Evaluating Dembski's ID

[reggieM]

You say “from the human perspective”. Why? One of the aims of ID is to extend the scientific perspective to include the supernatural. From the ID point of view the inclusion of the supernatural should be welcomed - why are you limiting your perspective to the purely human?

I think we might be mixing up some categories here.
In the Catholic view, a human being is composed of both natural and supernatural aspects. But I think what you’re getting at is seeing thing from a “divine perspective” – in other words, trying to view things from God’s point of view.
That is not as far-fetched as it might seem. Early scientists sought to understand what God was doing in the universe and they made assumptions based on what they knew about God.
The notion that the universe is consistent with human rationality is something that was tied to “how God did it” - since God is the author of human reason as well as of the universe.
But I mentioned “from the human perspective” because ultimately we can’t enter fully into the Divine Perspective.
As I said, for God, there is no “randomness”.
But from the human perspective, our knowledge is limited and we perceive randomness.
Design detection is an attempt to perceive the “effects of an intelligent agent”. There are lots of assumptions attached to that. The first, importantly, is that “the intelligent agent” has an intelligence that is compatible with human intelligence, and desires to communicate Himself to human beings.
Those are theological starting points. Some of that was assumed in the development of science itself.
So, when it comes to detecting design, it’s a question of probability based on what can reasonably be produced by a random process versus what requires some intelligence to produce.

The detection or not of design rests crucially on the existence or not of a specification, and we have no objective way to determine whether or not there is such a specification.

From what I’ve seen from Dembski and the people pursuing his ideas, objective definitions of specification are investigating function or relatedness and then measuring probability. Things like the measure of mutation rates plus time and a measure of the final output provides a baseline on what stochastic processes can create and the rate that they can create them.

In such a universe a working design detector will register “designed” whatever you point it at. Such a detector may well be correct, but it is also useless since it cannot distinguish anything.

Ok, thanks for that additional detail. Yes, that would be true depending on how one understands the meaning of the term “design” in this case. From the Catholic view, while it’s true that God “designs” everything in the universe, there is also the paradox of evil which is not directly designed by God but which is “permitted as a consequence” by God.

But measuring design from a human perspective means that we measure what human beings can detect and not what angels or God Himself can see underneath and beyond the material reality.

To put it another way, given an omnimax designer is there anything at all that such a designer could not have designed?

Again, that’s a good theological discussion point.
It certainly would be extremely important to answer that first before doing much of anything regarding society, laws, education, the judicial system, medicine, etc.

In other words, does God do everything?

Again, that’s a paradox that Catholicism answers in the teaching on free-will and the consequences of sin or virtuous acts (reward or punishment).

If we asserted that God directly creates and “does” everything in the universe, then human beings could not be held responsible for any of their actions. We could rightly say “God made me do it” when ever we were caught in any criminal activity (if that should happen to any of us, sadly).

We would need such an object to be able to test the “not designed” part of our design detector. What could we use to complete our testing?

That’s right. We choose “not designed” for those things that are reasonably understood as the result of natural processes.

If everything can be explained by natural processes alone, then one could say that there is no evidence that an intelligence was required to create anything (outside of the natural laws themselves but that is a different topic that requires other comparisons).

So, we test something like water and put it in the freezer and it comes out as ice cubes. Any person can do that with water and it is a repeatable process. It still may exhibit great intelligence or design principles in the substance of water itself and the properties of water (how can nature create hydrogen and oxygen molecules and what will water evolve into?)-- but the experiment can be produced through the course of nature.

 

[rossum]

I think we might be mixing up some categories here.

Very probably. ID is a mixture of science, politics and theology. It is often difficult to determine where the boundaries lie when discussing ID. For myself I am not very interested in the politics (I do not live in the USA) nor am I much interested in the theology. The scientific part of ID, and ID’s impact on the rest of science, are of more interest.

From what I’ve seen from Dembski and the people pursuing his ideas, objective definitions of specification are investigating function or relatedness and then measuring probability. Things like the measure of mutation rates plus time and a measure of the final output provides a baseline on what stochastic processes can create and the rate that they can create them.

Here I disagree. Probability calculations are useful to determine the Dembskian complexity of an object, they are of no use in determining whether or not a specification exists. In order to detect Complex Specified Information, the probability calculations are used in determining the “complex” part, not the “specified” part.

But measuring design from a human perspective means that we measure what human beings can detect and not what angels or God Himself can see underneath and beyond the material reality.

Science tries to be objective, that is it tries to remove the individual human perspective. Science contains many mechanisms intended to remove human biases - peer review, reproducibility and so forth. In that sense, science is looking for something beyond a purely human perspective.

That’s right. We choose “not designed” for those things that are reasonably understood as the result of natural processes.

One of the admitted problems with Dembski’s EF is that it can be prone to false negatives - registering “no design” when in fact design is present. This is because a designer can design something to look exactly as if it was the result of natural processes. We cannot rely on the detector in such cases.

Even if we have an object which we know is the result of natural processes, then from a theological point of view all of those natural processes were themselves designed by God, so there is at the least indirect design present in the object. I can design a computer program that produces some output. Is that output designed? I do not design the output directly, but surely there is an element of my design in the output. Dembski and others use this argument when discussing computer simulations of evolution. The same argument applies to a uiverse in which all of the natural processes were designed by God. The design detector should still detect design.

rossum

 

[FrankSchnabel]

Mornin Rossum,

The detection of specified complexity depends crucially on the existence of a specification and so far Dembski has been unable to supply an objective definition of what constitutes a specification. Without an objective definition then what constitutes a specification becomes dependent on which individual is performing the test, and that disqualifies it as science. Science must be reproducible by different researchers. The statememt may be possibly true but we cannot currently test it because we have no objectively tested examples of specified complexity to try it out on.

Well let’s take a look at specification. You have raised the issue of objectivity. Dembski in Ch 13 of The Design Revolution addresses the problem this way: “Does specified complexity describe an objective feature of the world or merely a subjective state of ignorance about the functioning of the world?”

SC is alleged to be a property of things. Properties come in two types: 1) objective properties that obtain regardless of who attributes them (e.g. solidity or fluidity) or 2) subjective properties that depend crucially on who attributes them

Dembski sez (p. 100-101): “The worry, then, is tht specified complexity may be entirely a subjective property, with no way of grasping nature at its ontological joints and thus no way of providing science with a valid tool for inquiry.”

Then, drawing on John Searles work, Dembski points out that often there is no neat distinction between objective and subjective. Sometimes there are elements of both. e.g. When we say X is married. Married is a social reality which is objective in one sense (presence of a marriage certificate) and subjective in another (depends on a stipulated agreement of subjects as to what is married).

Searle says we can not only distinguish properties in terms of objective-subjective, but also in terms of ontological-epistemic, with ontology referring to what exists and epistimology referring to what we know.

“Specifications, by being conditionally independent of the outcomes they describe, are, within Searle’s scheme, epistemically objective. Moreover, once a specification is given and the event it represents is identified, the probability of that outcome is ontologically objective.” p. 102

 

[FrankSchnabel]

So while the pattern (specification) under examination is ontologically subjective (depends on a human subject to perceive it), it is epistemically objective. e.g. we can agree that the bacterial flagellum functions like an outboard motor. Furthermore, once the pattern is identified, its associated probability is ontologically objective.

I think, then, that we can dispense with the criticism that specifications are purely subjective.

But has Dembski failed to provide “supply an objective definition of what constitutes a specification”?

According to Dr. Bill, specifications are conditionally independent patterns. “A pattern is conditionally independent of an event if adding our knowledge of the pattern to a chance hypothesis does not alter the event’s probability under that hypothesis.” p. 82.
Dr. Bill uses the term “detachability” for this conditional independence.

From No Free Lunch p. 15: “Detachability can be understood as asking the following question: given an event whose design is in question and a pattern describing it, would we be able to explicitly identify or exhibit the pattern if we had no knowledge which event occurred? Here is the idea. An event has occurred. A pattern describing the event is given. The event is one from a range of possible events. If all we knew was the range of possible events without any specifics about which event actually occurred (e.g. we know that tomorrow’s weather will be rain or shine, but we do not know which), could we still identify the pattern describing the event? If so, the pattern is detachable from the event.”

Now, we can agree on this definition or not. If not, we have no “objective” definition on which to proceed. IOW, we have no inter-subject agreement. The question then becomes, why can’t folks agree on this definition of specification?

 

[FrankSchnabel]

Hey SpiritHound,

I sed: Even if we were able to conceive of possible evolutionary explanations, isn’t it reasonable to reject them if they are too improbable?

To which the Hound responded: A phrase I like to use is “Every week some idiot wins the lottery.” It means we have a singularly improbable event, winning the lottery, but if millions of people play, somebody is going to win. similarly, have our statisticians taken into account all the times this evolution has failed to take place? Maybe now, 20,000 years with our 1B bacteria later, it’s due for something crazy to happen?

But let’s say that my chance of winning the lottery is 1 out of 10 to the 150th power. This is Dr. Bill’s proposed Universal Probability Bound. He sez at p. 85 of the TDR: “A universal probability bound is impervious to all available probalistic resources that can be brought against it. Indeed, all the probabilistic resources in the known physcial world cannot conspire to render remotely probable an event whose probability is less than this universal probability bound.”

So Dr. Bill is telling us that we cannot assume that chance resources are unlimited. We live in a finite universe, after all. If the probability of anyone winning the lottery is less than the UPB, then no one will win.

Somewhere Dr. Bill illustrates the problem this way. Let’s say Obama proposes we change the federal sentencing statutes. To not totally deprive even heinous criminals hope of escaping incarceration, Obama sez we should let a convict get out early if he can flip X heads in row using a fair coin. X will be set according to the number of years which in turn is determined by the normal sentencing guidelines (based on the severity of the crime, mitigating circumstances, etc.)

But let’s say we have an unrepentant mass murderer, KKK, serial rapist, baby seal abuser who, under the guidelines, merits a 100 year sentence. Under this new scheme, if he tosses 100 heads in a row he walks. So our convict on day one, filled with the audacity of hope, starts flipping his coin. Is his hope justified? Nope. Not enuff time, even in a hundred years, to generate enuff of a chance of pulling it off.

Some improbabilties are too large to overcome.

 

[FrankSchnabel]

Rossum sed earlier:

I think that it is not possible to calculate such a probability with any degree of accuracy. Indeed I would suspect that the error bars on such a calculation would be so wide as to prevent it being used for anything.

Biological forms arise from their parent or parents. You would need to calculate the probabilites for every generation back to the origin of life 4.5 billion years ago. That is not a practical proposition given the amount of information that has been lost over time.

But what if we look at a more isolated case? Let’s say we are looking at bacterium X and specifically a change in one aspect of the organism. We know pre-X’s DNA was configured thus and so. After the mutation, it is now different. In order to achieve this change, the DNA code has to be altered at certain points. Knowing the rate at which mutations can occur, can we not calculate the probability of this change occuring in one step. As the Germans say, “So weit, so gut.” Maybe we even could isolate structures or functions that would have to happen in multiple steps, in which case the improbabilities escalate. And, in theory, if everyone is comfy with the way we calculated the improbabilities, we can then make judgments regarding the presence of SC.

Merely looking at probabilities is not sufficient. If I shuffle two packs of non-identical cards together then there are 104! = 1.03 x 10166 possible arrangements of those cards. Despite the probability of any particular arrangement being below Dembski’s UPB, one of those arrangements is seen in the shuffled pack. Extremely improbable events do happen, and can happen very easily in some circumstances. How improbable is the precise current arrangement of all of the sandgrains on Bondi Beach?

Yes, merely looking at probabilities is not sufficient to warrant a design inference. There must also be a specification before the EF will crank out a design inference.

 

[Spirithound]

So while the pattern (specification) under examination is ontologically subjective (depends on a human subject to perceive it), it is epistemically objective. e.g. we can agree that the bacterial flagellum functions like an outboard motor.

Does it matter to our concept of “pattern” that in making an outboard motor, we copied the flagellum?

But let’s say that my chance of winning the lottery is 1 out of 10 to the 150th power. This is Dr. Bill’s proposed Universal Probability Bound. He sez at p. 85 of the TDR: “A universal probability bound is impervious to all available probalistic resources that can be brought against it. Indeed, all the probabilistic resources in the known physcial world cannot conspire to render remotely probable an event whose probability is less than this universal probability bound.”

So Dr. Bill is telling us that we cannot assume that chance resources are unlimited. We live in a finite universe, after all. If the probability of anyone winning the lottery is less than the UPB, then no one will win.

But let’s say we have an unrepentant mass murderer, KKK, serial rapist, baby seal abuser who, under the guidelines, merits a 100 year sentence. If he tosses 100 heads in a row he walks. So our convict on day one, filled with the audacity of hope, starts flipping his coin. Is his hope justified? Nope. Not enuff time, even in a hundred years, to generate enuff of a chance of pulling it off.

Some improbabilties are too large to overcome.

Some improbabilities are too large to overcome individually. But isn’t this the point for stipulating that we need 1B bacteria and 20,000 years? The position is not that there are unlimited resources which generate change, but just a really really large amount of these resources.

 

[buffalo]

Does it matter to our concept of “pattern” that in making an outboard motor, we copied the flagellum?

Some improbabilities are too large to overcome individually. But isn’t this the point for stipulating that we need 1B bacteria and 20,000 years? The position is not that there are unlimited resources which generate change, but just a really really large amount of these resources.

We had a propeller before we could see the flagellum, so how could we copy it?

 

[Spirithound]

We had a propeller before we could see the flagellum, so how could we copy it?

Point taken!

 

[FrankSchnabel]

Hey Hound,

Some improbabilities are too large to overcome individually. But isn’t this the point for stipulating that we need 1B bacteria and 20,000 years? The position is not that there are unlimited resources which generate change, but just a really really large amount of these resources.

I think Dr. Bill’s point is that some probabilities are too large to overcome for even brazillions over a very long (but finite) time.

 

[buffalo]

Point taken!

Let’s explore -

We know we designed the propeller. We know it is man made.
We know the processes it takes to produce it.

We see the flagellum. Dembski thought it to be designed, others that it came about.

Can we draw any conclusions from this?

That for something to propel itself through a medium it needs a propeller. So we see it in nature and we see it in man made design.

I don’t know where I am going with this.

 

[The_Barbarian]

Let’s explore -

We know we designed the propeller. We know it is man made.
We know the processes it takes to produce it.

Yes. We can see that it is designed.

We see the flagellum. Dembski thought it to be designed, others that it came about.

Can we draw any conclusions from this?

With a little evidence we can. The first key point is that “the flagellum” is not irreducibly complex. There are several versions of the bacterial flagellum, each a bit more complex than the next.

And there is a structure, containing most of the elements found in bacterial flagella, in a bacterial structure that isn’t for movement at all. The Type III Secretory Apparatus is quite similar, has many of the same components in the same assemblies, but has no role in motion:
millerandlevine.com/km/evol/design2/article.html

That for something to propel itself through a medium it needs a propeller.

Actually not. Many bacteria move about without any propeller at all. Because of the very low Reynolds numbers, propellers are not nearly as efficient at that size as they are for larger propellors. It’s not very efficient, but it works well enough to survive.

So we see it in nature and we see it in man made design.

Except, of course that in man-made designs, it’s optimally efficient.

And, in nature we see evidence of evolution from other structures not related to movement.

I don’t know where I am going with this.

I don’t either, but it seems that you helped bring up an important point.

 

[buffalo]

Yes. We can see that it is designed.

With a little evidence we can. The first key point is that “the flagellum” is not irreducibly complex. There are several versions of the bacterial flagellum, each a bit more complex than the next.

And there is a structure, containing most of the elements found in bacterial flagella, in a bacterial structure that isn’t for movement at all. The Type III Secretory Apparatus is quite similar, has many of the same components in the same assemblies, but has no role in motion:
millerandlevine.com/km/evol/design2/article.html

Actually not. Many bacteria move about without any propeller at all. Because of the very low Reynolds numbers, propellers are not nearly as efficient at that size as they are for larger propellors. It’s not very efficient, but it works well enough to survive.

Except, of course that in man-made designs, it’s optimally efficient.

And, in nature we see evidence of evolution from other structures not related to movement.

I don’t either, but it seems that you helped bring up an important point.

Is the man made propeller irreducibly complex?

 

[Sideline]

It’s interesting to me that Intelligent Design has often been compared to archaeology in that they both try and discover design. But actually, archaeology tries to discover “cultural modification”. They try not only to distinguish things that people have designed, but things that people have modified with a definite purpose or not.

What is interesting is that they are entirely different approaches to the issue. One of the most important features of archaeology is looking at the context of an artifact. In other words, what’s around this thing you are trying to study.

Complexity doesn’t factor into it at all. We are forever finding things that look like they were done on purpose, but aren’t. What is more important is where you found something and, what else was around it. If you find several pieces of obsidian miles away from any source of obsidian, you can guess it is culturally modified. If you find some pieces of obsidian lying in amongst the bones of a human being, several miles from a source of obsidian, then you can be almost certain that it has been modified.

Dembski’s test, oddly enough, would say these weren’t designed. He’d be right, of course, they aren’t designed, but they do show evidence of human interference. Humans moved them. It is interesting to note here, that they are no more complex than the obsidian found near the source, but they are still culturally modified.

We can also determine whether these pieces of obsidian were further modified by comparing them to obsidian found in context. If we find marks on obsidian that is associated with human bones, and we don’t find those marks on naturally occurring obsidian, we can somewhat safely conclude that those marks are the result of cultural modification. Again, this is due to the context of the marks. It is not whether they are more complex. Nor whether they are complex and useful. It is whether they are unique to a culturally modified context.

Consider the bacterial flagellum. It is found in the context of a cell. It is only found in that context. It is made of the same material as the rest of the cell. It doesn’t have tool marks on it. It doesn’t have anything out of place on it at all, in the context of the cell. It is as if we found a piece of obsidian, in a source of obsidian, that resembles an arrowhead. Unless we can also find marks that are unique to cultural modification, we can’t assume anything other than that it looks like an arrowhead. It may even be able to function like an arrowhead without any modification. We can not assume that it was designed to be an arrowhead.

The bacterial flagellum operates as a propeller. It looks like a human propeller. It certainly is NOT a human designed propeller. Could some creature have designed the propeller? If some being or beings designed all life on earth, then yes. If not all of life was designed, then probably not.

SETI actually works on exactly the same principles as archaeology, not the same way as ID. They are looking for a pattern of waves that are different than the background waves. It isn’t the complexity of the information, it is the context. The only known sources of those waves on Earth are by humans. The idea is that that if they are found in space, they could also be produced by intelligent life. If they find natural sources of those waves, then the experiment would be useless.

None of this proves ID wrong, it only shows that ID is NOT similar to the sciences with which it claims to be similar. ID is trying to establish a way of detecting design that is unique, and has in no way been shown to be effective.

 

[rossum]

Is the man made propeller irreducibly complex?

No, it has at least one other function - to act as an ornament on the walls of pubs and bars near airports, especially long established airports used in either of the world wars.

rossum

 

[rossum]

We had a propeller before we could see the flagellum, so how could we copy it?

The flagellum existed long before the man-made propellor so how can the man-made version act as a prior specification? We are not allowed to mark our targets after we have shot the arrows.

rossum

 

[buffalo]

No, it has at least one other function - to act as an ornament on the walls of pubs and bars near airports, especially long established airports used in either of the world wars.

rossum

Yeah and on the back of trucks!

 

[rossum]

Well let’s take a look at specification. You have raised the issue of objectivity. Dembski in Ch 13 of The Design Revolution addresses the problem this way: “Does specified complexity describe an objective feature of the world or merely a subjective state of ignorance about the functioning of the world?”

SC is alleged to be a property of things. Properties come in two types: 1) objective properties that obtain regardless of who attributes them (e.g. solidity or fluidity) or 2) subjective properties that depend crucially on who attributes them

Dembski sez (p. 100-101): “The worry, then, is tht specified complexity may be entirely a subjective property, with no way of grasping nature at its ontological joints and thus no way of providing science with a valid tool for inquiry.”

Then, drawing on John Searles work, Dembski points out that often there is no neat distinction between objective and subjective. Sometimes there are elements of both. e.g. When we say X is married. Married is a social reality which is objective in one sense (presence of a marriage certificate) and subjective in another (depends on a stipulated agreement of subjects as to what is married).

Searle says we can not only distinguish properties in terms of objective-subjective, but also in terms of ontological-epistemic, with ontology referring to what exists and epistimology referring to what we know.

“Specifications, by being conditionally independent of the outcomes they describe, are, within Searle’s scheme, epistemically objective. Moreover, once a specification is given and the event it represents is identified, the probability of that outcome is ontologically objective.” p. 102

Is Searle’s scheme objective or subjective?

I suspect that for most scientists, qua scientists, would expect something a little more definite than Dembski is currently providing. It is good to see that Dembski is recognising the problem, but I do not think that many will accept his attempt to effectively define it away. A working design detector will require something a little more definite - the sort of thing that a programmer can code into the microprocessor included in the detector. I do not expect it to be easy to do that with Searle’s definition.

rossum

 

[rossum]

So while the pattern (specification) under examination is ontologically subjective (depends on a human subject to perceive it), it is epistemically objective. e.g. we can agree that the bacterial flagellum functions like an outboard motor. Furthermore, once the pattern is identified, its associated probability is ontologically objective.

I do not agree that a flagellum functions “like” an outboard motor, at least not for all allowed definitions of “like”. For instance an outboard motor can be moved inboard if required, a flagellum cannot. In some aspects a flagellum is more llike an inboard motor in that the actual “motor” part resides inside the bacterium’s membrane rather than outside it, only the “propellor” protrudes outside the “hull” as with an inboard motor.

If a design detector is required to take a vote of the local human population before making a decision, the result may well be democratic but I do not think that it would neccessarily be scientific.

I think, then, that we can dispense with the criticism that specifications are purely subjective.

I am not asserting that they are purely subjective, merely that he has not yet eliminated all subjectivity from his proposed definition. Dr Dembski has not yet given us an objective definition of what is, and what is not, a specificaition. Science needs such an objective definition, one with no subjective component.

But has Dembski failed to provide “supply an objective definition of what constitutes a specification”?

According to Dr. Bill, specifications are conditionally independent patterns. “A pattern is conditionally independent of an event if adding our knowledge of the pattern to a chance hypothesis does not alter the event’s probability under that hypothesis.” p. 82.
Dr. Bill uses the term “detachability” for this conditional independence.

Dembski has certainly provided some pointers but these are insufficient - witness my points about outboard or inboard motors as a specification for the flagellum. Both are “detached” in Dembski’s sense so detachment gives us no objective way to determine if neither, one or both constitute a correct specification.

Now, we can agree on this definition or not. If not, we have no “objective” definition on which to proceed. IOW, we have no inter-subject agreement. The question then becomes, why can’t folks agree on this definition of specification?

As I see it detachment is neccessary, but not sufficient. It is better than nothing but more is required.

rossum

 

[Spirithound]

Is the man made propeller irreducibly complex?

Well, I don’t think so…Think “steel propeller shaved down to almost nothing, and remove one arm at a time”.