Dr. Anthony Fauci is a real time illustration on how absurdity becomes mainstream "science".

There is something of a bait and switch in your mt Rushmore example.

One can look at Mt Rushmore and know it was deliberately ~made~ by humans. We can see the tool marks, estimate its age of construction, etc …

As as we ~know~ that the things that are deliberately constructed by humans are designed, we can conclude Mt Rushmore was designed.

This is not the case at all with the machinery of the cell. The machinery in this case was constructed by the cell itself, which was constructed from the cell prior, and the cell prior, etc all the way back billions of years.

You have replicating systems that are billions of years old. Concluding design in this is not at all the same as looking at a human built artifact.

I agree that it is not a human construction, and it is self-replicating (which is remarkable in itself). The cell is also not hard stone. My point was to show that it is possible to detect design using objective criteria without knowing the history of the object under construction.

Do you suppose that with an organic, self-replicating cell, we could use an objective criterion to detect design? Or should we not even consider that a possibility?
 
Yes, of course they would think it was designed. And, yes, the video of the process of mitosis looks amazingly designed.Thanks for posting it.

Thanks for the feedback. These sorts of videos are fascinating. I would like to see more of these, or even continued improvement in their visuals. The video still oversimplifies what is happening in the cell because there is a lot more things happening, the nanomachines are not hard but soft and can change shape to multitask. Can you imagine a human factory with nanobots that are floating around changing shape and performing different functions, and then the whole factory organizes and replicates into two factories? It's breathtaking.
 
I agree that it is not a human construction, and it is self-replicating (which is remarkable in itself). The cell is also not hard stone. My point was to show that it is possible to detect design using objective criteria without knowing the history of the object under construction.

Do you suppose that with an organic, self-replicating cell, we could use an objective criterion to detect design? Or should we not even consider that a possibility?


I will never say it is not possible, and best of luck if you want to try, but I do see some pretty fundamental issues with trying to do this. Ones that could be extremely hard to overcome.

What you are talking about is a classification system - a heuristic that takes info about an object and assigns it a ‘class’. In your suggestion, it would be something like “was designed” vs “was not designed”. Similar types of classification systems are commonly built, and have very wide utility, think for example medical diagnosis, spam filtering, image segmentation, etc…

The application of such systems comes down to its accuracy, both in terms of its detection of true positives (sensitivity) and non-detection of true negatives (specificity). These rarely, if ever, are perfect (100%), and whether a designed classifier is useful or not depends on these characteristics relative to application it is being used for.

For example, in spam filtration, you would favor better specificity, so that you very rarely send good emails to the spam box, even if you ended up letting a couple emails from a desperate Nigerian prince through. For early cancer screening, you want the opposite, very high sensitivity - you don't want to miss a single case, even if you end up sending some people who are completely fine for follow up.

And the key here is these properties of the classifier need to be determined before its utility can be established and any weight put to an answer it gives when challenged with a new test.

The way that these properties are established is through validation testing using objects with known classification. For example, a user could go through their email and tag emails individually as spam or not, and then run the classifier. You could then calculate the sensitivity and specificity from the results and know how much better than a random guess your system was, and furthermore, whether you would trust it to make the classification for you.

So how would you validate a “design detector”? One thing you could do is run the algorithm on human constructed objects which we know were designed. Maybe you could run it on a range of objects, human built and natural, and show that it was able to accurately classify human designed objects as “designed” 99% of the time, and natural objects as “not designed” 98% of the time. That would be a fantastic classifier.

From this you could conclude that if you took a new object that you didn’t know what human made or not, if your classifier said “design!” you could say it was highly likely that the system under consideration was human designed. The heuristic had been previously validated for finding signs of human design, and you know although it can be wrong, the likelihood based on previous testing, is very small.

And so the problem though is the heuristic here has only been ‘validated’ for the classification you have done the controlled tested on, namely detecting human design in constructed objects. If you go and apply it to other situations, you ~will have no idea~ how much trust to put that classification. For example, I might have great confidence in the test results of a validated prostate cancer diagnostic, but what do I expect if I try to use it detect breast cancer? Do I expect the same accuracy that I found before? It is a “cancer” test after all.

Your proposed system would need relevant known test data that I just don't see exists. If you want to ask a classifier if a billion year old self replicating system was “designed” or not, without some way of establishing how good the system is at detecting design in ancient self-reproducing systems instead of human constructed objects, I just don't see how you could ever the trust the answer it gives.
 
Your proposed system would need relevant known test data that I just don't see exists. If you want to ask a classifier if a billion year old self replicating system was “designed” or not, without some way of establishing how good the system is at detecting design in ancient self-reproducing systems instead of human constructed objects, I just don't see how you could ever the trust the answer it gives.
Thanks for an excellent post.

There is a further problem for the ID side, in that most of then think that the universe was designed by some deity. Hence is is not possible to find any undesigned things in the universe to test any proposed design detector. If everything in the universe is designed, then a detector only needs a green light, you don't even need to connect the red light at all. Very simple to build, but completely unreliable, since the design assumption is built in.

For amusement, I applied Dr. Dembski's Explanatory Filter, which he proposed as a design detection methodology, to God. Due to the biases in Dembski's idea, it turns out that God is also designed. See here for the details.
 
Since we are discussing design detectors, I remembered my modest Proposal for a Theistic Design Detector from some time ago. When I started designing it, I thought that the calculations would be a lot more difficult than they turned out to be. Many potential inputs were not relevant after all, and the detector resolved down to something much simpler than I had expected.
 
I will never say it is not possible, and best of luck if you want to try, but I do see some pretty fundamental issues with trying to do this. Ones that could be extremely hard to overcome.

What you are talking about is a classification system - a heuristic that takes info about an object and assigns it a ‘class’. In your suggestion, it would be something like “was designed” vs “was not designed”. Similar types of classification systems are commonly built, and have very wide utility, think for example medical diagnosis, spam filtering, image segmentation, etc…

The application of such systems comes down to its accuracy, both in terms of its detection of true positives (sensitivity) and non-detection of true negatives (specificity). These rarely, if ever, are perfect (100%), and whether a designed classifier is useful or not depends on these characteristics relative to application it is being used for.

For example, in spam filtration, you would favor better specificity, so that you very rarely send good emails to the spam box, even if you ended up letting a couple emails from a desperate Nigerian prince through. For early cancer screening, you want the opposite, very high sensitivity - you don't want to miss a single case, even if you end up sending some people who are completely fine for follow up.

And the key here is these properties of the classifier need to be determined before its utility can be established and any weight put to an answer it gives when challenged with a new test.

The way that these properties are established is through validation testing using objects with known classification. For example, a user could go through their email and tag emails individually as spam or not, and then run the classifier. You could then calculate the sensitivity and specificity from the results and know how much better than a random guess your system was, and furthermore, whether you would trust it to make the classification for you.

So how would you validate a “design detector”? One thing you could do is run the algorithm on human constructed objects which we know were designed. Maybe you could run it on a range of objects, human built and natural, and show that it was able to accurately classify human designed objects as “designed” 99% of the time, and natural objects as “not designed” 98% of the time. That would be a fantastic classifier.

From this you could conclude that if you took a new object that you didn’t know what human made or not, if your classifier said “design!” you could say it was highly likely that the system under consideration was human designed. The heuristic had been previously validated for finding signs of human design, and you know although it can be wrong, the likelihood based on previous testing, is very small.

And so the problem though is the heuristic here has only been ‘validated’ for the classification you have done the controlled tested on, namely detecting human design in constructed objects. If you go and apply it to other situations, you ~will have no idea~ how much trust to put that classification. For example, I might have great confidence in the test results of a validated prostate cancer diagnostic, but what do I expect if I try to use it detect breast cancer? Do I expect the same accuracy that I found before? It is a “cancer” test after all.

Your proposed system would need relevant known test data that I just don't see exists. If you want to ask a classifier if a billion year old self replicating system was “designed” or not, without some way of establishing how good the system is at detecting design in ancient self-reproducing systems instead of human constructed objects, I just don't see how you could ever the trust the answer it gives.

I appreciate your response. You make some valid points. I like the direction of this.

AI for example, learns to distinguish a picture of a bird from something else through testing. Humans judge the AI choice as to whether it is correct or not using many pictures of birds and AI actually begins to build its own methodology that becomes increasingly accurate through this trial-and-error procedure with the assistance of humans. I suspect it will become better at this thus requiring less human intervention when learning in the next generations of AI. It is classification as you say that is dependent on previous validation.

There are many examples in modern technology for detecting patterns and you give some examples. Another would be casino's being able to detect a fraudulent activity, or banks sending you a notification when a suspicious transaction occurs. While they may not be 100% accurate, they are far more accurate than a random selector would be. Keep in mind that detection of intelligent design in life really only has to be 100% accurately correct 1 time to throw the entire evolutionary paradigm on its head. Charles Darwin put it this way: "If it could be demonstrated that any complex organ existed, which could not possibly have been formed by numerous, successive, slight modifications, my theory would absolutely break down. But I can find no such case".

So, I am in agreement with you on most of what you said, but what about detecting intelligent design when we don't know enough to validate based on experience?

Such is the case for the old SETI alien life searcher. A bit of a different approach had to be taken. Such is the case brought forward by the Intelligent Design Discovery Institute scientists and their use of the term "specified complexity". In other words, the quest is to find signals or patterns that could only be the result of intelligence because the best evidence available shows that no natural processes have been found to produce such signals or patterns. At a minimum, if our current best evidence that nature hasn't been found to produce specific signals or patterns, then intelligence ought to be at least a possible alternative. Are you willing to go that far? Why or why not?
 
Last edited:
Since we are discussing design detectors, I remembered my modest Proposal for a Theistic Design Detector from some time ago. When I started designing it, I thought that the calculations would be a lot more difficult than they turned out to be. Many potential inputs were not relevant after all, and the detector resolved down to something much simpler than I had expected.

Can you clarify what this is? I read your link a couple times but must have missed something. Thanks
 
Can you clarify what this is? I read your link a couple times but must have missed something. Thanks
I was working from Dr Dembski's Explanatory Filter, and his idea of Complex Specified Information. He covers this in reference 1 of my piece.

I work through the mathematics of CSI, to determine the minimum requirement for something to be complex. Another of Dembski's ideas is the Universal Probability Bound, which says that anything with a probability of 1 in 10^150 cannot happen by chance, and hence must have some other explanation. Given the size of the universe, the UPB is passed when 3 or more molecules, or atoms, are found in a 1 cubic metre volume. Since Uranium is the heaviest stable atom, that gives a mass above which three or more atoms or molecules must be present.

That result surprised me since I had expected something much more complex to work out.

Given that mass and volume were the only relevant parameters, the Detector is just a set of sensitive scales to determine the mass of the contents of a 1 cubic metre box. Something as simple as a set of scales can effectively detect design! At least according to Dembski's method.

For the Specification part of CSI, I rely on the attributes of a deity: omniscience and atemporality. That eliminates problems from an incomplete specification and a post-hoc specification. Obviously a correct detection of CSI relies on a correct specification being used.

As you can see from the date of my piece, I have been thinking about this for some time. Indeed that is why I have pushed for some evidence of any proposed Design Detector being thoroughly tested, as @sobchak also mentioned in post #866.

This result is one of the reasons I pointed out that it is impossible to test a Design Detector in a created universe, since if the whole universe is designed, there is nothing at all that is undesigned. This is a problem in any universe created by a deity: "In the beginning Vishnu created the heavens and the earth." :)
 
I was working from Dr Dembski's Explanatory Filter, and his idea of Complex Specified Information. He covers this in reference 1 of my piece.

I work through the mathematics of CSI, to determine the minimum requirement for something to be complex. Another of Dembski's ideas is the Universal Probability Bound, which says that anything with a probability of 1 in 10^150 cannot happen by chance, and hence must have some other explanation. Given the size of the universe, the UPB is passed when 3 or more molecules, or atoms, are found in a 1 cubic metre volume. Since Uranium is the heaviest stable atom, that gives a mass above which three or more atoms or molecules must be present.

That result surprised me since I had expected something much more complex to work out.

Given that mass and volume were the only relevant parameters, the Detector is just a set of sensitive scales to determine the mass of the contents of a 1 cubic metre box. Something as simple as a set of scales can effectively detect design! At least according to Dembski's method.

For the Specification part of CSI, I rely on the attributes of a deity: omniscience and atemporality. That eliminates problems from an incomplete specification and a post-hoc specification. Obviously a correct detection of CSI relies on a correct specification being used.

As you can see from the date of my piece, I have been thinking about this for some time. Indeed that is why I have pushed for some evidence of any proposed Design Detector being thoroughly tested, as @sobchak also mentioned in post #866.

This result is one of the reasons I pointed out that it is impossible to test a Design Detector in a created universe, since if the whole universe is designed, there is nothing at all that is undesigned. This is a problem in any universe created by a deity: "In the beginning Vishnu created the heavens and the earth." :)

I'm not totally following you on this one. Your statement, "if the whole universe is designed, there is nothing at all that is undesigned" is false.

For example, an ice crystal is not designed. It is specified because it has a repeating pattern of water molecules. Ice crystals are not complex because they are ordered by simple natural laws. A list of letters like 010101010101 is not complex either. It is specified because the text matches a simple repeating pattern.

Mt Everest is complex with all its shapes, but not specified. Its shapes can be easily explained by natural processes. Even though the chances of another mountain looking exactly like Everest are nil, a form that is unlikely is not enough to say it was designed.

Mt. Rushmore is complex with all its shapes, but it is also specified because it matches an independent pattern. "In the beginning God created the heavens and the earth" is complex and specified because the precise ordering of the letters as such is unlikely and it matches a known independent pattern of a readable English sentence. "dfadfarewarwaojasfdm2314%lm" is complex but not specified. It doesn't match an independent pattern.

I'm haven't researched the mathematical aspect of specified complexity, but from the non-mathematical statements you made avoid, it seems you are on the wrong track.

Your statement, "if the whole universe is designed, there is nothing at all that is undesigned" sounds like some sort of drug induced metaphysical statement. A Tesla car is designed but that in itself can't be evidence for whether the universe was designed or not. A rock that breaks off of Mt. Everest because of the wind is not designed and that in itself can't be evidence for whether the universe was designed or not.
 
Last edited:
For example, an ice crystal is not designed.
Then ice crystals are not created/designed by any deity, unlike most of the rest of the universe.

That goes against the basic assumption of the Abrahamic religions, and of many other religions. A created universe is obviously designed by its creator. If some component of the universe is not designed, then it was not created by any deity, and so no deity created the entire universe. At most that deity can only have created part of the universe.
 
Then ice crystals are not created/designed by any deity, unlike most of the rest of the universe.

That goes against the basic assumption of the Abrahamic religions, and of many other religions. A created universe is obviously designed by its creator. If some component of the universe is not designed, then it was not created by any deity, and so no deity created the entire universe. At most that deity can only have created part of the universe.
Couldn't some things be directly created, and other things be indirectly created?

For instance, God created the Universe could mean that God had an intended plan for the universe and set it into motion with a bang. He determined the laws of nature and constants that were required to get the intended result...a planet which would be habitable for creatures made in his image. He was able to watch all of it play out over billions of earth years. In effect, God directly created the universe and indirectly created everything in it.
 
Couldn't some things be directly created, and other things be indirectly created?

For instance, God created the Universe could mean that God had an intended plan for the universe and set it into motion with a bang. He determined the laws of nature and constants that were required to get the intended result...a planet which would be habitable for creatures made in his image. He was able to watch all of it play out over billions of earth years. In effect, God directly created the universe and indirectly created everything in it.
Sure, but where do you draw the line between directly and indirectly created?
I'm sure a lot of Christians take a similar stance in trying to reconcile the Biblical account, with what science tells us.
I doubt the YECs will get on board with that idea and neither will atheists.
 
Sure, but where do you draw the line between directly and indirectly created?
When God begins a process, like biogenesis followed by evolution, and includes self-replication, he is directly creating the process and indirectly creating all the follows or is produced by it.

Does that example answer your question?
I'm sure a lot of Christians take a similar stance in trying to reconcile the Biblical account, with what science tells us.
The natural world that science investigates is God's handiwork. It's sometimes questionable whether their interpretation of the world is correct.
I doubt the YECs will get on board with that idea and neither will atheists.
So? Their approval or disapproval is not what I'm pursuing.
 
When God begins a process, like biogenesis followed by evolution, and includes self-replication, he is directly creating the process and indirectly creating all the follows or is produced by it.
Does that example answer your question?
Ok, so did God kick of the universe and then only jump back in once the earth was ready for life?
Did he intervene in evolution to nudge it toward producing man?
You don't believe consciousness could have arisen from that process, so you'd have to invoke a further intervention.
So, No, it doesn't make anything clearer.

The natural world that science investigates is God's handiwork. It's sometimes questionable whether their interpretation of the world is correct.

So? Their approval or disapproval is not what I'm pursuing.
Can you give an example of a questionable interpretation?
Fair enough, I'm just saying, that idea won't win many fans on this forum.
 
I appreciate your response. You make some valid points. I like the direction of this.

AI for example, learns to distinguish a picture of a bird from something else through testing. Humans judge the AI choice as to whether it is correct or not using many pictures of birds and AI actually begins to build its own methodology that becomes increasingly accurate through this trial-and-error procedure with the assistance of humans. I suspect it will become better at this thus requiring less human intervention when learning in the next generations of AI. It is classification as you say that is dependent on previous validation.

There are many examples in modern technology for detecting patterns and you give some examples. Another would be casino's being able to detect a fraudulent activity, or banks sending you a notification when a suspicious transaction occurs. While they may not be 100% accurate, they are far more accurate than a random selector would be. Keep in mind that detection of intelligent design in life really only has to be 100% accurately correct 1 time to throw the entire evolutionary paradigm on its head. Charles Darwin put it this way: "If it could be demonstrated that any complex organ existed, which could not possibly have been formed by numerous, successive, slight modifications, my theory would absolutely break down. But I can find no such case".

So, I am in agreement with you on most of what you said, but what about detecting intelligent design when we don't know enough to validate based on experience?

Such is the case for the old SETI alien life searcher. A bit of a different approach had to be taken. Such is the case brought forward by the Intelligent Design Discovery Institute scientists and their use of the term "specified complexity". In other words, the quest is to find signals or patterns that could only be the result of intelligence because the best evidence available shows that no natural processes have been found to produce such signals or patterns. At a minimum, if our current best evidence that nature hasn't been found to produce specific signals or patterns, then intelligence ought to be at least a possible alternative. Are you willing to go that far? Why or why not?

Sure, I am willing to consider a hypothesis that some intelligent agent performed an action that left a pattern in the sum our observations, especially when that patten that has no known currently viable natural explanation. That seems completely reasonable.

And I would also say that a classification heuristic of unknown accuracy could certainly be a ~starting point~ here. Let’s say you have your system to detect design and intention, and that has it has been properly validated in human made systems vs naturally produced objects. You then apply it to an object that is very different than those contained in the validation set, and you see a ‘hit’. It would not be unreasonable to think there is the ~possibility~ of some aspect of human-like design in that system. It cannot be a ~conclusion~ of course, and as you do not know the accuracy of your classification heuristic in that scenario, you cannot even say anything about ~likelihood~, but if you were inclined, you could use it as a hint for further hypothesis based testing.

The search for intelligent life outside of earth is an excellent example of this. They are looking for specific signs of intelligence, and build classifiers to ‘find’ possibilities to further explore. A couple of specific examples -
  • Intelligent life may be trying to communicate with other civilizations
    • Build a heuristic to look for ‘human-like’ modulation of electromagnetic radiation.
  • An advanced civilization may be doing astroengineering
    • Build a classifier to detect potential dyson spheres through anomalies in stellar radiation.
  • A civilization will have a planet with atmospheric signs of industry
    • Use a test system to look for spectroscopic signals of specific molecules in an exoplanet atmosphere
In each case, a heuristic is developed is to identify ‘hits’. But as we have no idea about the accuracy of the heuristic for its intended purpose, each ‘hit’ is incredibly questionable. The system cannot be used to draw conclusions, but maybe, it gives a hint that there ~could~ be the possibility that there is something worth investigating further.
 
Last edited:
Ok, so did God kick of the universe and then only jump back in once the earth was ready for life?
That is one possibility to consider.
Did he intervene in evolution to nudge it toward producing man?
Maybe. My experience with God is that he is interactive with the world he designed down to the small details. Thinking that he set processes in motion and let them play out without intervention is a radically different way for me to think about him.
You don't believe consciousness could have arisen from that process, so you'd have to invoke a further intervention.
Yes. That would be an example of direct creation.
So, No, it doesn't make anything clearer.
My point is God would be credited as creator indirectly with the things that come about from the processes he set in place.
Can you give an example of a questionable interpretation?
The cause of the big bang. The cause of life on earth. The cause of consciousness. If one limits their view of these things to only material/natural causes their interpretation is skewed.
Fair enough, I'm just saying, that idea won't win many fans on this forum.
I don't know why you think that would matter to me.
 
Back
Top