say they were not designed.

    Here is where functional complexity becomes part of the procedure. We can infer design for a machine only if it is complex enough, so complex that no “natural” system could have created it.

    This is the procedure: functional specification + functional complexity = design inference.

    And it is still an inference, not a mathematical or logical demonstartion. An inference by analogy. Not by metaphor. By analogy.

    It’s the same process by which we infer consciousness in other humans. An inference by analogy. We know we are conscious. We observe others who are similar to us and who have a similar behaviour, and we infer that they are conscious too. Analogy. Not metaphor.

    ID is an inference. It requires both functional specification and complexity. It has no false positives, but a lot of false negatives.

    A lot of simple things are functionally specified. Some are designed, some are not. We can never say, unless we have evidence of the designer, because they are simple.

    But no non designed thing is both functionally specified and complex.

    Like

    • Now, let’s say that our point is to be able to distinguish machines from pseudo-machines.

      That’s one of the points of disagreement. Why would we even care to distinguish between machines (as designed systems) and pseudo-machines (as undesigned natural systems)?

      To me, what matters is what it does, and how it does it. Whether or not it was designed is of at most very minor interest.

      But no non designed thing is both functionally specified and complex.

      And that’s the other point of disagreement. You want to say that complex implies design. But we often find that an elegant simplicity in products of human design.

      But it’s worse. It isn’t clear that “complex” means anything. The Ptolemaic universe, with planets moving in cycles and epicycles, was far more complex than the Copernican/Keplerian universe with planets moving in ellipses. But they are both the same universe.

      Like

  18. “For the sake of discussion, I ask you to accept my operative definition of “machine” as something which has been designed by a conscious intelligent being for a purpose, to effect a function.”

    I can accept that, but then things like DNA do NOT qualify as machines since I have no evidence that they have been designed by a conscious intelligent being.

    That is precisely the debate between our sides.

    The EVO position is that DNA and life processes are NOT specified.

    We humans specify functionality that we implement with our designs, but in our case we can prove that specification by showing the documentation.

    That is proof that those functions are specified as a desired output.

    In order for you to claim specified functionality, you have to show me the spec or an equivalent.

    You have to show that the result is what the designer intended.

    How are you going to show me what the designer intended if the designer, and therefore his “specific” intentions, are off-limits?

    “This is the procedure: functional specification + functional complexity = design inference.”

    Show me the specification or the “this process is just like when humans build machines” analogy doesn’t hold up for life processes.

    Like

  19. Toronto:

    In my definition of “functional specification”, it is enough that an observer can explicitly define a function for the object. So, your “natural pseudo machines” are specified, even if not designed.

    I agree with you that DNA or proteins are the object of the debate. We don’t know in advance that they have been designed. We try to infer that from their characteristics.

    So, let’s try that again. We take a functional protein, an enzyme. It is specified, because we can define explicitly a function for it: any molecule which can accelerate reaction such at least of such and such. We can not only define a function, but also measure it according to our definition.

    So, we know that our enzyme is functionally specified. Still, we don’t know if we can infer that it is designed (at least according to the ID inference procedure). It could be a machine or a pseudo-machine.

    To go on with the inference, we need to measure its functional complexity in Fits (as the ratio of the functional target to the search target). There are ways to do that: the Durston metrics is at present the best way.

    So, if our functionally specified object (the enzyme) has a functional complexity higher than a threshold appropriate for the context (I have suggested 150 bits for a generic biological context), we infer that it is designed.

    That is the procedure, in short. There are more details in it. You can find a more detailed description in one of my posts in response to Neil, here:

    http://www.uncommondescent.com/intelligent-design/are-machine-information-metaphors-bad-for-science/#comment-366674

    Like

    • “In my definition of “functional specification”, it is enough that an observer can explicitly define a function for the object. So, your “natural pseudo machines” are specified, even if not designed.”

      I can’t accept that since I cannot deliver a product to my customer before I get his specifications. You are using the term “functional specification” to mean “functional description”.

      In order to design something you have to give me a spec first before I can even imagine how to build it, what materials to use, how much it’s going to cost and how long it’s going to take.

      I have an old PC that I use as a doorstop. Since that is what I see it doing now, can I consider that function to be specified?

      Specification is not a process done in hind-sight.

      Like

  20. Toronto:

    Maybe I have not been clear. My definition of functional specification is purely descriptive. So you are correct, it is a “functional description”. In the inference procedure, “functional description” is what I intend when I speak of “functional specification”.

    Your old PC is certainly functionallly specified for the function of doorstop. But it is not complex for that function (though it is certainly complex for the function of computing).

    Any solid body with some minimal properties of weight and form would do.

    I hope I have been more clear now. The possibility to describe a function in an objective way is only the first step in the procedure. After you have described a function, any function you like, then you have to show that a certain amount of complexity is necessary to achieve that function. That’s the way you assess dFCSI.

    That dFSCI is constantly an expression of design is not an a priori statement. It is an inference which derives from observation (it is always an expression of design in all the cases where we know directly or indirectly that a designer was involved, and it is never observed in all the cases where we know directly or indirectly that a designer was not involved). IOWs, it is a general empirical fact that dFSCI requires a conscious designer to be generated.

    Biological information, obviously, is the object of discussion. The premise is that we do not know, either directly or indirectly, if a designer was involved there.

    Like

    • “In the inference procedure, “functional description” is what I intend when I speak of “functional specification”.

      Your old PC is certainly functionallly specified for the function of doorstop. ”

      Here exactly is where our problem appears.

      First you agree that you mean “functional description” when you say “functional specification” and then in the very next sentence, instead of “description” you use the term “specification” again.

      Could you explain why you can’t use the term which you admit is actually clearer in describing what you mean?

      I believe that if you use the term description, you can no longer claim that a designer was involved. A designer has a clearly specified functional goal that he wishes to achieve with his designs, while a description does not require a designer at all.

      Dembski’s use of CSI is flawed in the same way. It is a description of a structure with no proof or even attempt at addressing the actual goal specified by a designer.

      Without addressing designer issues, how will you determine his specifications?

      If I walk into a room full of milk cartons full of vinyl albums, should I assume that’s what milk cartons were specifically designed for?

      Using your definitions, the CSI of a milk carton can never be different to that of the same box when it holds vinyl records, yet only the milk carton was specified by the designer.

      Like

  21. Neil:

    You say:

    “That’s one of the points of disagreement. Why would we even care to distinguish between machines (as designed systems) and pseudo-machines (as undesigned natural systems)?

    To me, what matters is what it does, and how it does it. Whether or not it was designed is of at most very minor interest.”

    That’s your choice. To me it is one of the most important topics in science to explain the causal origin of what we observe. A lot of people seems to believe that way. But you are entitled to your personal preferences.

    You say:

    “And that’s the other point of disagreement. You want to say that complex implies design. But we often find that an elegant simplicity in products of human design.”

    True. But you are making a very banal error in logic. What I say is that “functionally complex” implies design, but not that design implies “functionally complex”. It is obvious that many designed things are simple. They are designed (a conscious intelligent being outputted them for a function), they are functionally specified, but still they cannot be identified as designed by the ID procedure, because their function is simple. They are part of the false negatives of the procedure. The complexity threshold is fixed at a very high level in the procedure (150 – 500 bits) exactly because we want a procedure which empirically has no false positives. That implies, for well known reasons in empirical statistics, that we must have a lot of false negatives.

    You say:

    “But it’s worse. It isn’t clear that “complex” means anything. The Ptolemaic universe, with planets moving in cycles and epicycles, was far more complex than the Copernican/Keplerian universe with planets moving in ellipses. But they are both the same universe.”

    That’s why I have given a very explicit and empirical definition of functional complexity in my procedure to assess dFSCI. You can find it in the first of my many part answer to you. Perhaps you can comment on that, and show why it does not “mean anything”. I believe that it does mean only one thing, and a very clear thing. It applies to all digital strings, and is perfectly appropriate to evaluate the digital strings we observe in biological information.

    Like

    • “That’s why I have given a very explicit and empirical definition of functional complexity in my procedure to assess dFSCI.”
      …………………………………….

      “I believe that it does mean only one thing, and a very clear thing. It applies to all digital strings, and is perfectly appropriate to evaluate the digital strings we observe in biological information.”

      I am going to put on my ID hat and claim that there is no known natural process that will result in a vinyl record that contains a performance of Elvis Presley.

      1) The problem for you is that vinyl records are analog and yet very complex in their specified output. Since you are only concerned with digital strings, your process cannot be used here or anywhere non-digital structures appear.

      2) How do you determine whether biological structures are a code for a process, ( a digital string), or actually a physical part of a process?

      Like

    • To me it is one of the most important topics in science to explain the causal origin of what we observe.

      ID does not provide any kind of causal account.

      That’s why I have given a very explicit and empirical definition of functional complexity in my procedure to assess dFSCI.

      However, dFSCI is a criterion for strings. To apply it to physical things, you first have to provide a description of the physical thing, then apply dFSCI to that description. Description systems are human inventions, and different description systems can give different complexity answers.

      You then use Kolmogorov complexity to assess the complexity of the string, and personally I have never been persuaded of the value of Kolmogorov complexity. The answer that you get depends on a choice of encoding scheme, so it is really only useful asymptotically, where coding differences diminish in importance.

      Like

    • But you are making a very banal error in logic. What I say is that “functionally complex” implies design, but not that design implies “functionally complex”.

      That comment on “very banal error in logic” suggests that you misunderstood the point. In our experience, design is correlated with simplicity, not with complexity. Thus it seems odd to use complexity as a criterion for design.

      Like

  22. My choice of discussing dFSCI, a subset of CSI, is not due to any substantial difference between the two cases. but only to a couple of considerations:

    a) digital information is easier yo deal with

    b) biological information (or at least the basic biological information I want to discuss, that is protein coding genes and proteins) is digital.

    Regarding your second question, I paste here the first point in my definition of dFSCI:

    a) Any string which can be read as a series of digital values

    Now, there is no doubt, I believe, that the DNA in a protein coding gene can be read as a series of digital values. It is a series of digital values, based on a code in base 4, and three letter words. That’s exactly how it is translated in the cell.

    It is in no way a physical part of the process, because the information in the gene is completely symbolic: it is translated into the correct output (the functional protein) only because there is a translation system which couples each codon to the right aminoacid. Biochemical laws have nothing to do with that result. It is only the information system in the cell which interprets correctly the information stored in the DNA gene, and copied in the mRNA.

    Like

  23. Toronto:

    I answer now your previous post because I have read it only now.

    I use the term “specification” because it is the commonly used term in ID. But it is a description.

    You say:

    “I believe that if you use the term description, you can no longer claim that a designer was involved.”

    I am afraid you still don’t get the sense of ID. I say that if there is a functional description, plus a sufficient complexity necessary to achieve the described function, then we have dFSCI, and we can infer that a designer is involved.

    For the nth time, this is not a logical deduction, but an empirical inference, basewd on how dFSCI is distributed in known cases. Am I clear on that?

    Like

    • Then let’s use the proper terms. Instead of CSI and it’s subsets, let’s use CDI for “Complex Described Information” in those desriptions.

      Once we use the term description, we can no longer claim specified function.

      The function simply works with no designer intent. In order to get to a design inference, you need intent and a description doesn’t get you one.

      Without a pre-specified intended function, you cannot claim a designer regardless of how complex you may imagine something.

      Neil is correct when he says the term complexity is meaningless to us in this argument. Pi is very complex but that doesn’t mean that nature can’t make circles very easily.

      Like

  24. Toronto:

    You can use the terms you like, the reasoning does not change.

    I have never claimed the designer intent as a known fact. I have to infer it from the properties of the object. That’s why it is called design inference, and not deign deduction or design observation.

    Please, try to follow the reasoning. An observable function is a function which an observer can objectively describe for the object. Can we agree on that? We don’t know a priori if it has been designed with intent, or not. That is exactly what we want to find.

    OK?

    So, when you say:

    “Once we use the term description, we can no longer claim specified function.”

    That does not mean anything. I am not claiming anything at this point. I say that the object has a specified function, because I can specify a function for it. You call it as you like.

    Then I measure the complexity necessary to achieve that function, on Fits. For the complexity, I invite you (and Neil) to at least read what I have written at the link I gave above. And then discuss that, and not your generic idea of complexity.

    I will come back to pi in a moment.

    Then I define a property, which I call dFSCI. I define it so that we can in principle assess if it is present in any given string for any described function (although there may be empirical difficulties in the measurement).

    In brief, dFSCI is present if a digital string has a describable function, and if at least 150 bits of functional information are necessary to achieve it. (150 bits is my personal threshold for a biological environment: there are reasons for it, but I will not discuss them now). There is a supplementary requirement: the information must be scarcely compressible (Kolmogorov information), and no known necessity algorithm must be capable to generate it (which is more or less the same concept).

    The functional complexity is measured as the ratio between the functional space and the search space, and expressed in bits, exactly like Shannon’s H. We usually call them Fits (functional bits).

    According to that definition, we can in principle asses for any given string if dFSCI is present or not. Using Durston’s method, we can make that for proteins, with some reasonable approximation. The result is that most protein families exhibit dFSCI.

    The design inference comes only at this point. It is an inference by analogy, as I have said thousands of times.

    It is based on the following premises:

    a) dFSCI is observed often in human designed things. In particular, with great abundance in language and software programs.

    b) dFSCI is never observed in any object for which we are reasonably sure that a designer was not involved in its creation.

    c) The only category of things where we observe abundant dFSCI, and for which we have no certain notion about the possible involvement of a designer, is biological information.

    On these premises, we reasonable argue that design is a good explanation (indeed, the best and only available) for biological information.

    Now, like all scientific inferences, nobody is obliged to accept it. But the criticism which have been given here are completely out of order.

    Let’s go to pi. Pi is a number, and as such it is a creation of our mind (Neil should agree on that). We don’t see pi written on the sand, or in the sky. What we see are circles. Pi is a measure. It has a human origin. As a number which corresponds to a measure, I would say that a string expressing, say, the first 30 figures of pi is a designed object. I don’t expect to find such an object in a natural system.

    And yet, would that string exhibit dFSCI? I am not sure, but probably not. The reason is simple. The string is compressible. It can be calculated. Its Kolmogorov complexity is the complexity of the system which can compute the result.

    Once I have the system, I can compute 30 figures or 300. The a bsolute complexity grow, but the Kolmogorov complexity doesn’t.

    Obviously, the complexity of the computing system is probably higher than 150 bits. But by definition, I have included the clause for dFSCI that “no known necessity algorithm must be capable to generate it”. That is not true for pi, so pi is not dFSCI.

    I will try to give the meaning of all that in a more intuitive form, after having given rigorous definitions.

    dFSCI is a property of pseudo-random strings, which are similar to random strings formally (scarcely compressible, with the highest Shannon’s information value), but which carry the information for a function in their sequence. The functional information is not the total information if the sequence, but the information necessary for the function (it corresponds to the probability of finding the functional space through a random search of the search space).

    A protein coding gene is such a sequence. It has the properties of a quasi random string (except for some minor constraints due to the nature of the code), but it carries the information for a functional protein. That information has nothing to do with the biochemical properties of the molecule: it depends only on the sequence of the nucleotides, in the light of the genetic code. It is “superimposed” on the quasi random structure of the sequence. That is the functional information. The number of bits which are really necessary to have a protein with that function.

    A final invitation: if you fell like it, please give a single example of any string which exhibits dFSCI, which is not designed by humans, and which is not biological information. Just one.

    Like

    • Pi is a side issue, but I’ll comment anyway.

      Let’s go to pi. Pi is a number, and as such it is a creation of our mind (Neil should agree on that).

      That raises interesting issue.

      Yes, pi is a number, and numbers are a creation of the mind. But that number is used to describe a ratio, and whether ratios are a creation of the mind is a different issue.

      Of course, we could say that pi is a ratio for ideal circles, and ideal circles do not exist in nature so they too are an invention of the mind.

      That’s what makes it an interesting issue.

      While many mathematicians might agree that our mathematical systems are inventions of the mind, most will also say that pi is forced on them, so is discovered in their invented system of numbers (as in discovered rather than created).

      I suspect that it shows that our distinction between “invented” and “discovered” is a bit too limited to adequately cover mathematics.

      Like

    • “There is a supplementary requirement: the information must be scarcely compressible (Kolmogorov information), and no known necessity algorithm must be capable to generate it (which is more or less the same concept). ”

      This is an argument from the ID side I simply do not understand.

      Why can’t information be compressible? If we can show that what we consider very complex can be generated by something very simple, how can that not be valid?

      If a string of DNA results in something called Dolly, the cloned sheep, what does not contain CSI, the DNA or Dolly.

      If you could show more evidence for the compressibility of CSI, you would have more positive evidence for ID, not less.

      I believe good design means a simple way of solving a complex problem which Neil has also touched on.

      Show us evidence of the designer compressing data, like human designers do.

      That might be something that both sides could investigate.

      Like

  25. “You can use the terms you like, the reasoning does not change.”

    I think the more subtle point of this whole post is exactly that. Our reasoning changes because of our metaphors, analogies and descriptions.

    Dembski and others need that specific label because of its implications. A simple description stands on its own, while a specification needs a specifier.

    Your are right about Pi. It is simply our way of trying to describe the relationship between a circle and its diameter.

    By dropping toothpicks on a gridded floor, I can generate a digital string approximating Pi.

    I could randomly drop a another pile of toothpicks beside it with no intentions whatsoever.

    One contains dFSCI, my specific intent as the designer to generate a digital string, while the other doesn’t.

    Either one however, results in a value close to Pi.

    How would you determine which one contains dFSCI?

    Like

  26. Gpuccio,

    You could eliminate all the confusion by simply demonstrating, in detail, the calculation of the CSI in a real world biological system or component. That would allow others to do the same with other systems or components.

    Until you show the math, you’re just spewing the same nonsense as every other IDist, including Dembski and Marks.

    Let’s see what you’ve got. I’m predicting more blather and no details, like we’ve come to expect from the ID camp.

    Like

  27. Neil:

    I agree with your comments on pi. I suppose that my personal view of mathemathics is of the neoplatonist kind, but I understand that that is not the only possible interpretation.

    Whatever the philosophical approach, I remain amazed at how mathematics can effectively help explain many aspects of the real world. That’s probably because I consider anyway mathemathics (and philosophy) an indirect approach to knowledge. But that would lead us too far, I suppose.

    Like

  28. Toronto:

    The question of compressibility. It is not that information cannot be compressible. You must remember that, when I define dFSCI, I am not describing the general properties of information, not even of functional information. I am creating a property which can be applied to detect design in the cases in which it is detectable.

    Think of it as a clinical test. We are only building a test which will allow us to diagnose a condition, when possible. That test is set so that it gives no false positives, at the cost of having many false negatives.

    So, dFSCI is defined so that we are quite sure that, in the model of human design, if we find it we can infer design with certainty. But, if it is not present, we cannot neither affirm design nor rule it out.

    Non compressibility derives from that requirement. Complex, but compressible, information can be generated in a system out of necessity.

    For instance, let’s consider a system which generates strings by tossing a coin. If the coin is fair, we have probability 0.5 at each attempt to have “0” or “1” in the string. the resulting strings are truly random. And they remain random also if probabilities are different.

    But if the probability of having “1” is 1 (the coin is such that 0 can never happen), then only a long string of 1s can be generated. We are no more in a random system, but in a necessity algorithm.

    Now, a string of 100 “1s” is certainly peculiar. If it is generated in a truly random system, it could be considered specified (Dembski would certainly consider it that way), because it is “different” in a very specific way from all the others. And its complexity is very high (2^100). So, it could be considered as an example of CSI (even if not of dFSCI, unless we can find a functional specification for it).

    But in the second system, the probability of the string is 1. It is the only 100 fugures string which can be generated.

    The compressibility, as you can see, is equivalent to the possibility of being generated by necessity. If a short algorithm can describe a long string, then we can probably obtain the string through a simple implementation of that short algorithm, that is in a mechanical way. The kolmogorov complexity of the string is, in that case, much lower than its true complexity. A low kolmogorov complexity means that the output could have been generated by a simple algorithm, and is not necessarily designed.

    That’s why, in my definition of dFSCI, I require that the string must be scarcely compressible. IOWs, it must be pseudo-random.

    It’s not only me to think that way. The same concepts have been clearly exposed, for instance, by Abel and Trevors, Durston, and many others.

    The pseudo-randomness is fundamental also for another reason: it allows the flexibility which is necessary to encode symbolic information.

    Take the case of the DNA code. If DNA strings were not flexible, IOWs if some biochemical laws determined severe restrictions to the sequence (such as: after an A there must always be a T, or similar), then protein sequences could not be written. Because the information in protein sequences has no relationship with the biochemical nature of DNA. So, if I am a designer and I need to write a particular sequence of AAs in a gene, I need to be able to write any codon in any position. IOWs, I must be able to create pseudo-random sequences, unconstrained by strong rules.

    Finally, to give the simple idea of what dFSCI is: we need to recognize in the string a sequence which cannot emerge by necessity (no known algorithm can simply describe it; to get the information in the string, we need the string itself); and which has so low probabilities of being generated in a random system (the tossing of a fair coin, for instance) that we can certainly refute that possibility.

    In that case, and only in that case, we infer design.

    Like

  29. Toronto:

    You say:

    “By dropping toothpicks on a gridded floor, I can generate a digital string approximating Pi.

    I could randomly drop a another pile of toothpicks beside it with no intentions whatsoever.

    One contains dFSCI, my specific intent as the designer to generate a digital string, while the other doesn’t.

    Either one however, results in a value close to Pi.

    How would you determine which one contains dFSCI?”

    I am not sure I understand.

    In the first case, you generate a digital string approximating pi by intent. IOWs, you know the sequence of pi, and you guide the dropping so that it represents the figures of pi. That is certainly design, if I see you making it.

    In the second case, what happens? You drop the toothpicks and they form the figures of pi by chance?

    Well, first of all I would change the output in the example. I have stated that pi is not really an example of dFSCI, because it can be computed by an algorithm.

    So, if you allow, let’s say that you build the sequence of myoglobin with intention, in the first case, and that it comes out by chance in the second. Are you OK with that? The sequence of myoglobin cannot be computed by any known algorithm, and is functionally specified.

    So, it is simple. If I see a series of toothpicks which describe the correct sequence of myoglobin, and I know nothing of how they were generated, I will say that that string exhibits dFSCI (I assume that the complexity is higher than 150 bits here; we could discuss that separately).

    So I infer design.

    You can say: in the first case I am right, but in the second case I would be wrong.

    But that is exactly the point of ID: the second case will never happen. Its probability is too low.

    Show me a case where such a string comes out of a truly random system, and you will have falsified the basic assumption of ID: not of the application of ID to biology, but indeed of the principle of design detection itself. If that kind of thing happens, then design detection is no more valid even for human artifacts.

    But that kind of thing does not happen. It is logically possible, but it is empirically impossible.

    It does not happen, exactlyt as it never happens that all the molecules of a gas be contained, for a short time, in half the volume of its container: logically possible, but empirically impossible. IOWs, of no relevance to empirical science.

    So, if we observe dFSCI, we are certainly in the forst scenario (design), and not in the second (randomness).

    This is the point of ID. A purely empirical point. No metaphors. No philosophy. No religion. Just empirical science.

    Like

    • This is the point of ID. A purely empirical point. No metaphors. No philosophy. No religion. Just empirical science.

      Pick a name for it other than “design”, so as to avoid evoking religious connotations.
      Show that this newly named “replacement_for_designed” property can be reliably identified by empirical means, and without dependent of religious or other such assumptions.
      Demonstrate that, using this newly defined property, you are able to make some useful empirically testable predictions, that are not currently being made by science.

      If you can do all of that, and assuming that the predictions stand up to the testing, then you will have some demonstrated science.

      Like

  30. Maya:

    It’s easy.

    Please go to the Durston paper:

    http://www.tbiomed.com/content/4/1/47

    and look at table 1.

    You will find the numbers for the functional complexity in 35 protein families.

    Let’s take one, for example.

    Ribosomal S2 is a ribosomal protein. Its length is of 197 AAs. It cooperates in the ribosome function (translation).

    As you can see in the table, the computation ahs been made on 605 different sequences of that protein, in different species. The null state for the protein (the totally random configuration of that length) has a complexity of 851 bits (20^197).

    The functional information for that protein is calculated at 462 functional bits (Fits). IOWs, of the 851 bits in a purely random sequence of that length, only 462 are really necessary to have the protein function.

    Well, 462 bits are many. Many more than my threshold of 150 bits.

    Therefore, according to my definition, ribosomal s2 exhibits (largely) dFSCI.

    Like

    • The Durston paper does not support your claims. The calculations are nothing like what Dembski describes, although does not calculate for a real world biological system, nor does Durston even attempt to demonstrate that “fits” cannot be generated via natural processes. Basically, you’re waving the Durston paper around as a distraction.

      Unless and until you demonstrate exactly how to calculate CSI, or your variant thereof, for a real world biological system or component, there is no reason to take anything you say on the topic seriously. It’s long past time for you to put up or shut up.

      Like

    • Further, you are making the common ID creationist mistake of assuming that the protein you are discussing came into existence in one fell swoop. You are completely ignoring known evolutionary mechanisms. That makes your claims regarding ribosomal s2 baseless.

      Like

  31. “I am not sure I understand.

    In the first case, you generate a digital string approximating pi by intent. IOWs, you know the sequence of pi, and you guide the dropping so that it represents the figures of pi. That is certainly design, if I see you making it.

    In the second case, what happens? You drop the toothpicks and they form the figures of pi by chance?”

    If you google +”toothpicks” +”pi” you will see references to what I mean.

    The grid lines will have a length relative to the toothpick, I think about 1.5 times. The number of toothpicks that fall across 1 or two lines of the grid is then compared and the ratio ends up being very close to the value of pi to a few digits.

    The toothpicks don’t literaly spell out anything as Pi is derived statistically.

    Toothpicks just fall across the grid due to gravity and yet we have the value of Pi.

    The point is that we have an intentional value that is derived randomly whether by intent of the designer or by accident and that is what is interesting.

    You can’t tell even between two piles of toothpicks that I show you which is specific and we aren’t talking a lot of data.

    That is ID’s problem. That what appears designed isn’t and what appears random, is.

    You have to come up with a test that proves design at a very small scale first, otherwise your method is no better than a guess.

    If you can show evidence of intentional compression, that would be a huge step for ID, as that is when you would expect a designer to be working behind the scenes.

    Like

  32. Toronto:

    Your example is a very good demonstration of my premise: compressable data, data which can be calculated, are not an example of dFSCI, The toophpick example shows that there is a physical model which can approximate pi. That’s because pi can be derived from a physical system, although I believe with some computing work.

    Can you output one of Shakespeare’s sonnets in the same way? Or the sequence of myoglobin?

    Can you see the difference?

    Like

    • “Can you output one of Shakespeare’s sonnets in the same way? ”

      Can DNA output one of Shakespeare’s sonnets?

      If no, does that mean that DNA isn’t an example of dFSCI?

      That again is the point of the post, the misuse of metaphors, analogies and any other method of coupling two disparate ideas, which leads people to conclusions that are not warranted.

      My toothpick example can’t generate a life form any more than dropping the much more complex DNA on the floor will get you the value of Pi.

      As far as compression goes, if I could compress DNA down to 1,000th of its size, I would be an example of a very good designer.

      Compression is difficult to do. Find an example of that good designer and then you might win people over.

      Like

  33. Neil:

    “Design” does not evoke any religious connotation.What is religious in me “designing” this post?

    The property we “reliably identify by empirical means, and without dependent of religious or other such assumptions” is not design, but dFSCI. dFSCI is obsewrved and identified. Design is inferred.

    Where have I used religious connotations in my discourse?

    A testable prediction which will be confirmed as soon as possible: as proteins are designed, our growing understanding of the proteome and of its natural history will reveal that basic protein domains emerged rather suddenly, beyond the resources of any possible chance + necessity scenario. The only possible explanation for their emergence will be an input of information from a designer.

    Remember that, at present, the accepted “scientific” explanation of how proteins emerged is completely vague and inconsistent, and has not been able to generate a single detailed macroevolutionary scenario for the emergence of any significant biologic information. I am curious to know if that is demonstrated science, for you.

    Like

  34. Toronto:

    Here you are not fair. My phrase was:

    “Can you output one of Shakespeare’s sonnets in the same way? Or the sequence of myoglobin?”

    Why have you quoted only the first part?

    Moreover, agaiin you are pretending there is a problem of metaphors where there is none.

    The meaning of my argument should be clear. Your example of the toothpicks and pi is an example of an approximate calculation of a mathematical constant in a system. The properties of the system allow the computation, although approximate.

    Shakespeare’s sonnet and the sequence of hemoglobin, instead, are both examples of symbolic information. If a sonnet were written in DNA, and the translation system was setted to translate codon into letters, it could be read exactly in the same way that the sequence of myoglobin is read. In bioth cases, the physical or biochemical properties of the system have no relationship with the information which is written in it.

    I can’t understand why such a simple concept must be so difficult for spome people. Believe me, it is a very obvious concept. Most biologists understand it very well.

    There is absolutely no metaphor here. DNA genes are writteni a code. That code can be read in the translation system. The information that each protein gene conveys to the translation system is the specific sequence of a specific functional protein. That sequence, that information, can in no way be derived form the physical or biochemical properties of DNA. It is symbolic information, superimposed to the DNA strucutre in the form of a specific sequence of nucleotides.

    Is it so difficult to understand this? That the information written is a sonnet or the sequence of a protein does not change the substance of what I am saying (it would obviously change the conclusions we can make| It would be interesting to see what theory darwinists would create to explain a sonnet in a gene. I have faith they would be up to the task).

    Like

    • “There is absolutely no metaphor here. DNA genes are writteni a code.”

      For the purposes of understanding the workings of certain life processes, it is convenient to think of DNA as a code. To then take that idea and extend it to claim that where there is a code, there is a code-maker, is invalid.

      DNA is in a ready-to-use form, codes need to be decoded first.

      The phrase, “To be or not to be, that is the question”, is ready to be used.

      However, a code, “2b|~(2b)i.e.?”, needs to be decoded first before it is useful.

      You wil also notice, it is compressed. Why have a code the same length as it’s required ready to be used form?

      Like


  35. “Here you are not fair. My phrase was:
    Can you output one of Shakespeare’s sonnets in the same way? Or the sequence of myoglobin?”

    Why have you quoted only the first part?

    Moreover, agaiin you are pretending there is a problem of metaphors where there is none.”

    The point I was making was that each system has its own form of output. The toothpicks, with the help of gravity generate the only output they are capable of, an approximation of Pi.

    You asked if that system could generate something completely different, that of a human, Shakespeare. I said no, and neither could DNA, the point being that you can’t compare the two systems at all.

    You were trying to show the capability of one system was superior to another and the only way to do that would be to consider both of them to be members of a set X.

    ” Shakespeare’s sonnet and the sequence of hemoglobin, instead, are both examples of symbolic information.”

    That is where we disagree. DNA is part of a process, not a code. DNA is usable as is, it does not have to be converted into something else first.

    Neil made an analogy that DNA is like the parts of a key that open a lock. They are directly used to work the lock assembly.

    To claim that DNA is a code begs the question, “If an intelligence codified the instructions, why not compress them at that time also?”

    The code could be uncompressed before it ran and save a lot of energy when duplicating cells since the amount of power required would be cut down down relative to the amount of matter that you are no longer duplicating each time the cell divides.

    If it is permissible to equate what a human designer would do when evaluating what the designer of life would, I suggest he would have done that.

    That is however, the whole point of this post. The designer of life, if there is one, does not belong to the same set that I do. I can’t equate my experiences to his and neither can ID.

    ID’s whole argument is based on relating the actions of human designers to that of a ultimate designer. In effect, ID’s claim is, humans and the ultimate designer are all members of the same set when it comes to design and that’s because of misleading analogies and metaphors.

    DNA is not a code, it’s a ready-to-use part of a process.

    Like

  36. Toronto:

    I am afraid that your las post betrays a severe misunderstanding of biology and of logic.

    You say:

    “The point I was making was that each system has its own form of output. The toothpicks, with the help of gravity generate the only output they are capable of, an approximation of Pi.

    You asked if that system could generate something completely different, that of a human, Shakespeare. I said no, and neither could DNA, the point being that you can’t compare the two systems at all.”

    Biut you completely miss the point. Different systems can have different forms of output, also in informatics. That does not mean that we cannpot descrbe similar characteristics in the system. You seem to mistake abstarct reasoning with the creation of metaphors, and that is a big logical mistake indeed.

    Let’s go to our example. I explained that the difference between the toothpick system output and outputs like the sonnet on a sheet of paper or the protein coding gene is that in the toothpick case the result is computed from the system. I quote myself:

    “The meaning of my argument should be clear. Your example of the toothpicks and pi is an example of an approximate calculation of a mathematical constant in a system. The properties of the system allow the computation, although approximate.”

    On the contrary, bioth in the case of a sonnet written on paper and in the case of the AA sequence written in the DNA gene, the information has nothing to do with the physical system, and is written in a symbolic form.

    Can you understand that? It is not a metaphor. It is a concept. It is abstract thinking.

    Please note that I am not even saying here that a designer wrote the gene. Darwinists think that it was RV + NS to write that specific information through millions of years, as a result of a chance and necessity algorithm. The fact is, however, that the information for the correct sequence is written there.

    But you say:

    “That is where we disagree. DNA is part of a process, not a code. DNA is usable as is, it does not have to be converted into something else first.”

    How can you say such a thing? That is completely wrong! I wonder if you at least understand the basic facts of biology.

    DNA is exactly the contrary of what you say: it is not usable as it is. It is a mass memoty molecule, which would be completely useless if ot were not transcribed and translated. It is certainly part of the process: without the information stored in it, the protein could not be made. In the same way, the sonnet written is part of the process of reading that sonnet. What do you believe you are saying when you say that it is “part of the process”?

    DNA is not usable as it is. It does have to be converted into something else first. Indeed, it is converted twice. First its information is copied into
    an mRNA molecule. DNA as it is is a rather inert molecule confined to the nucleus (in eukaryotes). It could never initiate the process of translation. So, it is transcribed into a more flexible molecule containing the same information: the mRNA.

    But the true conversion takes place after. It is called translation. The very comlex process of translation transforms the information in the mRNA into a functional protein.

    And again, translation is not the result of any physical property of the system: it is the result of its informational structure. There are 20 very complex emzymes, one for each aminoacid, which couple the correct aminoancid to the correct tRNA, IOWs to the tRNA which, in another part its molecule, has the correct anticodon whioch will read the right codon in the mRNA. Finally, the ribosome, a very complex biochemical machine, reads the mRNA and builds the protein taking the correct AAs from the correct tRNAs.

    The functional protein os the purpose of the process. The original dNA gene is the information: iti is a mass mempory, and its role is to preserve the information and to pass it to daughter cells. But that information is completely useless if it is not converted to a functional protein,

    You say:

    “Neil made an analogy that DNA is like the parts of a key that open a lock. They are directly used to work the lock assembly.”

    And I have already said that Neil is wrong. First of all the key is analogic information, and DNA is digital information. Moreover, the key is in itself functional. The final protein can be better compared to a key. Indeed, many enzymes work in some way as a key.

    But not the DNA gene. The DNA gene can at most be compared to a sheet of paper where the digital information to create a specific key is written as coded numbers which can be used to guide a machine to build the key with the right form.

    You say:

    “To claim that DNA is a code begs the question, “If an intelligence codified the instructions, why not compress them at that time also?””

    First of all, you are wrong again. It is not DNA which is a code. It is the information in protein coding genes which is written according to a precise code, which is called the genetic code.

    Regarding your strange questions about compressing, or about what the designer should do, or not do, according to your personal opinions, I will not even try to answer them, becasue they are obviously senseless.

    First we have to infer design (whioch requires not a specific model of the designer). Then we can legitimately start to form possible models of the designer, but we must be guided in that exclusively by facts (the observed design itself), not certainly by the personal likes and dislikes or preconceived ideas of anyone.

    You say:

    “That is however, the whole point of this post. The designer of life, if there is one, does not belong to the same set that I do. I can’t equate my experiences to his and neither can ID.”

    If you are saying that the designer of life is not Toronto, I can agree. In the same way, also the author of the Tao Te King is not Toronto. Or myself. I doubt that you, or I, would have written the Tao Te King as it is.

    Your idea that you can judge the strategies and purposes and capacities of the designer of biological information starting from your personal views is frankly funny at best.

    Like

    • First of all the key is analogic information, and DNA is digital information.

      That is not correct. DNA is a chemical compound and is chemical reactions are analog, not digital, processes. DNA can be modeled as a store of digital information, for some purposes, but the model is not the reality. Failure to understand this distinction is quite common among ID creationists.

      Like

  37. Toronto:

    Regarding your previous post, which as usual I read after, a few comments.

    You say:

    “For the purposes of understanding the workings of certain life processes, it is convenient to think of DNA as a code. To then take that idea and extend it to claim that where there is a code, there is a code-maker, is invalid.”

    First, again I do not believe that it is us who think that the genetic code is a code because it is convenient. The genetic code is a code. Again, I insist that it is the genetic code which is a code, and not DNA. The genetic code is the symbolic correspondence of each possible codon of 3 nucleotides to a specific aminoacid, or to a stop signal. DNA is a molecule, not a code.

    Second, I have never said, least of all claimed, that “where there is a code, there is a code-maker”. Where can you find that in my posts?

    It’s funny, one writes so many things to receive comments on things one has never said.

    If you remember, the design inference is based on the observation of dFSCI, not on the observation of a code.

    I have alredy commented on the “ready-to-use” concept, and on your “compression obsession”.

    I don’t agree, however, that:

    “The phrase, “To be or not to be, that is the question”, is ready to be used.”

    That’s not true. It needs to be decoded according to the rules of english language and of syntax, and maybe accroding to more fundamental linguiostic rules, before it can become a meaningful conscious representation in the consciousness of the reader.

    Which is, I suppose, its function.

    By the way, strange that Shakespeare has not compressed the phrase…

    Like

    • …the design inference is based on the observation of dFSCI…

      That is not true. Thus far you haven’t been able to define CSI or your own dFSCI with enough rigor for anyone else to measure it. There is no way that something so ill-defined can be said to have been observed.

      Like

  38. “Second, I have never said, least of all claimed, that “where there is a code, there is a code-maker”. Where can you find that in my posts?”

    That is the whole point of the ID movement.

    If you are saying to me now, that DNA was not designed, UD may as well shut it’s site down.

    The implication of all of ID’s arguments, is that life is so complex, that it must have been designed. At the very heart of this, is ID’s most sophisticated exhibit, DNA.

    Are you saying that the most complex code your side points at, did not need a designer?

    Like

  39. “Your idea that you can judge the strategies and purposes and capacities of the designer of biological information starting from your personal views is frankly funny at best.”

    But that is my point, that neither of us can. We don’t all belong to the same set of designers, and that includes Dembski and Behe.

    They are no more qualified to comment on the conscious designer of life, if there is one, than you or I are.

    Is there a designer? We can’t determine that with anologies related to our design experiences.

    It may be funny that I don’t think I’m qualified to judge him, but it’s terrifying to think that the ID movement, which wants it’s concept adopted as science, thinks it can.

    We can’t map our concept of what design is on the ultimate designer.

    Do you agree?

    Like

  40. Maya:

    You say:

    “DNA is a chemical compound and is chemical reactions are analog, not digital, processes. DNA can be modeled as a store of digital information, for some purposes, but the model is not the reality. Failure to understand this distinction is quite common among ID creationists.”

    I am afraid you are really confused. Have you read my last posts to Totonto? How many times must I repeat things that are well lnown by any biologist?

    Well, maybe I expressed myself too shortly (I believe these things should be well known by anybody who talks of biological information).

    DNA as a molecule is just a molecule. The way it reacts with other molecules is regulated by biochemical laws. OK with that?

    The information in DNA protein coding genes (that is the correct sequence of aminoacids for the protein) is written in the sequence of nucleotides in DNA. We have 4 types of nucleotides. The genetic code is therefore in base 4. The words are the codons. As a codon is made of three “letters”, we have 64 possibilities. Through a redundant symbolic code, those 64 codons correspond in the translation process to the 20 AAs plus some stop signals.

    This is digital information. Each aminoacid is coded by one or more of 64 “numbers”, written in base four.

    How can you ignpore these simple things, or just misunderstand them?

    And againd, it’s not us who “model” the genetic code. The genetic code (the table of correspondences between codons and amonpoacids) is written in the translation system of all cells. We have just discovered it through observation.

    I am really surprised that I have to spend so many words to convince all of you that the information in protein coding genes is digital, a fact which is well known to all.

    Do you at least understand what “analogic information” means? A magnetic registration of the old type is analogic information, because the magnetic waves which are registered have the same “form” of the sound waves. Analogic information retains the form of the object it describes.

    If you code that form in a series of numbers, then you have digital infornation.

    For the DNA gene to be analogic, the form of the aminoacids or of the final protein should be in some way imprinted in the DNA molecule. But as we all know that is not the case. The only thing we find in the DNA gene is the coded information of the right sequence of aminoacids, expressed in digital form. That DNA molecule would bear absolutely no relationship to the protein, if it were not for the translation effected in the cell.

    Like

    • DNA as a molecule is just a molecule. The way it reacts with other molecules is regulated by biochemical laws. OK with that?

      With the caveat that “biochemical laws” are, in fact, stochastic in nature and that chemical reactions have equilibrium states that are affected by heat, pressure, catalysts, and other aspects of the reaction environment, sure.

      Chemistry is a messy, analog process.

      The information in DNA protein coding genes (that is the correct sequence of aminoacids for the protein) is written in the sequence of nucleotides in DNA. We have 4 types of nucleotides. The genetic code is therefore in base 4. The words are the codons. As a codon is made of three “letters”, we have 64 possibilities. Through a redundant symbolic code, those 64 codons correspond in the translation process to the 20 AAs plus some stop signals.

      This is digital information.

      No, this is a digital model of the analog process of transcription and expression. “Information” doesn’t do anything, the compounds that make up the transcription mechanism do, based on relatively well understood biochemistry.

      Perhaps if you would ever provide a detailed example of how to calculate CSI for a real world biological system, we could make some progress in understanding exactly what you mean by “information” and if your inability to distinguish the map from the territory has led you to deeper errors.

      Like

  41. gpuccio,

    Our host on this blog has summarized the problem with your claims and detailed the solution above in a post you may have missed:

    This is the point of ID. A purely empirical point. No metaphors. No philosophy. No religion. Just empirical science.

    Pick a name for it other than “design”, so as to avoid evoking religious connotations.
    Show that this newly named “replacement_for_designed” property can be reliably identified by empirical means, and without dependent of religious or other such assumptions.
    Demonstrate that, using this newly defined property, you are able to make some useful empirically testable predictions, that are not currently being made by science.

    If you can do all of that, and assuming that the predictions stand up to the testing, then you will have some demonstrated science.

    I would very much like to see the results of such an effort. You would be the first person to apply the scientific method to ID.

    Like

  42. Maya:

    You say:

    “The Durston paper does not support your claims. The calculations are nothing like what Dembski describes, although does not calculate for a real world biological system, nor does Durston even attempt to demonstrate that “fits” cannot be generated via natural processes. Basically, you’re waving the Durston paper around as a distraction.

    Unless and until you demonstrate exactly how to calculate CSI, or your variant thereof, for a real world biological system or component, there is no reason to take anything you say on the topic seriously. It’s long past time for you to put up or shut up.”

    Well, I am afraid you don’t understand the Durston paper, neither you understand my reasoning. I really don’t know if it is worthwhile to even try to explain it to you, given your general attitude. Anyway, I will try briefly. Read
    carefully, becasue it’s not too simple.

    The measurement of functional complexity is defined in principle as the rate of the functional space to the search space.

    For one protein of a certain length, we can assume the total number of combinations of that length as the search space.

    In the case of Ribosomal S2, that is the “null state value” in table 1 of the Durston paper: 851 bits (20^197).
    Now, it would be wrong to state that the complexity of the protein is 851 bits, because there are ceratinly many sequences which have the function. We have to make an approximate extimation of the functional space. Then we can compute the ratio.

    Durston utilizes the concept of Shannon’s H and applies it to a whole protien family. In the case of Ribosomal S2, he has analyzed 605 different sequences with the same function. In the completely random state, the uncertainty is maximum at each site, and it is of 4.32 bits. He then computes the uncertainty in the family for each aminoacid at each site, according to Shannon’s formula. IOWs, if at a site one aminaocid is always the same the uncertainty is 0, and the reduction of uncertainty is maximum (4.32 bits). On the contrary, if at a site any aminoacid has the same probability to occur, at that site the reduction of 0 bits. The sun of the values for all site give the functional information in the molecule. And it corresponds exactly to an estimate of the rate of the functional space to the search space (but I will not even try to show you why).

    Durston is very clera about the fact that what he is measuring is functional information. I quote from the paper:

    “Eqn. (6) describes a measure to calculate the functional information of the whole molecule, that is, with respect to the functionality of the protein considered. The functionality of the protein can be known and is consistent with the whole protein family, given as inputs from the database”

    And if you have heard Durston on video, or read his posts at UD, you can have no doubts that his views are perfectly in line with ID.

    Like

    • Thus far you have completely failed to define your terms rigorously, so discussing how Durston’s paper supports them is premature. You have also failed to address the fact that you are assuming that proteins come into existence ex nihilo. That is not in accordance with known mechanisms of evolution, so your claims have no basis in reality.

      Let’s see an example of how to calculate CSI for a real world biological system. Once you have that, we can determine if Durston’s work supports your claims or not.

      Like

  43. Maya:

    You say:

    “I would very much like to see the results of such an effort. You would be the first person to apply the scientific method to ID.”

    But I have already answered that post and those requests. I obviously have no hope that you may be satisfied by my answers…

    Like

    • No, you have not answered that post at all. If you disagree, please provide a reference to where you have rigorously defined CSI (or your variant of it), demonstrated how it can be “reliably identified by empirical means” (in our host’s words), and made “useful empirically testable predictions” based on that work.

      You haven’t even gotten to the point where you have your terms defined unambiguously.

      Like

  44. Maya:

    You say:

    “No, this is a digital model of the analog process of transcription and expression. “Information” doesn’t do anything, the compounds that make up the transcription mechanism do, based on relatively well understood biochemistry.”

    I habve no hope to bring you to reason. I suppose that for you all computers are analogic, because electromagnetic laws are analogic. What in the world do you mean which such nonsense?

    Like

    • I am simply pointing out that you continue to confuse the map with the territory, like many other creationists who attempt to use math or science to support their pre-conceived notions. Insulting me may make you feel better, but it doesn’t address the flaws in your argument.

      Like

  45. Toronto:

    You say:

    “If you are saying to me now, that DNA was not designed, UD may as well shut it’s site down.”

    Please, try to be more reasonable. I didn’t say that. What I said is:

    “Second, I have never said, least of all claimed, that “where there is a code, there is a code-maker”. Where can you find that in my posts?

    It’s funny, one writes so many things to receive comments on things one has never said.

    If you remember, the design inference is based on the observation of dFSCI, not on the observation of a code.”

    The information in DNA is designed, but I have never inferred that from the existence of the genetic code. I have inferred that from the observation of dFSCI in proteins.

    I don’t knw if your misunderstanding and misquoting of what I write is intentional or simply due to distraction, superficiality or prejudioce. Eiither way, I will not go on this way for long. Trying to understand what the other says is the minimal form of respect in a discussion.

    Like

  46. Toronto:

    You say:

    “It may be funny that I don’t think I’m qualified to judge him, but it’s terrifying to think that the ID movement, which wants it’s concept adopted as science, thinks it can.

    We can’t map our concept of what design is on the ultimate designer.

    Do you agree?”

    No, I don’t agree. I just think that your confusion, intentional or not, is supreme.

    ID doesn’t map anything on the designer. ID detects design. It’s as simple as that.

    I have said clearly that, after having accepted the design inference, we can try to create some model of the designer form his design. That is a scientific process, and it is worthwhile to try.

    I would never try to model a map of the designer from ny personal prejudices, as you have tried to do.

    Like

  47. Let’s try and disengage in a manner that won’t leave lasting wounds, so that we can take another attempt at understanding each other in the future.

    Here’s my last point.

    “ID doesn’t map anything on the designer. ID detects design. It’s as simple as that.”

    The design you detect is based on what you consider constitutes design.

    This means that the designer designs in a way that you can detect.

    There is only way you can equate the two, and that is to use a common standard.

    Since the designer is unknown at this point, and yet to be examined, that common standard you are using, comes from you.

    You have mapped your definition of design onto your designer.

    Like

  48. I see that things have been busy today.

    Thanks for keeping it civil.

    A couple of comments. On the key issue, claiming that a key is analogic is a bit doubtful, when lock manufacturers often provide a digital code that can be used to grind a new key.

    I keep wondering why this emphasis on codes. Discussion of evolution should be, at least in part, a discussion of biology. The DNA is an implementation detail, but there is more to biology than that. Yes, I’ll grant that Dawkins sometimes overdoes that with his “selfish gene” metaphor. But if you could come up with a complete refutation based on codes, you would only be refuting that use of the metaphor. Evolution would still be occurring, and there would still be an abundance of evidence supporting it. What would have been demonstrated, is that the traditional explanation has some minor flaws. All scientific explanations have minor flaws, so that’s no big deal.

    In the original post, I indicated that one of my concerns about metaphors was the way they were used by AI. There, I see too much emphasis on computation, and on the brain as computer, but too little attention paid to the interaction between an intelligent agent (such as a person) and the environment. The discussion here on codes and on DNA as information seems to be making the same mistake.

    Like

  49. Maya and Toronto:

    OK, I think that’s enough. Thank you for all your contributions. It has been fun, for a while.

    Maya, no intention to offend you. I still wonder if you consider computers analogic. Thank you for quoting tha map and territory issue. It’s one of my favourite issues, absolutely. I quote it often. I had never been accused before, of confounding the map with the territory, but there is always a first time. I go back to the “UD cesspool” with a new inner satisfaction.

    Neils,

    thank you for having hosted me here, and for your kind (although few) interventions.

    Bye bye guys. Have a good time.

    Like

  50. I’m disappointed in you, gpuccio. You impressed me as one of the more rational and intelligent participants at UD (not a high bar to clear, admittedly), but you’ve chosen to run away and hide behind Clive Hayden’s skirts rather than continue the conversation in an open forum.

    You have not defined your terms rigorously. You have not demonstrated how to calculate CSI for a real biological system. You have presented no evidence in support of your claim that only intelligence can generate CSI. In short, you’ve failed utterly to support your claims and have joined the long line of ID creationists who refuse to do any science.

    I expected greater intellectual honesty from you. Very disappointing indeed.

    Like

  51. gpuccio,

    I’m disappointed that you didn’t address my last point.

    I hope you come back and try.

    Thanks for the discussion though and maybe we’ll see you again.

    Like

  52. Toronto:

    OK, I think after all I owe you an answer on that point, but then I would really stop my contribution here.

    You say:

    “The design you detect is based on what you consider constitutes design.”

    No. Design is simply defined as any output where a conscious intelligent being contributes to the final form for a purpose. That is a very simple and empirical definition, and it implies no special theory about what “constitutes design”. It simply describes what I intend for design.

    Now please, don’t ask me what I mean with “conscious intelligent being”, and so on. I have already dealt with those things in my multipart answer to Neils. I have nothing to add.

    With that definition, design detection needs only answer the question: given this output, is the best explanation that a conscious intelligent being was involved in its production?

    “This means that the designer designs in a way that you can detect.”

    This means only that I observe from all the cases where the involvement or non involvement of a conscious intelligent being can be directly or indirectly known, the presence of a special property in the object, dFSCI, which I have clearly defined in the mentioned multipart answer, is always, invariably, associated with the involvement if a designer.

    Therefore, dFSCI is empirically used as a reliable indicator of design.

    The only implication here is that design is the work of a conscious intelligent being (a definition. more than an implication), and that in all known cases dFSCI is invariably associated to the work of a designer (a fact).

    You say:

    “Since the designer is unknown at this point, and yet to be examined, that common standard you are using, comes from you.

    You have mapped your definition of design onto your designer.”

    I believe you are rather confused about what “mapping” means. I have given a definition of design, and expressed an empirical procedure to detect design. It is rather obvious that by “detecting design” I mean what I have defined, that is “detecting the involvement of a conscious intelligent being”.

    That is not mapping. It is only “giving a definition and using it”.

    The only thing I am “mapping” is that by designer I mean a conscious intelligent being who contributes to an output for a purpose.

    And that is all.

    Like

    • “I believe you are rather confused about what “mapping” means. I have given a definition of design…” ….SNIP….

      Yes you have given “your” definition of design, not the “designer’s”, and without consideration that the designer may have an unknown method of designing that is not detectable by you.

      You wouldn’t know this without investigating the designer which your side claims is not necessary.

      Thanks for answering.

      Like

Trackbacks

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 137 other followers

%d bloggers like this: