Home » Darwinism, Design inference, News » Bill Dembski answers, How do we explain bad design?

Bill Dembski answers, How do we explain bad design?

Continuing with James Barham’s The Best Schools interview with design theorist Bill Dembski – who founded this blog – on who bad design is a problem for and how.

TBS: How do you explain suboptimal or bad design? Do you have a scientific explanation for such instances of design?

First off, let’s be clear that design is rarely, if ever, optimal. The problem is that all designs involve compromise among competing objectives.

WD: The reason we put the adjective “intelligent” in front of the noun “design” is not to stress that the design we find in nature is optimal or good or morally acceptable. Rather, it is to underscore that the design we find in biology and in the universe more generally is actual. Richard Dawkins opens his book The Blind Watchmaker by stating “Biology is the study of complicated things that give the appearance of having been designed for a purpose.”

For Darwinian biologists, all such design is merely an appearance. The “intelligent” in “intelligent design” underscores that we’re not just dealing with an appearance of design, but rather with actual design.

So while the question of suboptimal or bad design may be interesting, it is not central to intelligent design as a scientific program, which in the first instance is interested in looking for evidence of design überhaupt. That said, it will be helpful to bring some clarifications to this discussion, especially since the problem of bad, and even malevolent design, is such a stumbling block for many people in accepting ID.

First off, let’s be clear that design is rarely, if ever, optimal. The problem is that all designs involve compromise among competing objectives. They are multicriteria optimization problems, and the problem with multiple criteria is that there is no unique way to rank criteria.

Take a coat hanger. What is the best coat hanger? One that is strong, resilient, and extremely light. Okay, try a titanium coat hanger. But now you’re paying a lot of money for the coat hanger. If one of your criteria of optimality is cost, then you’ll probably forgo titanium and go with the plastic Walmart specials.

Leaving aside the issue of multicriteria optimization, one might still point to certain biological systems and argue that they could have been designed better. But even this is typically far from clear.

One Darwinian favorite is the inverted retina of vertebrates. The wiring is backward, and any self-respecting designer, we are told, would have designed it differently.

Whenever I hear such criticisms, however, what I don’t hear is a concrete redesign plan that, when implemented, actually demonstrates the superiority of the new design. It’s one thing to speculate about how to make something better. It’s another thing to actually do it. Evolutionary biologists are notorious for mounting arguments from imagination, where it’s enough to imagine some improvement without actually implementing it. And for them, such an argument always trumps design.

Evolutionary biologists are notorious for mounting arguments from imagination, where it’s enough to imagine some improvement without actually implementing it.

With the inverted retina, there are actually good functional reasons for it. I recount that in my book The Design of Life, coauthored with Jonathan Wells (Foundation for Thought and Ethics, 2007). Briefly, a visual system needs three things: speed, acuity, and sensitivity. To achieve sensitivity, retinal cells need a copious blood flow. Putting nerves and blood vessels in front of the light sensitive cells allows for just that. Nor does this block light, because Müller glial cells serve as fiber optics that bring the light without distortion to where it needs to be.

Okay, what about parasites and nasty critters that inflict pain on others? Even here, one finds that the designs are quite remarkable—the parasites seem designed to do a number on their hosts. Yes, but what sort of designer would have done this? Read my book, The End of Christianity. Natural evil is a problem, but it is a problem for theology and not for intelligent design per se.

Next: Dembski’s conception of ID

See also:

Bill Dembski on the problem of good

Bill Dembski on young vs. old Earth creationists, and where he stands

Bill Dembski on the Evolutionary Informatics Lab – the one a Baylor dean tried to
shut down

Why Bill Dembski took aim against the Darwin frauds and their enablers #1

Why Bill Dembski took aim against the Darwin frauds and their enablers Part 2

Bill Dembski: The big religious conspiracy revealed #3

Bill Dembski: Evolution “played no role whatever” in his conversion to Christianity #4

So how DID Bill Dembski get interested in intelligent design? #5b – bad influences, it seems

So how DID Bill Dembski get interested in intelligent design? #5a

So how DID Bill Dembski get interested in intelligent design? #5b – bad influences, it seems

Bill Dembski: Trouble happens when they find out you mean business

What is Bill Dembski planning to do now?

What difference did Ben Stein’s Expelled film make? Dembski’s surprisingly mixed review

Bill Dembski on the future of intelligent design in science

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

24 Responses to Bill Dembski answers, How do we explain bad design?

  1. A lot of ink has been spilled on the “bad design” arguments. First of all, we have to distinguish between “bad design” in the engineering sense, and “bad design” in the evil sense.

    Bad (Engineering) Design
    This is largely an empirical question. However, it is often a subjective assessment, in that there are design tradeoffs. Further, things can break and degrade. Finally, we don’t really fully understand the systems that are being critiqued. So the key points to keep in mind regarding bad engineering design are:

    1. We know for a fact that there are all kinds of things designed that have “bad” design characteristics. Doesn’t mean they aren’t designed. (iPhone antenna anyone?)

    2. There are constraints and tradeoffs in any design. No design meets some undefined, theoretical, ephemeral idea of “perfect” design. It has to exist in the real world with real-world constraints.

    3. There are extremely few living systems that exhibit poor engineering. Virtually all the examples of allegedly poor design have turned out, on further inspection, to be exquisite design. The argument of bad design in biology has an extremely poor track record.

    4. Even if something is well-designed, it can still break down or degrade over time.

    5. Those who complain about bad design in biology have never put forward an engineering quality assessment of the system in question and a careful analysis of how the system should be improved, taking into account engineering constraints and real-world application. Instead, people who don’t understand engineering tend to seize on something that looks unusual to them and start shouting from the rooftops that ‘no designer worth their salt would do it that way.’ This is why we get the stupid “backwards wiring in the retina” kind of nonsense.

    Bad (Evil) Design

    This is a subjective, even theological, argument. It typically rests on things like disease, parasites, bugs that eat each other, and similar observations that make us queasy or uncomfortable or hurt. The responses here include:

    1. This is not an empirical, but a philosophical and theological argument. A judgment of “bad” in this context depends on one’s expectations about what the allegedly-non-existent designer would have done if the designer existed. It essentially amounts to nothing more than “I don’t like this” or “If I were in charge, I wouldn’t do it this way.” In a slightly more substantive sense, it is really just another example of the “Why is there evil in the world?” question. This is a deep and important philosophical question that deserves reflection. But it is not an empirical objection to design.

    2. Bad (evil) design typically rests on some idealized idea of divine creation, in which creatures are all kind to each other, no-one ever gets hurt, life is blissful and peaceful, and everyone gets along. Is there any reason to think that the designer, as Cornelius Hunter aptly asked, intended that kind of utopic existence for us? Further, as long as this is a theological question, those who are so inclined can offer, at least in part, a theological answer: there was something akin to an idealized creation and will be again; the current state of affairs represents a fallen, temporary state during which we can learn, and grow, and experience, and come to understand what it means to have good and evil.

    3. Finally, and post pointedly for ID, we know for a fact that many things that hurt, do damage, and cause pain are designed. Thus, even if an objective assessment leads us to conclude that some biological system is “evil,” it doesn’t mean it wasn’t designed.

  2. “Natural evil is a problem, but it is a problem for theology and not for intelligent design per se.”

    Steady, there, Boss. You’ll be talking above their heads.

  3. I think it was Paul Nelson who once made an argument along the lines of: If a car has rust, the rust does not preclude the reality of an engineer who designed the car. The car may have been well designed but it is still subject to (possibly unforeseen) factors in the environment.

    Does anyone know the exact quote/citation?

  4. YEC teaches there is no bad design.
    After the fall the universe changed.
    We got disease and remedy to it. Before the fall there was no disease and no ability to deal with it.
    Creatures completly changed and therefore the changes might not be as great as the original creation.
    Anything that is actually badly designed is simply showing its a change from the original plan.
    YEC teaches this.
    ID folk do get tripped up be not accepting Genesis.

  5. ” ID folk do get tripped up be not accepting Genesis.”

    ID folks certainly do not get tripped up by not accepting Genesis. Even if Genesis taught that there is no bad design (which is arguable), it would only be relevant as a theological response, which although perhaps interesting, doesn’t address the other issues at hand.

    Some of us are quite happy to address the issue on scientific and engineering grounds, and on those grounds ID does just fine without any extraneous references to Genesis or any other sacred text.

    We’re not getting tripped up at all. We’re being careful and scrupulous about making the appropriate distinctions and responses.

    As I explained, a reference to the Fall is an appropriate response to a limited aspect of the bad design argument, and I don’t want to disrespect anyone’s views on that point, but it doesn’t address the primary issues raised.

  6. Hey do you guys want to read something that is hilarious?

    Elizabeth Liddle thinks she can create CSI with Natural selection:

    Creating CSI with NS

    Of course when I told her she was misrepresenting Dembski she got all defensive and wouldn’t have any of it.

  7. Eric Anderson.
    YEC would say its not a theological reply but a true reply of real nature and the origin of nature.
    The fall described in genesis makes room and demands post fall bad design reality.
    if ID teaches a creator did the great creations then ID would need to explain the reasons for inferior design matters in biology.
    otherwise a great creator would of blundered in his product.
    How else?

  8. Joe @6:

    Good luck. Elizabeth is convinced that evolutionary algorithms create something from nothing, the informational equivalent of perpetual motion. Incidentally, I’m not sure why she is still trying to demonstrate this, given the passionate arguments made that evolutionary algorithms such as Avida have already demonstrated this. Or maybe she finally realized they haven’t demonstrated any such thing. :)

    —–

    Robert @7:

    I have outlined in #1 above several perfectly objective responses to the ‘bad design’ line of argumentation, without any need to refer to the fall.

    otherwise a great creator would of blundered in his product

    This is precisely the theological argument made by materialist evolutionary proponents against design. The important thing here is to understand the nature of the argument, so that one is not misled and can respond appropriately. By accepting the faulty evolutionary argument, you leave yourself in the untenable position of (i) having to provide a (questionable/arguable) theological response and (ii) undercutting your own scientific grounds for explaining why there are perfectly good reasons for “imperfect” design.

    The ‘bad design’ line of argumentation is wrong on multiple fronts. It needs to be recognized as such and responded to clearly and forcefully on all fronts. Don’t limit yourself to a single theological response, particulary when that response happens to be the one that is least likely to be taken seriously by the person making the bad design argument.

  9. Hi Eric-

    It is useless as Lizzie thinks that all one needs to know about CSI is in Dembski’s paper on “Specification”. And now she thinks that Meyer and Dembski stand in contradiction pertaining to CSI.

  10. Joe:

    I presume you are referring to this paper?

    http://www.designinference.com.....cation.pdf

    Do you think she is right, in the limited sense that if we look at this specific paper Dembski is making some mistake and/or is in contradiction to Meyer? Or is she not understanding this paper either?

    In my experience, when people don’t understand what information is (for example, don’t understand that it is separate from and exists independent of the particular physical instantiation), then it doesn’t much matter how many papers they read. They’re never able (or willing) to grasp the basic concept and end up hopelessly confused, thinking that somehow information will just arise once we have a physical system in place.

  11. Hi Eric-

    Yes, that is the paper. She thinks it replaces “No Free Lunch” even though Dembski says, in an addendum, that the paper is in addition to it.

    The problem, as far as I can see, is she is stuck on compressibility- she is equating compressibility with CSI and she says that is what Dembski says.

    However Dembski says compressibility = specification.

    So she starts out with a random sequence of heads and tails, reproduces that intial string, with variation, checks it with some “fitness function” and continues. Then if the string becomes compressible via some algorithm she is declaring CSI is produced via Darwinian processes.

    But anyway the alleged contradiction between Meyer and dembski stems from Meyer saying that complexity cannot be algorithmically compressed. And I used examples of computer programs, assembly instructions and encyclopedia articles as examples of CSI that cannot be algorithmically compressed.

    All that said she has never read “No Free Lunch” or “Signature in the Cell”.

  12. 12

    Eric Anderson
    If they invoke the God wouldn’t do it that way thing then they started it.
    They have to accept why God would of done it.
    There would be bad design if everything was redesigned because of the story in gEnesis.
    This doesn’t help ID but it does give YEC a great answer.
    I understand you have other excellent points.

  13. “However Dembski says compressibility = specification.”

    Well, I doubt he says this. At least not without some qualification.

    It is true that the classic example of a complex string that is not subject to meaningful compression is a purely random string. And it is also true that a specified string will often be more compressible than a purely random string. So perhaps that was the point he was trying to make.

    Indeed, on the other end of a spectrum, we could have a specified string of 1000 ones or zeroes, which would be highly compressible. Yet at that point the string would not be complex. Further, if we are talking about information in the Shannon sense of compressibility, we need to understand that it does not tell us anything whatsoever about the underlying meaning or function of the compressed string.

    Any time we talk about these issues we have to take into account both specificity and complexity. And when applied to real-world applications, we have to consider meaning/function as well. This is a very straightforward issue, which Dembski certainly understands. However, many people seem to miss the issue, which is probably what is happening with Elizabeth’s attempt.

  14. Eric-

    The paper has a section titled “Specification via Compressibility”.

    To sum up, the collection of algorithmically compressible (and therefore nonrandom) sequences has small probability among the totality of sequences, so that observing such a sequence is reason to look for explanations other than chance.

    I told her that law/ regularity can also produce compressibility and that is an explanation other than chance.

    My point is she is conflating “specification” with CSI

  15. Eric:

    “However Dembski says compressibility = specification.”

    Well, I doubt he says this. At least not without some qualification.

    Indeed, on the other end of a spectrum, we could have a specified string of 1000 ones or zeroes, which would be highly compressible. Yet at that point the string would not be complex.

    May I be so bold as to recommend that you actually read Dembski to learn what he means by specification and complexity. You’ll also learn that we don’t necessarily have to consider meaning/function in real-world applications, and you might learn a few things about info theory and realize that “information in the Shannon sense of compressibility” doesn’t make sense.

  16. Right- we don’t have to consider meaning/ function, but if it is present we don’t ignore it.

  17. RObb @15:

    LOL! May I be so bold as to recommend you read what I actually wrote! :) I have read a great deal of Dembski’s work and am quite familiar with complexity and specification.

    Always happy to learn, however, so perhaps you can explain to me where I’ve gone wrong in the qualifications I provided to the blanket statement that “compressibility = specification.”

    A slightly separate issue, but one related and raised by Meyer, which is why I brought it up: I’d also be interested in real-world examples of complex specified information in which meaning/function is wholly absent. And even if we can come up with a few scattered examples, are they the exception or the rule?

  18. Always happy to learn, however, so perhaps you can explain to me where I’ve gone wrong in the qualifications I provided to the blanket statement that “compressibility = specification.”

    Certainly. Your qualification is, “Indeed, on the other end of a spectrum, we could have a specified string of 1000 ones or zeroes, which would be highly compressible. Yet at that point the string would not be complex.” But Dembski’s most commonly used example of specified complexity is the sequence DDDDDDDDDDDDDDDDDDDDDDRDDDDDDDDDDDDDDDDDD, and the complexity of that sequence is a function of its length, not its incompressibility. So it’s incorrect to say that a string 10000 ones or zeros would not be complex.

    A slightly separate issue, but one related and raised by Meyer, which is why I brought it up: I’d also be interested in real-world examples of complex specified information in which meaning/function is wholly absent.

    I don’t see any meaning or function in a sequence of 1000 coin tosses that all come up heads. To be specified, something must be easily describable (which is an informal way of saying compressible) — it may be described in terms of its function or meaning, but it doesn’t have to be.

  19. But Dembski’s most commonly used example of specified complexity is the sequence DDDDDDDDDDDDDDDDDDDDDDRDDDDDDDDDDDDDDDDDD, and the complexity of that sequence is a function of its length, not its incompressibility.

    It’s also the context of that string- it isn’t just the string in isolation.

    I don’t see any meaning or function in a sequence of 1000 coin tosses that all come up heads.

    The meaning would be it was a fixed toss.

  20. R0bb:

    It occurs to me that this blog comment format may have caused us to misunderstand each other. If it was my poor wording, I apologize. What I stated is that a purely random string would not be very compressible. At the other end of the spectrum would be a string made up of all one digit, such as 1000 ones or 1000 zeroes. (If my statement was unclear and you thought I was talking about a string of 1000 ones and zeroes strung together in a particular way or in a random way, I apologize for the ambiguity.) My example relates to the compressibility of a string (of whatever length) that is made up of a single digit/character. It is the simplicity of the string that determines its compressibility, not so much its length.

    Are you saying Dembski says this string:

    DDDDDDDDDDDDDDDDDDDDDDRDDDDDDDDDDDDDDDDDD

    is complex? It would seem to me it is only ‘complex’ in comparison to the same string with all D’s, instead of one R thrown in the middle. Can it be specified? Sure. Could it also have independent meaning in the particular context in which it is used? Sure. But that doesn’t mean the string itself is complex. Rather that the string, taken together with available information and context, could point to design.

    I don’t see any meaning or function in a sequence of 1000 coin tosses that all come up heads.

    Where is the complexity in that? It could have meaning in some contexts (gambling, etc.), but I’m willing to stipulate it doesn’t for purposes of discussion, because it is irrelevant: we were talking about complex specified information. I’m curious to know if there are examples of complex specified information that just exist — no meaning or function associated with them.

    To be specified, something must be easily describable (which is an informal way of saying compressible) — it may be described in terms of its function or meaning, but it doesn’t have to be.

    Are you saying that the plans for a rocket engine are easily describable? In comparison to what? In comparison to the parts lying on the floor? What is it about a specification that makes it easily describable? That doesn’t make any sense to me as a blanket statement. It totally depends on what is being specified.

    —-

    I think we’ve gone a bit off the rails here as to what complexity and specification are. I’m going to read Dembski’s paper again to see if he says what you seem to be claiming he says. If you think I’m off base, please let me know where.

  21. I don’t see any meaning or function in a sequence of 1000 coin tosses that all come up heads.

    Where is the complexity in that?

    The complexity is in the improbability of the given outcome — it’s 1000 bits regardless of the resulting sequence.

    This is a prime example of the confusion sown by Dembski’s terminology. He dropped the term saturated probability in favor of the term specified complexity over a decade ago, and people are still confused by it. (Here‘s a recent example.) You’ve read a great deal of Dembski’s work and you’ve talked about CSI for years, but your question above shows that you still don’t know what complexity means in the context of specified complexity or CSI. That doesn’t speak well of the clarity of Dembski’s work.

  22. R0bb:

    Let me make sure I am understanding you correctly. You are saying that 1000 ones in a row is complex? That is a pretty remarkable statement. Particularly when the string could easily be produced by a law-like process. So according to you, a string of 1000 identical digits and a similar length string with the digits of pi are both complex?

    Apparently we’ve just eviscerated every possible meaning that could have existed in the word “complex.”

    But since you assert I don’t understand what complexity means in the context of CSI perhaps you can kindly clarify.

    What is complexity?

  23. But since you assert I don’t understand what complexity means in the context of CSI perhaps you can kindly clarify.

    What is complexity?

    Dembski defines complexity, when measured in bits, as log2(P(T|H)). But in practice, ID proponents don’t include the qualifier “under null hypothesis H” when they say that X has N bits of complexity. The de facto null hypothesis for CSI is a uniform distribution — that is what is always used in actual calculations, and in Dembski’s latest incarnation of CSI at the Evo Info Lab, he codifies that practice.

    So the complexity of 1000 coin flips is 1000 bits, regardless of the resulting sequence. And using the uniform null hypothesis, we can calculate the complexity of any outcome from any sample space, even if we don’t know what physical phenomenon is being modeled. Thus Dembski can say, as he often does, that a single letter is not complex, and he can calculate the complexity of a phrase like “METHINKS IT IS LIKE A WEASEL”, without reference to a null hypothesis. And the phrase “AAAAAAAAAAAAAAAAAAAAAAAAAAAA” has exactly the same complexity as “METHINKS IT IS LIKE A WEASEL”.

    If you have a hard time believing that easily described outcomes can be complex in Dembski’s parlance, consider that his examples of CSI include the plain rectangular monolith in “2001: A Space Odyssey”, a narrowband signal (epitomized by a pure sinusoidal), and the Caputo sequence DDDDDDDDDDDDDDDDDDDDDDRDDDDDDDDDDDDDDDDDD.

  24. And the phrase “AAAAAAAAAAAAAAAAAAAAAAAAAAAA” has exactly the same complexity as “METHINKS IT IS LIKE A WEASEL”.

    LoL! The two may have the same number of characters but not the same complexity- the string of A’s can be compressed whereas the “Methinks” phrase cannot be compressed.

    And again you are ignoring CONTEXT- why is that?

Leave a Reply