Home » Intelligent Design » Irreducible complexity is all around us

Irreducible complexity is all around us

I gave a talk at the beginning of this year to a group of students at Biola University [1].  In the talk I described just how revolutionary ID is compared to the current scientific paradigm of chance and necessity.  But, such a talk is likely to go over students heads if there aren’t concrete examples.  How could I show them everyday instances of intelligent agents creating information?  Then it struck me just how pervasive the notion of irreducible complexity is.  Just about everything we make as humans is a form of irreducible complexity.  All the machinery and technology of our modern lives are very evidently irreducibly complex, especially considering how often we have to repair them…But we can see irreducible complexity in our less technical tools and creations.

For instance, part of what makes a great work of art so great is that it is a unified integration of a very large number of very carefully crafted parts, often integrated to such a degree that they join together seamlessly.  On the other hand, when elements are just jumbled together our instinctive reaction is that the “art” is a load of junk, regardless of whether some conniving artist has been able to defraud the government of our tax dollars.  Perhaps ID can help us define good art here?

At the organizational level, one of the hallmarks of a good leader is someone who can bring together a disparate group of talented individuals to accomplish a difficult goal.  Steve Jobs is an excellent example of a leader who brought together extraordinary technical and design geniuses to create an irreducibly complex organization, which in turn created numerous irreducibly complex products that have helped ease the technical burden faced by many of us in the IT age, while still enjoying its many benefits (he writes on his MacBook).

In the realm of academia, one of my favorite examples of irreducible complexity is a carefully crafted argument, such as David Stove’s argument against Hume’s inductive skepticism [2], or Euclid’s argument for an infinite number of primes [3].  In fact, it appears to me that irreducible complexity is responsible for most of what I like in life.

[1] http://www.box.net/shared/68puolysda

[2] http://nekhbet.com/hume/index.html

[3] http://en.wikipedia.org/wiki/Euclid%27s_theorem

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

31 Responses to Irreducible complexity is all around us

  1. As I read this I looked down at my keyboard. It’s functionally useless unless nearly all of the keys are present and labeled. The labeled keys are useless without the underlying circuitry. And completed keyboard is useless without the hardware and software needed to recognize keystrokes.

    But then it all fell apart. First, my telephone has a keypad, so the buttons on my keyboard could have been co-opted.

    Second, the local beaches are covered with sand which is where silicon comes from. With that much sand arguments based on probability lose their strength.

  2. The simplest demonstration that a keyboard is not irreducible is the fact that numeric keypads such as on older phones are routinely used to send messages.

    You can also type anything on a computer keyboard using only the Alt key and the digits 0-9.

    Even more damaging to your argument is the fact that language and spelling have been simplified by texting.

    Computers have a log evolutionary history going back to typesetting (which also has a history), through typewriters and mechanical teletypes. Computing itself goes back to automated weaving machines.

    Long histories of incremental change.

  3. And with all the piano tuners out there, arguments based on fine tuning also lose their strength!

  4. Could you list things that have CSI? e.g.

    – Music (patterns)
    – Letters in the alphabet (shapes)
    – Paintings (most)
    – Building designs
    – Life
    – Flowers

    What about things that do not have CSI?

    – Electrical noise
    – Falling rain
    – Sea waves(?)
    – Desert sand dunes(?)

  5. Irreducible complexity is the “nature” of things. It is everywhere to be found in the laws of physics, and even more obviously to be found in biological systems. IC is the rule, not the exception.

  6. Your post got me to thinking, and I realised that something irreducibly complex can have a quality. This quality can be more or less, but is independent.
    The example of reducing a high quality keyboard to a low quality keyboard is not an example of reducing the complexity of a keyboard; it is an example of reducing the speed and comfort at which someone can type.

    This reminded me of the jumbled letters myth “Aoccdrnig to rscheearch at Cmabrigde uinervtisy, it deosn’t mttaer waht oredr the ltteers in a wrod are, the olny iprmoetnt tihng is taht the frist and lsat ltteres are at the rghit pclae.”
    See link for more details.
    The complexity of the idea the words communicate has not been reduced, only the quality of the information. Reducing the quality has the affect of slowing reading speed. If we carry on reducing the quality of the information, at some point the idea will be lost, but at no point will the complexity of the idea be reduced.

  7. Computers have a log evolutionary history going back to typesetting (which also has a history), through typewriters and mechanical teletypes. Computing itself goes back to automated weaving machines.

    Yes, trees, the building blocks of automated weaving machines, were everywhere.

  8. The complexity of the idea the words communicate has not been reduced, only the quality of the information.

    My simple Alt+three digit keyboard has 100 percent of the functionality of a full keyboard. It sacrifices no information, just part count.

    The original statement of irreducible complexity asserted that function would be lost if parts were removed. The simple fact is that many if not most complex objects can have parts reduced without eliminating all function.

    For many decades a simple on off switch served the function of a keyboard. It was called a telegraph. It sacrificed nothing in the way of ability to transmit information.

    The same is true of biological mechanisms, including the iconic flagellum. There are dozens of bacterial structures simpler than the E.coli flagellum that produce locomotion, and there are dozens of partial flagellum sequences performing real functions in real bacteria. The fact that some of these functions are not locomotion has no bearing on their usefulness.

    The whole notion of irreducibility is fraught with argument from ignorance. It only survives where the possibility of simpler or partial structures has not been investigated.

  9. But we were discussing function, not parts lists.

  10. It seems to me that the most frequent objection to these types of examples of known intelligence creating obvious irreducible complexity (or complex specified information) is not an objection to the IC or CSI but to the very existence of actual intelligence, mindfulness. And that is coming from a devout yet reasonable Darwinist (i.e. willing to try to understand the arguments of ID, not try to extinguish the fire with a relentless cloud of hate and defamation [i.e. Pharyngula]). It typically goes something like:

    “Well our ‘intelligence’ is simply an output from our brains, which are simply dictated by law and chance, so this proves nothing except that law and chance are capable of producing CSI/IC”.

    There is an obvious philosophical difference between “mind is matter” and “mind above matter” which are probably not going to be overcome in a discussion of this nature, but I still think there is common ground that could further meaningful reason. I would say something like:

    “Okay, let’s assume mind is simply an output of law and chance in the electrochemical machine in our skull. We don’t yet know exactly how consciousness can arise from the behavior of matter, but let’s assume there is a way. You at least have to concede that this output we call ‘intelligence’ or ‘consciousness’ is clearly different than the typical behavior of law and chance acting on matter in this universe. It is even remarkably differentiable from natural selection acting on self-replicating entities with variable, heritable traits, which is due simply to the speed and efficiency of the ‘design’ process of intelligent agents; I could design a functional screw driver in minutes with maybe one or two attempts but it would take the Darwinian mechanisms perhaps thousands or millions of iterations (assuming we had a reproducing population of screwdrivers whose reproductive success was in direct relation to the efficacy of their screw turning). So even though we’re assuming intelligence is a fully natural process, it is fair to consider it distinguishable from any other natural process. The next step is to decide the best way to objectively distinguish it.”

  11. You both seem to be arguing that ideas can only be developed as a step-by-step gradual process and therefore the complexity of every idea can be reduced.
    The history of ideas is more like slowly improving the quality of an idea, before a brand new idea comes along.
    As an analogy we could consider a child being born, slowly growing up into a man, then fathering a new child.
    When von Neumann described the modern computer in 1945 it’s clear he was building on the work of others (notably Turing in 1936), but a new irreducibly complex idea had been created. You can pick 1945 or 1936 as the date, but this was the beginning of computing.

  12. It is rather amazing that a human can start from a zygote and expand to a full-grown man or woman.
    Now try that with a computer. We’re talking about growth, not evolution, but it shows the vast gulf between the example and the reality.
    How did we go from the beginning of computing to what sits in front of us now? Didn’t every single step involve teams of people drawing schematics for overall systems and the tiniest components, testing, solving problems, and then building what they designed using tools that were similarly designed?

    It doesn’t make a good analogy to something that grew through the tiniest of incremental changes and without any intentional guidance.

  13. The simplest demonstration that a keyboard is not irreducible is the fact that numeric keypads such as on older phones are routinely used to send messages.

    You can also type anything on a computer keyboard using only the Alt key and the digits 0-9.

    Even more damaging to your argument is the fact that language and spelling have been simplified by texting.

    Computers have a log evolutionary history going back to typesetting (which also has a history), through typewriters and mechanical teletypes. Computing itself goes back to automated weaving machines.

    Long histories of incremental change.

    I do not find this demonstration to be terribly demonstrable towards showing the reducibility of keyboards.

    You are talking about simplification, which no one denies can be done with most “irreducibly complex” systems. Like with a mouse trap, maybe the spring could have a few less turns, or maybe the hammer could work with one 60º bend instead of two 90º bends. But you still need a hammer, you still need a spring, etc.

    With keyboards, heck you could reduce it down to one unlabeled button and use a Morse-code-like system for the entire alphanumeric and symbolic collection, including:

    ,./;’[]\?:”{}|-=_+`~!@#$%^&*()

    You could vastly simplify the signal transmission sequence from something like (and I’m leaving tons of stuff out; never mind any technical problems since I’m not a computer guy):

    key –> key wire/circuit –> keyboard circuit board –> USB cable –> CPU –> RAM/HDD, Monitor, modem –> Internet (servers, trillions of miles of fiber optic wires –> your modem –> your ethernet cable –> your CPU –> your monitor cable –> your monitor

    …down to…

    my button –> wire –> your listening device

    But it’s still irreducible because you need a mechanism to input information, a mechanism to transmit the information and a mechanism to interpret, store or print the information. A keyboard (or functionally-equivalent machine) is pointless without these three entities. I could tap away all day on a Morse transmitter, but what’s the point if the transmission cable isn’t connected to anything?

    I don’t understand the discrepancy ID skeptics have of the concept in the OP. Irreducible complexity obviously exists in our universe (in our creations, perhaps most obviously) and we are obviously intelligent. How anyone could disgree with these two statements amazes me; they just seem so obvious to me, so much so that I believe you have to be forced to disagree with them by an a priori philosophical assumption

  14. Perhaps, but the computer is much more than a teletype machine. Contrary to the telephone, which is merely a communication channel and not Turing machine, a computer does require communication channels but has new functionality in its ability to process information.

    So, it wouldn’t be accurate to say that the computer is merely a co-opted teletype machine. If it was just an incremental addition in functionality, then that’d be correct. But I have no idea how one could incrementally generate a Turing machine from an information channel. Merely focussing on the keyboard is like merely focussing on the wire in a mousetrap and saying the mousetrap is merely an incremental improvement on a coat-hangar. It’s mistaking the material for the function, as I believe another commentator here pointed out.

  15. I respectfully disagree with that notion, since I believe the original objection you laid out is legit, as long as its plausible that ID can be the output of a mechanical process. B/c if it is accepted that a mechanical process in the brain can produce ID, then this would imply that mechanical processes could eventually produce ID, which undermines the ID argument.

    But, that’s not to say undermining ID is the problem, perhaps ID is not well founded in that regard. I’m merely pointing out that such a premise is not consistent with mainstream ID claims.

  16. Folks:

    That you may build a different kind of keyboard or the like has little or nothing to do with whether this particular keyboard in front of us requires a certain core of components to be in place, to be properly organised relative to one another, and to all be working, for it to work at all.

    Pardon, but — for months and years now — I keep getting the strong impression that many strident and vociferous objectors to the IC concept have never had to design, build or repair any fairly complex functional system, or if they have, have never seriously thought about what is going on.

    The very existence of knockout studies as a genetic research tool speaks to the place of IC in biosystems.

    And, the notion that “Hello World” can transform itself incrementally by lucky errors detected through trial and error success into an operating system is laughably naive on its face.

    Something is very wrong indeed with evolutionary biology in our day, if that sort of thinking is what is being held up as a paradigm of how evolution created biofunctional complexity.

    Please, let us not cling to absurdities.

    GEM of TKI

  17. It’s not just a question of simplification of existing function. It’s a matter of looking at the history of complex objects. The more history you know, the more obvious it is that complex designs made by humans are the result of incremental changes plus co-opting features from other objects. Horizontal transfer.

    Computing has been around since people carved lines in wood or clay to keep a count of things. The abacus is a re-usable tablet for counting. The relay is an electrical abacus.

    The Jacquard loom is generally considered to be the first implementation of an information storage and retrieval system, but it had predecessors in the work of Basile Bouchon, Jacques de Vaucanson and others. Long before that the Greeks and Romans were building automata.

    The point is, that the more you know about the history of an object, the more incremental becomes the invention. It’s only when you don’t know the full history that it looks like gaps.

  18. That you may build a different kind of keyboard or the like has little or nothing to do with whether this particular keyboard in front of us requires a certain core of components to be in place, to be properly organised relative to one another, and to all be working, for it to work at all.

    But that’s simply wrong. We have been talking about a particular keyboard, and it is easy to demonstrate that one can remove whole sections of keys without reducing function (other than typing speed). There’s an enormous amount of functional redundancy in a keyboard.

    And even if some electrical flaw caused the keys to change function occasionally, a simple feedback mechanism would discover and correct the flaw.

  19. First, I’m assuming you mean IC (irreducible complexity) instead of ID when you talk about producing it. Just want to make sure we’re on the same page

    ID does not say there must be a non-mechanical, non-material explanation for IC/CSI. It says that intelligence is the best known, observed cause of these phenomena. We observe our own intelligent actions on a daily basis. What does it matter, for the sake of this argument, if that intelligence is a mechanical kludge, a “ghost in the machine” or something else? How is it not still distinguishable from the rest of the known universe? Unless someone is theorizing that a Darwinian process in our brain is creating our thoughts (i.e. it’s generating millions of thoughts and only the “logical” ones are selected, or something), it would be quite a unique mechanism which clearly behaves radically differently than the rest of the matter in the universe. This almost seems undebatable to me.

    Also, I don’t understand this sentence:

    I’m merely pointing out that such a premise is not consistent with mainstream ID claims.

    In what way?

  20. And, the notion that “Hello World” can transform itself incrementally by lucky errors detected through trial and error success into an operating system is laughably naive on its face.

    kairosfocus, I completely agree with your sentiment. But for the sake of argument I feel like we do need to consider, in this example, a “progression of keyboards” to have a useful analogy to the evolution of life. Organisms are created through self-replication, unlike how keyboards are assembled. So if we’re going to compare the two there needs to be something to make them comparable.

    I said that the simplest keyboard-like mechanism still has at least three 100% necessary parts (input, transmission, output). In reality, the “evolution” from Morse code “straight key” to a digital keyboard has IC at every progressive step; I’m an engineer, and I know if I want to add one simple feature to an existing design (of anything) I almost always need to modify/add several other features. I’m in the filtration industry. If I wanted to add an access door to a dust collector, the steps to create this “evolutionary step” (in the evolution of this dust collector) is not simply:

    STEP 1: Add door

    It is more like (for bare minimum functionality):

    STEP 1: Remove material for door
    STEP 2: Add hinges
    STEP 3: Add door panel
    STEP 4: Add latch system to keep door closed
    STEP 5: Add gasket to prevent dust bypass
    And typically there are many more steps than that, and even each one of those steps are irreducibly complex, anyway. But all of it is necessary for a functional door.

    I have seen absolutely zero reason to believe that the evolution of novel biological features is any different than the design process I have to go through.

  21. I need to elaborate on that. If the wires connecting a keyboard to a computer are cut, there will be no functionality. If the transistors acting as switches fail it is equivalent to cutting the wires.

    But I have had keyboards that were damaged by water and which had sections fail, large blocks of keys, but which could be used (slowly), by holding the alt key and typing the ascii codes.

    This is an example of functionality that is redundant, and the complexity of the object is therefore reducible.

    We are discussing the keyboard, not the invention of the computer. It is trivial to demonstrate that a keyboard could be reduced to an on/off switch and still input ascii code.

  22. I have seen absolutely zero reason to believe that the evolution of novel biological features is any different than the design process I have to go through.

    Well take a look at the evolutionary sequence of the mammalian middle ear, from jawbone to the inner ear bones. I won’t cite that as “proof” in the mathematical sense, but it is evidence that your doors and hinges can be incrementally made from other parts. At least in biology.

  23. Is there a multipart core, where each is necessary and together the organised parts are sufficient for function?

    Plainly, yes.

    Cut power, cut or scramble data connexions, cut controlled switching that triggers code, or just a stuck key and bye bye function.

  24. Cut power, cut or scramble data connexions, cut controlled switching that triggers code, or just a stuck key and bye bye function.

    I mentioned that the cord is essential.

    The only essential component of a keyboard is an on/off switch. Everything else is convenience.

    I’ve already described in some detail how to work around missing or non-functional keys. Alt + ascii code.

  25. But it’s still irreducible because you need a mechanism to input information

    I think you’ve missed the point of irreducibility. It’s just the keyboard under discussing, not the universe of computing. You’ve correctly noted that the keyboard can be reduced to a switch capable of transmitting Morse code, which is a sequence of pulses. But even without modifying the computer a bit, a keyboard with half the keys missing or disabled can still be fully functional. There are a number of alternate ways around the missing buttons.

    The iconic flagellum is also pointless without a cell to propel, and all the supporting structures. But Behe hasn’t tied the irreducibility of the flagellum to the entire cell. He simply argues that the flagellum loses its most obvious function if you knockout some of it’s structure.

  26. If you don’t know the full history the gaps may look larger than they really are, but the gaps are still there. But maybe your point about “horizontal transfer” is a big factor in this. Consider the history of the electronic computer keyboard, starting in the 1970s the keyboard went through a series of incremental changes developing from this to this. Meanwhile, starting in the 1990s Bluetooth short-wave radio transmission technology was being developed gradually. Then in 2003 Apple combined the two and created a new irreducibly complex product; a bluetooth keyboard.
    The relevant point is that there is no intermediary step between a wired keyboard and a wireless keyboard. So there might be a less complex keyboard than a wireless keyboard, but there is no such thing as a less complex wireless keyboard.

  27. Actually there are degrees of complexity in wireless keyboards, and vast degrees of complexity in wireless transmission in general.

    Early wireless transmission was commercially successful before transistors or vacuum tubes. All you needed for a transmitter was a large spark generator.

    A receiver was just a large antenna attached to a tiny gap so you could see the induced spark. That was the kind of wireless used by ships at sea when the Titanic went down. It worked. The Titanic was able to radio for help using this system.

    The more you know the history of things,the smaller the gaps.

    That’s really the central weakness of the ID program. It’s vulnerability to knowledge. Without knowing the history of things, you don’t know the size of the gaps. Without knowing the history, you have no basis for calculating probabilities.

    Regardless of who “wins” this argument, the only way to settle it is to explore history. If you can’t find the complete actual history, you can still do what forensics does in criminal cases. You can investigate the processes.

  28. I think you’re missing the point. If we removed all the redundancy from a keyboard, the keyboard would still have the same basic level of complexity.

    Less redundancy, same complexity.

  29. I think you miss the point of IC. If you can remove parts and still have the same function, it’s not IC.

    The only part necessary for data transmission is an on/off switch.

  30. There are two points here.
    1 A keyboard with all redundancy removed is IC.
    2 Redundant features (namely more keys) exist on a keyboard.

    Your argument is, because 2 therefore not 1.
    This is clearly wrong.

    You need to explain which parts could be removed from a single button keyboard which would not break the function.

  31. A single button keyboard (a switch) may be irreducible, but it is no longer complex.

Leave a Reply